methods in bioengineering - muhammad1988adeel · 2012-02-15 · 11.2.3 preparation of the...

Methods in BioengineeringNanoscale Bioengineering and Nanomedicine

The Artech House Methods in Bioengineering Series

Series Editors-in-Chief

Martin L. Yarmush, M.D., Ph.D.Robert S. Langer, Sc.D.

Methods in Bioengineering: Biomicrofabrication and Biomicrofluidics,Jeffrey D. Zahn and Luke P. Lee, editors

Methods in Bioengineering: Microdevices in Biology and Medicine,Yaakov Nahmias and Sangeeta N. Bhatia, editors

Methods in Bioengineering: Nanoscale Bioengineering and Nanomedicine,Kaushal Rege and Igor Medintz, editors

Methods in Bioengineering: Stem Cell Bioengineering,Biju Parekkadan and Martin L. Yarmush, editors

Methods in Bioengineering: Systems Analysis of Biological Networks,Arul Jayaraman and Juergen Hahn, editors

Series Editors

Martin L. Yarmush, Harvard Medical SchoolChristopher J. James, University of Southampton

Advanced Methods and Tools for ECG Data Analysis,Gari D. Clifford, Francisco Azuaje, and Patrick E. McSharry, editors

Advances in Photodynamic Therapy: Basic, Translational, and Clinical,Michael Hamblin and Pawel Mroz, editors

Biomedical Surfaces, Jeremy Ramsden

Intelligent Systems Modeling and Decision Support in Bioengineering,Mahdi Mahfouf

Translational Approaches in Tissue Engineering and Regenerative Medi-cine, Jeremy Mao, Gordana Vunjak-Novakovic, Antonios G. Mikos, andAnthony Atala, editors

Methods in BioengineeringNanoscale Bioengineering and Nanomedicine

Kaushal RegeDepartment of Chemical EngineeringArizona State University

Igor L. MedintzCenter for Biomolecular Science and EngineeringU.S. Naval Research Laboratory

Editors

a r techhouse . com

Library of Congress Cataloging-in-Publication DataA catalog record for this book is available from the U. S. Library of Congress.

British Library Cataloguing in Publication DataA catalogue record for this book is available from the British Library.

ISBN-13: 978-1-59693-410-8

Text design by Darrell JuddCover design by Igor Valdman

© 2009 Artech House. All rights reserved.

Printed and bound in the United States of America. No part of this book may be reproduced orutilized in any form or by any means, electronic or mechanical, including photocopying, record-ing, or by any information storage and retrieval system, without permission in writing from thepublisher.

All terms mentioned in this book that are known to be trademarks or service marks have beenappropriately capitalized. Artech House cannot attest to the accuracy of this information. Use ofa term in this book should not be regarded as affecting the validity of any trademark or servicemark.

10 9 8 7 6 5 4 3 2 1

Contents

Preface xv

CHAPTER 1Preparation and Characterization of Carbon Nanotube-Protein Conjugates 1

1.1 Introduction 2

1.2 Materials 3

1.3 Methods 3

1.3.1 Physical Adsorption of Proteins on Carbon Nanotubes 3

1.3.2 Protein Assisted Solubilization of Carbon Nanotubes 4

1.3.3 Covalent Attachment of Proteins onto Carbon Nanotubes 5

1.4 Data Acquisition, Anticipated Results, and Interpretation of Data 7

1.4.1 Characterization of Proteins Physically Adsorbed onto1.4.1 Carbon Nanotubes 7

1.4.2 Characterization of Protein-Solubilized Carbon Nanotubes 11

1.4.3 Characterization of Covalently Attached Carbon1.4.1 Nanotube-Protein Conjugates 13

1.5 Discussion and Commentary 18

1.6 Applications Notes 19

1.7 Summary Points 21

Acknowledgments 21

References 21

CHAPTER 2Peptide-Nanoparticle Assemblies 25

2.1 Introduction 26

2.2 Materials 27

2.3 Methods 28

2.3.1 Coil-Coil Peptide Mediated NP Assembly 28

2.3.2 Synthesis of Hybrid Structures Using Multifunctional Peptides 31

2.4 Assembly Mediated by Metal Ion-Peptide Recognition 32

2.5 Peptides as Antibody Epitopes for Nanoparticle Assembly 33

2.6 DATA Acquisition, Anticipated Results, and Interpretation 34


v

2.8 Application Notes 36


Acknowledgments 36

References 37

CHAPTER 3Nanoparticle-Enzyme Hybrids as Bioactive Materials 39

3.1 Introduction 40

3.2 Materials 40

3.3 Methods 41

3.3.1 Enzyme-Attached Polystyrene Nanoparticles 41

3.3.2 Polyacrylamide Hydrogel Nanoparticles for3.3.2 Entrapment of Enzymes 41

3.3.3 Magnetic Nanoparticles with Porous Silica Coating for3.3.3 Enzyme Attachment 42

3.3.4 Enzyme Loading and Activity Assay 42

3.4 Results 44

3.4.1 Polystyrene-Enzyme Hybrid Nanoparticles 44

3.4.2 Polyacrylamide Hydrogel Nanoparticles with3.4.2 Entrapped Enzymes 45

3.4.3 Magnetic Nanoparticles for Enzyme Attachment 46


3.6 Troubleshooting 49



Acknowledgments 50

References 50

CHAPTER 4Self-Assembled QD-Protein Bioconjugates and Their Use in FluorescenceResonance Energy Transfer 53

4.1 Introduction 54

4.2 Materials 56

4.2.1 Reagents 56

4.2.2 Equipment 56

4.3 Methods 56

4.3.1 Quantum Dot Synthesis 56

4.3.2 Surface Ligand Exchange 58

4.3.3 Biomolecule Conjugation 61

4.3.4 Fluorescence Measurements 65

4.4 Data Analysis and Interpretation 66

4.4.1 Calculating Donor-Acceptor Distances 68

4.4.2 Calculating Reaction Rates of Surface-Bound Substrates 70

Contents

vi


4.6 Conclusions 72

References 72

Annotated References 74

CHAPTER 5Tracking Single Biomolecules in Live Cells Using Quantum Dot Nanoparticles 75

5.1 Introduction 76

5.2 Materials 78

5.2.1 Reagents 78

5.2.2 Imaging Equipment 79

5.3 Methods 79

5.3.1 Forming QD Bioconjugates 79

5.3.2 Treating Cells with QD Bioconjugates 79

5.4 Data Acquisition, Anticipated Results, and Interpretation 79

5.4.1 Imaging QD-Bound Complexes in Cells 79

5.4.2 Analysis of the Real-Time QD Dynamics 80


References 82

CHAPTER 6Nanoparticles as Biodynamic Substrates for Engineering Cell Fates 85

6.1 Introduction 86

6.2 Experimental Design 88

6.3 Materials 88

6.3.1 Cell Culture, Fixing, Staining, and Analysis Reagents 88

6.3.2 Nanoparticle Fabrication and Functionalization 89

6.3.3 Microscale Plasma Initiated Patterning 89

6.4 Methods 89

6.4.1 Albumin Nanoparticle Fabrication 89

6.4.2 Albumin Nanoparticle Functionalization 91

6.4.3 Albumin Nanoparticle Pattern Creation—Microscale6.4.3 Plasma Initiated Patterning (μPIP) 93

6.4.4 Cell Culture 94

6.4.5 Keratinocyte Morphology and Migration 94

6.4.6 Fibroblast Extracellular Matrix Assembly 94

6.4.7 Cell Attachment Assay 95

6.5 Results 95

6.5.1 Enhanced Cell Migration 95

6.5.2 Enhanced Extracellular Matrix Assembly 97

6.6 Discussion of Pitfalls 100

6.6.1 Spatial Guidance of Cell Attachment—Microscale Plasma6.6.1 Initiated Patterning 100

Contents

vii

6.6.2 Three-Dimensional Presentation of Albumin Nanoparticles 101


Acknowledgments 103

References 103

CHAPTER 7Magnetic Cell Separation to Enrich for Rare Cells 107

7.1 Introduction 108

7.1.1 Principle 110

7.1.2 Examples of Cell Magnetic Separation Applications 115

7.2 Materials and Methods 116

7.2.1 Enrichment Process 116

7.2.2 Red Cell Lysis Step 117

7.2.3 Immunomagnetic Labeling 117

7.2.4 Magnetic Cell Separation Step 117

7.3 Data Acquisition, Results, and Interpretation 117


7.5 Summary Points to Obtain High-Performance,7.5 Magnetic Cell Separations 120

Acknowledgments 120

References 121

CHAPTER 8Magnetic Nanoparticles for Drug Delivery 123



8.3 Materials 126

8.3.1 Reagents 126

8.3.2 Facilities and Equipment 127

8.4 Methods 128

8.4.1 Synthesis of Magnetic Nanoparticles 128

8.4.2 Physical Characterization of Magnetic Nanoparticles 129

8.4.3 Conversion of DOX•HCl 129

8.4.4 Drug Loading and Release Kinetics 129

8.4.5 Kinetics of DOX Release from Magnetic Nanoparticles 130

8.4.6 Antiproliferative Activity of Doxorubicin Loaded Magnetic8.4.6 Nanoparticles on MCF-7 Cells 131

8.4.7 Antiproliferative Activity of Doxorubicin Loaded Magnetic8.4.6 Nanoparticles on MCF-7 Cells in the Presence of a8.4.6 Magnetic Field 131




Contents

viii


Acknowledgments 135

References 135

CHAPTER 9Imaging and Therapy of Atherosclerotic Lesions with Theranostic Nanoparticles 137



9.3 Materials 140

9.3.1 Reagents 140

9.3.2 Facilities/Equipment 140

9.3.3 Animal Model 141

9.3.4 Alternate Reagents and Equipment 141

9.4 Methods 141

9.4.1 Synthesis of Theranostic Nanoparticles 141

9.4.2 Intravital Fluorescence Microscopy 143

9.4.3 Light-Based Therapy 144


9.5.1 Characterization of Theranostic Nanoparticles 145

9.5.2 Animal Experimentation 146

9.5.3 Intravital Fluorescence Microscopy 146

9.5.4 Statistical Analyses 147

9.5.5 Anticipated Results 148



Acknowledgments 150

References 150

CHAPTER 10Biomedical Applications of Metal Nanoshells 153


10.1.1 Biomedical Applications of Metal Nanoshells 154

10.1.2 Nanoshells for Combined Optical Contrast and10.1.2 Therapeutic Application 155


10.3 Materials 156

10.3.1 Nanoparticle Production 156

10.3.2 Protein Conjugation to Nanoshells Surface 156

10.3.3 Cell Culture 157

10.3.4 In Vitro Assays 157

10.4 Methods 157

10.4.1 Fabrication of Gold/Silica Core Nanoshells 157

10.4.2 Nanoshells for Combined Imaging and Therapy In Vivo 158

Contents

ix

10.4.3 Passivation of Nanoshells with PEG 159

10.4.4 Conjugation of Biomolecules to Nanoshells 160

10.4.5 Quantification of Antibodies on Nanoshells 160

10.5 Results 161

10.5.1 Gold/Silica Nanoshells Allow Both Imaging Contrast Increase10.5.1 and Therapeutic Benefit 161

10.5.2 Evaluation of Antibody Concentration per Nanoshell 163

10.6 Discussion of Pitfalls 163

10.7 Statistical Analysis 165

Acknowledgments 166

References 166

CHAPTER 11Environmentally Responsive Multifunctional Liposomes 169


11.1.1 Cis-Aconityl Linkage 171

11.1.2 Trityl Linkage 172

11.1.3 Acetal Linkage 172

11.1.4 Polyketal Linkage 172

11.1.5 Vinyl Ether Linkage 172

11.1.6 Hydrazone Linkage 173

11.1.7 Poly(Ortho-Esters) 173

11.1.8 Thiopropionates 173

11.2 Materials 174

11.2.1 Chemicals 174

11.2.2 Syntheses 175

11.2.3 Preparation of the TATp-Bearing, Rhodamine-Labeled11.2.3 Liposomal Formulations 175

11.2.4 Preparation of the TAtp-Bearing, Rhodamine Labeled,11.2.3 pGFP Complexed Liposomal Formulations 175

11.3 Methods 176

11.3.1 Synthesis of Hydrazone-Based mPEG-HZ-PE Conjugates 176

11.3.2 Synthesis of PE-PEG1000-TATp Conjugate 183

11.3.3 In Vitro pH-Dependant Degradation of PEG-HZ-PE11.3.3 Conjugates 184

11.3.4 Avidin-Biotin Affinity Chromatography 184

11.3.5 In Vitro Cell-Culture Study 184

11.3.6 In Vivo Study 185

11.3.7 In Vivo Transfection with pGFP 185


11.4.1 Synthesis of Hydrazone-Based mPEG-HZ-PE Conjugates 185

11.4.2 Synthesis of PE-PEG1000-TATp Conjugate 186

11.4.3 In Vitro pH-Dependant Degradation of PEG-HZ-PE11.4.3 Conjugates 186

Contents

x

11.4.4 Avidin-Biotin Affinity Chromatography 188

11.4.5 In Vitro Cell Culture Study 188

11.4.6 In Vivo Study 188

11.4.7 In Vivo pGFP Transfection Experiment 189

11.5 Conclusion 191


Acknowledgments 192

References 192

CHAPTER 12Biodegradable, Targeted Polymeric Nanoparticle Drug Delivery Formulationfor Cancer Therapy 197


12.2 Materials 200

12.2.1 Polymer Synthesis of PLA-PEG and PLGA-PEG 200

12.2.2 Nanoparticle Formation 201

12.2.3 Ligand Conjugation 201

12.2.4 Quantification of Drug Encapsulation 201

12.2.5 Release Experiments 202

12.2.6 Postformulation Treatment 202

12.2.7 Cell Binding and Uptake Experiments 202

12.2.8 Cytotoxicity Experiments 203

12.3 Methods 203

12.3.1 Polymer Synthesis of PLA-PEG and PLGA-PEG 204

12.3.2 Nanoparticle Formation 207

12.3.3 Conjugation of Targeting Ligand 209

12.3.4 Quantification of Drug Encapsulation 211

12.3.5 Drug Release Studies 212

12.3.6 Postformulation Treatment 213

12.3.7 In Vitro Experiments: Cell Binding and Uptake Studies 214

12.3.8 In Vitro Experiments: Cytotoxicity Studies 215

12.4 Data Acquisition, Results, and Interpretation 216

12.4.1 Polymer Characterization 216

12.4.2 Nanoparticle characterization 217

12.4.3 In Vitro Experiments 220


12.5.1 Particle Size 222

12.5.2 Particle Shape 224

12.5.3 Surface Chemistry 224

12.5.4 Drug Loading 225

12.5.5 Drug Release 226

12.5.6 Active Targeting and Ligand Conjugation 228

12.6 Troubleshooting Tips 230

Contents

xi



Acknowledgments 231

References 231

CHAPTER 13Porous Silicon Particles for Multistage Delivery 237


13.2 Fabrication of PSPs 245

13.2.1 Materials 245

13.2.2 Methods 247

13.2.3 Characterization 251

13.3 Oxidation and Surface Modification with APTES of PSPs 252

13.3.1 Reagents 252

13.3.2 Methods 252

13.4 Fluorescent Dye Conjugation of PSPs 254

13.4.1 Reagents 254

13.4.2 Methodology 254

13.5 Zeta Potential Measurement 254

13.5.1 Equipment 254

13.5.2 Reagents 254

13.5.3 Methodology 254

13.5.4 Results 255

13.6 Count and Size Analysis of PSPs 255


13.6.2 Methods 255

13.6.3 Data Acquisition, Anticipated Results, and Interpretation 256

13.7 Using Inductively Coupled Plasma–Atomic Emission Spectroscopy13.7 (ICP-AES) to Determine the Amount of Degraded Silicon in Solution 257


13.7.2 Methods 258


13.8 Flow Cytometry to Characterize PSP Shape, Size, and13.8 Fluorescence Intensity 260


13.8.2 Methods 262


13.9 Loading and Release of Second-Stage NPs from PSPs 264

13.9.1 Loading of NP into PSPs 264

13.9.2 Release of NPs from PSPs 265



Acknowledgments 271

Contents

xii

References 271

CHAPTER 14Mathematical Modeling of Nanoparticle Targeting 275


14.2 Molecular/Cellular Scale 277

14.2.1 Methods 277


14.2.3 Discussion and Commentary 280

14.3 Tissue Scale 282

14.3.1 Methods 282



14.4 Organism Scale 285

14.4.1 Methods 285



14.5 Model Validation and Application 287

14.5.1 Statistical Guidelines 287


Acknowledgments 290

References 290

CHAPTER 15Techniques for the Characterization of Nanoparticle-Bioconjugates 293


15.2 Methods 296

15.2.1 Separation-Based Techniques 296

15.2.2 Scattering Techniques 300

15.2.3 Microscopy 308

15.2.4 Spectroscopic 312

15.2.5 Mass Spectroscopy 317

15.2.6 Thermal Techniques 318


Acknowledgments 320

References 321

About the Editors 333

List of Contributors 334

Index 337

Contents

xiii

Preface

As a research field, nanotechnology is already spinning off numerous stand-alonesubdisciplines including bionanotechnology, nanomedicine, nanophotonics, nano-plasmonics, and nanotoxicology. Concomitant with this, the materials, especially thenanoparticles, utilized in these fields are steadily moving into the mainstream andbecoming known to researchers pursuing other endeavors including most particularlythe myriad areas of biological research. For example, biologists who commonly utilizefluorescent or molecular imaging techniques have heard of quantum dots and are mostlikely curious if these nanocrystalline fluorophores can further enhance their capabili-ties. Alternatively, many in the pharmaceutical industry are excited by the potentialbenefits offered by nanoparticle-mediated drug delivery which may help improvedrug-targeting and potentially mitigate systemic toxicity issues. Although there aremany more examples, the common thread among all the researchers is the need for asource of methods to synthesize, characterize, biofunctionalize, and apply the nano-material that is most suitable to tackle the problem at hand. They may wonder how hardit would be to make and characterize a particular nanoparticle or attach a biomolecule toa nanoparticle. How will they know if the materials they have prepared have the proper-ties they would like? This method-based focus of this book serves to fill this critical gap.

Following the cross-disciplinary nature of nanotechnology itself, the contributors ofeach of the chapters found in this book are drawn from among many different fieldsincluding materials science, chemistry, chemical engineering, molecular biology, phys-ics, imaging, and medicine to name but a few. They represent the best scientists andengineers in their respective fields and have been drawn together in this book to providebiomedical scientists and others with the tools and methods they need to pursue the fur-ther biological applications of nanoparticles.

This book describes many of the methods needed to synthesize, biofunctionalizeand apply nanoparticles at bimolecular, cellular, and tissue/organism scales. Chapters 1through 4 describe the interface between nanoparticles including quantum dots andcarbon nanotubes with biomolecules such as peptides and proteins for biosensing andbiocatalytic applications. Chapters 5 through 8 describe the use of nanoparticles andnanoassemblies for cellular applications including intracellular trafficking, engineeringcell fates, tissue engneering, and cell separations. Chapters 9 through 14 focus on theemerging field of nanomedicine and focus on the use of magnetic, polymeric, metal,and multifunctional nanoparticles as potential therapeutics and imaging agents for dev-astating diseases including cancer and atherosclerosis. Chapter 15 focuses on the model-ing of interactions between nanoparticles and cells and tissues. We have also asked agroup at the US FDA to put together a comprehensive review of the available methods

xv

for characterizing nanoparticle-bioconjugates for inclusion in this book (Chapter 16).The pressing need for the method described in this book is intended to be of use to allwho already use or are planning to use nanoparticles in their respective applications.

We hope that well-worn copies of this book will find a place in your laboratory.

Kaushal Rege and Igor L. Medintz

Preface

xvi

C H A P T E R

1Preparation and Characterization of CarbonNanotube-Protein Conjugates

Jonathan S. Dordick,* Dhiral A. Shah, Ravindra C. Pangule, Shyam Sundhar Bale,Prashanth Asuri, Amit Joshi, Akhilesh Banerjee, David Vance, and Ravi S. Kane*

Department of Chemical and Biological Engineering, Rensselaer Polytechnic Institute, Troy, NY

*Corresponding Authors: Prof. Ravi S. Kane, Department of Chemical and Biological Engineering,Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY 12180, Phone: 518-276-2536, Fax:518-276-4030, e-mail: [email protected]; Prof. Jonathan S. Dordick, Department of Chemical and BiologicalEngineering, Rensselaer Polytechnic Institute, 110 8th Street, Troy, NY 12180, Phone: 518-276-2899, Fax:518-276-2207, e-mail: [email protected]

1

Abstract

This chapter describes methods of immobilizing proteins on carbonnanotubes, using two different routes—physical adsorption and covalentattachment. We also provide an overview on how such conjugates can be char-acterized with the help of various techniques, such as Raman, Fourier trans-form infrared (FT-IR), circular dichroism (CD), and fluorescence spectroscopies,in addition to the standard enzyme kinetic analyses of activity and stability.Both the attachment routes—covalent and noncovalent—could be used to pre-pare protein conjugates that retained a significant fraction of their native struc-ture and function; furthermore, the protein conjugates were operationallystable, reusable, and functional even under harsh denaturing conditions. Thesestudies therefore corroborate the use of these immobilization methods to engi-neer functional carbon nanotube-protein hybrids that are highly active andstable.

Key terms enzyme immobilizationcarbon nanotubesphysical adsorptioncovalent attachmentnanotube solubilization

1.1 Introduction

Nanomaterials, such as carbon nanotubes (CNTs) offer a unique combination of electri-cal, mechanical, thermal, and optical properties [1] that make them promising materialsfor various applications ranging from sensing [2] and diagnostics to biotransformationsand the cellular delivery of peptides and proteins [3, 4]. For instance, Barone et al. [5]have developed carbon nanotube-glucose oxidase conjugates that can act as glucose sen-sors. Recently, Dai and coworkers [2] demonstrated the recognition of monoclonal anti-bodies by a recombinant human antigen immobilized onto carbon nanotubes. Carbonnanotubes have also been used for both biomolecule delivery and targeted therapy.Pantarotto et al. [3] demonstrated that carbon nanotubes functionalized with peptidescan penetrate cell membranes of human and murine fibroblasts, and serve as carriers forbiomolecule delivery. Dai and coworkers [4] observed internalization of nanotube-pro-tein conjugates in nonadherent human cancer cells as well as adherent cell lines. Kam etal. [6] demonstrated that functionalized CNTs could be used to selectively target cancercells and destroy them by irradiating CNTs with near-infrared (NIR) light. These studiesrepresent a fraction of the exciting opportunities at the interface of nanotechnologyand biotechnology. It is, however, important to interface carbon nanotubes withbiomolecules, such as proteins, to realize some of these applications. As a result, variousmethods of functionalization have been developed recently to functionalize CNTs withproteins. In this chapter, we describe three methods of preparing carbon nanotube-pro-tein conjugates, each of them possessing distinct structural, mechanical, and functionalcharacteristics.

Noncovalent attachment is probably the simplest technique for attaching proteinsonto carbon nanotubes. The adsorption of proteins onto CNTs is hypothesized to be aresult of the attractive hydrophobic interactions between carbon nanotubes and pro-teins [7]. This method has been found to preserve a significant fraction of the nativestructural and functional properties of several proteins as well as the physicochemicalproperties of nanotubes [2, 8–10]. The resulting formulations prevail in the form ofaggregates, which can be easily separated from other solution components. However,the limited solubility of these conjugates in water limits their attractiveness for manyapplications in biotechnology [11, 12]. Nevertheless, such conjugates have been used forbiosensing, diagnostics and preparing antifouling nanocomposites films [13].

To overcome the aforementioned limitation of water solubility, Karajanagi et al.have described a simple method that uses proteins to solubilize single-walled carbonnanotubes (SWNTs) in water [14]. Efficient solubilization of SWNTs has previously beenachieved using surfactants [15, 16], polymers [17, 18], single stranded DNA [19], pep-tides [20], and polysaccharides [12, 21]. The direct solubilization of SWNTs using a vari-ety of proteins differing in size and structure is a simple and scalable alternative thatenables the generation of individual nanotube solutions. Moreover, proteins are rich instructure and function and have numerous reactive groups, such as hydroxyls, amines,thiols, carboxylic acids, and others, which can be used as orthogonal reactive handlesfor further functionalization of SWNTs.

Finally, Asuri et al. have developed an alternative method of preparing water-solubleconjugates of carbon nanotubes with a broad range of proteins [22]. CNTs can be acidoxidized to produce hydrophilic carboxylic acid and hydroxyl groups along their side-walls [23, 24], thereby leading to water solubility. Proteins can then be covalentlyattached to oxidized water-soluble CNTs using carbodiimide activation of the carboxylic

Preparation and Characterization of Carbon Nanotube-Protein Conjugates

2

acid groups. These water-soluble conjugates not only display low diffusional resistance[25] and high activity with stable protein attachment [26], but also have added advan-tages of high stability and reusability, thereby overcoming the traditional limitations ofwater-soluble proteins. Though the covalent immobilization of proteins onto CNTsleads to stable protein attachment, the chemical modification of the CNTs surface maycompromise the desirable electronic properties of CNTs. Such water-soluble CNT-pro-tein conjugates may find application in fields other than biosensing, for example,biotransformations, biomaterials, medicine, and self-assembled materials.

It is, therefore, clear that many methods have been explored to prepare functionalnanotube-protein conjugates. Each of these methods possesses its own unique set ofadvantages and disadvantages, and the best choice of the method depends on thedesired end application of the hybrid conjugates.

1.2 Materials

Raw and purified HIPCO single-walled carbon nanotubes (SWNTs) (1–1.5 nm diameter,ca. 10 μm length, <35 wt% ash content) were purchased from Unidym (Houston, TX).Multiwalled carbon nanotubes (MWNTs) (10–20 nm diameter, 5–20 µm length, 95wt% purity) were purchased from Nanolab, Inc. (Newton, MA). Enzymes—soybeanperoxidase (SBP), horseradish peroxidase (HRP) and Mucor javanicus lipase (MJL)—werepurchased from Sigma-Aldrich (St. Louis, MO) as salt-free, dry powders. Bicinchoninicacid (BCA) assay kit for determining solution phase protein concentrations waspurchased from Pierce Biotechnology, Inc. (Rockford, IL). Guanidine hydrochloride(GdnHCl), sodium dodecylbenzene sulfonate (NaDDBS), and all other chemicals wereobtained from Sigma. Bovine serum albumin (BSA) was purchased from Fisher ScientificInternational, Inc. (Hampton, NH). All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO). All enzymes and chemicals were used as received without anyfurther purification.

1.3 Methods

1.3.1 Physical Adsorption of Proteins on Carbon Nanotubes

Attachment of proteins to carbon nanotubes via physical adsorption represents a facilemethod of preparing nanotube-protein conjugates, wherein an aqueous dispersion ofSWNTs is mixed with a protein solution to achieve adsorption. The unadsorbed proteinis washed off and nanotube-protein conjugates are then resuspended in aqueoussolution. The detailed procedure is described below:

1. Sonicate a fixed amount of raw SWNTs in dimethylformamide (DMF) at aconcentration of 1 mg/mL for 30 minutes using a bath sonicator (Model 50T, VWRInternational, West Chester, PA) with rated power of 45W to obtain a uniformdispersion in solution.

2. Dispense 1 mL of the resulting SWNT dispersion into an Eppendorf microcentrifugetube, and centrifuge the solution at 8,000 rpm for 1 minute. Remove the supernatantand resuspend the settled SWNTs in an aqueous buffer (50 mM phosphate buffer, pH

1.2 Materials

3

7.0) by vortexing the solution. Repeat this wash procedure at least five more times toremove any residual organic solvent. This gradual change from organic phase to anaqueous phase renders unfunctionalized SWNTs more dispersed in buffer. Finally,disperse 1 mg SWNTs in 500 μL aqueous buffer (15 minutes).

NOTE: The selection of buffer for preparing conjugates is dependent upon the choiceof protein and the retention of its native function in that buffer.

3. Prepare a fresh solution of protein in the aqueous buffer. Add the aqueous dispersionof SWNTs (500 μL, 2 mg/mL) to the protein solution (500 μL), and shake the mixtureon a platform shaker for 2 hours at 200 rpm and room temperature (2 hours 15minutes).

NOTE: In the case of thermally unstable proteins or enzymes undergoing autolysis,such as trypsin, shaking should be carried out at 4°C to prevent deactivation duringincubation.

4. After incubation, centrifuge the SWNT dispersion at about 8,000 rpm for 1 minute tosettle the SWNT-protein conjugates. Carefully decant the supernatant (ca. 800 μL)without loss of any conjugates. Typically, perform six such washes, with fresh bufferadded each time to remove unbound protein (20 minutes).

NOTE: While collecting the supernatant, tilt the microcentrifuge tube and gentlypipette out approximately 800 μL supernatant from close to the tube walls without dis-turbing the settled SWNTs, so that the supernatant does not contain SWNTs, whichcan interfere with BCA assay for protein content determination in supernatants. Aswinging bucket microcentrifuge is ideal for this step, as the SWNTs settle at the bot-tom and not on the walls of the tube after centrifugation. Do not resuspend theSWNT-protein conjugates by vortexing, as this can lead to desorption of the proteinfrom SWNT surface. Resuspend the conjugates by gently inverting and tapping themicrocentrifuge tube containing the conjugates.

5. Analyze the supernatants for protein content using the BCA assay (for proteinconcentration range of 20–2,000 μg/mL) or the micro-BCA assay (for proteinconcentration range of 0.5–20 μg/mL). Determine the amount of protein attachedonto the SWNTs by mass balance. Use the SWNT-protein dispersion for furtheranalysis and characterization. (1 hour 15 minutes) The process flowchart is shown inFigure 1.1.

The total time to carry out the procedure is approximately 5 hours.

1.3.2 Protein Assisted Solubilization of Carbon Nanotubes

For preparation of protein solubilized carbon nanotubes, an aqueous dispersion ofnanotubes is dispensed in a concentrated protein solution and exposed toultrasonication for a predetermined time period. The supernatant, collected aftersequential steps of ultracentrifugation of the CNT-protein dispersion, containsolubilized carbon nanotubes that are stable at room temperature and show no signs ofaggregation. The detailed procedure is described below:


4

1. Disperse purified SWNTs in DMF at a concentration of 1 mg/mL by sonication,and replace the organic phase gradually with an aqueous phase through repeatedwashing with milliQ water (as stated in section 1.3.1) (45 minutes).

2. Disperse 200 μg of SWNTs in 4-mL protein solution (10 mg/mL) and sonicatethe dispersion of SWNTs for 2 hours using a bath sonicator (Model 50T, VWRInternational, West Chester, PA) with rated power of 45W (2 hours).

3. Ultracentrifuge the dispersed solution at 123,000g for 30 minutes.

4. Carefully collect 60% of the supernatant and ultracentrifuge at 185,000g for 30minutes.

5. Collect 75% of the supernatant that contains protein adsorbed SWNTs. Use thissolution for further analysis and characterization.


NOTE: The sonication efficiency and, hence, the quality of the dispersion varies withthe volume of the solution sonicated. For best dispersions, use 4 mL volume forsonication.

1.3.3 Covalent Attachment of Proteins onto Carbon Nanotubes

For covalently attaching proteins onto carbon nanotubes, the carbon nanotubes are firstfunctionalized with carboxylic acid groups by acid treatment. The carboxylic acidgroups are then “activated” to form succinimide esters using carbodiimide chemistry[23]. These activated carboxylic groups react with amine groups on proteins enablingthe covalent attachment of proteins onto carbon nanotubes. A detailed description ofthe procedure is given below:

1.3 Methods

5

Calculate the proteinloading on SWNTs by

mass balance

Shake the mixture on aplatform shaker for 2

hours at 200 rpm

Sonicate a 1 mg/mLdispersion of SWNTs

in DMF for 30 min

Mix the SWNTssolution with a freshly

prepared proteinsolution

Centrifuge the SWNT -protein mixture at 8000

rpm for 2 min

Remove thesupernatant and

resuspend the SWNTsin aqueous buffer

Centrifuge 1 ml ofSWNTs dispersion at8000 rpm for 2 min

Disperse the SWNTs in500 µL aqueous buffer

Collect the supernatantand resuspend the

SWNT -proteinconjugates in buffer

Repeat the steps 5 times to removeresidual organic solvent

Repeat the steps 6 times to removeunadsorbed enzyme from solution

Calculate the proteinloading on SWNTs by

mass balance

Shake the mixture on aplatform shaker for 2

hours at 200 rpm

Sonicate a 1 mg/mLdispersion of SWNTsin DMF for 30 min

Mix the SWNTssolution with a freshly

prepared proteinsolution

Centrifuge the SWNT-protein mixture at 8000

rpm for 2 min

Remove thesupernatant and

resuspend the SWNTsin aqueous buffer

Centrifuge 1 ml ofSWNTs dispersion at8000 rpm for 2 min

Disperse the SWNTs in500 L aqueous bufferμ

Collect the supernatantand resuspend the

SWNT-proteinconjugates in buffer

Repeat the steps 5 times to removeresidual organic solvent

Repeat the steps 6 times to removeunadsorbed enzyme from solution

Figure 1.1 Process flowchart for the physical adsorption of proteins onto SWNTs.

1. Sonicate a fixed amount of multiwalled carbon nanotubes (MWNTs) in a mixtureof concentrated sulfuric acid and nitric acid (3:1, v/v; 400 mL/100 mg carbonnanotubes) using a bath sonicator with a rated power of 45W for 3 hours.Periodically replace the contents of the bath with ice cold water to ensure that theMWNT suspension does not get heated up during sonication (3 hours).

NOTE: Acid oxidation not only leads to the functionalization of MWNTs withcarboxylic acid groups, but also causes cutting of MWNTs. Longer sonication timesresult in finer oxidized MWNTs.

2. Add the nanotube-containing acid solution (400 mL) to an ice-cold solution ofmilliQ water (3,600 mL) gradually with constant swirling. Allow 10 to 15 minutes fordissipation of the heat generated on diluting the acid mixture.

3. Filter the solution through a 0.22-μm polycarbonate filter membrane (Isoporemembrane, Millipore) in batches of approximately 200 mL to remove the acid. Aftereach filtration, disperse the nanotube film or bucky paper (a mass of carbonnanotubes tangled with each other to form a film or mat) thus formed in 50-mLmilliQ water by ultrasonication in the bath sonicator for approximately 10 minutes,until the nanotubes are dispersed entirely in solution. Dilute this suspension with150-mL milliQ water (40 minutes).

4. Repeat the ultrasonication/filtration step at least three times until water-solubleMWNTs are obtained and the pH of the filtrate becomes neutral (2 hours).

NOTE: These oxidized and “cut” nanotubes can be stored in aqueous solution(1 mg/mL) at room temperature.

5. After the final filtration, disperse the oxidized nanotubes (2 mg/mL) in MES (2-(N-Morpholino) ethanesulfonic acid) buffer (50 mM, pH 6.2), and add an equalvolume of 400-mM N-hydroxysuccinimide (NHS) in MES buffer (5 minutes).

6. Sonicate the mixture for 30 minutes in a bath sonicator.

7. Add N-ethyl-N’-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) (20mM in MES buffer) to the nanotube solution to initiate the coupling of NHS to thecarboxylic groups on the oxidized nanotubes, and stir the mixture at 400 rpm for 30minutes at room temperature.

8. Filter the activated nanotube solution through a polycarbonate filter membrane(0.22 μm) and rinse thoroughly with MES buffer to remove excess EDC and NHS (20minutes).

9. Transfer the nanotube film immediately into a freshly prepared protein solution (2mg/mL, 10 mM phosphate buffer, pH 8.0), and sonicate for few seconds to dispersethe nanotubes in solution.

NOTE: Do not allow the nanotube film to dry out completely on the filter membraneas it may lead to hydrolysis of the active ester and hence decreased attachment of theprotein.

10. Shake the mixture on an orbital shaker at 200 rpm for 3 hours and at roomtemperature to allow the attachment of proteins to the nanotubes.


6

NOTE: Proteins such as proteases or thermally unstable proteins require that this stepbe carried out at 4°C to prevent protein deactivation.

11. Filter the nanotube-protein suspension and wash it three times with milliQ water(5-mL/mg nanotubes) and once with 1% Tween-20 (5-mL/mg nanotube) to removeany nonspecifically bound protein (2 hours).

12. Allow flocculates of nanotube-protein conjugates, if any, to settle overnight and usethe supernatant for further experiments.

13. Quantify the amount of immobilized protein by elemental analysis of the oxidizednanotubes and the nanotube-protein conjugates. The schematic of the process ofprotein functionalization on nanotubes is shown in Figure 1.2.


1.4 Data Acquisition, Anticipated Results, andInterpretation of Data

We have employed various techniques, such as Raman, FT-IR, CD, and fluorescencespectroscopies in addition to the standard enzyme kinetic analyses of activity and stabil-ity, to understand how the attachment onto CNTs influences protein structure andfunction. The choice of technique depends on the method used for protein attach-ment and the resulting characteristics of the formulation (e.g., protein loading anddispersibility). In this section, we discuss these characterization techniques and includeour own data of experiments to enable the reader to evaluate the CNT-protein conju-gates prepared by the previously described methods of protein attachment onto carbonnanotubes.

1.4.1 Characterization of Proteins Physically Adsorbed onto CarbonNanotubes

We have used enzymes as probes of protein structure and function. To measure theretention of enzyme activity upon attachment, it is necessary to quantify the amount ofenzyme physically adsorbed onto carbon nanotubes. Measuring enzyme activity anddetecting the change in its secondary structure by FT-IR spectroscopy before and after

1.4 Data Acquisition, Anticipated Results, and Interpretation of Data

7

H SO : HNO2 4 3

3:1, 3h

CNT

(a) (b)

EDC/NHS

pH 6.2

Protein

NH2

CNT-Protein

Figure 1.2 CNT-protein composites. (a) Schematic of protein functionalization of carbon nanotubes.(b) Photograph of water-soluble MWNTs. (Adapted from Asuri et al. [22].)

adsorption is useful for studying the influence of the hydrophobic nanoscale environ-ment of carbon nanotubes on protein structure and function. Also, enzyme activitymeasurement can be used to determine the stability of enzymes adsorbed ontonanotubes under harsh conditions.

1.4.1.1 Measurement of Loading of Proteins on Carbon Nanotubes by the BCA Assay

The Pierce BCA Protein Assay uses a detergent-compatible formulation based onbicinchoninic acid (BCA) for the colorimetric detection and quantification of total pro-tein. It involves the reduction of Cu2+ to Cu1+ ions by proteins to form a water solublecomplex with BCA that strongly absorbs at 562 nm. Using this assay, the loading of pro-teins on carbon nanotubes is calculated as follows:

Amount of protein loaded per mg SWNT = ⋅ − ⋅∑C V C Vi i j jn

(1.1)

where,

Ci = Initial concentration of protein before exposing it to SWNTs;

Vi = Initial volume of protein solution added to SWNT dispersion;

Cj = Concentration of supernatant in jth wash;

Vj = Volume of supernatant in jth wash;

n = Number of washes performed.

Representative data for the loading of SBP on SWNTs is shown in Figure 1.3(a). Theadsorption of SBP followed a pseudosaturation behavior, with a maximum loading of575 μg SBP/mg SWNTs (Figure 1.3(a)). We observed that adsorbed SBP has a strong affin-ity for the SWNTs, with almost complete adsorption observed within the first minute(data not shown). Protein adsorption was irreversible at lower loadings. For example, at aloading of 250-μg protein/mg SWNT, essentially no protein desorption was observed(Figure 1.3(b)). The AFM images of SWNT-SBP conjugates are shown in Figure 1.3(c) and(d). The globular structures seen on the wire like SWNTs represent SBP molecules. Linescans reveal that a region on the SWNTs that does not contain SBP has a height of 3.7nm, while a region containing SBP has a height of 9.6 nm, the difference (5.9 nm) beingthe height of adsorbed SBP molecules.

1.4.1.2 Retention of Protein Activity Upon Physical Adsorption

Adsorption onto CNTs can influence the structure, function, and stability of proteins.Since the catalytic activity of proteins relies on the retention of their native structure,measurement of catalytic activity can be used to evaluate the influence of the nanoscaleenvironment of a CNT on protein properties. Using enzymes as highly sensitive probesof protein function, we studied the strong influence of the CNT surface on proteinfunction and stability in harsh environments.


8

Determination of Protein Activity Upon Physical AdsorptionThe structure, function, and spatial orientation of proteins attached onto carbonnanotubes strongly depends on the interactions of the nanotube surface with proteins.Since the catalytic activity and exquisite selectivity of proteins requires the near com-plete retention of native structure, measurement of enzyme activity can be used to eval-uate the influence of the hydrophobic nanoscale environment of nanotubes on enzymestructure and function. To that end, comparison of the activity of native and immobi-lized enzyme can provide insight into the influence of carbon nanotubes on the reten-tion or loss of native-like enzyme properties.

As an example, the activity of native SBP was measured using p-cresol as the substrate[27]. SBP catalyzes the oxidation of p-cresol in the presence of H2O2 to form fluorescentoligo- and polyphenol products. The initial reaction rates were measured by tracking theincrease in fluorescence of the reaction mixture at excitation and emission wavelengthsof 325 nm and 402 nm, respectively, using an HTS 7000 Plus Bio Assay Reader(Perkin-Elmer, Wellesley, MA). For a typical solution-phase assay, 0.15-µg/mL SBP wasused with 20-mM p-cresol and 0.125-mM H2O2 in a volume of 200 μL. To measure the


9

(a) (b)

(d)(c)

1000

800

600

400

200

00 500

Amount of SBP exposed to SWNTs( g SBP/mg SWNTs)μ

1000 1500 2000 250

300

250

200

150

100

50

00 1 2 3 4 5 6 7

Number of washes

50100

150200

250 nm

0 100 200 300 400nm

0 100 200 300 400nm

nmnm

Figure 1.3 Loading of SBP on SWNTs: (a) Protein loading as a function of amount of SBP exposed toSWNTs. (b) Protein loading as a function of washing with fresh buffer. (c) AFM images of SBP adsorbedonto SWNTs. (d) Surface plot of height image for SBP adsorbed onto SWNTs revealing SBP moleculeson the SWNTs. (Reprinted with permission from Karajanagi et al. . Copyright (2004) American Chemi-cal Society [7].)

activity of SBP adsorbed onto SWNTs, a well-mixed dispersion of SWNT-SBP at a concen-tration of 1.0 mg/mL was prepared in aqueous buffer. For a typical experiment, 0.2–2.5μg of SWNTs were used on the basis of the loading of SBP. It was found that SBP retainedsignificant specific activity at all loadings (Figure 1.4) ranging from 18 to 280 μg SBP/ mgSWNTs. The specific activity of SBP was strongly dependent on the loading; up to 28% ofnative solution activity was obtained at 50% of maximal surface coverage, and this valuedropped to ca. 10% at 3% of maximal surface coverage (Figure 1.4). The increase in spe-cific activity of adsorbed SBP with an increase in the surface coverage on SWNTs may bedue to a higher retention of native structure at higher surface loadings.

Protein Stability under Harsh ConditionsPhysical adsorption of proteins to carbon nanotubes enhances the stability of proteinsin strongly denaturing environments where native proteins undergo substantial deacti-vation. To determine protein stability at elevated temperatures (Figure 1.5(a)), thenanotube-protein conjugates were subjected to these temperature conditions for differ-ent periods of time and cooled in an ice bath. Initial enzymatic reaction rates were thendetermined at room temperature. To determine rate constants in approximately 100%methanol (Figure 1.5(b)), the initial rates were measured in methanol as a functionof incubation time. The deactivation constant was determined from the slope of astraight-line fit through the plot of loge (% activity retained) versus time.

1.4.1.3 Determination of Protein Secondary Structure Using Fourier TransformInfrared (FT-IR) Spectroscopy

FT-IR spectroscopy is an established tool for the structural characterization of proteins[29]. The secondary structure of a protein can be quantitatively determined from a spec-trum by considering the amide I region, between 1,600 and 1,700 cm-1. This region,which consists mainly of the C-O stretching vibration of the backbone peptide bonds inproteins, was used to obtain the α helix and β sheet contents of the protein [30, 31]. We


10

Fraction of maximal coverage ofenzymes on SWNTs

0.0 0.1 0.2 0.3 0.4 0.5 0.60

5

10

15

20

25

30

35

Perc

ent

nativ

esp

ecifi

cac

tivity

reta

ined

Figure 1.4 Enzymatic activity retained as a function of the surface coverage of SBP adsorbed onSWNTs (▲). (Reprinted with permission from Karajanagi et al. [7]. Copyright (2004) AmericanChemical Society.)

used FT-IR spectroscopy to compare the secondary structure of proteins before and aftertheir adsorption onto carbon nanotubes. The differences in secondary structure betweenthe soluble and adsorbed states are represented by the simple sum of magnitudes ofchanges in α helix and β sheet contents. For example, SBP showed a total change in α

helical and β sheet content of 13% (Table 1.1), which suggests that SBP retains much ofits native structure and activity upon absorption onto SWNTs.

1.4.2 Characterization of Protein-Solubilized Carbon Nanotubes

The aggregation state of the protein solubilized carbon nanotube dispersions can becharacterized by ultraviolet-visible (UV-Vis) and Raman spectroscopy. These methodscan be used effectively to distinguish between solubilized and nonsolubilized carbonnanotubes. We describe the use of these two techniques to characterize the solubilizedCNTs.

1.4.2.1 Characterization of Carbon Nanotube Dispersions Using UV-Vis Spectroscopy

The UV-Vis absorption spectra of dispersions of SWNTs are known to be sensitive totheir aggregation state [15]. The UV-Vis spectrum for SWNTs in water in the absence of adispersing agent was essentially featureless, which indicates the presence of aggregatesof SWNTs (data not shown). In contrast, the UV-Vis spectra for solutions of SWNTs


11

Table 1.1 Secondary Structure Percentages of SBP in Solution andAbsorbed onto SWNTs, as Determined by FT-IR Spectroscopy Calculatedfrom the Amide I Spectra

Sample % α Helix % β Sheet

Native solution of SBP 36.1 ± 1.2 25.1 ± 2.5SBP absorbed onto SWNTs 27.9 ± 4.1 20.6 ± 6.9

(Adapted from Karajanagi et al. [7])

0 501.0

10.0

100.0

Perc

ent

activ

ityre

tain

ed

100 150 200 250

Min

0 501.0

10.0

100.0

Perc

ent

activ

ityre

tain

ed

100 150 200 250

Min

(a) (b)

Figure 1.5 Time-dependent deactivation of native SBP (�) and SBP on SWNTs (�) (a) at 95°C and (b)in 100% methanol. The activities are normalized relative to the initial activity (activity at t = 0 min).Figure 1.5(b) does not contain % activity data for native SBP as it shows no activity in 100% methanol.(Reprinted with permission from Asuri et al. [28]. Copyright (2006) American Chemical Society.)

obtained using the proteins BSA and MJL exhibited sharp and well-resolved peaks(Figure 1.6). These sharp van Hove peaks are a characteristic of aqueous solutionscontaining debundled, individually dispersed SWNTs. The UV-Vis spectrum for SWNTsdispersed in water using NaDDBS also shows similar sharp features (Figure 1.6). Wenote that the spectra for SWNT-BSA and SWNT-MJL show peaks in the region beyond900 nm that are red-shifted by approximately 10 to 15 nm with respect to those forSWNT-NaDDBS. This shift may be attributed to the greater accessibility of water to theSWNT surface for SWNT-BSA and SWNT-MJL than for SWNT-NaDDBS. It has beenshown [32] that proteins can form a more porous layer on the SWNT surface than surfac-tants, thereby permitting water and other small molecules to associate with the surface.

1.4.2.2 Raman Spectroscopy to Probe Aggregation State of SWNTs

Raman spectroscopy is a versatile tool, which enables us to probe the aggregation state ofSWNTs in solutions. In the case of protein-solubilized SWNT dispersions, the radialbreathing mode (150–350 cm–1) and tangential mode observations can be used asindicators of the quality of nanotube dispersion. To obtain the Raman spectra of thesolubilized SWNT-BSA conjugates, 10 μg of the conjugates were placed on a cleaned sili-con substrate and samples were analyzed using a laser excitation at 785 nm at a power of10 mW, with a 50x lens. (Spectra were recorded from 0–3000 cm–1 for 4 minutes). Thewavenumber calibration was carried out using the 521-cm–1 line of silicon substrate as areference. The relative intensities of Raman peaks in the region between 230 and 270cm–1 were found to be good indicators of the nanotube dispersion because of the changein Raman spectrum depending on their dispersed state. Specifically, in the aggregatedstate, (10,2) nanotubes are in resonance and (10,5) nanotubes are off resonance, whilewhen SWNTs are dispersed, (10,5) nanotubes are in resonance and (10,2) nanotubes areoff resonance. Accordingly, we see a peak at 267 cm–1 for SWNT aggregates, whereassolubilized SWNT-BSA conjugates show no peak at 267 cm-1 but a prominent peak at 234cm–1 (Figure 1.7(a)). Furthermore, the peak corresponding to the tangential mode (cen-tered at 1,591 cm–1) for soluble SWNT-BSA was narrower than that for the aggregated


12

400 600

Wavelength (nm)

Nor

mal

ized

abso

rban

ce

800 10000.0

0.2

0.4

0.6

0.8

1.0

Figure 1.6 UV-Vis absorption spectra of SWNTs dispersed in water using NaDDBS (dashed line), BSA(solid line), and MJL (dash-dot line) normalized at 410 nm. (Reprinted with permission fromKarajanagi et al. [14]. Copyright 2006 American Chemical Society.)

SWNTs, with a decrease in the full width at a half-maximum of approximately 5 cm–1

(Figure 1.7(b)), which is in agreement with similar observations for solutions containingindividually dispersed SWNTs [20, 33, 34].

1.4.3 Characterization of Covalently Attached Carbon Nanotube-ProteinConjugates

We determined the retention of protein structure and function upon covalent attach-ment to carbon nanotubes using Hammett analysis of protein activity as well as spectro-scopic techniques, such as CD and fluorescence spectroscopies. Structural analysis byCD or fluorescence spectroscopy is not possible for conjugates prepared by physicaladsorption of proteins onto bundles of nanotubes or for covalent MWNT-protein conju-gates because of interference from the carbon nanotubes. On the other hand, because ofthe higher solubility and higher protein loading obtained in case of covalent attachmentof proteins to oxidized SWNTs, CD, and fluorescence measurement-based structuralstudies are possible. The activity measurements of the nanotube-protein conjugatesindicated that the conjugates demonstrated not only enhanced stability in harshconditions, but also operational and storage stability.

1.4.3.1 Hammett Analysis for Protein Structure-Activity Relationship

It is often important to study the structural perturbations of the protein to further probethe effects of immobilization. However, analyses of proteins on MWNTs, such as CD andFT-IR spectroscopy, are hindered by the strong absorbance and intrinsic fluorescence ofnanotubes. Hammett analysis, on the other hand, is a well-established kinetic techniqueto probe an enzyme’s transition state structure [27]. In the case of SBP catalysis, theHammett coefficient ρ provides a measure of the sensitivity of SBP’s catalytic efficiencyto the electronic nature of substituents on phenolic substrates (electron-donating orelectron-withdrawing), as reflected in the values of their substituent electronic parame-ter σ. Positive values of σ represent electron withdrawal by the substituent from the aro-matic ring, whereas negative values indicate electron release to the ring. Deviation in ρ

values for SBP bound to a support from that for the native enzyme in aqueous buffer


13

0.0160 180 200 220 240 260 280 300

0.2

0.4

0.6

0.8

1.0

1.2

Wavenumber (cm )−1

−1 −1

(a) (b)

234 cm 267 cm

0.0

1560 1580

Nor

mal

ized

inte

nsity

Nor

mal

ized

inte

nsity

1600 1620

0.2

0.4

0.6

0.8

1.0

1.2

Wavenumber (cm )−1

−11591 cm

Figure 1.7 Raman spectroscopic analysis of SWNT-BSA conjugates (solid line) and SWNTs (dashedline) in (a) Radial breathing mode, (b) Tangential mode at 785-nm excitation. (Reprinted with permis-sion from Karajanagi et al. [14]. Copyright 2006 American Chemical Society.)

would indicate that the active site structure of the enzyme is perturbed by adsorptiononto the support.

logVKM

max constant⎛⎝⎜

⎞⎠⎟ = ⋅ +σ ρ (1.2)

To that end, we determined the Hammett coefficient, ρ, based on a modified form ofthe Hammett equation for SBP (1.2), using a series of phenolic substrates, p-OC2H5,p-CH3, p-CH2OH, and p-Cl with different values of the electronic parameter (σ) varyingfrom -0.24 to +0.23 [27].1 The standard kinetic parameters—maximum reaction rate(Vmax) and Michaelis constant (KM)—were determined for the different substrates usingnonlinear Michaelis-Menten fits. Figure 1.8 depicts the Hammett analysis for native SBPand MWNT-SBP in aqueous buffer. Interestingly, the Hammett coefficients for native

and immobilized SBP are essentially identical. The comparable values of ρ indicate thatthe differences in the active site structure for native and immobilized SBP are minimal;therefore, the mechanism of catalysis for MWNT-SBP is similar to that for native SBP.Thus, the high retention of catalytic activity for the MWNT-SBP conjugates is consistentwith the enzyme retaining its intrinsic active site structure throughout the attachmentprocess.

1.4.3.2 Determination of Protein Secondary Structure Using Circular Dichroism (CD)Spectroscopy

CD spectroscopy is used for studying the conformational stability of a proteinunder harsh conditions—thermal stability, pH stability, and stability against chemicaldenaturants. CD measures the difference in absorbance of a sample between left-hand polarized light and right-hand polarized light; these differences arise because of


14

−0.3 −0.2−6.2

−6.0

−5.8

−5.6

−5.4

−5.2

−5.0

−4.8

−4.6

−4.4

log

(V/K

)m

axM

−0.1 0.0 0.1 0.2 0.3σ

Figure 1.8 Influence of the substituent electronic parameter, σ, on the catalytic efficiency of nativeSBP (�) and MWNT-SBP conjugates (�) in aqueous buffer. Slope of the lines gives the Hammett coeffi-cient in each case: ρ for native SBP = –1.6 ± 0.1; ρ for MWNT-SBP = –1.5 ± 0.2. (Adapted from Asuri etal. [22].)

1 The values of the electronic parameter (σ) for p-OC2H5, p-CH3,p-CH2OH, and p-Cl are –0.24, –0.17,0.00, and 0.23 respectively.

structural asymmetry in a molecule. Secondary protein structure is usually comprised ofα helices and β sheets, each producing a characteristic spectrum in the far-UV range(190–250 nm). α helices produce a spectrum with valleys around 208 and 222 nm, whileβ sheets show a single valley around 215 nm. As proteins lose their native structure andbecome less ordered, the absence of regular structure is reflected in zero CD intensity.Thus, by measuring the far-UV CD spectrum of a protein before and after attachmentonto nanotubes, one can get an idea of how the structure of the protein has been altered.Using data processing software that can analyze a CD spectrum and determine the rela-tive content of α helix and β sheet, we determined that HRP attached to SWNTs retained68% of its native α helix content (Figure 1.9).

We used CD spectroscopy to monitor the change in the secondary structure of HRPupon exposure to varying concentrations of GdnHCl denaturant and high tempera-tures. The secondary structure of HRP and SWNT-HRP conjugates were thus monitoredby CD, using a protein concentration of 0.05 mg/mL, in the presence or absence ofdenaturant. After equilibrating the samples with GdnHCl for 24 hours, CD spectra weremeasured (Figure 1.10(b)). The concentration of GdnHCl required to denature the pro-tein by 50% in the sample (Cm) increased from 1.6 to 2.4M as a result of conjugation. Forthermal denaturation, the temperature was slowly raised (0.5°C/min) from 20°C to 99°Cwhile spectra were taken (Figure 1.10(a)). The temperature required to unfold the pro-tein by 50% in the sample (Tm) increased from 79°C to 92°C for the SWNT-HRP conju-gate. Characterization by CD spectroscopy therefore revealed a substantial increase inprotein stability under stronger denaturing conditions and higher temperatures whencovalently attached to SWNTs.

1.4.3.3 Characterization of Protein Tertiary Structure Using Tryptophan Fluorescence

Proteins contain three aromatic amino acid residues (tryptophan, tyrosine, andphenylalanine), which contribute to their intrinsic fluorescence. In particular, the polar-ity and charge densities surrounding tryptophan residues influence both the fluores-cence intensity and maximal emission fluorescence wavelength (λmax). As the proteindenatures, losing its tertiary structure, the environment around buried tryptophan resi-


15

200 210 220 230 240 250

−60

−40

−20

0

Ellip

ticity

(deg

rees

10)

×3

20

260

Wavelength (nm)

Figure 1.9 Far-UV CD spectra of native HRP (�), SWNT-HRP (�), and bare SWNTs (�). (Reprintedwith permission from Asuri et al. [35]. Copyright 2007 American Chemical Society.)

dues changes drastically, eventually leading to their exposure to solution. Thus, uponprotein denaturation, the fluorescence intensities and tryptophan emission wavelengthtend toward those of free tryptophan in solution; structural changes can thus be inferredfrom alteration of the tryptophan’s microenvironment.

As an example, the protein HRP contains one buried tryptophan residue at position117 [36]. When GdnHCl was used as the denaturant (Figure 1.11), the fluorescenceintensities and λmax values for both native HRP and SWNT-HRP conjugate were lowerthan those for free L-tryptophanamide when excited at 283 nm at lower Gdn HCIconcentrations [27, 36]; however, at higher GdnHCl concentrations, both the valuesapproached those of L-tryptophanamide, indicating that HRP’s tryptophan residue wasnow more accessible to the solvent due to protein denaturation. The SWNT-HRP conju-gates showed a more gradual increase toward the values of L-tryptophanamide thannative HRP, indicating that they are more stable under denaturing conditions thannative HRP.


16

0.0

0.2

0.4

0.6

0.8

1.0

0 1[GdnHCI] (M)

Frac

tion

dena

ture

d

2 3 4 50.0

0.2

0.4

0.6

0.8

1.0

20 40Temperature (°C)

Frac

tion

dena

ture

d

60 80 100

(a) (b)

Figure 1.10 Fraction of HRP denatured determined by monitoring the CD signal at 222 nm of nativeHRP (�) and SWNT-HRP (�) as a function of (a) GdnHCl concentration and (b) solution temperature.(Reprinted with permission from Asuri et al. [35]. Copyright 2007 American Chemical Society.)

20

30

40

50

60

0 1[GdnHCI] (M)

Flur

esce

nce

inte

nsity

2 3 4 5330

335

340

345

350

355

360

0 1[GdnHCI] (M)

λ(n

m)

max

2 3 4 5

(a) (b)

Figure 1.11 (a) Fluorescence intensity (excitation at 283 nm and emission at λmax) and (b) λmax ofL-tryptophanamide (�), native HRP (�), and SWNT-HRP (�) as a function of GdnHCl concentration.(Reprinted with permission from Asuri et al. [35]. Copyright 2007 American Chemical Society.)

1.4.3.4 Thermostabilization of Proteins Via Covalent Attachment onto CarbonNanotubes

Exposure of proteins to high temperatures can lead to irreversible unfolding and deactiva-tion, posing a critical limitation to their commercial use. We have found that covalentattachment of proteins onto MWNTs leads to thermostabilization of the protein. There isa certain optimal temperature (Topt) at which the protein’s catalytic activity is at itsmaximum, beyond which the protein unfolds and gets deactivated irreversibly. Thethermostabilization of proteins upon immobilization causes an elevation in Topt values.While the Topt for native SBP was found to be approximately 75°C, the MWNT-SBP conju-gates displayed a Topt of approximately 90°C (Figure 1.12), which is close to the nativemelting temperature of SBP (Tm = 90.5°C) [37]. This enhanced stability of MWNT-SBPleads to a 2.5-fold increase in the maximal initial reaction rate at 90°C as compared to thatof native SBP at 75°C, thus rendering the protein formulation well suited for applicationswhere harsh conditions are required.

1.4.3.5 Operational and Storage Stability of Carbon Nanotube-Enzyme Conjugates

Two other issues concerning the commercial use of native enzymes in biocatalysis are dif-ficulty of enzyme reuse and loss of enzymatic activity on prolonged storage. While mac-roscopic supports provide ease of separation and reusability of immobilized enzymes,stabilization provided by such supports is significantly less compared to nanoscale sup-ports [28]. On the other hand, use of inherently long oxidized MWNTs for attachingenzymes not only stabilizes enzymes under different reaction and storage conditions butalso allows easy recovery of conjugates from reaction mixture through filtration. Amat-like film forms after filtering the reaction mixture through the filter membrane; thisfilm can be redispersed in an aqueous buffer by minimal sonication. For example, SBP,which was covalently attached to MWNT, retained about 70% of its initial activity evenafter being reused over 100 times (Figure 1.13(a)). Additionally, such conjugates werefound to be stable for an extended period. Even after 30 days, the MWNT-SBP conjugatesretained ca. 70% of their initial activity (Figure 1.13(b)). On the other hand, native SBPretained only about 30% of its initial activity. These observations indeed suggest that


17

0

100

200

300

400

500

600

20 40 60Temperature (°C)

Initi

alra

te(m

Mm

gs

)−1

−1

80 100

Figure 1.12 Influence of temperature on the kinetics of native SBP (�) and MWNT-SBP (�) in aque-ous buffer. (Adapted from Asuri et al. [22].)

nanoscale supports, such as MWNTs, make the enzyme formulation reusable andstorage compatible.

1.5 Discussion and Commentary

As discussed in the previous sections, characterization of CNT-protein conjugates exhib-its the subtle differences observed due to these different methods of protein immobiliza-tion on CNTs. Biofunctionalization of CNTs and characterization of resultant hybridmaterial has been carried out for various reasons. First, it is of great interest to study howdifferent biomolecules interact with carbon nanotubes compared to conventional microor macroscale supports. For this study, biomolecules were interfaced with nanotubesthrough physical adsorption or covalent attachment. Second, use of CNTs, as drug deliv-ery vehicles or for construction of self-assembled nanoscaled superstructures, requiresthem to be water soluble. The amphiphilic nature of biomolecules can be exploited insolubilizing CNTs. Additionally, these biofunctionalized CNTs can be used in a widerange of applications, which include biosensing, bioelectrochemistry, biomedicine,and intracellular delivery of peptides and proteins. In the course of preparing suchnanobiocomposite materials and realizing their potential applications, we have criti-cally optimized our protocols to overcome some of the problems that can occur withthese techniques. In this section, we discuss some precautions to take while preparingnanotube-protein conjugates.

CNTs, being in the form of clumpy or fluffy black powder, should be handledusing personal protective equipment and in safety hoods with adequate ventilation. Ifinhaled, remove to fresh air. If breathing difficulties persist, get medical attention. Incase of contact, immediately flush eyes or skin with plenty of water for at least 15 min-utes. If irritation develops or persists, get medical attention. During physical adsorptionof proteins, we have observed that protein adsorption occurs in the initial 5 to 10 min-utes of mixing CNTs and protein solutions. Therefore, before mixing these solutions, itis necessary to achieve a uniform dispersion of CNTs through effective sonication. As anadditional precaution, it is advisable to thaw the protein-containing vial after removing


18

0

20

40

60

80

100

1 4 7 10

Average of every 5 cycles

Perc

ent

activ

ityre

tain

ed

13 16 19 220

20

40

60

80

100

0 5 10

Days

Perc

ent

activ

ityre

tain

ed

13 20 25 30

(a) (b)

Figure 1.13 Operational and storage stability of MWNT-SBP. (a) Reusability of MWNT-SBP conju-gates. (b) Retention of enzymatic activity in aqueous buffer at room temperature—native SBP (�) andMWNT-SBP (�). (Adapted from Asuri et al. [22].)

it from storage temperature of 4°C or –20°C. This is to ensure that the protein powderdoes not pick up unwanted moisture on exposure to air.

As a common safety practice, it is recommended to handle acids in fume hoods whilecarrying out acid oxidation of CNTs. During the sonication step, intermittent swirling ofnanotube-acid mixture would maintain well-mixed reaction conditions. The rise intemperature due to exothermic acid oxidation and sonication leads to heating up ofwater bath, which could cause excessive oxidation of CNTs and hence formation of fine,non-recoverable particles. Therefore, periodic replacement of water in the bath withice-cold water is necessary for maintaining desired operating conditions. Filtration ofCNTs after acid oxidation leads to formation of a densely packed nanotube mat. A sim-plistic approach to get uniform nanotube suspension would be to disperse this nanotubefilm in enough volumes of milliQ water by sonication. After ester functionalization andfiltration of oxidized CNTs, ensure that the filtered CNTs film does not dry outcompletely; disperse the film immediately in protein solution. The pH of the bufferused for carrying out EDC-NHS chemistry was found to govern the extent of esterfunctionalization onto carbon nanotubes. While a pH range of 4 to 7 is suitable, lowerpH conditions results in higher functionalization and hence higher protein attachment.

There are a few notable differences between the three methods of nanotube-proteinpreparation, and the choice of one over the other is governed by the properties of conju-gates desired and their end application. Some of the differences between physicaladsorption of proteins onto CNTs, protein solubilization of CNTs, and covalent attach-ment methods of proteins onto CNTs are listed in Table 1.2.

Troubleshooting Table

Problem Explanation Potential Solution

Low catalytic activity of proteinsupon covalent attachment.

Low protein attachment onto oxidized CNTsin EDC-NHS reaction steps.

Prolonged exposure of active ester to air canlead to its hydrolysis. Add protein solutionimmediately after nanotube activation.

Presence of CNT aggregates afteracid oxidation and filtration steps.

Inefficient oxidation of CNTs. Increase acid treatment duration.

1.6 Applications Notes

The methods of noncovalent and covalent functionalization of carbon nanotubes withproteins have been used in numerous applications two of which are highlighted in this

1.6 Applications Notes

19

Table 1.2 General Comparison Between the Three Methods of Protein Attachment onto CNTs

Physical Adsorption of Proteins onto CNTs Protein Solubilization of CNTsCovalent Attachment of Proteins ontoCNTs

Ease of attachment Ease of attachment Cumbersome with many steps involvedFacile method of attachment preservesnative structural and functional properties ofboth CNTs and proteins

Ultrasonication can lead to proteindenaturation

Chemical modification of CNTs can com-promise its native electronic andmechanical properties

Leaching of proteins upon agitation andstorage

Leaching of proteins upon agitationand storage

No leaching effect observed

Conjugates present in aggregate form Conjugates are water-soluble Conjugates are water-solubleConjugates can be separated fromsolution by centrifugation

Conjugates can be separated fromsolution by filtration

Conjugates can be separated from solu-tion by filtration

section. As our first example, we consider the role of nanobiocomposites in carrying outbiotransformation in biphasic medium wherein phase transfer biocatalysis involves themass transfer of water-insoluble substrates from organic to aqueous phase. Therefore,interfacial adsorption of enzymes is desired to carry out biotransformations at the aque-ous-organic interface. We have demonstrated that SWNTs along with the attached pro-tein can be directed to aqueous-organic interfaces with the aid of surfactants [38].SWNTs as a protein support not only provide high intrinsic surface area but also over-come any intraparticle diffusional limitations that restrict use of enzymes in biphasicsystem. We showed that physical adsorption increased specific enzyme activity by threeorders of magnitude as compared to native enzymes in aqueous phase, with enhancedstability at high temperatures. Thus, the nanotube-mediated interfacial assembly ofenzymes can be very advantageous in directing greater amounts of enzymes from thebulk aqueous phase to the interface and in increasing the stability of enzymes againstinactivation.

The procedure of covalent attachment of proteins onto carbon nanotubes has beensuccessfully employed to produce highly active and stable DNAzyme-carbon nanotubehybrids. Certain small single-stranded DNA fragments possess catalytic activity (e.g.,endonuclease-type activity) and are known as DNAzymes [39]. Yim et al. covalentlyattached streptavidin to acid-treated MWNTs using EDC-NHS chemistry, followedby the binding of biotinylated DNAzyme to yield MWNT-DNAzyme conjugates thatwere soluble in aqueous buffer [40]. The MWNT-DNAzyme conjugates followedMichaelis-Menten kinetics under the conditions where substrate concentration ishigher than that of DNAzyme (Figure 1.14). Additionally, this hybridization led to aformulation providing very high turnover numbers, without the need for substrate-DNAzyme hybridization between each catalytic event. Conjugating such DNAzymeswith nanomaterials can be of potential use in the development of biosensors to detectmetal ions and nucleic acids as well as in designing strategies for directing nanoparticleassembly [41, 42].


20

0 2 4

0

20

40

60

80

6 8 10 12 14 16[Substrate] ( M)μ

Initi

alra

teof

clea

vage

reac

tion

(nm

ol/m

in/m

g)

Figure 1.14 Catalytic activity of MWNT-DNAzyme conjugates. The line represents a nonlinear fit ofthe Michaelis-Menten expression to the data. Inset shows analysis of extent of conversion offluorescently labeled substrate DNA by polyacrylamide gel electrophoresis (PAGE), with upper bandrepresenting uncleaved DNA and lower band representing cleaved fragments. (Reprinted with permis-sion from Yim et al. [40]. Copyright 2005 American Chemical Society.)

1.7 Summary Points

Physical adsorption of proteins onto CNTs is a simple and effective method for prepar-ing nanotube-protein conjugates without any modification of electronic and mechani-cal properties of CNTs.

Protein assisted solubilization of CNTs can be important for biomedical applications,such as biomedical devices, cellular delivery; besides, the wide variety of functionalgroups on adsorbed proteins can act as orthogonal reactive handles for thefunctionalization of CNTs.

Water-soluble CNT-protein conjugates, prepared by acid oxidation of CNTs andcovalent attachment of proteins, possess high enzyme activity, high stability andreusability, and low diffusional resistance; these conjugates can find application inbiomaterials, biotransformations, medicine and self-assembled materials.

Acknowledgments

The methods presented here would not have been possible without the dedicated workof Dr. Sandeep S. Karajanagi, Dr. Tae-Jin Yim, and Dr. Dae-Yun Kim who took part in theoriginal investigations. We also thank Dr. Cerasela Zoica Dinu and Dr. Guangyu Zhu forinsightful discussions and comments.

References

[1] Ajayan, P. M., “Nanotubes from carbon,” Chemical Reviews Vol. 99, No. 7 1999, pp. 1787–1799.[2] Chen, R. J., Bangsaruntip, S., Drouvalakis, K. A., Kam, N. W. S., Shim, M., Li, Y. M., Kim, W., Utz, P.

J., and Dai, H. J., “Noncovalent functionalization of carbon nanotubes for highly specific elec-tronic biosensors,” Proceedings of the National Academy of Sciences of the United States of America Vol.100, No. 9 2003, pp. 4984–4989.

[3] Pantarotto, D., Briand, J. P., Prato, M., and Bianco, A., “Translocation of bioactive peptides acrosscell membranes by carbon nanotubes,” Chemical Communications, No. 1 2004, pp. 16–17.

[4] Kam, N. W. S., Jessop, T. C., Wender, P. A., and Dai, H. J., “Nanotube molecular transporters: Inter-nalization of carbon nanotube-protein conjugates into mammalian cells,” Journal of the AmericanChemical Society Vol. 126, No. 22 2004, pp. 6850–6851.

[5] Barone, P. W., Parker, R. S., and Strano, M. S., “In vivo fluorescence detection of glucose using a sin-gle-walled carbon nanotube optical sensor: Design, fluorophore properties, advantages, and disad-vantages,” Analytical Chemistry Vol. 77, No. 23 2005, pp. 7556–7562.

[6] Kam, N. W. S., O’Connell, M., Wisdom, J. A., and Dai, H. J., “Carbon nanotubes as multifunctionalbiological transporters and near-infrared agents for selective cancer cell destruction,” Proceedings ofthe National Academy of Sciences of the United States of America Vol. 102, No. 33 2005, pp.11600–11605.

[7] Karajanagi, S. S., Vertegel, A. A., Kane, R. S., and Dordick, J. S., “Structure and function of enzymesadsorbed onto single-walled carbon nanotubes,” Langmuir Vol. 20, No. 26 2004, pp. 11594–11599.The authors investigate the structure and function of proteins immobilized onto SWNTs todevelop a better understanding of SWNT-protein interactions.

[8] Panhuis, M. I. H., Salvador-Morales, C., Franklin, E., Chambers, G., Fonseca, A., Nagy, J. B., Blau,W. J., and Minett, A. I., “Characterization of an interaction between functionalized carbonnanotubes and an enzyme,” Journal of Nanoscience and Nanotechnology Vol. 3, No. 3 2003, pp.209–213.

[9] Carrillo, A., Swartz, J. A., Gamba, J. M., Kane, R. S., Chakrapani, N., Wei, B. Q., and Ajayan, P. M.,“Noncovalent functionalization of graphite and carbon nanotubes with polymer multilayers andgold nanoparticles,” Nano Letters Vol. 3, No. 10 2003, pp. 1437–1440.

[10] Shim, M., Kam, N. W. S., Chen, R. J., Li, Y. M., and Dai, H. J., “Functionalization of carbonnanotubes for biocompatibility and biomolecular recognition,” Nano Letters Vol. 2, No. 4 2002, pp.285–288.

1.7 Summary Points

21

[11] Pompeo, F., and Resasco, D. E., “Water solubilization of single-walled carbon nanotubes byfunctionalization with glucosarnine,” Nano Letters Vol. 2, No. 4 2002, pp. 369–373.

[12] Star, A., Steuerman, D. W., Heath, J. R., and Stoddart, J. F., “Starched carbon nanotubes,”Angewandte Chemie-International Edition Vol. 41, No. 14 2002, pp. 2508.

[13] Asuri, P., Karajanagi, S. S., Kane, R. S., and Dordick, J. S., “Polymer-nanotube-enzyme composites asactive antifouling films,” Small Vol. 3, No. 1 2007, pp. 50–53.

[14] Karajanagi, S. S., Yang, H. C., Asuri, P., Sellitto, E., Dordick, J. S., and Kane, R. S., “Protein-assistedsolubilization of single-walled carbon nanotubes,” Langmuir Vol. 22, No. 4 2006, pp. 1392–1395.The authors report a simple method that uses proteins to solubilize SWNTs in water.

[15] O’Connell, M. J., Bachilo, S. M., Huffman, C. B., Moore, V. C., Strano, M. S., Haroz, E. H., Rialon, K.L., Boul, P. J., Noon, W. H., Kittrell, C., Ma, J. P., Hauge, R. H., Weisman, R. B., and Smalley, R. E.,“Band gap fluorescence from individual single-walled carbon nanotubes,” Science Vol. 297, No.5581 2002, pp. 593–596.

[16] Islam, M. F., Rojas, E., Bergey, D. M., Johnson, A. T., and Yodh, A. G., “High weight fractionsurfactant solubilization of single-wall carbon nanotubes in water,” Nano Letters Vol. 3, No. 2 2003,pp. 269–273.

[17] O’Connell, M. J., Boul, P., Ericson, L. M., Huffman, C., Wang, Y. H., Haroz, E., Kuper, C., Tour,J., Ausman, K. D., and Smalley, R. E., “Reversible water-solubilization of single-walled carbonnanotubes by polymer wrapping,” Chemical Physics Letters Vol. 342, No. 3–4 2001, pp. 265–271.

[18] Huang, W. J., Fernando, S., Allard, L. F., and Sun, Y. P., “Solubilization of single-walled carbonnanotubes with diamine-terminated oligomeric poly(ethylene glycol) in differentfunctionalization reactions,” Nano Letters Vol. 3, No. 4 2003, pp. 565–568.

[19] Zheng, M., Jagota, A., Semke, E. D., Diner, B. A., McLean, R. S., Lustig, S. R., Richardson, R. E., andTassi, N. G., “DNA-assisted dispersion and separation of carbon nanotubes,” Nature Materials Vol.2, No. 5 2003, pp. 338–342.

[20] Dieckmann, G. R., Dalton, A. B., Johnson, P. A., Razal, J., Chen, J., Giordano, G. M., Munoz, E.,Musselman, I. H., Baughman, R. H., and Draper, R. K., “Controlled assembly of carbon nanotubesby designed amphiphilic peptide helices,” Journal of the American Chemical Society Vol. 125, No. 72003, pp. 1770–1777.

[21] Numata, M., Asai, M., Kaneko, K., Bae, A. H., Hasegawa, T., Sakurai, K., and Shinkai, S., “Inclusionof cut and as-grown single-walled carbon nanotubes in the helical superstructure of schizophyllanand curdlan (ss-1,3-glucans),” Journal of the American Chemical Society Vol. 127, No. 16 2005, pp.5875–5884.

[22] Asuri, P., Karajanagi, S. S., Sellitto, E., Kim, D. Y., Kane, R. S., and Dordick, J. S., “Water-soluble car-bon nanotube-enzyme conjugates as functional biocatalytic formulations,” Biotechnology and Bio-engineering Vol. 95, No. 5 2006, pp. 804–811.The authors present a method of preparing water-soluble carbon nanotube-enzyme conjugates,which retain high activity and are stable at high temperatures.

[23] Jiang, K. Y., Schadler, L. S., Siegel, R. W., Zhang, X. J., Zhang, H. F., and Terrones, M., “Proteinimmobilization on carbon nanotubes via a two-step process of diimide-activated amidation,” Jour-nal of Materials Chemistry Vol. 14, No. 1 2004, pp. 37–39.

[24] Vix-Guterl, C., Couzi, M., Dentzer, J., Trinquecoste, M., and Delhaes, P., “Surface characterizationsof carbon multiwall nanotubes: Comparison between surface active sites and Raman Spectros-copy,” Journal of Physical Chemistry B Vol. 108, No. 50 2004, pp. 19361–19367.

[25] Wang, P., “Nanoscale biocatalyst systems,” Current Opinion in Biotechnology Vol. 17, No. 6 2006,pp. 574–579.

[26] Huang, W. J., Taylor, S., Fu, K. F., Lin, Y., Zhang, D. H., Hanks, T. W., Rao, A. M., and Sun, Y. P.,“Attaching proteins to carbon nanotubes via diimide-activated amidation,” Nano Letters Vol. 2,No. 4 2002, pp. 311–314.

[27] Ryu, K., and Dordick, J. S., “How Do Organic-Solvents Affect Peroxidase Structure and Function,”Biochemistry Vol. 31, No. 9 1992, pp. 2588–2598.

[28] Asuri, P., Karajanagi, S. S., Yang, H. C., Yim, T. J., Kane, R. S., and Dordick, J. S., “Increasing proteinstability through control of the nanoscale environment,” Langmuir Vol. 22, No. 13 2006,pp. 5833–5836.The authors report that carbon nanotubes may be used to enhance protein stability in harshenvironments.

[29] Jackson, M., and Mantsch, H. H., “The use and misuse of FTIR Spectroscopy in the determination ofprotein-structure,” Critical Reviews in Biochemistry and Molecular Biology Vol. 30, No. 2 1995,pp. 95–120.

[30] Dong, A., Huang, P., and Caughey, W. S., “Protein secondary structures in water from 2nd-deriva-tive amide-I infrared spectra,” Biochemistry Vol. 29, No. 13 1990, pp. 3303–3308.


22

[31] Vedantham, G., Sparks, H. G., Sane, S. U., Tzannis, S., and Przybycien, T. M., “A holistic approachfor protein secondary structure estimation from infrared spectra in H2O solutions,” AnalyticalBiochemistry Vol. 285, No. 1 2000, pp. 33–49.

[32] Barone, P. W., Baik, S., Heller, D. A., and Strano, M. S., “Near-infrared optical sensors based on sin-gle-walled carbon nanotubes,” Nature Materials Vol. 4, No. 1 2005, pp. 86–U16.

[33] Dalton, A. B., Stephan, C., Coleman, J. N., McCarthy, B., Ajayan, P. M., Lefrant, S., Bernier, P., Blau,W. J., and Byrne, H. J., “Selective interaction of a semiconjugated organic polymer with single-wallnanotubes,” Journal of Physical Chemistry B Vol. 104, No. 43 2000, pp. 10012–10016.

[34] Rao, A. M., Chen, J., Richter, E., Schlecht, U., Eklund, P. C., Haddon, R. C., Venkateswaran, U. D.,Kwon, Y. K., and Tomanek, D., “Effect of van der Waals interactions on the Raman modes in singlewalled carbon nanotubes,” Physical Review Letters Vol. 86, No. 17 2001, pp. 3895–3898.

[35] Asuri, P., Bale, S. S., Pangule, R. C., Shah, D. A., Kane, R. S., and Dordick, J. S., “Structure, function,and stability of enzymes covalently attached to single-walled carbon nanotubes,” Langmuir Vol.23, No. 24 2007, pp. 12318–12321.The authors report that enzymes covalently attached to SWNTs retain a high fraction of theirnative structure and function, with enhanced stability in harsh environments.

[36] Welinder, K. G., “Amino-Acid Sequence Studies of Horseradish-Peroxidase .4. Amino and CarboxylTermini, Cyanogen-Bromide and Tryptic Fragments, the Complete Sequence, and Some StructuralCharacteristics of Horseradish Peroxidase-C,” European Journal of Biochemistry Vol. 96, No. 3 1979,pp. 483–502.

[37] McEldoon, J. P., and Dordick, J. S., “Unusual thermal stability of soybean peroxidase,” Biotechnol-ogy Progress Vol. 12, No. 4 1996, pp. 555–558.

[38] Asuri, P., Karajanagi, S. S., Dordick, J. S., and Kane, R. S., “Directed assembly of carbon nanotubes atliquid-liquid interfaces: Nanoscale conveyors for interfacial biocatalysis,” Journal of the AmericanChemical Society Vol. 128, No. 4 2006, pp. 1046–1047.The authors demonstrate that SWNT-enzyme conjugates can be directed to aqueous-organic inter-faces and used for interfacial biocatalysis.

[39] Li, Y. F., and Breaker, R. R., “Deoxyribozymes: new players in the ancient game of biocatalysis,”Current Opinion in Structural Biology Vol. 9, No. 3 1999, pp. 315–323.

[40] Yim, T. J., Liu, J. W., Lu, Y., Kane, R. S., and Dordick, J. S., “Highly active and stable DNAzyme - Car-bon nanotube hybrids,” Journal of the American Chemical Society Vol. 127, No. 35 2005, pp.12200–12201.

[41] Liu, J. W., and Lu, Y., “A colorimetric lead biosensor using DNAzyme-directed assembly of goldnanoparticles,” Journal of the American Chemical Society Vol. 125, No. 22 2003, pp. 6642–6643.

[42] Sando, S., Sasaki, T., Kanatani, K., and Aoyama, Y., “Amplified nucleic acid sensing using pro-grammed self-cleaving DNAzyme,” Journal of the American Chemical Society Vol. 125, No. 51 2003,pp. 15720–15721.

References

23

C H A P T E R

2Peptide-Nanoparticle Assemblies

Joseph M. Slocik and Rajesh R. Naik*

Nanostructured and Biological Materials Branch, Materials and Manufacturing Directorate,Air Force Research laboratory, Wright-Patterson AFB, OH 45433-7750

*Corresponding author e-mail: [email protected]; phone: 937-255-3808

25

Abstract

Unlike material science where there is a general lack of control, poorly assem-bled structures, and high levels of impurities; biology uses precise molecularand genetic control to guide the assembly of complex nanostructures andshapes via proteins/peptides. Experimentally, this is appealing given thatproteins or peptides can be used to functionalize a nanoparticle surface andpromote assembly through peptide-peptide interactions, formation of supra-molecular structures, and/or recognition of specialized targets. In this chapter,we describe the use of peptides for the controlled assembly of nanoparticleswith regard to different types of interfaces, resulting nanostructures, andenhanced properties.

Key terms peptidesbiomimetic synthesisassemblynanoparticle

2.1 Introduction

Nanoparticle assembly is critical in the fabrication of complex devices, hybrid struc-tures, and biosensors; where the collective properties, functionality, and efficiency (elec-tronic, optical, mechanical, and catalytic) are ultimately determined by how well it isassembled. For example, depending on the interparticle spacing and aggregation size ofassembled gold nanoparticles; a multitude of different colors can be produced [1], whileenhancement in catalytic activity can be achieved by controlling the orthogonal assem-bly of gold and palladium [2]. As a consequence, this necessitates a demand for increasedcontrol over nanoparticle synthesis, processing, and especially, assembly. Currently, thesynthesis and assembly of nanomaterials using conventional material science tech-niques (high temperatures, pressures, organic solvents, harsh reducing agents, andextreme pH) often results in poorly defined structures and high levels of impurities [3].

In nature, however, biological systems exhibit precise control over the assembly oforganic and inorganic materials through the use of diverse biomolecule building blocksand intrinsic biomolecular interactions at ambient conditions (low pressure and temper-ature). For instance, the marine diatom utilizes biominerlization peptides/proteins thatare genetically encoded by their amino acid sequences to synthesize and assemble sil-ica nanoparticles into micron-sized silica cell walls (frustules) with exquisite control.Similarly, magnetotactic bacteria manufacture highly organized chains of magnetitenanoparticles for geomagnetic navigation [4, 5].

Biomimetically, these same interactions and templates can guide the temporal andspatial assembly of nanoparticles, thereby offering an unparalleled level of controlunattainable by the use of nonbiological routes. Biomolecules in the form ofDNA, streptavidin or biotin, amino acids, and/or peptides can be functionalized ontonanoparticle surfaces and assembled via specific biological interactions [6–8]. For exam-ple, hybridization of nanoparticles functionalized with complementary single strandsof thiolated DNA has resulted in elaborate 2-D nanoparticle lattices, discreteheterostructures comprised of gold and quantum dots, and extended gold nanoparticleassemblies [9, 10]. Also, the three-dimensional architecture of proteins, viruses, and cellstructures can be used to organize and/or template metal nanoparticles in specificlocations along the bioscaffold resulting in unique structures (nanowires, patternedspherical cages) [11, 12].

Alternatively, peptides are appealing for nanoparticle synthesis and assembly giventheir simplicity, ability to be chemically synthesized, high affinity to nanoparticlesurfaces (equivalent to thiols), self-assembly into supramolecular structures (helical bun-dles, protein cages, viral capsids, filaments, nanotubes), response to thermal and chemi-cal stimuli, abundance in nature, and ability to be engineered for a specific material byuse of phage displayed peptide libraries [4, 7, 13]. For the latter, phage display enablesthe rapid screening and identification of peptide templates that can be extended toinclude virtually any material or nonnaturally occurring nanoparticle such as hightemperature ceramics [4].

To date, peptide functionalized nanoparticles have been assembled via coil-coilinteractions [14], addition of metal ions [15], changes to solution pH, exploiting themultifunctionality of peptide coat to template two different metal nanoparticles [16,17], and peptide-antibody recognition [18]. In these examples, the choice of peptideinterface greatly influenced how the nanoparticles were assembled into the finalstructure.

Peptide-Nanoparticle Assemblies

26

In this chapter, we present several examples of nanoparticle assembly usingmultifunctional peptides derived from phage display, coiled-coil peptide motifs, andpeptide-antibody interactions to assemble metal nanoparticles, bimetallic nano-particles, and metal-semiconductor quantum dot structures. For each of the peptideassembled structures, we discuss the peptide assembly interface, method of assembly,the resulting material properties, and physical structure.

2.2 Materials

Peptides: A set of complementary antiparallel coil peptides (E5 and K5) was used inthe assembly of gold nanoshell dimers and gold-quantum dot structures. Coil peptideswere modified to include a cysteine residue at the N-terminus of each peptide in order topromote binding of peptides to the gold nanoshell surface. The gold-thiol interaction isthe preferred means to attach polymers, organic molecules, and/or biomolecules to agold surface because of the strong affinity of thiols for gold [19]. Additional peptidesinclude a bifunctional FlgA3 peptide for the synthesis and assembly of bimetallic Au-Pdparticles and a Flg peptide epitope for the assembly of gold with antibody functionalizedquantum dots. The FlgA3 peptide was selected from a phage display peptide library bypanning against a gold surface using a commercially available phage display kit fromNew England BioLabs. All peptides were chemically synthesized using an automatedpeptide synthesizer by New England Peptides. Peptides are dissolved in double deionizedwater to make a 10 mg/mL stock solution of peptide. The peptides have the followingsequences below:

(E5) CGGEVSAALEKEVSALEKEVSALEKEVSALEKEVSALEK

(K5) CGGKVSALKEKVSALKEKVSALKEKVSALKEKVSALKE

(FlgA3) DYKDDDDKPAYSSGAPPMPPF

(Flg) DYKDDDDK

1. Nanoparticles: The synthesis of gold nanoshells was reported previously [20].Nanoshells were tuned to have a plasmon resonance at 810 nm by producing a13-nm thick gold shell (total diameter ~176 nm) around a spherical silica particle.Concentration of nanoshells is 3.7 x 109 particles/mL in water [21]. Also, CdSe/ZnSEvitag-Fort orange carboxyl quantum dots (QD-COOH) (emission 605 nm) werepurchased from Evident Technologies, catalog # ET-C11-CB1-0600 at aconcentration of 0.25 mg/mL, while QDot 605 (emission 605 nm) streptavidincoated conjugates were obtained from Quantum Dot Corp at a concentration of 1μM for antibody functionalization and are composed of a CdSe core and a ZnS shell.

2. Nanoparticle synthesis precursors: 0.1 M stock solutions of Au3+ and Pd4+ metal ions

were prepared by dissolving 17.0 mg of HAuCl4⋅3H2O (Fisher Scientific) and 19.9 mg

of K2PdCl6 (Aldrich) in 500 μL of doubly deionized water in a microfuge tube. Metal

ion solutions were stored at 4°C and covered in foil. Sodium borohydride reductant

was prepared by dissolving 1.9 mg of NaBH4 (Aldrich) in 500 μL of double deionizedwater in a microfuge tube. (Note: Prepare fresh sodium borohydride daily as it losesits reducing strength over time.)

2.2 Materials

27

3. Buffers: 0.1 M HEPES buffer (hydroxyethylpiperazine-N’-2-ethanesulfonic acid), pH7.4 was made by diluting 1 mL of a sterile 1 M HEPES solution (Amresco) with 9 mLof double deionized water.

2.3 Methods

2.3.1 Coil-Coil Peptide Mediated NP Assembly

Coil peptides represent the simplest assembling structure in biology beyond DNA. Itsuses include stabilizing structural components in proteins, triggering viral fusion to cellsurfaces, assembling multimeric protein structures, joining biomolecules, and signalingbiochemical events by means of forming a helical coiled-coil peptide complex [22, 23].In vivo, two complimentary peptides self-assemble in a parallel or antiparallel arrange-ment to form a heterodimeric coiled-coil peptide complex and an interface between twolike or unlike biomolecules. Similarly, coil-coil formation presents an excellent meansfor linking inorganic nanoparticles together and obtaining different structures [14].Here, we detail the modification of coil peptides for nanoparticle binding, nanoparticlefunctionalization, and formation of nanoparticle assemblies (Figure 2.1). Specifically,this type of interface was used to assemble gold nanoshell extended networks anddiscrete gold nanoshell-quantum dot structures [21].

2.3.1.1 Gold Nanoparticle Assembly

1. Directly functionalize gold nanoshells resonating at 810 nm with each cysteinemodified coil peptide (E5 and K5, 10 mg/mL in double deionized water) according toFigure 2.1. Incubate 200 μL of gold nanoshells (1.1 x 1010 particles/mL) with 10 μL ofeach antiparallel coil peptide (10 mg/mL in double deionized water) in 200 μL of 0.1M phosphate buffer pH 9.0 for 2 hours. Coil peptides bind to gold nanoshell surfacethrough the cysteine residue at the N-terminus forming a gold-thiol bond [19].

2. After incubation, purify each coil functionalized nanoshell (E5 and K5) from theunbound coil peptides by centrifugation at 450 rcf for 10 minutes. Remove thesupernatant and redissolve the green nanoshell pellet in 200 μL of deionized waterand centrifuge again at 450 rcf for 10 minutes (see Troubleshooting Table). Repeat


28

Size profile

AssembledFree

0 100 200 300 4000

20

40

60

80

100

Size (nm)

Abs

orp

tion

Assembly

4hK5-NS

E5-NS

2hpH 9

Figure 2.1 Assembly of gold nanoshells and gold-quantum dot nanoparticles using coil-coil peptideformation. Size profile of peptide assembled nanoshells.

three times in total to ensure removal of most of the unbound peptide. Upon thefinal centrifugation, dissolve the peptide functionalized nanoshells in 200 μL ofwater.

3. Upon functionalization of nanoshells with each coil peptide, add 200 μL ofE5-nanoshells to 200 μL of K5-nanoshells and incubate the set for 4 hours to induceformation of the coil-coil peptide complex and nanoshell dimers.

4. Characterize coiled-coil assembled nanoshell structures by transmission electronmicroscopy (TEM) and particle size analysis based on sedimentation of particles on asucrose gradient. For TEM, dropcast 10 μL of particle solution onto a 200 meshcopper grid with carbon substrate, Ted Pella Inc., and dried. Obtain size distributionsof the assemblies using a CPS disc centrifuge particle size analyzer DC240000 (CPSInstruments) operating at 24,000 rpm with a sucrose gradient of 24% to 8% andcompare against free unassembled E5-nanoshells (size profile is represented in Figure2.1). 100 μL of sample is injected onto sucrose gradient and scanned over a particlesize range of 50 to 500 nm.

2.3.1.2 Assembly of Gold-Quantum Dot Heterostructures

The emission properties of quantum dots are very sensitive to the presence of goldnanoparticles and depend largely on how their assembled into a hybrid structure. Forexample, the quantum dot fluorescence can either be enhanced or suppressed whenassembled with gold nanoparticles. A 40-fold enhancement of fluorescence wasachieved for a CdSe nanowire coated with small gold nanoparticles via streptavidin-bio-tin binding, but quenched when assembled with DNA functionalized gold into smalldiscrete structures [24, 25]. For each, the selection of biomolecule interface controlledthe arrangement, proximity of quantum dots to gold, and overall geometry. Here,we use coil peptides to decorate the nanoshell surface with quantum dots byfunctionalizing each particle with a complementary coil peptide (Figure 2.2). Uponassembly, the resulting quantum dot-gold nanoshell (QD-NS) structure shows modifiedoptical properties.

1. To functionalize QD-COOH with the N-terminal cysteine modified E5 coil peptide;the carboxyl surface of the quantum dot is converted to a thiol surface usingEDC/NHS (Pierce) and cysteamine (Aldrich). Activate 50 μL of QD-COOH with 20 μLof 0.1 M EDC/NHS and couple with 50 μL of 0.1 M cysteamine for 2 hours.

2.3 Methods

29

Cysteamine(Cys)

Cysteaminecoupled QD

Coil peptides Peptidefunctionalized QD

QD-COO− E5

Phosphate

pH 9, 2hrDisulfidelinkage

−SH

EDC/NHS

2 hr

Figure 2.2 Functionalization of quantum dots with coil peptide. Carboxylated quantum dot is acti-vated, converted to an ester, and derivatized with cysteamine. The cysteamine coupled quantum dotsare then functionalized with the cysteine modified coil peptide via a disulfide linkage.

Cysteamine contains only a primary amine for coupling to the carboxyl surface ofthe quantum dots.

2. Confirm coupling by FT-IR on a Perkin-Elmer FT-IR microscope. Dropcast 5 μL ofpeptide functionalized quantum dot onto a double-sided polished silicon wafer.(Note: silicon is transparent to IR.) FT-IR spectrum is collected from 4,000 to 500 cm-1

and shows vibration associated with cysteamine (S-H stretch).

3. After coupling, add 10 μL of E5 coil peptide with cysteamine coupled quantum dotsin 200 μL of 0.1 M phosphate buffer pH 9.0 for 4 hours to induce a disulfide linkagebetween the cysteamine functionalized quantum dots and the cysteine residue ofthe E5 coil peptide. Confirm peptide coupling by FT-IR. The spectrum of the coupledpeptide will show the absence of the characteristic S-H stretch indicating theformation of the disulfide bond between the cysteamine coupled quantum dots andthe cysteine modified coil peptide. Additional vibrations associated with the peptidewill be present (N-H bend, C=O stretch, O-H stretch).

4. Add E5 coil functionalized quantum dots to K5 coil functionalized gold nanoshellsand incubate for 4 hours to promote coiled-coil formation. Monitor assembly in situusing a Cary eclipse fluoromoter. Quantum dots were pulsed every 1 sec with a400-nm excitation and monitored over time for fluorescence at 605 nm. Over 30minutes, fluorescence of quantum dot was quenched by assembly with gold.

2.3.1.3 Disassembly of Coiled-Coil Mediated Nanostructures

The reversible unfolding of coil peptides can be exploited for disassemblingnanostructures. For example, the coil-coil peptide interface of the assemblednanoparticle structure can unfold and dissociate by means of heating, addition of dena-turing agent (quanidinium hydrochloride), change in pH, or remotely viaphotoillumination, into separated nanoparticle components [14, 21]. Consequently,the ability to disassemble nanostructures is attractive for actuation, switching, andmodulating nanoparticle properties. Below, we describe the disassembly of nanoshelldimers and nanoshell-quantum dot hybrids upon illumination with near-IR light as inFigure 2.3.


30

Size profile

Disassembled

Disassembled 0 100 200 300 4000

20

40

60

80

100

Size (nm)

Abs

orp

tion

Assembled

LED’s810 nm(20 mW)

Figure 2.3 Near-IR mediated disassembly of peptide assembled nanoshells. Included is representativesize profile demonstrating change in particle size.

1. Illuminate the peptide assembled nanoshell structures (NS-NS and QD-NS)with 810-nm near-IR light to induce disassembly for 15 minutes as in Figure 2.3.Photoillumination sources consisted of an array of three 810-nm, 20-mW LEDmolded lamps (Marubeni Inc.) or a mercury lamp of a Cary Eclipse fluorometerwhich pulses at 810 nm.

2. Monitor disassembly by CPS particle size analysis using sedimentation as describedin Section 2.1.1 for NS-NS and QD-NS structures (Figure 2.3) (see TroubleshootingTable). For QD-NS, measure the fluorescence of complex after irradiation onfluorometer by immediately exciting NS-QD structure at 400 nm and measuringemission at 605 nm every 1 sec over 60 minutes. Over time the fluorescence willdecay as the QD becomes quenched upon returning to nonirradiated state.

2.3.2 Synthesis of Hybrid Structures Using Multifunctional Peptides

Peptides can be designed to impart multifunctionality. In this capacity, peptides cancarry out nanoparticle synthesis, analyte binding/sensing, recognition of two differentmaterials, assembly, and/or any other biological function depending on the amino acidsequence. To achieve multifunctionality; the peptide is programmed to contain two ormore distinct sequences capable of performing specific functions corresponding tothose domains [4]. Appealingly, each domain can be replaced with an entirely newdomain capable of performing a different function. In the following, we describe the useof a bifunctional peptide to synthesize gold nanoparticles decorated with palladium andthe assembly of peptide coated gold structures by the addition of various metal ions. Ineach case, the two domains function codependently towards the assembly of bimetallicnanostructures and recognition of metal ions.

The association of gold with palladium produces a bimetallic structure withenhanced catalytic and electrical properties; but again, is dependent upon the methodof assembly as mentioned above. To date, the assembly of bimetallic Au-Pd particles hasbeen limited to nonbiomimetic routes involving large dendrimer hosts [26], polymers[27], and micelles [28]. Unfortunately, these lack chemical and structural control byserving as bulk containers, and invariably lead to undefined structures and deficientproperties. Alternatively, the multifunctionality of peptides is attractive for the synthe-sis and assembly of gold-palladium hybrid particles as shown in Figure 2.4. Here, thebifunctional FlgA3 peptide is used to direct the synthesis of peptide coated goldnanoparticles (A3 domain) and assemble Pd clusters via binding and nucleation at theFlg peptide domain.

1. Synthesize peptide coated gold nanoparticles using the FlgA3 bifunctional peptide(DYKDDDDKPAYSSGAPPMPPF) [29]. Add 10 μL of FlgA3 peptide (10 mg/mL indouble deionized water, New England peptide) to 500 μL of 0.1 M HEPES buffer pH7.1 in a microfuge tube. To the peptide solution, introduce 2.5 µL of 0.1 M AuCl4

-

by micropipette. After 4 hours, the solution slowly turns red indicative of goldnanoparticle formation. Presumably, the tyrosine residues of the FlgA3 peptidein conjunction with the HEPES buffer contribute to reduction of Au3+ to Au0

nanoparticles.

2. Purify FlgA3 peptide coated gold nanoparticles by repeated centrifugation andwashing with deionized water as described above in step 2 of Section 2.1.1.

2.3 Methods

31

3. Upon purification, add 2.5 μL of 0.1 M K2PdCl6 to the FlgA3 coated gold nano-particles and incubate for 10 minutes to promote binding of Pd4+ ions at the FlgA3peptide surface of gold (see Troubleshooting Table). Addition of Pd4+ to FlgA3-NP’scauses a color change from red to purple indicative of metal ion binding to peptidecoat.

4. Following binding of Pd4+ ions, reduce ions with 20 μL of 0.1 M NaBH4. Addition ofNaBH4 instantly reduces the bound Pd4+ ions to Pd0. Over 30 minutes, Pd clustersform at the gold-peptide interface evenly distributed over the surface.

5. Confirm the Au-Pd bimetallic structure by TEM as performed above in Section 2.1.1and as seen in Figure 2.4.

2.4 Assembly Mediated by Metal Ion-PeptideRecognition

Metal ion-peptide interactions are important for the folding and function of bio-molecules in nature; in addition to regulating enzyme activity by cycling through redoxstates [30]. Similarly, these interactions are useful for controlling the extent of assemblyof peptide functionalized nanoparticle networks via the addition of different metal ions.In the following, we used the (Flg) palladium binding domain of the FlgA3-NP’s to bindalternate metal ions (Zn2+, Pb2+, Cu2+, Hg2+, Pt2+, Pd4+) and form various gold nanoparticleassemblies through metal ion-peptide coordination of functionalized nanoparticles asin Figure 2.3 [15]. Given the different metal ion-peptide binding affinities, each metalion results in the assembly of an optically different nanoparticle network with a distinctaggregate size. In total, a matrix of gold nanoparticle assemblies spanning the visiblecolor spectrum is collected with addition of metal ions (Figure 2.5).

1. Synthesize FlgA3 peptide coated gold nanoparticles as above in Section 2.2.1 andpurify to yield a stable nanoparticle solution.

2. To promote assembly, add 2 μL of metal ions (0.1 M) to the solution of nanoparticles.Metal ions include Ag+, Zn2+, Ni2+, Co2+, Hg2+, Pb2+, Pd4+, and Pt2+. Addition of metal ionsinstantly causes a color change to nanoparticle solution.

3. Characterize each metal ion/NP assembly by UV-Vis spectroscopy on a VarianUV-Vis-NIR spectrophotometer. Dilute 100 µL of metal-ion/FlgA3-NP with 400 μL ofdeionized water in a 750 μL quartz cuvette and scan from 200 to 800 nm. Theplasmon peak can shift from ~530 nm to 550 to 730 nm depending on response to


32

Figure 2.4 Assembly of palladium decorated gold nanoparticles using a bifunctional peptide display-ing affinities for gold and palladium as demonstrated by representative TEM micrograph.

metal ion and the extent of assembly. Determine size distributions for the nano-particle assemblies by sedimentation on a sucrose gradient as described above inSection 2.1.1.

2.5 Peptides as Antibody Epitopes for NanoparticleAssembly

Depending on the peptide sequence, the repeating nature of peptides along a nano-particle surface can create an immunomolecular interface for antibody recognitionand/or assembly with antibody functionalized nanoparticles. This was first demon-strated by the antibody recognition of nanoparticles coated with a histidine-rich peptideepitope [31]. Unlike the assembly of peptide functionalized nanoparticles describedabove, antibody-peptide interactions offer higher specificity, increased rigidity of thebiomolecule pair, and larger size of antibody interface. The large size of the antibodyaffects interparticle distance, geometry, and stoichiometry of the components. With thelatter, the number of assembled particles is limited primarily by the number of antibod-ies available on the quantum dot surface. Here, we use the recognition of antibodyfunctionalized quantum dots for assembly with Flg peptide coated gold nanoparticles asa means to produce simple hybrid structures consisting of a single quantum dot with 1to 6 gold particles attached (Figure 2.6). Through assembly of these structures, weobtained sequential increases in fluorescence quenching.

1. Synthesize peptide coated gold nanoparticles with the Flg antibody binding epitopeas above in Section 2.2.1 by using 10 μL of Flg peptide (Sigma, DYKDDDDK).

2. Functionalize quantum dots with Anti-Flg antibodies by adding 0.5 μL of QDot 605streptavidin coated conjugates (Quantum Dot Corp.) with 10 μL of diluted anti-FlgBioM2 (Conjugated with biotin, diluted 10-fold to yield a concentration of 100μg/mL in Tris buffered saline, Sigma) in 200 μL of Tris buffered saline (0.5 M Tris, pH7.4, 0.15 M NaCl). Incubate antibodies with QDot for 1 hour at room temperature for

2.5 Peptides as Antibody Epitopes for Nanoparticle Assembly

33

[M ]n+

Figure 2.5 Assembly of FlgA3 peptide coated nanoparticles in the presence of metal ions. Bottom:Image of nanoparticle assemblies with different metal ions.

functionalization. Purify excess antibodies from functionalized quantum dots byplacing solution on a streptavidin coated glass microscope slide (Xenopore) for 15minutes and removing solution. Unbound antibodies bind to streptavidin surface ofglass slide.

3. Assemble Flg coated gold particles with anti-Flg conjugated quantum dots byaddition of 0-80-mM gold-peptide to a fixed quantum dot concentration of 2 nMand incubate for 4 hours. Varying the concentration of gold results in differentnumbers of gold particles bound to the quantum dot center. (Note: the assembly ofgold and quantum dots result in a distribution of structures; i.e., 4 gold particles + 1QD, 3 gold + 1 QD, 2 gold + 1QD, or 1 gold + 1 QD.)

2.6 DATA Acquisition, Anticipated Results, andInterpretation

1. The functionalization of nanoparticles with peptides following the method abovewas determined qualitatively by FT-IR, UV-Vis, and circular dichroism (CD)spectroscopies. Each technique provides a characteristic spectrum that containsspecific absorptions, electronic transitions, and vibrations associated with thefunctionalized peptide. For instance by FT-IR, the presence of a set of Amide I (C=Ostretch) at ~ 1650 cm-1 and Amide II (NH bend, ~1550 cm-1) vibrations indicates thebound peptide. Also, if the peptide contains a cysteine residue; the S-H stretch can beused to confirm binding and will be absent if it is bound to the nanoparticle in theFT-IR spectrum. Additional evidence of functionalization can be obtained bycollecting a UV-Vis absorbance spectrum and looking for absorbance peaks at 210nm for the peptide backbone and/or 280 nm for the aromatic rings in tryptophan


34

Biotinylatedanti-Flg

antibody

StreptavidinconjugatedCdSe/ZnS

quantum dot Antibody-quantum dot-Auassembly

Antibody-quantum assembly

Flg-synthesizedgold nanoparticle

Figure 2.6 Assembly of quantum dot-gold heterostructures using antibody functionalized quantumdots. Quantum dots are functionalized with Anti-Flg antibodies while gold nanoparticles are synthe-sized with the Flg peptide epitope.

and tyrosine. Circular dichroism spectroscopy (CD) can be used to supplement theseother techniques and/or as a means to monitor structural changes to the peptideupon nanoparticle binding. CD provides information regarding the secondarystructure and conformation (α-helix, β-sheet, 310 helix, unordered, random coil,turns) of the biomolecule bound to the nanoparticle surface. For example, a peptideadopting an α-helical secondary structure on the nanoparticle surface would reveal a

positive peak at 190 nm (n→π*) and two negative peaks at 212 nm and 222 nm (π→π*)in the CD spectrum.

2. After nanoparticles are functionalized with selected peptide; the peptidefunctionalized components are assembled via one of the approaches described aboveand characterized by transmission electron microscopy (TEM), UV-Vis spectroscopy,atomic force microscopy (AFM), dynamic light scattering (DLS), and/or particle sizeanalysis using a centrifugal particle sizing system. From the TEM micrographs and/orAFM images; assembly can be determined by examining changes in interparticlespacing and the presence of unique geometries (dimers, trimers, extended structures,raspberry decorated particles). DLS and centrifugal particle sizing are an excellentmeans to confirm assembly by looking for a change in the size profile as illustratedabove in Figures 2.1 and 2.3. Assembly can also be monitored optically by means ofUV-Vis or fluorescence spectroscopy.


Peptides offer tremendous control in the synthesis and assembly of nanoparticles via adiverse assortment of specific interactions that can be tailored to bind a desiredmaterial and/or respond to the presence of an external agent. As exemplified above,nanoparticles with different structures and material properties were obtained dependingon the type of peptide interface used for assembly. Consequently, as with anynanoparticle functionalization technique; the functionalization efficiency of peptidescan be low resulting in a limited number of peptide copies on nanoparticle surface per agiven area. For example, for a complete functionalized 2 nm particle, there are approxi-mately 22 7-mer peptides predicted to decorate the nanoparticle surface [31]. In thiscase, the peptide remarkably stabilizes the nanoparticle in solution for extended periodsof time (years) with no ripening or precipitation of particles. Also, peptides can adoptmultiple different conformations and structures when constrained on a nanoparticlesurface, unlike monolayer protected gold clusters where the alkanethiol ligands aredensely packed and aligned on the surface. For sensing or recognition, the likelihoodof binding to a target molecule or assembling with other peptide functionalizednanoparticles are increased, since binding is dependent a lot upon the structure of thepeptide epitope (antibody-antigen binding) in addition to its amino acid sequence.


35


Problem Explanation Potential Solutions

Purification of nanoshells If particles are centrifuged fasterthan 450 rcf, they will irreversiblyaggregate.

Centrifuge lower than 450 rcf.

Assembly of QD-Au hybrids Results in a distribution of struc-tures (1+1, 1+2, and 1+4).

Separate structures on acolumn or using a sucrosegradient.

Low peptidefunctionalization

Some peptides may have lowerfunctionalization efficiencies.

Incubate peptide withnanoparticle for longer time ormake modifications tosequence.

Precipitation of gold NPsinduced by metal ions

Incubation of metal ions withFlgA3 coated gold for longer than10 min. will cause ppt to form.

Incubate for less time.

2.8 Application Notes

The peptide-nanoparticle assemblies presented above are appealing as materials forchemical and biological sensing platforms, catalysis, biological labeling, actuation, opti-cal devices, thermal management, and as lubricants, to name a few. For example, theFlgA3 peptide coat of gold nanoparticles has a high affinity to metal ions (Hg2+, Pb2+, andAg+) and could serve as a colorimetric sensor for the detection of metal ions in toxicologyor the environment at ppb sensitivities. Other potential uses of peptide assemblednanoparticles include the lubrication of RF based MEMS switches [32], catalysts forhydrogenation and/or hydrodechlorination reactions, and photoinduced actuators.Notably as nanoparticle lubricants, the implementation of peptide assembled Au-Pdnanoparticles prevents shorting of the MEMS switch while significantly increasing itsdurability, performance, and the length of operation over 106 cycles [32].

2.9 Summary Points

Peptides offer excellent control for nanoparticle assembly via an assortment of highlyspecialized interactions. Consequently, these ultimately determine the final structure(extended or discrete), geometry, and properties (optical, mechanical, and electronic) ofthe assembled material whether it be enhancement or suppression of an individualproperty. For a given material, the peptide interface also affords a means to modulateproperties by exploiting the responsive nature of peptides to various stimuli as demon-strated by the disassembly of coil-coil gold structures. In total, the peptides presentedabove represent a fraction of potential interfaces, but highlight their importance incontrolling assembly.

Acknowledgments

We thank the Air Force Office of Scientific Research for funding.


36

References

[1] Hayat, M.A., Colloidal Gold: principles, Methods, and Applications, San Diego, CA: Academic Press,1989.

[2] Toshima, N. and Yonezawa T., “Bimetallic Nanoparticles-Novel Materials for Chemical and Physi-cal Applications,” New Journal of Chemistry, Vol. 22, 1998, pp. 1179–1201.

[3] Xie, J., Lee, J.Y., Wang, D.I.C., and Ting, Y.P., “Identification of Active Biomolecules in theHigh-Yield Synthesis of Single-Crystalline Gold Nanoplates in Algal Solutions,” Small, Vol. 3, 2007,pp. 672–682.

[4] Dickerson, M.B., Sandhage, K.H., and Naik, R.R., “Protein- and Peptide-Directed Syntheses of Inor-ganic Materials,” Chemical Reviews, Vol. 108, 2008, pp. 4935–4978.

[5] Crookes-Goodson, W.J., Slocik, J.M., and Naik, R.R., “Bio-Directed Synthesis and Assembly ofNanomaterials,” Chemical Society Reviews, Vol. 37, 2008, pp. 2403–2412.

[6] Katz, E., and Willner, I., “Integrated Nanoparticle-Biomolecule Hybrid Systems: Synthesis, Proper-ties, and Applications,” Angewandte Chemie International Edition, Vol. 43, 2004, pp. 6042–6108.

[7] Sarikaya, M., Tamerler, C., Candan, S., Daniel, T., Baneyx, F., “Materials Assembly and FormationUsing Engineered Polypeptides,” Annual Review of Materials Research,” Vol. 34, 2004, pp. 373–408.

[8] Li, M., Wong, K.K.W., and Mann, S., “Organization of Inorganic Nanoparticles Using Bio-tin-Stretavidin Connectors,” Chemistry of Materials, Vol. 11, 1999, pp. 23–26.

[9] Storhoff, J.J., and Mirkin, C.A., “Programmed Materials Synthesis with DNA,” Chemical Reviews,Vol. 99, 1999, pp. 1849–1862.

[10] Kanaras, A.G., Wang, Z., Hussain, I., Brust, M., Cosstick, R., and Bates, A.D., “Site-Specific Ligationof DNA-Modified Gold Nanoparticles Activated by the Restriction Enzyme Styl,” Small, Vol. 3,2007, pp. 67–70.

[11] Uchida, M., Klem, M.T., Allen, M., Suci, P., Flenniken, M., Gillitzer, E., Varpness, Z., Liepold,L.O., Young, M., and Douglas, T., “Biological Containers: Protein Cages as MultifunctionalNanoplatforms,” Advanced Materials, Vol. 19, 2007, pp. 1025–1042.

[12] Berry, V., and Saraf, R.F., “Self-Assembly of Nanoparticles on on Live Bacterium: An Avenue to Fab-ricate Elecrtonic Devices,” Angewandte Chemie International Edition, Vol. 44, 2005, pp. 6668–6673.

[13] Banerjee, I.A., Yu, L., and Matsui, H., “Cu Nanocrystal Growth on Peptide Nanotubes byBiomineralization: Size Control of Cu Nanocrystals by Tuning Peptide Conformation,” Proceedingsof the National Academy of Sciences, U.S.A., Vol. 100, 2003, pp. 14678–14682.

[14] Stevens, M.M., Flynn, N.T., Wang, C., Tirrell, D.A., and Lnger, R., “Coiled-Coil Peptide-BasedAssembly of Gold Nanoparticles,” Advanced Materials, Vol. 16, 2004, pp. 915–918.

[15] Slocik, J.M., Zabinski, J.S., Phillips, D.M., and Naik, R.R., “Colorimetric Response of PeptideFunctionalized Gold Nanoparticles to Metal Ions,” Small, Vol. 4, 2008, pp. 548–551.

[16] Slocik, J.M., and Naik, R.R., “Biologically Programmed Synthesis of Bimetallic Nanostructures,”Advanced Materials, Vol. 18, 2006, pp. 1988–1992.

[17] Banerjee, I.A., and Regan, M.R., “Preparation of Gold Nanoparticle Templated GermaniaNanoshells,” Materials Letters, Vol. 60, 2006, pp. 915–918.

[18] Slocik, J.M., Govorov, A.O., and Naik, R.R., “Optical Characterization of Bio-Assembled HybridNanostructures,” Supramolecular Chemistry, Vol. 18, 2006, pp. 415–421.

[19] Templeton, A.C., Wuelfing, W.P., Murray, R.W., “Monolayer-Protected Cluster Molecules,”Accounts of Chemical Research, Vol. 33, 2000, pp. 27–36.

[20] Pham, T., Jackson, J.B., Halas, N.J., Randal Lee, T., “Preparation and Characterization of GoldNanoshells Coated with Self-Assembled Monolayers,” Langmuir, Vol. 18, 2002, pp. 4915–4920.

[21] Slocik, J.M., Tam, F., Halas, N., and Naik, R.R., “Peptide Assembled Optically ResponsiveNanoparticle Assemblies,” Nano Letters, Vol. 7, 2007, pp. 1054–1058.

[22] Chao, H., Bautista, D.L., Litowski, J., Irvin, R.T., and Hodges, R.S., “Use of a HeterodimericCoiled-Coil System for Biosensor Application and Affinity Purification,” Journal of ChromatographyB, Vol. 715, 1998, pp. 307–329.

[23] Litowski, J.R., and Hodges, R.S.,“Designing Heterodimeric Two-Stranded α-Helical Coiled-Coils.Effects of Hydrophobicity and α-Helical Propensity on Protein Folding, Stability, And Specificity,”Journal of Biological Chemistry, Vol. 4, 2002, pp. 37272–37279.

[24] Lee, J., Govorov, A.O., Dulka, J., and Kotov, N.A., “Bioconjugates of CdTe Nanowires and GoldNanoparticles: Plasmon-Exciton Interactions, Luminescence Enhancement, and CollectiveEffects,” Nano Letters, Vol. 4, 2004, pp. 2323–2330.

[25] Oh, E., Hong, M.-Y., Lee, D., Nam, S.-H., Yoon, H.C., and Kim, H.-S., “Inhibition Assay ofBiomolecules Based on Fluorescence Resonance Energy Transfer (FRET) Between Quantum Dotsand Gold Nanoparticles,” Journal of the American Chemical Society, Vol. 127, 2005, pp. 3270–3271.

Acknowledgments

37

[26] Knecht, M.R., Weir, M.G., Frenkel, A.I., and Crooks, R.M., “Structural Rearrangement of BimetallicAlloy PdAu Nanoparticles Within Dendrimer Templates to Yield Core/Shell Configurations,”Chemistry of Materials, Vol. 20, 2008, pp. 1019–1028.

[27] Garcia-Gutierrez, D.I., Gutierrez-Wing, C.E., Giovanetti, L., Ramallo-Lopez, J.M., Requejo, F.G.,Jose-Yacaman, M., “Temperature Effect on the Synthesis of Au-Pt Bimetallic Nanoparticles,” Jour-nal of Physical Chemistry B, Vol. 109, 2005, pp. 3813–3821.

[28] Chen, C.-H., Sarma, L.S., Chen, J.-M., Shih, S.-H., Wang, G.-R., Liu, D.-G., Tang, M.-T., Lee,J.-F., Hwang, B.-H., “Architecture of Pd-Au Bimetallic Nanoparticle in SodiumBis(2-ethylhexyl)sulfosuccinate Reverse Micelles as Investigated by X-Ray Absorption Spectros-copy,” ACS Nano, Vol. 1, 2007, pp. 114–125.

[29] Slocik, J.M., Stone, M.O., and Naik, R.R., “Synthesis of Gold Nanoparticles Using MultifunctionalPeptides,” Small, Vol. 1, 2005, pp. 1048–1052.

[30] S.J. Lippard and J.M. Berg, Principles of Bioinorganic Chemistry, University Science Books: Mill Valley,CA, 1994.

[31] Slocik, J.M., Moore, J.T., and Wright, D.W., “Monoclonal Antibody Recognition of Histidine-RichPeptide Encapsulated Nanoclusters,” Nano Letters, Vol. 2, 2002, pp. 169–173.

[32] Patton, S.T., Slocik, J.M., Campbell, A., Hu, J., Naik, R., and Voevodin, A.A., “BimetallicNanoparticles for Surface Modification and Lubrication of MEMS Switch Contacts,”Nanotechnology, Vol. 19, 2008, pp. 405705


38

C H A P T E R

3Nanoparticle-Enzyme Hybrids as BioactiveMaterials

Xiaodong Tong, Songtao Wu and Ping Wang*

Biotechnology Institute and Department of Bioproducts and Biosystems EngineeringUniversity of Minnesota, St Paul, MN 55108

*Corresponding author

39

Key terms bioactive nanoparticlesenzyme immobilizationbiocatalysisbiotechnologycomposite materials

Abstract

Since the large-scale application of immobilized enzymes in 1960s, there havebeen substantial R&D efforts to optimize their structures for better catalytic effi-ciency. The unique properties and behaviors of nanostructured materials madeit possible for the development of a new class of biocatalyst systems that differfrom traditionally immobilized enzymes in terms of preparation, catalytic effi-ciency, and application potentials. Beyond their high surface area-to-volumeratios, nanoscale biocatalysts system also offer some unique features such asBrownian motion of nanoparticles, high degree of curvature of nanofibers andspatial confining effect of nanopores, which bring about both opportunities fordevelopment and new phenomena for understanding. This chapter reviewsmethodologies for preparation of nanoparticle-enzyme hybrid materials,which are probably the most extensively examined nanoscale structures withbioactivities.

3.1 Introduction

Enzymes are proteins that catalyze reactions of organic matters. Their unparalleled spec-ificity, high efficiency, and mild reaction conditions make them particularly advanta-geous over traditional chemical catalysts for many applications. Immobilized enzymesare highly preferred for most industrial applications because they promise easy catalystrecycling and product purification, continuous operation, and high thermal and opera-tional enzyme stabilities. However, the performance of immobilized enzymes has to beimproved substantially for many large-scale bioprocessing applications to competeagainst traditional chemical catalyst processes [1, 2].

One approach to improve the efficiency of immobilized enzymes is to manipulatethe structure of carrier materials. Reducing the size of carrier materials can provide largersurface area per unit mass for enzyme attachment, and shorten the diffusion path forsubstrates and products [3–7]. In this regard, nanostructured materials provide the upperlimits in balancing the key factors that determine the efficiency of biocatalysts, includ-ing surface area, mass transfer resistance, and effective enzyme loading. For example, theeffective enzyme loading can reach over 10% (wt/wt) with particles smaller than 100 nmwith monolayer attachment [8]. In addition to polymeric materials, nanoparticles madeof silica, magnetite, and gold have also been applied for biocatalysis [9–12]. In thischapter, we review the preparation of such nanoparticle-enzyme hybrids with a focuson three types of nanoparticles: (1) hydrophobic polymeric nanoparticles with sur-face-attached enzymes, (2) hydrophilic hydrogel particles with enzymes physicallyentrapped, and (3) magnetic nanoparticles with shell coatings for enzyme attachment.

3.2 Materials

Most of the key chemicals were purchased from Sigma-Aldrich Chemical Co. (St. Louis,MO), including ammonium persulfate, ammonium hydroxide (28%), α-chymotrypsin(α-CT) from bovine pancreas, Bradford reagent, divinylbenzene (DVB), iron (II)chloride, iron (III) chloride, dimethyl sulfoxide (DMSO), docusate sodium salt (AOT),n-succinyl-ala-ala-pro-phe-p-nitroanilide (SAAPPN), n-acetyl-L-phenylalanine ethylester (APEE), Oleic acid, polyvinylpyrrolidone (PVP, MW 29 kDa),(3-Mercaptopropyl)-trimethoxysilane (MPTOS), and tetraethyl orthosilicate (TEOS).Succinimidyl 4-[N-malemidomethyl] cyclohexane-1-carboxylate) (SMCC) was providedby Fisher Scientific. 2, 2-V-Azobis [2-methyl-N-(2-hydroxyethyl) propionamide](VA-086) was provided by Wako Chemicals USA, Inc. (Richmond, VA). 2-Sulfoethylmethacrylate (2-SEM) was purchased from Monomer-Polymer & Dajac Labs, Inc.(Feasterville, PA). N-Acryloxysuccinimide (NAS) was obtained from Acros Organics (Bel-

gium). α-Amylase (KLEISTASE SD80) and lipase PS were kindly provided as gifts fromAmano Enzyme Inc. 30% (w/v) Acylamide/Bis (19:1) solution and N, N, N’,N’-tetramethyleethylenediamine (TEMED) were the products of Bio-Rad Laboratories(Hercules, CA). Styrene, ethanol (HPLC grade), and n-propyl alcohol (n-PrOH, HPLCgrade) were obtained from EM (Gibbstown, NJ). Unless specially mentioned, all otherreagents and solvents used in the experiments were of the highest grade commerciallyavailable.

Nanoparticle-Enzyme Hybrids as Bioactive Materials

40

3.3 Methods

3.3.1 Enzyme-Attached Polystyrene Nanoparticles

Polystyrene nanoparticles have been prepared via an emulsion polymerization pro-cess [13, 14]. Typical procedures included the preparation of a stock solution of apolymerizable surfactant, 2-sulfoethyl methacrylate (2-SEM) by dissolving 5g of 2-SEMinto 50g DI water. The stock solution was diluted to 100g using DI water while pH of thesolution is adjusted to 3.5 by using a solution of sodium hydroxide (10 wt-%). The emul-sion solutions were prepared by dissolving certain amount of NAS (ranging from 98–196mg) in the mixture of styrene (0.6 ~ 1.2 ml) and DVB (8.2 ~ 16.0 μl) in a 20-ml scintilla-tion vial, followed by mixing with the aqueous phase, which contains the stabilizer(PVP, up to 5.5 mg/ml), ethanol (0.125 ~ 0.50 ml/ml), and 2-SEM (25 ~ 75 μl/ml). Thepolymerization reaction was initiated by adding 50 mg of VA-086 under N2 (1 min) andheating the system to 70ºC in a water bath with stirring. The reaction was stopped after10 hours and the particles were then washed with ethanol and DI water in a stirredultrafiltration cell with a polyethersulfone membrane (cut-off MW: 300 kDa) for at least6 hours. The yield of polystyrene nanoparticles could reach over 90%.

Variation in recipes and emulsion conditions have been performed to prepare parti-cles with sizes ranging from 0.1 ~ 1 μm. In particular, the diameter of polystyrenenanoparticles could be effectively controlled by varying volume ratio of water and oilphases, and higher w/o ratios generally led to smaller particles.

The enzyme can be covalently attached onto polymeric particles via the couplingreaction between the succinimide ester group of NAS and amino groups of enzyme. TheNAS groups are expected to be exposed to the outer surface of the particles due to theirhydrophilicity. Typically, 100 mg of α-amylase was dissolved in 2 ml of PBS (pH6.0,0.1M), and the insoluble power was removed by a 0.22-μm syringe filter. Then, 100 mgpre-cleaned nanoparticles were dispersed into 0.1 M pH 6.0 phosphate buffer at a con-

centration of 50 mg/ml and are then mixed with the α-amylase solution in a 20-ml glassvial. The reaction mixture was stirred at 4oC for 10 hours. pH 6 has been reported in liter-ature as the optimum condition for the enzyme attachment reaction and it was alsoinhibitory to the hydrolysis reaction of the functional group of NAS [8, 15]. The result-ing enzyme attached particles were purified and washed using an ultrafiltration unitwith a membrane of cutoff Mw of 300 kDa for at least 3 hours. Between filtration steps,45 ml of fresh pH 7.8 buffer (0.05M phosphate) containing 0.2 M NaCl was used to washthe particles with 20-minute sonication (using Branson 5510, Brandon). The particleswere washed for at least 5 times till the filtrate solution showed no absorbance at 280nm. About 120 mg α-amylase attached polystyrene nanoparticles were recovered, andthen stored at 4oC for further use.

3.3.2 Polyacrylamide Hydrogel Nanoparticles for Entrapment of Enzymes

Enzymatic polyacrylamide hydrogel beads can be prepared via reverse emulsion poly-merization reaction [16]. The continuous organic phase was prepared by dissolving 3.97g of AOT into 40 ml of toluene. Ten μl of TEMED was then added and stirred by a mag-netic stirrer for 10 min at 4oC under nitrogen. The aqueous phase was prepared by add-ing 1.2 ml of 20% acrylamide/bis (19:1) solution containing 2.5 mg α-chymotrypsin(α-CT) and 0.1 ml of 10% ammonium persulfate. After being purged with N2 for 10 min-

3.3 Methods

41

utes, the aqueous phase was slowly added into organic phase in droplets within 15 min.The polymerization reaction was allowed to continue for additional 2h. About 1gpolyacrylamide particles were thus obtained, and those nanoparticles were screened bytwo different sizes of filtration membranes, and the particles between the desired diame-ter of 400 to 1,000 nm were recovered. The particles were then washed three times with50 ml of toluene followed by three washes with DI water. The particles were put in afume-hood overnight to remove the residual organic solvent. The cleaned particles werestored in a 20-ml glass vial at 4oC for further use.

3.3.3 Magnetic Nanoparticles with Porous Silica Coating for EnzymeAttachment

A coprecipitation method was employed to prepare nano-sized Fe3O4 particles [17]. Afresh aqueous solution of ferric chloride (FeCl3) (15 ml, 0.38 g) was first prepared with DIwater (degassed) via sonication. The solution was filtered with a 0.22-μm syringe filter.The solution was then mixed with a fresh ferrous chloride (FeCl2) (15 ml, 0.24g) solu-tion, the mixture was added into a glass reactor of 120 ml. The solution was stirredmechanically at 1,500 rpm and 80oC under N2. A solution of 28% ammonium hydroxidewith a total volume of 6 ml was slowly added into the reaction system, followed by theaddition of 0.9 ml of oleic acid. During this procedure, the color of the suspension solu-tion changed gradually from red to black. The black suspension was continuously stirredat 80oC for 30 min. About 0.5g magnetic nanoparticles were obtained and washed withDI water for three times before being recovered with a magnet.

Magnetic nanoparticles prepared through the abovementioned procedure werefurther coated with silica for enzyme immobilization. Typically, 30 ml of toluene con-taining 1.5g of AOT was mixed with 1.5 ml of magnetic fluid containing 0.3g of mag-netic nanoparticles in a glass vessel. Subsequently, 1.5 ml of silica solution (TEOS:MPTOS=10:1, v/v) and 3 ml of 28% ammonium hydroxides were added into the glassvessel. The solution was mechanically stirred at 300 rpm for 4 hour. About 1g magneticspheres were recovered by using a magnet and were then washed sequentially with etha-nol, 50% ethanol, DI water, and PBS (0.1 M, pH7.0). Each wash lasted 4 hours in theshaker at 300 rpm. The clean particles were stored at room temperature.

An enzyme, lipase PS, was covalently attached to the silica-coated magnetic particlesby first functionalizing the silica coating with SMCC. Briefly, 5 ml of 2 mg/ml SMCC inPBS (0.1 M, pH7.0) containing 1 ml of DMSO was incubated with a solution of 20 mg/ml

lipase PS of the same volume at 4oC for 2 hours. One hundred μl of precleaned magneticporous silica particles was adjusted to 10% solid content, and was then mixed with 1 mlof 3 mg/ml DTT for 1 hour. The particles were washed with a PBS buffer solution (pH 7.0,0.1 M), and were then mixed with SMCC-modified lipase PS. The coupling reaction wasallowed to last for 12 hours at 4oC. The particles were further washed using fresh buffersolution and stored in desired buffer at 4oC for further use.

3.3.4 Enzyme Loading and Activity Assay

Protein loadings of nanoparticle–enzyme hybrids were determined by the reverse biuretmethod [18]. Two analytical reagent solutions were first prepared and stored at roomtemperature: reagent A was prepared by dissolving 15 mg CuSO4 · 5H2O, 45 mg of potas-


42

sium sodium tartrate, and 2.4g NaOH into 100 ml of DI water, while reagent B was pre-pared by dissolving 25 mg of ascorbic acid and 37 mg of bathocuproinedisulfonic aciddisodium salt in 100 ml DI water. Typically, 50 ml of 10 mg of nanoparticle–enzyme

hybrid solution were added to 200-μl reagent A and the mixture was incubated at 37°Cfor 5 minutes. Then 1,000 μl of reagent B was added and incubation was continued foran additional 0.5 minutes. The solution was filtrated through 0.22-μm syringe filters andthe absorbance at 485 nm was measured for the filtrate on a Shimadzu UV-1601 UV–visspectrophotometer. The protein content was calculated using BSA calibration deter-mined through the same procedure except that the 50 μl of nanoparticle–enzyme hybridsolution was replaced with protein solutions of the same volume.

The hydrolytic activity of polystyrene α-amylase hybrid nanoparticles was measuredusing 1.0% (w/v) soluble potato starch as substrate in pH 6.9, 20 mM sodium phosphatebuffer containing 6.7 mM sodium chloride. In brief, 1 ml of potato starch solution wasmixed with 1 ml of 1 mg/ml polystyrene α-amylase hybrid nanoparticles in 20 ml ofglass vial. The mixture was incubated at room temperature for 3 minutes. Then, 1 ml ofcolor reagent solution containing 48 mM 3, 5-dinitrosalicylic acid solution, 0.25g ofsodium potassium tartrate, and 0.167 ml of 2 M NaOH was added into the mixture, andthus put into the boiling water bath for 15 minutes. The reaction was stopped by puttingthe vial into the ice bath. An additional 9 ml of DI water was used to dilute the coloredsolution, and the absorbance was recorded at 540 nm. One unit of α-amylase activity isdefined as liberating 1.0 mg of maltose from starch in 3 minutes at pH 6.9 at roomtemperature.

The hydrolytic activity of hydrogel entrapped α-CT was measured using SAAPPN assubstrate in pH 7.5, 0.1 M sodium acetate buffer. In a typical measurement, hydrogelparticles entrapped α-CT (~1–2 mg) were added into 5 ml of buffer in a 20-ml glass vialfor preswollen, and then was mixed with 25 μl of SAPPN stock solution (160 mM) atroom temperature under shaking at 200 rpm. Aliquots of 1 ml of the supernatant solu-tion were taken periodically to monitor the concentration of the hydrolysis product,p-nitroaniline, by following the absorbance at 410 nm. One unit of hydrogel entrappedα-CT is defined as the amount of enzyme hydrolyzing SAAPPN to produce absorbance

equivalent to 1.0 μmole of p-nitroaniline per minute at pH 7.5 at room temperature.The transesterification activity of α-CT in organic solvents was measured at room

temperature in hexane or isooctane containing 20 mM APEE and 0.5 M n-PrOH. The sol-vents received from the suppliers were stored with 3 Å molecular sieves for at least 24hours before being used. Typically, 5 mg of native CT or 50 mg of hydrogel enzyme wasadded to 10 ml of reaction solution to initiate the reaction. The reaction system wasshaken at 200 rpm, and the enzyme was removed by filtration using a 0.22-μm PTFEsyringe filter. The product concentration was monitored by using a gas chromatograph

equipped with a FID detector and a RTX-5 capillary column (0.25 mm × 0.25 μm × 10m,Shimadzu). A temperature gradient from 100 oC to 190oC at a heating speed of 20oC/min,followed by 5-minute retention at 190oC was used. The initial reaction rate for theformation of n-acetyl-L-phenylalanine propyl ester (APPE) was calculated using datacollected before the conversion reached 5%.

The immobilized lipase activity was measured by using 0.5% p-nitrophenylpalmitate (p-NPP) as substrate in ethanol. Typically, 200 mg of immobilized lipase wasadded to a mixture of 1 ml of 0.5% (w/v) p-NPP solution and 1 ml of 0.05 m PBS (pH 7.0).The mixture was incubated for 5 minutes at room temperature, at 200 rpm. The reaction

3.3 Methods

43

was terminated by adding 2 ml of 0.5 N Na2CO3 solution followed by centrifuging for 10minutes at 10,000 rpm. The supernatant of 0.5 ml was diluted 10-folds with distilledwater, and measured at 410 nm in a UV/VIS spectrophotometer. One unit of lipaseactivity was defined as the release of 1 mmol p-nitrophenol per minute under theexperimental conditions.

The transesterification reactions were conducted in 20-ml glass scintillation vialsunder shaking (200 rpm). Soybean oil and ethanol were mixed in a molar ratio of 1:12unless specified otherwise. Lipases magnetic silica hybrid particles were used with theconcentration of 20 mg/ml. The reaction was analyzed by monitoring changes in theconcentration of product. Typically, aliquots of 200-μl samples were taken from the well

mixed reaction medium and were centrifuged at 14,000 × g for 10 minutes. Two hun-dred microliters of supernatant was diluted four times with iso-octane and was analyzedby using gas chromatography (Shimadzu GC-17A, Kyoto, Japan) equipped with a RTX-5

column (0.25 mm × 0.25 μm × 10m, Shimadzu). The column temperature was kept at200°C, whereas the injector and detector were kept at 270°C. Ethyl oleate was chosen asbiodiesel standards for GC analyses. The conversion of transesterification reaction wascalculated in this work by taking the actual amount of ethyl ester produced and the the-oretical value calculated from initially added soybean oil.

3.4 Results

3.4.1 Polystyrene-Enzyme Hybrid Nanoparticles

Figure 3.1 illustrates the chemical route for synthesis of functionalized polystyreneand subsequent attachment of enzymes. Following the procedure as described earlier,polystyrene particles with average diameters of 150 and 300 nm were prepared. The SEM


44

CH2

n* C2H

C2H

N

O

O

OC

O

CH

CH2CH2CH2

CHCHCH

CH2

CH

70°C, 10h

Polymerization

4°C, pH 6.0

Enzyme

CH2

n*CH

CH2CH2CH2

CHCHCH

N

O

O

OC

O

2H C HC

Figure 3.1 Synthetic route for functionalized polystyrene and subsequent attachment of enzyme.

image of the 300-nm particles is shown in Figure 3.2. A model enzyme, α-amylase, wascovalently attached to the nanoparticles with protein loadings of up to 2.4% (wt/wt)with the 150-nm particles. An average enzyme loading of 0.6% (wt/wt) was realized withthe particles of 300 nm with the activity of 0.92 U/mg solid.

Unlike large-sized solid materials, nanoparticles dispersed in a solution are mobile inform of Brownian motion. In other words, the particles functioned as “nanomotors”carrying the enzymes moving around in the reaction media. It was found very interest-ingly that the reactions catalyzed by enzymes may also drive the motion ofnanoparticles (i.e., improve the mobility of nanoparticles) [19–21]. In a recently pub-lished work, the relationship between particle mobility and the activity of the carryonenzymes was examined through experimental measurements and theoretical modelingusing polystyrene nanoparticles ranging from 0.1 ~ 1 μm [13]. It was revealed that thecatalysis with nanoparticle-supported enzymes point to a transitional region betweenthe homogeneous catalysis with free enzymes and the heterogeneous catalysis withimmobilized enzymes.

3.4.2 Polyacrylamide Hydrogel Nanoparticles with Entrapped Enzymes

In most cases, time-dependent activity loss of enzymes is caused by conformationalchanges of the enzymes. Immobilization usually stabilizes enzymes since con-formational changes of enzymes become difficult once the enzymes are attached to thesurfaces of solid materials. It was also believed that multi-point covalent bonding ofenzymes to polymeric materials could improve the enzymes’ stability with much higherdegree than that achieved with physical adsorption [22–24]. It seems reasonable toimagine that confining an enzyme molecule into a space of comparable size may limitthe surrounding 3-D environment available for the enzyme to undergo unfolding, thusproviding a mechanism of enzyme stabilization that is different from what is involved inmacroscopic materials [7]. Except that enzymes were attached into the surfaces of

3.4 Results

45

Figure 3.2 SEM image of enzyme-carrying polystyrene nanoparticles.

nanostructures, another way to develop nanoscale biocatalyst is to entrap enzymes intonanopores. Mesoporous silica gels have been used to host enzymes via both physicaladsorption [25] and chemical binding [7], as shown in Figure 3.3. For example, it wasestimated that the half-life of such entrapped CT could reach up to 143 days, a bigimprovement from the one-day half-life of free CT. Through careful synthetic routes,enzymes were also entrapped into cores of discrete polymeric [24] and silica [26] parti-cles. More sophisticated structures, such as porous materials hosting enzyme-carryingnanoparticles [27] and cross-linked enzyme aggregates [28] have been developed byapplying enzyme modification and fabrication procedures.

Hydrogels are widely applied as entrapment materials to provide the nanoporousstructure for biocatalysis. Variations in the concentration, fraction, and functionality ofmonomer and crosslinker used in three-dimensional aqueous hydrogel would changethe gel structure and its porosity. For example, a higher percentage of cross-linker willlead to clumping during polymerization. Polyacrylamide hydrogel microbeads withdiameters ranging from 600 to 800 nm have been synthesized following the procedureas described earlier. The pore size of polyacrylamide is usually reported as a wide arrayfrom 20 to 200 nm according to different preparation recipes [29]. We estimated thepore size of hydrogel prepared in this work to be around 26 nm according to a methodprepared previously [30]. The hydrogel entrapped a-CT were found active (~ 1.4 U/g par-ticles with the protein loading of 7.8 mg protein/g particles) and stable for biocatalyticapplications, the activity remained as high as 80% after a couple of weeks’ storage. Itindicated that the confinement of polyacrylamide structure could prevent the confor-mation changes of protein against the surrounding environments. In addition, during a24-hour transesterification reaction in anhydrous isooctane with APEE and n-PrOH, theα-CT hydrogel particles was almost 2,000 times more productive than native CT.

3.4.3 Magnetic Nanoparticles for Enzyme Attachment

Although nanoparticles have demonstrated promising potentials in revolutionizing thepreparation and use of biocatalysts, it is difficult to recover nanoparticle catalysts fromreaction media. Magnetic nanoparticles were therefore examined for enzyme immobili-


46

(a) (b)

Figure 3.3 Enzyme entrapped in mesoporous silica gels. (a) Physical entrapment, and (b) entrapmentwith chemical binding.

zation [9, 22, 31, 32]. Figure 3.4 shows the TEM photographs of the magnetic nano-particles prepared according to the procedure described in this chapter. The diameter ofthe particles was read to be 15 nm from the picture. Lipase PS was attached onto themagnetic nanoparticles with porous silica coatings. A typical lipase PS loading of 190units per gram of magnetic silica particles was achieved. The enzyme appeared to behighly stable in that no activity loss was observed within one month storage. Thenanoparticle-attached lipase was applied successfully and repeatedly to catalyze the con-version of soybean oil to ethyl ester in the presence of ethanol.


Nanoparticles represent the physical limit in reducing the size of carrier materials forenzyme immobilization, and certainly provide the upper limit of the efficiency ofimmobilized enzymes. However, we are still at the beginning in using nanoparticles forbiocatalysis. In addition to the promising features of nanoparticles, an interesting sub-ject arises considering the unique physical behaviors of nanoparticles. Unlike large-sizedsolid materials, nanoparticles dispersed in a solution are mobile in form of Brownianmotion (dynamic thermal vibration) as small molecules. In that sense, the enzymesattached to the nanoparticles are not “immobilized.” On the other hand, according toEinstein-Stokes equation, the mobility (or diffusivity) of the nanoparticles has to besmaller than that of native free enzymes due to their relatively larger size. This mobilitydifference points to an interesting transitional region between the homogeneous cataly-sis with free enzymes and the heterogeneous catalysis with immobilized enzymes(Figure 3.5). To date, little is known about this transitional region [13].


47

Figure 3.4 TEM image of magnetic nanoparticles.

As the size of the biocatalyst increases from several nanometers as that of the nativeproteins to big particles visible to naked eyes, the mobility (or self diffusivity, D), drivenby Brownian motion (thermal vibration), also decreases to zero. It is generally acceptedthat the particle size determines the mobility through the Einstein-Stokes equation:

Dk T

rB=

6πη

where kB is Boltzmann constant, the viscosity of the solution, and r the radius of theparticle. Collision theory accounts the effects of mobility and size of both the catalystsand reagents on the reaction rates. According to the theory, the substrate and theenzyme must first collide and the product is then released. The rate constant predictedfrom collision theory, as usually referred as diffusion-limited reaction rate constant(from hereon, the term “collision-limited” is used for the same meaning as “diffu-sion-limited” in this proposal), for a bimolecular reaction can be expressed as:

( )k Zpe E RTcoll = −

where Z is the frequency of collisions, p a steric factor, and E the activation energy. Tak-ing the reacting species as spherical particles, Z can be expressed as:

( )( )Z

N D D r rAvo A B A B=+ +4

1000

π

Combining the above equations, the collision frequency may be calculated as:

( )Z

RT r r

r rA B

A B

=+⎛

⎝⎜⎜

⎞

⎠⎟⎟

23000

2

η

These equations correlate the collision-limited reaction rate constant to both the sizeof catalysts and the viscosity of the reaction media. A reaction system consisting a sub-strate with radius of 0.8 nm (~ Mw 400) and an enzyme of 2.2 nm in radius (the size of

α-chymotrypsin) will give a collision frequency of 9.5 × 109 M–1 S–1 at 298 K in water.


48

Free Enzymer = ~nm, D = ~10 cm S

−7 −12Nanopartical-attached enzymer < 1000 nm, decrease withD r

Immoblized enzyme

r , 0D∞ →(a) (b) (c)

Figure 3.5 (a–c) The transitional properties represented by nanoparticle biocatalysts in terms of size,mobility, and activity.

Many of the second-order rate constants (kcat/KM) of enzymatic reactions are found to beabout 108~9 M–1 S–1, and many more reactions are in the order of 105~7 M–1 S–1. Whether thereactions are subject to collision-limited may be largely determined by the values of pand E, to which experimental data is usually not available. Nevertheless, this modelpoints out the importance of particle size in determining the intrinsic activities ofnanoparticle-associated enzymes [13].

3.6 Troubleshooting

Enzyme activity is the primary concern in preparation of biocatalysts. There are stan-dard activity assay methods readily available for almost all the commercially availableenzymes, such as those from Sigma-Aldrich. One common method to verify the successin preparation of nanoparticle biocatalysts is to carry out those standard activity assaytests. One convenience for nanoparticle biocatalysts is that, if well dispersed, they pres-ent negligible interference for most photo-spectrometry analysis.

In addition to biochemical assays, a number of methods and devices that have beenroutinely practiced for nanoscale materials can be used to characterize the enzyme-car-rying nanoparticles for quality control. For example, nanoparticle morphology wasstudied with both scanning electronic microscopy (SEM) and transmission electronicmicroscopy (TEM). Study on enzyme-material interactions may benefit from infraredspectroscopy (FTIR) and nuclear magnetic resonance (NMR) analyses. Light scatteringanalysis can be applied to reveal solution behaviors including diffusivity and aggrega-tion status of enzyme-carrying nanoparticles.


The development of highly efficient biocatalysts has been driven largely by the increas-ing demands in biochemical analysis and “green chemistry.” It is expected thatnanotechnologies will generate revolutionary impacts to the course of advances in theseareas. Specific applications may range from environmental remediation, clinical bio-chemical analysis, immunoassay, therapeutic enzymes, organic synthesis, to biofuels,and bioenergy.

3.8 Summary Points

Nanoaprticles are an important class of materials promising a broad range of biotechno-logical applications. This chapter reviews the preparation of three typical structures ofenzyme-carrying nanoparticles, namely plastic, hydrogel and magnetic particles. In allthe cases examined, enzymes can be effectively loaded to the nanostructures and func-tion well under various conditions. Although the activity of the immobilized enzymesmay not be as high as free native enzymes, their improved stability and extended life-times combined with the easiness in catalyst recovering are expected to enable highlyefficient biocatalytic applications for both chemical processing and recognition.

3.6 Troubleshooting

49

Acknowledgments

The authors thank Dr. Hongfei Jia and Ms. Xueyan Zhao in helping reading the manu-script of this chapter.

References

[1] Caruana, C. M. “Enzymes Tackle Tough Processing,” Chem. Eng. Prog., Vol. 93, 1997, pp. 13–20.[2] Daubresse, C., Grandfils, C., Jerome, R., and Teyssie, P. “Enzyme Immobilization in Reactive

Nanoparticles Produced by Inverse Microemulsion Polymerization,” Colloid. Polym. Sci., Vol. 274,1996, pp. 482–489.

[3] Cantarella, M., Cantarella, L., and Alfani, F. “Entrapping of Acid Phosphatatse inPoly(2-Hydroxyethyl Methacrylate) Matrices: Preparation and Kinetic Properties,” Brit. Polym. J.,Vol. 20, 1988, pp. 477–485.

[4] Huang, F. C., Ke, C. H., Kao, C. Y., and Lee, W. C. “Preparation and Application of Partially PorousPoly(Styrene-Divinylbenzene) Particles gor Lipase Immobilization,” J Appl. Polym. Sci., Vol. 80,2001, pp. 39–46.

[5] Kajiwara, S., Maeda, H., and Suzuki, H. Enzyme Immobilization by Entrapment in a Polymer Gel Matrix,U.S. Patent 4,978,619, issued 1990.

[6] Martinek, K., Klibanov, A. M, Goldmacher, V. S, and Berezin, I. V. “The Principles of Enzyme Stabi-lization I. Increase in Thermostability of Enzymes Covalently Bound to a Complementary Surfaceof a Polymer Support in a Multipoint Fashion,” Biochim Biophys Acta, Vol. 485, 1977, pp. 1–12.

[7] Wang, P., Dai, S., Waezsada, S. D, Tsao, A. and Davison, B. H. “Enzyme Stabilization by CovalentBinding in Nanoporous Sol-Gel Glass for Nonaqueous Biocatalysis,” Biotechnol. Bioeng., Vol. 74,No. 3, 2001, pp. 249–255.

[8] Chen, J. P. and Su, D. R. “Latex Particles with Thermo-Flocculation and Magnetic Properties forImmobilization of Alpha-Chymotrypsin,” Biotechnol. Progr., Vol. 17, No. 2, 2001, pp. 369–375.

[9] Matsunaga, T. and Kamiya, S. “Use of Magnetic Particles Isolated from Magnetotactic Bacteria forEnzyme Immobilization,” Appl Microbial Biotechnol, Vol. 26, 1987, pp. 328–332.

[10] Shinkai, M., Honda, H. and Kobayashi, T. “Preparation of Fine Magnetic Particles and Applicationfor Enzyme Immobilization,” Biocatalysis Biotransform., Vol. 5, 1991, pp. 61–69.

[11] Crumbliss, A. L., Perine, S. C., Stonehuerner, J., Tubergen, K. R., Zhao, J. and Henkens, R. W. “Col-loidal Gold as a Biocompatible Immobilization Matrix Suitable for the Fabrication of Enzyme Elec-trodes By Electrodeposition,” Biotechnol. Bioeng., Vol. 40, 1992, pp. 483–490.

[12] Kondo, A., Murakami, F. and Higashitani, K. “Circular Dichroism Studies on ConformationalChanges in Protein Molecules Upon Adsorption on Ultrafine Polystyrene Particles,” Biotechnol.Bioeng., Vol. 40, 1992, pp. 889–894.

[13] Jia, H., Zhu, G. and Wang, P. “Catalytic Behaviors of Enzymes Attached to Nanoparticles: TheEffect of Particle Mobility,” Biotechnol. Bioeng., Vol. 84, No. 4, 2003, pp. 406–414.

[14] Jia, H., Zhu, G., Vugrinovich, B., Kataphinan, W., Reneker, D. H and Wang, P. “Enzyme-CarryingPolymeric Nanofibers Prepared via Electrospinning for Use as Unique Biocatalysts,” Biotechnol.Progr., Vol. 18, No. 5, 2002. pp. 1027–1032.

[15] Anjaneyulu, P. S. R. and Staros, J. V. “Reactions of N-hydroxysulfosuccinimide Active Esters,” InternJ Peptide & Protein Res., Vol. 30, 1987, pp. 117–124.

[16] Daubresse, C., Grandfils, C., Jerome, R. and Teyssie, P. “Enzyme Immobilization in NanoparticlesProduced by Inverse Microemulsion Polymerization,” J Colloid Interface Science, Vol. 168, 1994,pp. 222–229.

[17] Tong, X., Xue, B. and Sun, Y. “A Novel Magnetic Affinity Support for Protein Adsorption and Purifi-cation,” Biotechnology Progress, Vol. 17, No. 1, 2001, pp. 134–139.

[18] Matsushita, M., Irino, T., Komoda, T. and Sakagishi, Y. “Determination of Proteins by a ReverseBiuret Method Combined with the Copper–Bathocuproine Chelate Reaction,” Clin. Chim. Acta.,Vol. 216, 1993, pp. 103–111.

[19] Lee, B. S., Lee, S. C. and Holliday, L. S. “Biochemistry of Mechanoenzymes: Biological Motors forNanotechnology,” Biomed. Microdevices, Vol. 5, No. 4, 2003, pp. 269–280.

[20] Blum, D. J., Ko, Y. H., Hong, S., Rini, D. A. and Pedersen, P. L. “ATP Synthase Motor Components:Proposal and Animation of Two Dynamic Models for Stator Function,” Biochem. Biophys. Res.Commun., Vol. 287, 2001, pp. 801–807.

[21] Wada, Y., Sambongi, Y. and Futai, M. “Biological Nano Motor, ATP Synthase FoF1: From Catalysisto rec10-12 Subunit Assembly Rotation,” Biochim Biophys Acta, Vol. 1459, 2000, pp. 499–505.


50

[22] Mozhaev, V. V., Melik-Nubarov, N. S., Sergeeva, M.V., Siksnis, V. and Martinek, K. “Strategy for Sta-bilizing Enzymes. Part One: Increasing Stability of Enzymes via Their Multi-Point Interaction witha Support,” Biocatalysis, Vol. 3, No.3, 1990, pp. 179–187.

[23] Mozhaev, V. V., Sergeeva, M. V., Belova, A. B. and Khmel’nitskii, Y. L. “Multipoint Attachment to aSupport Protects Enzyme from Inactivation by Organic Solvents: Alpha.-Chymotrypsin in Aque-ous Solutions of Alcohols and Diols,” Biotechnol. Bioeng., Vol. 35, No. 7, 1990, pp. 653–659.

[24] Mozhaev, V. V. “Mechanism-Based Strategies for Protein Thermostabilization,” TIBTECH, Vol. 11,1993, pp. 88–95.

[25] Takahashi, H., Li, B., Sasaki, T., Miyazaki, C., Kajino, T. and Inagaki, S. “Catalytic Activity inOrganic Solvents and Stability of Immobilized Enzymes Depend on the Pore Size and Surface Char-acteristics of Mesoporous Silica,” Chem. Mater., Vol. 12, No. 11, 2000, pp. 3301–3305.

[26] Kim, J. and Grate, J. W. “Single-Enzyme Nanoparticles Armored by a Nanometer-Scale Organic/Inorganic Network,” Nano Let., Vol. 3, No. 9, 2003, pp. 1219–1222.

[27] Phadtare, S., Vinod, V. P., Mukhopadhyay, K., Kurnar, A., Rao, M., Chaudhari R.V. and Sastry, M.“Immobilization and Biocatalytic Activity of Fungal Protease on Gold Nanoparticle-Loaded ZeoliteMicrospheres,” Biotechnol. Bioeng., Vol. 85, No. 6, 2004, pp. 629–637.

[28] Dohnalkova, A., Park, H. G., Chang, H. N., Wang, P., Grate, J. W. and Hyeon, T. “Simple Synthesisof Hierarchically Ordered Mesocellular Mesoporous Silica Materials Hosting Crosslinked EnzymeAggregates,” Small, Vol. 1, No. 7, 2005, pp. 744–753.

[29] Holemes, D. L. and Stellwagen, N. C. “Estimation of Polyacrylamide Gel Pore Size From FergusonPlots of Linear DNA Fragments II. Comparison of Gels with Different Crosslinker Concentrations,Added Agarose and Added Linear Polyacrylamide,” Electrophoresis, Vol. 12, 1991, pp. 612–619.

[30] Lina, Y.-Z., Lib, Y.-G., and Lua, J.-F., J Colloid Interface Sci., Vol. 251 No. 2, pp. 256–262.[31] Hutten, A., Sudfeld, D., Ennen, I., Reiss, G., Hachmann, W., Heinzmann, U., Wojczykowski,

K., Jutzi, P., Saikaly, W. and Thomas, G. “New Magnetic Nanoparticles for Biotechnology,” J.Biotechnol., Vol. 112, No. 1–2, 2004, pp. 47–63.

[32] El-Zahab, B., Jia, H. and Wang, P. “Enabling Multienzyme Biocatalysis Using Nanoporous Materi-als,” Biotechnol. Bioeng., Vol. 87, No. 2, 2004, pp. 178–183.

References

51

C H A P T E R

4Self-Assembled QD-Protein Bioconjugates andTheir Use in Fluorescence Resonance EnergyTransfer

Aaron R. Clapp,1 Hedi Mattoussi,2 and Igor L. Medintz3

1Department of Chemical and Biological Engineering, Iowa State University, Ames, IA 50014, Phone:515-294-9514, Fax: 515-294-2689, e-mail: [email protected]. Naval Research Laboratory, Center for Bio/Molecular Science and Engineering, Code 6900, 4555Overlook Avenue, SW, Washington, D.C. 203753U.S. Naval Research Laboratory, Optical Sciences Division, Code 5611, 4555 Overlook Avenue, SW,Washington, D.C. 20375

53

Abstract

Due to their unique size, chemical composition, and optical properties, quan-tum dots (QDs) provide a flexible platform for developing FRET-based applica-tions in biology. In this chapter we present methods for preparing watersoluble QDs, stably attaching biomolecules to their surface, and performingexperiments that utilize fluorescence resonance energy transfer as a signaltransduction mechanism. The protocols are presented in a generalized formatthat are applicable to a variety of potential uses including sensitive detection ofanalytes in solution, measuring enzymatic activity, quantifying distances, andobserving molecular rearrangements. Special considerations for using QDs asFRET donors are highlighted throughout including unique features of dataanalysis and interpretation.

Key terms quantum dotsfluorescence resonance energy transferorganic dyesbiosensingfluorescence spectroscopy

4.1 Introduction

Luminescent quantum dots (QDs) have, in many ways, revolutionized fluorescencetechniques used in biology due to their unique electronic and optical properties [1–5]. Inthe simplest example, QDs demonstrate uncommon resistance to bleaching effects dueto chemical and photo-induced degradation which vastly improves imaging lifetimeand fidelity. QDs are ideal fluorescent tags for more complicated multiplexing applica-tions where a single wavelength can excite numerous colors simultaneously by virtue oftheir broad absorbance spectra and large effective Stokes shifts. Many of the noveltiesafforded by QDs cannot be replicated with organic dyes and have thus only recentlybeen accessible to researchers. An intriguing aspect of QDs is their size, typically on theorder of a few nanometers in diameter. Clearly larger than the molecular scale, thissize range allows patterning of the nanocrystal with a variety of surface ligands andbiomolecules. Rather than acting merely as a fluorescent tag, a QD functions more like acentral scaffold that can accommodate an assortment of molecules. The surface canbe tailored for numerous applications such as receptor-mediated cell delivery andimmunolabeling.

Fluorescence has become the dominant imaging modality of optical microscopy inpart due to its ability to label, track, and report upon conditions within microscopic bio-logical samples in a way that minimally disrupts their native state [6]. The sensitivity offluorescence microscopy is such that even single molecules can be observed with relativeease. In some instances, fluorescence allows an investigator to probe interactions anddynamics occurring at length scales far below the diffraction limit. One such method isfluorescence (or Förster) resonance energy transfer (FRET), which capitalizes on theexchange of energy between the electrical dipoles of fluorophores [6]. A donor species inan excited electronic state can induce transfer of its energy to a nearby acceptor underfavorable conditions. Perhaps unsurprisingly, QDs are similar to molecular dyes and flu-orescent proteins in their ability to function as FRET donors and are exceptionally effi-cient at executing this process. There are a number of appealing consequences of FRETwith biological relevance, but a particularly useful application is the ability to deduceextremely small distances (with angstrom resolution) between donor and acceptor spe-cies. The nature of the dipole-dipole interaction ensures a strong proximity-driveneffect, and this can be used to, in a sense, “measure” distances on the molecular scale [7].Consequently, FRET provides the ability to detect subtle changes in biomolecular struc-ture and orientation and is invaluable for studying the dynamics and interactions ofproteins, oligonucleotides, and phospholipids in a manner that is minimally invasive.

Over the past several decades, FRET has been used widely as a powerful bioanalyticaltechnique [6, 8–11]. In most cases the donor and acceptor fluorophores are organic dyesthat are chemically attached to biomolecules or genetically engineered fluorescent pro-teins. In general, these fluorophores have provided excellent results in FRET experi-ments and have demonstrated the utility and flexibility of the technique to generateunique insights. However, there are certain limitations that inhibit their expanded use.Even in traditional fluorescent tagging experiments used to label cellular components,most fluorophores suffer from bleaching due to an unfavorable local chemical environ-ment or continuous excitation from a high-intensity source. Inherent limitations suchas low donor quantum yield and pH sensitivity can also be troublesome during experi-ments. However, these issues are common even in standard fluorescence experiments.In the specific context of energy transfer, it is imperative to have a well-matched donor

Self-Assembled QD-Protein Bioconjugates and Their Use in Fluorescence Resonance Energy Transfer

54

and acceptor pair such that there is substantial overlap between the emission andabsorption spectra. Matching donor emission with acceptor absorption is typicallystraightforward; however it is challenging to sufficiently isolate the excitation sourcefrom the emission signals (confined by the Stokes shift of the donor). The convolvedoptical signals received by the detector can lead to quantitative errors in the assess-ment of energy transfer efficiency, which readily propagates to other calculations likedonor-acceptor distance. Additionally, in most cases the experiment is limited topair-wise interactions where one donor interacts with one acceptor. The process canoccur between countless pairs throughout a sample, but if the average efficiency is poorwithin each of these interacting pairs, there is no simple way to increase the efficiency.

In addition to the well-known benefits of QDs as fluorescent tags, these advantagesextend to FRET applications as well [12–14]. Most notably, the emission maximum ofQDs varies with size, which can be precisely tuned to match the absorption spectrum ofa particular acceptor molecule. This level of control is unique to nanocrystals andexploits the quantum confinement effect of charge carriers. Due to their size, which issimilar to that of proteins, QDs can accommodate multiple biomolecules on their sur-face, which in the context of FRET, can vastly improve the energy transfer efficiencyfrom individual QDs. Multiple acceptors positioned near a central fluorescent donor cansignificantly extend the effective interaction distance beyond the characteristic Försterdistance [12]. This has important implications in biosensing where enhanced energytransfer efficiency leads to an improved signal. Distance measurements relevant to struc-tural biology (Figure 4.1) also benefit from the ability to obtain multiple estimates usingvarying numbers of acceptors surrounding the donor [15]. In some cases the nanocrystalsurface functions to regulate the orientation of biomolecules, which can extend thecapabilities and relevant information obtained in some FRET measurements.

4.1 Introduction

55

λ

λ

λ

ex

em, QD

em, dye

Biomolecule

Distan

ce

Ligands

Shell

Core

Dye

Figure 4.1 Schematic of a core-shell quantum dot bioconjugate with a labeled biomolecule (protein)bound to the nanocrystal surface. Fluorescence in the QD is induced by an excitation source withwavelength λex. In this arrangement, the excited state energy of the QD is transferred nonradiatively tothe dye label on the protein at a rate determined in part by the QD-dye spectral overlap and distance.The emission maxima of the (quenched) QD and dye are represented by λem,QD and λex,dye, respectively.

In this chapter, we will describe methods for generating stable QD bioconjugatescapable of FRET interactions with proximal organic dyes. Starting with organic-solubleluminescent QDs, the critical steps include surface ligand exchange, biomolecule conju-gation, fluorescence measurements, and data analysis. There are a variety of possibleuses, and thus we describe methods generally such that they can be adapted for a partic-ular application. In particular we highlight FRET-based distance calculations [13] andmethods for estimating enzymatic activity [16, 17]. Although QD-based FRET is some-what similar to interactions with dye pairs, there are several notable differences thatrequire careful consideration and further discussion. Also note that while we are primar-ily considering in vitro applications, many of the same considerations hold for experi-ments within live cells or in vivo.

4.2 Materials

4.2.1 Reagents

• Luminescent core-shell QDs dispersed in organic solvent;

• Thioctic acid (DHLA precursor);

• Sodium borohydride (NaBH4);

• Deionized or ultrapure water;

• Potassium tert-butoxide (K[t-BuO]);

• Millipore hydrophilic and organic filters (disposable);

• Scintillation vials (glass, screw cap, 20 mL);

• Ultra-free centrifugal filters (Millipore);

• Sodium tetraborate buffer (10 mM, pH 9.5, Sigma);

• Microcentrifuge tubes, 1.5 mL (Eppendorf);

• Protein/peptide with a charged (basic/acidic) attachment domain;

• Protein/peptide engineered with a terminal polyhistidine domain.

4.2.2 Equipment

• Rotovap for purification of ligands;

• Centrifuge for solution purification;

• UV-vis absorption spectrophotometer;

• Fluorescence spectrophotometer or fluorescence plate reader;

• Handheld UV lamp (λ = 365 nm).

4.3 Methods

4.3.1 Quantum Dot Synthesis

The protocols and experiments described in this chapter are based on the use of crystal-line CdSe core nanoparticles having narrow size distributions with fluorescence emis-sion maxima from 480 to 640 nm. There are now innumerable published methods for


56

synthesizing high-quality CdSe QDs describing the respective processes in explicit detail[18–23], so rather than provide a particular method for synthesis, we will limit ourdiscussion to an overview of these procedures and highlight factors that lead to theproduction of suitable materials. Most well-established methods react organometallicprecursors (e.g., CdO, Se) in a hot coordinating solvent mixture under a dry inert gasatmosphere of Ar or N2 as depicted in Figure 4.2. The CdSe QDs are typically purifiedfrom unreacted precursors and reintroduced into a fresh coordinating solvent bath. TheCdSe cores are then overcoated with multiple layers of a wider band gap material (oftenZnS or CdS) to produce highly luminescent core-shell QDs [24]. The shell layer also pro-vides a surface for attaching different ligands subsequent to particle synthesis. We willfocus on QDs having a ZnS shell throughout this chapter; however, there are many pos-sibilities available. The overcoating step proceeds in a manner similar to the coregrowth, but at a lower temperature and with a slower addition of precursors to promotethe addition of layers to existing nanocrystals rather than nucleation of new particles.Each growth stage is followed by an extended annealing period where the core andcore-shell particles are stirred at moderate temperatures overnight to reduce defects inthe nanocrystals. Core-shell QDs are extremely stable if stored in their original “growthsolution” containing a mixture of nanocrystals, coordinating ligands, organic solvents(e.g., hexane, toluene, butanol), and residual unreacted precursors. Eventually the sam-ples are purified first by centrifugation to remove excess metals, and then using a solventthat is miscible with the growth solution, but unfavorable to nanocrystal solubility. Oneexample is to add methanol to a growth solution with toluene and butanol. Eventuallythe added methanol promotes aggregation of the nanocrystals such that they readilyprecipitate from the sample. The nanocrystals can then be resuspended in a minimumvolume of nonpolar organic solvent (e.g., hexane). The quality of a preparation is evalu-ated by a number of analytical techniques including absorption spectroscopy (UV-vis),fluorescence spectroscopy, transmission electron microscopy (TEM), small angle x-ray

4.3 Methods

57

Processing

CdSe diameter (nm)

2.5 2.9 3.2 3.6 4.1

Figure 4.2 Schematic of the high temperature synthesis of luminescent CdSe QDs. The size-depend-ent emission is shown for five separate populations (in toluene) on the right where all samples are illu-minated by a single UV source at 365 nm.

scattering (SAXS), and powder x-ray diffraction (XRD). A full characterization includingall of the preceding methods is typically not necessary for most well-documented syn-thetic methods. Absorption and fluorescence spectroscopy provide information regard-ing size polydispersity, luminescence quantum yield, and concentration. TEM is oftenused as a quality control method to confirm crystallinity, size, and morphology of anew preparation; however, the remaining methods are only used if more detailedinformation is desired.

The choice of specific organometallic precursors, coordinating ligands, and reactionconditions will ultimately influence nanocrystal size, brightness, crystal structure, andaspect ratio, but most published protocols used to form binary core-shell nanocrystalsconsisting of CdSe-ZnS are substantially similar. Ideally, the QDs should be bright, crys-talline, stable, and nearly spherical. These characteristics are important for subsequentprocessing steps to render the nanocrystals water soluble and are necessary for successfulFRET investigations.

4.3.2 Surface Ligand Exchange

As prepared, most common nanocrystal preparations will have surface ligands that areinsoluble in polar media (see Figure 4.3). In order for QDs to be used in biological investi-gations, the surface chemistry must be altered to be hydrophilic. There are numerousmethods available that render QDs water-soluble, however not all of these are suitablefor QD-based FRET experiments. Förster theory predicts characteristic interaction dis-tances ranging from about 2 to 10 nm depending on the specific donor-acceptor pair [6],which in turn places a limit on the overall size of the nanocrystal (the sum of core radius,shell thickness, and effective ligand thickness). The theory considers the interaction oftwo point dipoles, which is an accurate depiction for molecular dyes, but perhaps lessappropriate for luminescent nanoparticles. However, it is a reasonable approximation toassume point dipoles positioned (on average) at the QD center of mass. If a nanocrystalcapping ligand is too large, it is possible this could prohibit energy transfer altogetherdue to distance considerations. Therefore, we will restrict our attention to relativelycompact ligands that do not increase the overall nanocrystal size significantly. This pre-cludes larger amphiphilic ligands, such as phospholipids, that bind to the existinghydrophobic coordinating ligands that remain following synthesis [25]. While this is auseful strategy for preserving a hydrophobic layer around the nanocrystal and thusmaintaining a high quantum yield, the overall size of these capping groups is usually toobulky for FRET applications. With these considerations in mind, an ideal ligand will


58

P P

O

NH2

Figure 4.3 Typical coordinating ligands bound to the surface of hydrophobic CdSe-ZnS QDs. Fromleft to right: tri-n-octylphosphine (TOP), tri-n-octylphosphine oxide (TOPO), 1-hexadecylamine (HDA).

have the following properties: minimal length, high affinity for the nanocrystal surface,electronic passivation of the QD, and stability over a wide range of pH and salt concen-trations. The premise is to exchange existing hydrophobic ligands that populate thenanocrystal surface with new hydrophilic ligands.

A variety of core-only and core-shell QDs can readily bind ligands possessing thiolgroups. This is a common strategy used to form stable interactions between the newligand and the QD surface. Monothiols are the most common where a single thiol groupbinds to the nanocrystal surface upon exchange with the native ligand. A wide variety ofligands are available that exploit this strategy including peptides containing the aminoacid cysteine and ligands such as mercaptoacetic (thioglycolic) acid. Regarding the lat-ter, the alkyl linker that separates the thiol and carboxyl groups can be varied to alter thelength and stability of the ligand, as depicted in Figure 4.4. A common capping ligand ismercaptoundecanoic acid (MUA), which contains an 11 carbon spacer region [26].Dithiols in their reduced form have also been used to bind ligands to nanocrystal sur-faces. This improves the affinity of the ligand for the QD surface and ensures better sta-bility. In particular, a reduced form of thioctic acid called dihydrolipoic acid (DHLA) hasproved to be an excellent ligand for stabililizing nanocrystals in aqueous solutions hav-ing a basic pH [20]. More recently there have been reports of DHLA derivatives thatexploit the terminal carboxyl group as a covalent attachment point for other chemicalligands [27, 28]. Using simple carbodiimide chemistry (and formation of an amidebond), DHLA can be modified with terminal amine, alcohol, PEG, and biotin functionalgroups among others.

The cap exchange process involves precipitating QDs from a growth solution ornonpolar solvent and decanting off the liquid. This leaves behind nanocrystals withnative capping ligands on their surface. A solution containing the desired cappingligand (either neat or dissolved in an organic solvent) is then added to the QD vial alongwith a magnetic stir bar. The vial is capped with a rubber septum, sealed, backfilled withan inert gas, and placed into an oil bath. The mixture is stirred vigorously at a moderatetemperature (well below the boiling point of the ligand solution) overnight to facilitateexchange of the native ligands with the new ligands in solution. The concentratedligand solution coupled with convective mixing and an elevated temperature enhancesthe process to the point that a nearly complete exchange results. The nanocrystals arenext purified and resuspended in water. Depending on the ligand, the aqueous solutionmay need to be maintained a particular pH to ensure long-term stability. As an example,a protocol for DHLA preparation and ligand exchange is provided below. The ligandexchange details are largely similar for other thiolated molecules which have an affinityfor nanoparticle surfaces.

4.3 Methods

59

O

n

SHOH

OH

SHHS

OH

O

O

H N2

SH HS

HS

OHOH

Figure 4.4 Examples of mono- and dithiol ligands used for creating water-soluble CdSe-ZnS QDs.From left to right: mercaptoaceitc acid (where n = 1), dithiothreitol (DTT), dihydrolipoic acid (DHLA),cysteine (Cys).

4.3.2.1 DHLA Preparation

1. Prepare a 1L stock solution of 0.25M sodium bicarbonate.

2. Prepare a cold bath consisting of an ice water mixture in a flat bottom glass containerset on a magnetic stir plate. Mount a round bottom flask in the cold bath and add amagnetic stir bar to the flask. Add 1g of thioctic acid per 20 mL of the 0.25M sodiumbicarbonate solution to the flask. A standard preparation uses about 5 to 6g ofthioctic acid. We will subsequently consider a preparation using 6g of thioctic acid.

3. Begin continuously stirring the mixture vigorously with the stir bar.

4. Slowly add 1.4g of fresh NaBH4 to the stirring mixture. The additions should be madein steps consisting of about 20 mg of NaBH4 each.

5. Let the mixture stir for 2 hours or until it turns clear/cloudy white. Note: The qualityof the added NaBH4 is critically important to fully reducing the thioctic acid. If theeffectiveness of the reduction step is in question (e.g., the sample does not turnclear), replace the NaBH4 with fresh stock.

6. Add 100 mL of toluene to the mixture. This results in a two-phase mixture.

7. Acidify the aqueous phase to ~ pH 1. This can be checked using a glass pipet toextract a small sample and blotting on pH paper.

8. The reduced thioctic acid should transfer fully into the organic phase having a whitemilky appearance.

9. Collect the organic phase (containing the reduced product) using a separatoryfunnel.

10. Add magnesium sulfate drying agent to remove excess water. The solution shouldbecome clear.

11. Vacuum-filter the solution using a Büchner funnel and filter paper.

12. Boil off the solvent under modest vacuum and heat (0.2 atm, 120°C) to yield pureDHLA liquid. A rotovap is ideal for this step.

13. The DHLA liquid should be used as soon as possible. The shelf life of the neatsolution is on the order of weeks if kept sealed in the freezer (–5°C). Note: Pure DHLAis a transparent, colorless liquid. Any residual yellow color indicates that the thiocticacid was not fully reduced. A light yellow product will not bind as readily to thenanocrystal surface during the cap exchange process. It is therefore recommendedthat the entire process be repeated with new reagents until a clear product isobtained.

4.3.2.2 Ligand Exchange and Transfer of Hydrophobic CdSe-ZnS QDs into Water

1. Disperse dried hydrophobic QDs (usually capped with phosphines and amines) infreshly prepared DHLA in a ratio of about 200 mg QDs per mL of DHLA (assume 200mg of QDs in 1.0 mL DHLA for the purpose of this protocol). This is most easilycarried out in a 20 to 30 mL glass scintillation vial with a small magnetic stirrer. Capthe vial with a rubber septum, seal with a zip tie or wire, and purge the vial with inertgas.

2. Heat the mixture to 80°C in a mineral oil bath and stir the mixture overnight (~12hours). Shorter periods may suffice, however the exchange is often more complete ifleft for several hours.

3. Dilute the QD solution in 4 mL of dimethylformamide (DMF).


60

4. Slowly add excess potassium-tert-butoxide (K[t-BuO]). This basifies the mixture anddeprotonates the terminal carboxyl groups. A precipitate is formed consisting ofDHLA-capped QDs. K[t-BuO] is very hygroscopic and must be stored in a sealeddesiccator prior to use.

5. Sediment the precipitate by centrifugation (5 minutes, 3,000 rpm) and carefullydiscard the supernatant so as to retain the solid product at the bottom.

6. Add a minimum volume (a few mL) of deionized or ultrapure water to the precipitateto resuspend in liquid. The newly capped QDs should disperse readily in water.Moderate sonication may expedite this step. As a general rule, a more concentrated

solution is preferable (>10 M) to ensure long term stability in water.

7. Test the pH of the solution. The pH should be maintained around 10 to 12 to ensureQD stability in water. Add dilute NaOH to the sample if the pH is low.

8. Purify the aqueous suspension by using an ultra-free centrifugal filter (50 kDa MWcutoff).

9. Repeat the centrifugal filtration cycle using the above filtration device three to fourtimes and resuspend the QD solution in deionized or ultrafiltered water.

10. If the solution is slightly turbid, verify that the pH is basic. If the pH is 9 orabove (ideally 9–10), an additional filtration step may be required. Use a 0.45- mdisposable syringe filter to remove any large aggregates that may have formed duringthe cap exchange process.

11. The water-soluble QD solutions should be refrigerated in tightly sealed vials at 4°C.Over time, samples may begin to dry out, precipitate, and/or sustain bacterialgrowth. If properly handled, however, samples have been shown to last manymonths. It is recommended to periodically check samples to ensure they haven’tdegraded. Some of these issues can be resolved by simply adding base to the solution

(final pH >9), sonicating, or filtering with a 0.45- m disposable syringe filter. If thesolution is filtered, an absorption measurement should be taken to re-evaluate theconcentration. This is especially important for quantitative studies that rely onaccurate estimates of concentration as some QDs are invariably lost during filtration.

12. The quantum yield of the QDs should be measured after transfer into water. It isknown that a cap exchange using thiols will result in a decreased quantum yield(compared with the native hydrophobic sample); however, a value of ~5–15% isconsidered to be reasonable for water-soluble QDs. Higher values can be achievedwith better passivating ligands. Visual inspection of brightness is not a good indi-cator of the quantum yield since brightness is also a function of concentration.

4.3.3 Biomolecule Conjugation

There are several ways to stably attach biomolecules to QDs; however, we will highlighttwo specific methods here. The first is based on electrostatic self-assembly in whichbiomolecules having a net charge interact with oppositely charged QDs (e.g., avidinassociating with DHLA-capped QDs). This process typically results in rapid and stableself-assembly, even in high salt conditions. If direct protein self-assembly is desired, it isoften necessary to engineer a region of charged amino acids (e.g., lysine) that have anaffinity for the oppositely charged QD surface [20, 29]. If the biomolecule itself does nothave a dense region of charge, an alternative is to use avidin as a bridging molecule[30, 31]. Avidin has sufficient positive charge to associate with negatively charged QDs,

4.3 Methods

61

and additionally allows any biotinylated molecule (e.g., biotin-labeled antibody) tobe readily attached to an avidin-coated QD. This is a versatile means of buildingnanocrystal-biomolecule conjugates using noncovalent self-assembly; however, oneobvious drawback is the prohibitive size of these conjugates. The calculated Försterdistance for most QD-dye pairs is much too small (~30–60Å) to expect reasonableenergy transfer efficiency in these systems unless multiple acceptors are used or theenergy can be relayed to distal acceptor dyes via an intermediate dye (a two-step FRETmechanism) [12].

The second attachment method is specific to peptides and proteins that haveengineered terminal polyhistidine tags (His-tags, shown in Figure 4.5). EngineeredHis-tagged proteins are commonly purified using a nickel-nitrilotriacetic acid (Ni-NTA)resin where they are later eluted from the column with an imidazole solution. Histidinehas affinity for other metals as well (e.g., Zn, Co, Fe), which makes it a convenient routefor self-assembly on metallic and semiconductor nanoparticle surfaces [12, 32].Onecaveat, however, is that the His-tag must have access the metal surface and not be signifi-cantly impeded by a steric barrier due to ligands on the surface. For example, if thenanoparticle is solublized by a polyethylene glycol (PEG) ligand, it is not likely that aHis-tagged protein will be able to bind to its surface due to the bulkiness of the ligand.Peptide oligomers bearing terminal His-tags are more likely to bind to these surfaces,but in general the His-tag self-assembly method greatly favors an accessible surface.Extensive work with DHLA has shown it to be a suitable ligand for the His-tag self-assem-bly strategy suggesting that it does not sterically inhibit attachment or overpopulatenanocrystal binding sites. Other compact thiols have similar features to DHLA and areappropriate candidates for the His-tag procedure. Since the affinity of a His-tagged pro-tein or peptide depends on both surface coverage and size of the ligand, it is difficult todetermine a priori which ligands are compatible with this self-assembly procedure. How-ever, most thiols having short carbon chains appear to be appropriate candidates basedon previous studies. Either self-assembly procedure provides a general and flexible


62

OCH2

HN

CH2 CH2

NH

HN

H C2

HN

HN

NH

NH

O

O O O

OH

NH NH

N

N

N N

HN

N

CH2

Figure 4.5 C-terminal pentahistidine (5×His) tag of a protein or peptide. The multiple imidazolerings on the His residues provide high affinity for metals including Zn. The His-tag is a commonmethod for purifying engineered proteins on Ni-NTA resins.

platform for attaching biomolecules to the surface of QDs (see Figure 4.6). In thischapter, we will focus on the His-tag attachment method since it is a more directapproach and typically minimizes the overall size of the conjugate, which is critical forFRET applications.

Organic dyes are ideal acceptors for QD-based FRET experiments due to their smallsize, narrow absorption spectra, and versatile labeling schemes. QDs have nearly oppo-site features making them poor acceptor candidates. In the case of peptides and proteins,the location of the dye can be specified by using reactive derivatives that target certainamino acids. For example, dyes modified with maleimide groups preferentially reactwith thiol groups found on cysteine (Cys). If the peptide or protein has only one Cys res-idue in its structure, a dye can be labeled at unique locations within the primary struc-ture. Peptide synthesis and site-directed mutagenesis can therefore be used to control ofthe position of the dye. Other dye derivatives may be used to label proteins and peptidesmore generally where the location and number of dyes per molecule are less important.For example, dyes modified with N-hydroxysuccinimide (NHS) ester groups react readilywith primary amines found on lysine (Lys). There are many other options available, andthe choice of labeling method largely depends on the specific application. In some casesacceptors may be linked to molecules not directly bound to the QD. For example,Medintz et al. reported a nanosensor that generated enhanced fluorescence followingthe displacement of a sugar analog from QD-bound maltose binding protein (MBP) [12].In that assay designed to detect maltose in solution, acceptor dye was covalentlyattached to a cyclic oligosaccharide that was later preassociated to the binding pocket ofMBP prior to self-assembly on the nanocrystal surface. Depending on the application,labeling may need to be performed on separate molecules followed by an annealing orbinding reaction to establish the desired association.

Oligonucleotide labeling is achieved through similar covalent methods that attachdyes directly to bases within the sequence. Intercalating dyes such as ethidium bromide(EtBr) can also be used to label double stranded oligonucleotides that do not require

4.3 Methods

63

His-taggedprotein

Avidin

b-lgG

(b)

(a)

Figure 4.6 A schematic showing two examples for attaching biomolecules to negatively charged QDsurfaces. (a) Electrostatic self-assembly of avidin and attachment of a biotinylated antibody (b-IgG). (b)Metal affinity coordination using a His-tagged protein for direct attachment.

covalent attachment. Some of these dyes are highly sensitive to local conditions, whichcan drastically change their emission properties. EtBr in particular is only weakly fluores-cent in water but becomes ~30-fold brighter when bound to DNA [6]. While most accep-tor dyes are emissive, quencher dyes with very low quantum yields also have utilitywhere a secondary fluorescence signal is undesirable [33]. However, in most cases theenhanced acceptor emission provides conclusive evidence of a FRET interaction even ifthe acceptor signal is not directly used to quantify the efficiency. Although measuringchanges in donor fluorescence intensity and/or lifetime is preferable for estimating FRETefficiency, in some situations it is difficult or impossible to carry out necessary controlsthat are needed to evaluate the extent of quenching from the donor signal alone.

A generic protocol for generating dye-labeled QD bioconjugates is provided below.While applications will vary, many of the basic features for producing self-assembledbioconjugates are preserved regardless of use. Here we present a method for directself-assembly of labeled biomolecules (proteins, peptides, DNA, etc.) to a QD surface.Choice of the appropriate biomolecule-QD ratio is usually guided by steric limitations;however, the number acceptors associated to each QD depends on the desired FRET effi-ciency. In many cases the number of labeled biomolecules per QD is varied in each sam-ple in order to understand the relationship between FRET efficiency and the number ofdyes per donor. This information can later be used to understand how biomolecules arearranged on the nanocrystal surface or guide the choice of an optimal starting conditionfor a biosensing arrangement.

4.3.3.1 Biomolecule Conjugation Protocol

1. Prepare a sterile 1.5 mL microcentrifuge tube by adding 100 L of sodium tetraboratebuffer (pH 9.5). Note: While basic buffer solutions are recommended for DHLA-capped QDs, neutral buffers such as HEPES and PBS have been shown to work as well.The stability of these QDs in neutral buffers is probably on the scale of hours to a fewdays; however, this is often acceptable for many applications.

2. Add to the buffer the desired molar quantities of labeled and unlabeled biomoleculesappropriate for binding 20 pmol of QD. Ensure the solution is well-mixed prior tocontinuing.

3. Add 20 pmol of QD (determine the appropriate volume based on the QD stockconcentration). Mix the solution thoroughly with the pipetter.

4. Allow the biomolecules and QDs to self-assemble at room temperature (preferably inthe dark) for at least 15 minutes. A longer incubation time may be required to ensurecomplete association (30–60 minutes).

5. If desired, additional samples should be assembled in separate microcentrifuge tubesin parallel and allowed to incubate. For example, samples having varying numbers oflabeled proteins or peptides per QD might be useful. A good practice is to maintainthe average total number of biomolecules (labeled and unlabeled) per QD constant.

6. One of the samples should be a control containing a fixed average number ofunlabeled biomolecules per QD. This is critical for assessing the influence ofbiomolecule binding on the quantum yield of the QD and is required for subsequentFRET calculations.

7. Add buffer to the microcentrifuge tube(s) such that the final volume is appropriatefor sampling by a spectrofluorimeter or plate reader (typically 0.5–1 mL total


64

volume). If a larger volume is required (e.g., 3-mL cuvette), add additional buffer tothe cuvette.

8. Fluorescence measurements should be taken soon after the incubation period. Theself-assembled bioconjugates are stable for many hours (and perhaps longer) at roomtemperature; however, for consistency it is recommended to take fluorescencemeasurements as soon as possible.

4.3.4 Fluorescence Measurements

The spectroscopic measurements required to characterize FRET can be made in either atime-resolved mode or with continuous excitation to induce steady-state fluorescence.While the former is considered to be superior due to the ability to resolve the lifetimes offluorescent subpopulations, it requires more expensive and sophisticated equipmentincluding pulsed laser sources (on the order of ps or shorter) and time correlated singlephoton counting (TCSPC) detectors. The latter method is more common and yields sim-ilar results. Even within the context of steady-state measurements, fluorescence spectracan be measured in a variety of ways; however, for simplicity we will describe a protocolusing a dual monochromator spectrofluorimeter that allows precise control over theexcitation wavelength and produces high-resolution spectra. The basic considerationsfor spectral data acquisition are largely independent of the specific instrumentationused. However, the quantitative nature of the measurements requires careful selection ofdetector settings, appropriate control samples, background subtraction, and repeatableconditions. FRET measurements often rely on comparisons between samples, whichrequire careful sample preparation and subsequent measurements.

4.3.4.1 Fluorescence Spectra Acquisition Protocol

1. An appropriate excitation wavelength is chosen that efficiently excites the QD donoryet minimally excites the acceptor dye. Consult the QD and dye absorption spectrato identify a suitable excitation wavelength. As an example, a QD population havinga peak emission wavelength of 530 nm is a good donor for a Cy3 acceptor.An excitation wavelength of 400 nm will preferentially excite the QD donorwhile minimally exciting the Cy3 acceptor. Identifying the minimum in the dyeabsorption spectrum is a reasonable starting point. The broad absorption of QDsmakes this a relatively flexible choice as any wavelength shorter than the emissionmaximum wavelength will suffice.

2. Choose appropriate detector settings (such as slit width and integration time) toensure a high signal-to-noise ratio and high resolution. These settings shouldremain unchanged over the course of all measurements. The quantitative nature ofthese measurements requires consistent sampling conditions. Do not change thesesettings between measurements. If possible, the spectral resolution should be 2 nmor less to ensure accurate data analysis later.

3. Collect a fluorescence spectrum from a blank sample consisting of aqueous bufferonly and subtract this from subsequent spectral measurements. Also record spectrafrom two separate control samples containing only QD and only dye. The formershould contain a concentration exactly equal to the concentration of QDs insubsequent conjugate samples. The QD control is used to quantify the extent of QD

4.3 Methods

65

quenching due to FRET and calculate the energy transfer efficiency. The dye controlis used to determine the amount of dye fluorescence generated via direct excitation(in this case, excited at 400 nm). If using the dye signal to corroborate FRETefficiency estimates (not as common), a dye control spectrum should be subtractedfrom each subsequent QD-dye spectra. Since the concentration of dye can vary bysample, the control sample must match the concentration precisely.

4. Record a comprehensive emission scan that measures the fluorescence intensityfrom the QD and dye in a given bioconjugate. In the example of 530-nm emittingQDs and Cy3 (using 400-nm excitation), the recorded emission spectrum shouldextend from 450 to 700 nm. This captures the complete signal from both donor andacceptor and eliminates recording the excitation source entirely. The blank andcontrol spectra should also cover this range.

5. If measuring other samples later in the same vessel, rinse it thoroughly with waterand/or buffer before proceeding with the next sample. In particular, a quartz cuvetteshould be rinsed with water multiple times and lastly with buffer prior to dryingand subsequent use. Note: Glass cuvettes will often be stained by fluorescent dyesor proteins with repeated use. Occasionally the cuvette will need to be cleanedwith dilute HCl to remove contaminants. This condition can be monitored byperiodically measuring the fluorescence signal from a blank sample containing onlybuffer. The best way to avoid this problem is to clean the vessel immediatelyfollowing use.

6. Measured spectra are saved in a spreadsheet format (CSV or ASCII) and readied fordata analysis.

4.4 Data Analysis and Interpretation

In order to produce quantitative information, the measured spectra must be further pro-cessed. This can be accomplished in a number of ways ranging from a manual spread-sheet approach to an automated algorithm. In most cases the end goal is to obtaininformation such as donor-acceptor distance (r) or the number of digested biomoleculesfollowing an enzymatic reaction. It is also possible to generate more advanced measure-ments such as proteolysis rates [16] or protein orientation [15] by including additionalinformation and using proper models. To begin, consider the QD control spectrumwhere the nanocrystal is patterned with biomolecules lacking fluorescent tags (i.e.,unlabeled). If we numerically integrate this spectrum over all measured wavelengths,this provides a value for the starting intensity of QD photoluminescence (PL) to whichwe will compare other measurements. We expect the QD signal to be reduced whenthere is an efficient exchange of energy from the donor to acceptor.

Next, we need to spectrally separate (deconvolute) the individual signals consistingof the QD donor and dye acceptor from the measured composite spectrum. This is neces-sary to accurately quantify the individual signal changes in each, and ultimately calcu-late the FRET efficiency. For this we assume that the measured composite spectrum is thelinear combination of two individual signals that have the same shape as the pure QDand dye control samples [34]. Using these known shapes, we can use a simple fitting pro-cedure to identify the magnitude (or proportional contribution) of the donor and accep-tor to the composite spectrum. Figure 4.7 shows a deconvolution example where the QD


66

and dye spectra are isolated. The QD best fit (i.e., isolated QD signal) can then be numeri-cally integrated to give an overall intensity that is directly compared with the (dye-free)QD control sample. For the fit to work, we further allow for some slight spectral shiftingto occur when the conjugates form. This means that the location of the peak emissionfrom the donor and acceptor can independently translate slightly to the blue or red(usually only a fraction of a nanometer in wavelength):

( ) ( ) ( )I a I b a I bfit QD dyeλ λ λ= + + +1 1 2 2 (4.1)

where the various I( ) represent spectra for the QD and dye samples, respectively; a1 anda2 are proportionality constants for the overall intensity of QD and dye, respectively; b1

and b2 are spectral shifts. This linear four-parameter model typically provides an excel-lent fit to the data (using a least squares regression) and allows us to accurately estimatethe FRET efficiency. At this point, the energy transfer efficiency can be measured in sev-eral ways; however, the most common and straightforward method is to detect changesin the steady-state emission intensity of the donor fluorophore:

EII

aDA

D

= − = −1 1 1 (4.2)

where E is the observed FRET efficiency, and ID and IDA are the integrated intensities of thedonor alone and the donor in the presence of acceptor (i.e., dye), respectively. From theabove expression, it follows that a donor (QD) that is completely quenched indicates100% FRET efficiency. In order to calculate efficiency, we turn to the fit provided in


67

450 500 550 600 650 700Wavelength (nm)

Phot

olum

ines

cenc

e(a

.u.)

0

0.5

1

1.5

2

2.5× 10

6

Cy3

QD

Composite

Figure 4.7 Deconvolution of a composite QD-dye signal into its constituent spectra. In this example,the QD donor is significantly quenched by the Cy3 acceptor. Proper decoupling of donor and acceptorsignals is critical to accurately estimating FRET efficiency.

equation 1. The parameter a1 is an estimate of the fractional intensity the QD donormaintains in the presence of acceptor. As we see, parameter a1 provides a nearly directestimate of the FRET efficiency.

4.4.1 Calculating Donor-Acceptor Distances

A suitable deconvolution algorithm provides an accurate estimate of the FRET efficiencythat can then be applied to a particular application. For example, if a QD conjugate isformed with multiple acceptors surrounding it, this will have a profound effect on whatthe overall measured FRET efficiency means. If the goal is to estimate the averagedonor-acceptor distance, we must use a model that accounts for multiple interactions.In most FRET experiments, there is a pair-wise interaction between a donor and accep-tor, but with QDs it is entirely possible to have multiple acceptors per donor. The usualrelationship between FRET efficiency and distance for a donor-acceptor pair is:

( )E rR

R r=

+06

06 6

(4.3)

where R0 is the calculated Förster distance for the interacting pair [6]. Equation (4.3) sug-gests that there is a precipitous drop in the efficiency as the distance exceeds R0. TheFörster distance depends on a multitude of physical parameters and merits some furtherdiscussion in the context of QD-based FRET. The Förster distance (in angstroms) iscalculated as follows:

( ) ( ) ( )RQ

N nF dD

A DD A0

2

5 44

0

1 69000 10

128=

⎡

⎣⎢

⎤

⎦⎥

∞

∫ln κ

πλ ε λ λ λ (4.4)

where κ2 is a dipole orientation factor, QD is the quantum yield of the donor, NA is Avoga-dro’s number, nD is the refractive of the media between donor and acceptor, FD is theemission spectrum of the donor, and A is the absorption spectrum (expressed as theextinction coefficient) of the acceptor [6]. The integrated quantity is referred to as theoverlap integral J(λ) and is related to the spectral overlap between donor and acceptor asdepicted for a sample QD-dye FRET pair in Figure 4.8. The tunability of QD emissionallows optimization of the spectral overlap and is a convenient way to improve FRET

efficiency in a given system. Also note that the integrand in (4.4) is weighted by 4, whichmeans that donor-acceptor pairs with longer wavelength emission and absorptionspectra will increase R0, all else constant.

Accurate calculation of Förster distance requires careful consideration of each para-meter value. When using QDs as donors, the quantum yield should be measured justprior to an experiment to account for changes in the sample over time. This is far moreimportant than for typical dye donors where the quantum yield does not vary signifi-cantly. It is also critical to measure the quantum yield of QDs with biomoleculesattached to the surface as this can dramatically influence passivation and thus QD. Insome cases, the quantum yield can increase three-fold or more with attached proteins[29]. The orientation factor κ2 typically receives little scrutiny as most references suggestusing a standard value of 2/3 consistent with random dipole orientation [6]. This is likelya good estimate for QDs due to 2-D polarization at room temperature [35], and the ran-


68

dom orientation of labeled biomolecules on the QD surface. While the orientation can-not be ignored, it is probable that the random circular polarization of the QD and therandom positioning of the dye in an ensemble of QD bioconjugates will not exhibit anappreciable orientation effect. Lastly, we consider the refractive index which is mostoften estimated as 1.4 for biomolecules in aqueous solutions (slightly above the 1.33value for pure water). Likewise, more refined estimates of nD are rare even though theFörster distance is somewhat sensitive to this parameter (varying as nD

−2 3). The use of aQD donor complicates matters further if we consider that the refractive index accountsfor the average electrical permittivity of the material between the donor and acceptordipoles. In the case of most commonly used QDs (type I, core-shell), the exciton is effec-tively confined within the core and thus the dipole interactions occur through a crystal-line shell layer in addition to surface ligands, biomolecules, and the aqueous medium.This likely means that a generic refractive index value of 1.4 cited for biomolecules inwater underestimates the true value; however, there has been little effort given to deter-mining a more appropriate estimate of nD in these QD-based systems.

If we ignore the possibility of multiple acceptors, the variation of FRET efficiencywith distance described by (4.3) will certainly be in error. However, we can apply anintuitive modification, which accounts for multiple acceptors surrounding a central QDdonor:

( )E r nnR

nR r, =

+06

06 6

(4.5)

where n is the average number of acceptors surrounding each donor [13]. This accountsfor multiple energy transfer channels (each having the same transfer rate) between theQD donor and nearby acceptor dyes. Rearranging the above expression for the distance rgives:


69

400 450 500 550 600 650 700Wavelength (nm)

Nor

mal

ized

abs,

PL

0

0.2

0.4

0.6

0.8

1 Cy3 abs

QD PL

Figure 4.8 Spectral overlap between a 530-nm max emission QD sample and Cy3 dye.

( )r R

n E

E=

−⎡

⎣⎢

⎤

⎦⎥0

1 61

(4.6)

The efficiency estimate in (4.5) has two further assumptions inherent to it. First, itassumes that every donor has precisely n acceptors surrounding it, which is essentiallynever true. A self-assembly process will invariably lead to a Poisson distribution of accep-tors; however a plurality of conjugates will have n acceptors per QD. In many cases, the

potential error due to this distribution is small so long as r > R0 and diminishes quickly inall cases as n increases [16]. Second, the equation assumes that every acceptor is the exactsame distance from the donor’s center (as depicted in Figure 4.9). This may be nominallytrue for some bioconjugate systems, but for others it may be a poor assumption; thevalidity of the assumption must be individually assessed. Overall, (4.6) is a rather simpleyet powerful model for estimating distances for multiple acceptor systems, but onlyunder certain conditions. It is possible to develop more sophisticated models relatingdistance to efficiency that account for the complications described above, however thisis beyond the scope of this chapter.

4.4.2 Calculating Reaction Rates of Surface-Bound Substrates

In some applications, such as monitoring proteolysis, estimates of FRET efficiency canbe used to determine the rate at which QD-bound substrate is cleaved by solubleenzymes [16, 17]. This requires a quantitative relationship between efficiency (anobservable quantity) and the number of intact substrate molecules that is obtained bygenerating a standard curve where the average number of dye-labeled substrate mole-cules per QD is systematically varied. In this case it is not necessary to stipulate acentrosymmetric arrangement of dyes surrounding a central QD donor; here, we aremerely considering how the efficiency changes with n and assuming that this ensemblerelationship is repeatable for this bioconjugate system. Figure 4.10 shows such a


70

Figure 4.9 QD bioconjugate with six proteins (three dye-labeled and three unlabeled). This exampleshows an idealized uniform arrangement of the biomolecules around the central QD where the donoracceptor distance, r, is approximately uniform.

standard curve for a QD bioconjugate system that has numerous dye-labeled peptidesbound to its surface. Fitting the standard curve to a suitable interpolating function (ahyperbola is used in Figure 4.10) provides a means for estimating the average number ofintact dye-labeled biomolecules per QD given a measurement of the FRET efficiency.One of these arrangements (of n molecules per QD) is chosen as a starting experimentalcondition. A maximum change in efficiency following digestion is most desirable (largerslope); from the data in Figure 4.10, a starting point of five dyes per QD is a reasonablechoice. The bioconjugate (consisting of five labeled peptides per QD) is then brieflyexposed to excess enzyme (~10 minutes) in order to cleave some of the dyes from the QDsurface and obtain an estimate of the initial digestion rate (or “velocity”). Followingaddition of an inhibitor to arrest the reaction, the efficiency is again measured to deter-mine the average number of intact molecules per QD. Because the concentration andreaction time are known, this allows calculation the reaction rate (expressed in molL-1s-1). The preceding experiment is repeated over a range of substrate concentrations inorder to produce a saturation curve that shows the initial reaction rate (velocity) versussubstrate concentration. Analysis of this behavior using Michaelis-Menten or similarmodels reveals information about the mechanism of enzymatic activity and providesrelevant kinetic parameters.16 It should be noted that, since the substrate is bound to thesurface of a nanocrystal, the general assumptions inherent to homogeneous catalysismodels like Michaelis-Menten may not be strictly valid in these systems. Previous workhas shown that a homogeneous model appears to fit the data well when analyzing reac-tions occurring on QD-bound substrates, however the substrate diffuses as a confinedbundle rather than as individual molecules which, in a rigorous sense, requires a morecomplex model.


71

Dye-labeled peptides per QD ( )n0 5 10 15 20

0.0

0.2

0.4

0.6

0.8

1.0

Effic

ienc

y(

)E

Figure 4.10 Standard curve relating the average dyes per QD and FRET efficiency.

4.5 Summary Points

• Bright, stable nanocrystals are critical to the success of these experiments. Thequality of the QD preparation influences the ability of biomolecules to bind thesurface as well as the efficiency of energy transfer.

• QDs provide a unique and flexible platform for developing new classes of FRET-based biosensors. The design of these materials takes advantage of the nanocrystalsurface area which can accommodate a variety of ligands and multiplebiomolecules simultaneously.

• Suitable spectral deconvolution algorithms are necessary to separate compositesteady-state spectra into constituent donor and acceptor signals. The FRETefficiency is best estimated by calculating the intensity loss in the presence ofacceptors.

• Proper distance measurements require appropriate models that describe the FRETefficiency. The validity of these models depends principally on the particularbiomolecules used. Additionally, critical physical parameters must be accuratelymeasured or estimated.

4.6 Conclusions

QD-based FRET is a powerful spectroscopic technique that has many notable advantagesover more traditional donor-acceptor systems composed of organic dyes. In the contextof biological studies, the ability to easily generate self-assembled bioconjugates allows aversatile method for detecting enzymatic activity, biomolecule association/dissociation,structural rearrangements, and soluble analytes. In addition to well-known brightnessand stability benefits, QDs allow multiple interactions with surface-bound acceptors andcan significantly extend the effective interaction distance between donor and acceptors.By tuning surface ligands, QDs can be tailored for stability in a variety of environmentsand interfaced with nearly any functional biomolecule. While many of the applicationsoutlined in this chapter are carried out in vitro, there is a growing focus on live cell imag-ing and in vivo studies. The stability and functionality of QDs in these more complexenvironments is clearly a substantial challenge.

References

[1] Bruchez, M., Jr., Moronne, M., Gin, P., Weiss, S., and Alivisatos, A. P., “Semiconductor nanocrystalsas fluorescent biological labels,” Science, Vol. 281, 1998, pp. 2013-16.

[2] Chan, W. C. W., and Nie, S., “Quantum dot bioconjugates for ultrasensitive nonisotopic detec-tion,” Science, Vol. 281, 1998, pp. 2016-18.

[3] Michalet, X., Pinaud, F. F., Bentolila, L. A., Tsay, J. M., Doose, S., Li, J. J., Sundaresan, G., Wu, A. M.,Gambhir, S. S., and Weiss, S., “Quantum dots for live cells, in vivo imaging, and diagnostics,” Sci-ence, Vol. 307, 2005, pp. 538-44.

[4] Medintz, I., Uyeda, H., Goldman, E., and Mattoussi, H., “Quantum dot bioconjugates for imaging,labelling and sensing,” Nature Materials, Vol. 4, 2005, pp. 435-46.

[5] Klimov, V. I. Semiconductor and Metal Nanocrystals: Synthesis and Electronic and Optical Properties,New York: Marcel Dekker, 2004

[6] Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Singapore: Springer, 2006[7] Stryer, L., “Fluorescence Energy Transfer as a Spectroscopic Ruler,” Annual Review of Biochemistry,

Vol. 47, 1978, pp. 819-46.


72

[8] Sapsford, K. E., Berti, L., and Medintz, I. L., “Materials for Fluorescence Resonance Energy TransferAnalysis: Beyond Traditional Donor-Acceptor Combinations,” Angewandte Chemie InternationalEdition, Vol. 45, 2006, pp. 4562-89.

[9] Jares-Erijman, E. A., and Jovin, T. M., “FRET imaging,” Nature Biotechnology, Vol. 21, 2003,pp. 1387-95.

[10] Jares-Erijman, E. A., and Jovin, T. M., “Imaging molecular interactions in living cells by FRETmicroscopy,” Current Opinion in Chemical Biology, Vol. 10, 2006, pp. 409-16.

[11] Selvin, P. R., “The renaissance of fluorescence resonance energy transfer,” Nature Structural Biology,Vol. 7, 2000, pp. 730-34.

[12] Medintz, I. L., Clapp, A. R., Mattoussi, H., Goldman, E. R., Fisher, B., and Mauro, J. M., “Self-assem-bled nanoscale biosensors based on quantum dot FRET donors,” Nature Materials, Vol. 2, 2003,pp. 630-38.

[13] Clapp, A. R., Medintz, I. L., Mauro, J. M., Fisher, B. R., Bawendi, M. G., and Mattoussi, H., “Fluores-cence resonance energy transfer between quantum dot donors and dye-labeled protein acceptors,”Journal of the American Chemical Society, Vol. 126, 2004, pp. 301-10.

[14] Clapp, A. R., Medintz, I. L., and Mattoussi, H., “Förster resonance energy transfer investigationsusing quantum dot fluorophores,” ChemPhysChem, Vol. 7, 2006, pp. 47-57.

[15] Medintz, I. L., Konnert, J. H., Clapp, A. R., Stanish, I., Twigg, M. E., Mattoussi, H., Mauro, J. M., andDeschamps, J. R., “A fluorescence resonance energy transfer derived structure of a quantumdot-protein bioconjugate nanoassembly,” Proceedings of the National Academy of Sciences, Vol. 101,2004, pp. 9612-17.

[16] Medintz, I. L., Clapp, A. R., Brunel, F. M., Tiefenbrunn, T., Uyeda, H. T., Chang, E. L., Deschamps, J.R., Dawson, P. E., and Mattoussi, H., “Proteolytic activity monitored by fluorescence resonanceenergy transfer through quantum-dot-peptide conjugates,” Nature Materials, Vol. 5, 2006,pp. 581-89.

[17] Clapp, A. R., Goldman, E. R., Uyeda, H. T., Chang, E. L., Whitley, J. L., and Medintz, I. L., “Monitor-ing of Enzymatic Proteolysis Using Self-Assembled Quantum Dot-Protein Substrate Sensors,” Jour-nal of Sensors, Vol. 2008, 2008, pp. 10.

[18] Murray, C. B., Norris, D. J., and Bawendi, M. G., “Synthesis and characterization of nearlymonodisperse CdE (E = sulfur, selenium, tellurium) semiconductor nanocrystallites,” Journal of theAmerican Chemical Society, Vol. 115, 1993, pp. 8706-15.

[19] Dabbousi, B. O., Rodriguez-Viejo, J., Mikulec, F. V., Heine, J. R., Mattoussi, H., Ober, R., Jensen, K.F., and Bawendi, M. G., “(CdSe)ZnS Core-Shell Quantum Dots: Synthesis and Optical and Struc-tural Characterization of a Size Series of Highly Luminescent Materials,” Journal of Physical Chemis-try B, Vol. 101, 1997, pp. 9463-75.

[20] Mattoussi, H., Mauro, J. M., Goldman, E. R., Anderson, G. P., Sundar, V. C., Mikulec, F. V., andBawendi, M. G., “Self-Assembly of CdSe-ZnS Quantum Dot Bioconjugates Using an EngineeredRecombinant Protein,” Journal of the American Chemical Society, Vol. 122, 2000, pp. 12142-50.

[21] Peng, Z. A., and Peng, X., “Formation of high-quality CdTe, CdSe, and CdS nanocrystals using CdOas precursor,” Journal of the American Chemical Society, Vol. 123, 2001, pp. 183-84.

[22] Clapp, A. R., Goldman, E. R., and Mattoussi, H., “Capping of CdSe-ZnS quantum dots with DHLAand subsequent conjugation with proteins,” Nature Protocols, Vol. 1, 2006, pp. 1258-66.

[23] Talapin, D. V., Rogach, A. L., Kornowski, A., Haase, M., and Weller, H., “Highly LuminescentMonodisperse CdSe and CdSe/ZnS Nanocrystals Synthesized in aHexadecylamine-Trioctylphosphine Oxide-Trioctylphosphine Mixture,” Nano Letters, Vol. 1,2001, pp. 207-11.

[24] Hines, M. A., and Guyot-Sionnest, P., “Synthesis and characterization of strongly luminescingZnS-Capped CdSe nanocrystals,” Journal of Physical Chemistry, Vol. 100, 1996, pp. 468-71.

[25] Dubertret, B., Skourides, P., Norris, D. J., Noireaux, V., Brivanlou, A. H., and Libchaber, A., “In vivoimaging of quantum dots encapsulated in phospholipid micelles,” Science, Vol. 298, 2002,pp. 1759-62.

[26] Aldana, J., Wang, Y. A., and Peng, X., “Photochemical Instability of CdSe Nanocrystals Coated byHydrophilic Thiols,” Journal of the American Chemical Society, Vol. 123, 2001, pp. 8844-50.

[27] Uyeda, H. T., Medintz, I. L., Jaiswal, J. K., Simon, S. M., and Mattoussi, H., “Synthesis of compactmultidentate ligands to prepare stable hydrophilic quantum dot fluorophores,” Journal of the Amer-ican Chemical Society, Vol. 127, 2005, pp. 3870-78.

[28] Susumu, K., Uyeda, H. T., Medintz, I. L., Pons, T., Delehanty, J. B., and Mattoussi, H., “Enhancingthe Stability and Biological Functionalities of Quantum Dots via Compact Multifunctional Lig-ands,” Journal of the American Chemical Society, Vol. 129, 2007, pp. 13987-96.

[29] Mattoussi, H., Mauro, J. M., Goldman, E. R., Green, T. M., Anderson, G. P., Sundar, V. C., andBawendi, M. G., “Bioconjugation of highly luminescent colloidal CdSe-ZnS quantum dots with anengineered two-domain recombinant protein,” Physica Status Solidi B, Vol. 224, 2001, pp. 277-83.

References

73

[30] Goldman, E. R., Balighian, E. D., Mattoussi, H., Kuno, M. K., Mauro, J. M., Tran, P. T., and Ander-son, G. P., “Avidin: A Natural Bridge for Quantum Dot-Antibody Conjugates,” Journal of the Ameri-can Chemical Society, Vol. 124, 2002, pp. 6378-82.

[31] Jaiswal, J. K., Goldman, E., R., Mattoussi, H., and Simon, S. M., “Use of quantum dots for live cellimaging,” Nature Methods, Vol. 1, 2004, pp. 73-78.

[32] Sapsford, K. E., Pons, T., Medintz, I. L., Higashiya, S., Brunel, F. M., Dawson, P. E., and Mattoussi,H., “Kinetics of Metal-Affinity Driven Self-Assembly between Proteins or Peptides and CdSe-ZnSQuantum Dots,” Journal of Physical Chemistry C, Vol. 111, 2007, pp. 11528-38.

[33] Clapp, A. R., Medintz, I. L., Uyeda, H. T., Fisher, B. R., Goldman, E. R., Bawendi, M. G., andMattoussi, H., “Quantum dot-based multiplexed fluorescence resonance energy transfer,” Journalof the American Chemical Society, Vol. 127, 2005, pp. 18212-21.

[34] Goldman, E. R., Clapp, A. R., Anderson, G. P., Uyeda, H. T., Mauro, J. M., Medintz, I. L., andMattoussi, H., “Multiplexed Toxin Analysis Using Four Colors of Quantum Dot Fluororeagents,”Analytical Chemistry, Vol. 76, 2004, pp. 684-88.

[35] Chung, I. H., Shimizu, K. T., and Bawendi, M. G., “Room temperature measurements of the 3D ori-entation of single CdSe quantum dots using polarization microscopy,” Proceedings of the NationalAcademy of Sciences of the United States of America, Vol. 100, 2003, pp. 405-08.

Annotated References

Lakowicz, J. R. Principles of Fluorescence Spectroscopy, Singapore: Springer, 2006An indispensible and comprehensive resource for the field of fluorescence spectroscopy, thisbook covers the fundamentals of FRET interactions, novel fluorophores, labeling chemistry, anddata analysis.

Mattoussi, H., Mauro, J. M., Goldman, E. R., Anderson, G. P., Sundar, V. C., Mikulec, F. V., andBawendi, M. G., “Self-Assembly of CdSe-ZnS Quantum Dot Bioconjugates Using an EngineeredRecombinant Protein,” Journal of the American Chemical Society, Vol. 122, 2000, pp. 12142-50.

A seminal paper describing the use of DHLA as a capping ligand and electrostatic self-assemblyto build stable QD bioconjugates.

Medintz, I. L., Clapp, A. R., Mattoussi, H., Goldman, E. R., Fisher, B., and Mauro, J. M., “Self-assem-bled nanoscale biosensors based on quantum dot FRET donors,” Nature Materials, Vol. 2, 2003, pp.630-38.

An early paper that describes development of a FRET-based nanosensor sensitive for maltosesugar using a His-tagged protein self-assembly procedure. The paper also details an alternatebiosensing scheme using a two-step QD FRET process with Cy3 and Cy3.5 dyes.

Michalet, X., Pinaud, F. F., Bentolila, L. A., Tsay, J. M., Doose, S., Li, J. J., Sundaresan, G., Wu, A. M.,Gambhir, S. S., and Weiss, S., “Quantum dots for live cells, in vivo imaging, and diagnostics,” Sci-ence, Vol. 307, 2005, pp. 538-44.

An excellent review of the biological applications of quantum dots focusing on live cells.

Clapp, A. R., Goldman, E. R., and Mattoussi, H., “Capping of CdSe-ZnS quantum dots with DHLAand subsequent conjugation with proteins,” Nature Protocols, Vol. 1, 2006, pp. 1258-66.

A specific protocol for the synthesis of CdSe-ZnS QDs is presented with additional details of aDHLA cap exchange and conjugation of antibodies for a multiplexed detection assay.

Clapp, A. R., Medintz, I. L., and Mattoussi, H., “Förster resonance energy transfer investigationsusing quantum dot fluorophores,” ChemPhysChem, Vol. 7, 2006, pp. 47-57.

A recent review of QD-based FRET applications.


74

C H A P T E R

5Tracking Single Biomolecules in Live CellsUsing Quantum Dot Nanoparticles

Katye M. Fichter, Ardalan Ardeshiri, and Tania Q. Vu

Department of Biomedical Engineering, Oregon Health and Science UniversityPortland, OR 97239

75

Abstract

A monumental challenge of live-cell imaging is the ability to determine thetrafficking behavior of single biomolecules. Current methods allow tracking ofpopulations of molecules, but the subtle behavior of an individual moleculehas increased potential to reveal biology’s most well-guarded secrets. Here, weintroduce a versatile method for tracking single biomolecules using quantumdot nanoparticles. Because of their intense brightness, photostability, andunique blinking pattern, the identification of single molecules can be observed.An extensive variety of conjugation techniques exist to conjugate quantumdots with biochemical tags. Furthermore, although current research in this areafocuses on the tracking of receptors, an untapped well of potential exists forthe study of intracellular processes, including, but not limited to: apoptosis,nuclear import, and pathogenic responses. The application of single moleculetrafficking has potential for breakthroughs in many biomedical areas such asneurochemistry, cancer research, and drug delivery.

Key terms single particle trackingquantum dotsnanotechnologylive-cell imagingintracellular traffickingfluorescent bioconjugates

5.1 Introduction

The ability to track a single molecule inside a living cell is a highly sought-after tech-nique that may play a significant role in uncovering biology’s most elusive and funda-mental mechanisms. Considered a cutting-edge technique, the field of single moleculetracking is currently experiencing rapid growth. Using this technique may allow theresearcher to elucidate countless types of cellular mechanisms such as uptake, move-ment, and fate at a molecular level.[1–5] Moreover, in addition to basic biological appli-cations, initial strides have been made to apply single molecule tracking to biomedicalproblems. Furthermore, single molecule tracking has been used to visualize and measurethe efficacy of therapeutic agents such as nucleic acids, proteins, and other drugcompounds, based on dynamic uptake and intracellular fate [6, 7].

Fluorescence live cell imaging has revolutionized biology and medicine since it wasintroduced during the 1980s [8, 9]. Fluorescence microscopy overcomes the diffractionlimits of transmitted light microscopy and provides high contrast imaging of specificpopulations of cellular biomolecules. However, while the use of fluorescence live cellimaging has rapidly accelerated the understanding of many cellular processes in biologyand biomedicine, many challenges still remain. One prominent challenge is to gainfiner spatiotemporal resolution of individual molecules, or small groups of molecules.Currently, the averaged behavior of the total population of fluorescently tagged mole-cules is commonly observed and studied. A second challenge is to track an individualbiomolecule undergoing successive stages of its lifetime, such as membrane internaliza-tion, membrane recycling, or interorganelle transport. If these two main challengescan be overcome, then the subtle behavior of single biomolecules (currently unattain-able because of population averaging) can be examined. Such subtle behaviors mayyield critical information able to elucidate fundamental cellular mechanisms currentlyinaccessible to investigators.

Organic fluorescence dyes such as fluorescein and rhodamine were often used inearly fluorescence live cell imaging studies. However, investigators faced major prob-lems such as photobleaching and phototoxicity, especially in long-duration live cellexperiments. Such experiments routinely exposed these organic fluorophores inside livecells to repetitious pulses of light over long time points and, unlike fixed (dead)cell experiments, the use of fade-resistant mounting media was not possible. Morephotostable organic dyes, such as the Cyan and Alexa series dyes provided someimprovements [10] but were still troublesome in longer-term live cell imaging experi-ments. The development of “living” enhanced fluorescent proteins (EFPs) [11] was arevolutionary turn in the field of fluorescence live cell imaging [12–15]. EFPs permittedthe visualization of proteins transcribed in live cells. However, this technique also suf-fered from drawbacks. Photobleaching of the EFPs also remained an issue of concernover longer time-lapse experiments. Additionally, large amounts of non-endogenousEFPs can be cytotoxic to cells, as they accumulate in the cytosol and other intracellularcompartments [16]. Finally, while some researchers were able to use EFPs to measure themovements of single proteins in cells under short durations (minutes) [17, 18], their rel-atively dim fluorescence made single EFPs extremely difficult to visualize for practicaluse. As a result, photobleaching, phototoxicity, and photostability of fluorescent tagscontinue to be major hurdles in fluorescence live cell imaging. These limit the observa-tion of single molecules or small groups of molecules undergoing cellular processes overlonger time points.

Tracking Single Biomolecules in Live Cells Using Quantum Dot Nanoparticles

76

Fluorescent quantum dot nanoparticles provide promising potential and address thecall for a brighter, more photostable fluorophore for following the movements of singlemolecules. The most commonly used quantum dots consist of a core nanocrystalsemiconductor, CdSe, with a ZnS shell that enhances the optical properties of thefluorophore. Although not fully understood, these nanoparticles undergo intermittentnonradiant states, which cause them to “blink” [19]. For biological studies, the QDs aretypically covered with a layer of amphiphilic polymer to increase water solubility. Thesurface of the nanoparticle can then be conjugated to various molecules such aspoly(ethyleneglycol) (PEG), and/or chemical handles such as affinity tags and antibod-ies, making them versatile conjugation reagents [20] (Figure 5.1). Quantum dots havethe advantage of being extremely photostable, allowing for hours of imaging withoutphotobleaching [21]. Additionally, they are extremely bright, with extinction coeffi-cients that are about an order of magnitude higher than organic dyes [22]. Finally,because of their blinking patterns and brightness, the detection of single quantum dotsare possible [23], allowing researchers to follow the trajectories of single molecules inlive cells.

Quantum dots were introduced to biological applications as substitutes for organicfluorophores in immunofluorescence-type experiments using fixed samples [24–26].The benefits of the inherent multifunctional properties of QDs were, and continue to be,successfully demonstrated in a wide range of cellular applications. Quantum dots haveemission wavelengths that are dependent on their size, in narrow bandwidths approxi-mately ranging from 511 to 800 nm [27]. This huge selection of colors has opened afloodgate of multicolor experiments that were not previously possible. AdditionallyQDs have utility as detectors of pH and divalent cations [28], and long luminescent life-times similar to that of lanthanides [29]. It was not long before quantum dots were inves-tigated and used in live cell imaging experiments [30, 31]. Most of these studies focusedof the diffusion of lipids in membranes [32] as well as membrane receptors on cellsurfaces [33, 34].

5.1 Introduction

77

Chemical handle

PEG

Polymer coating

ZnS

CdSe

Figure 5.1 Exemplary structure of a functionalized quantum dot nanoparticle.

Although quantum dots are superior and versatile fluorophores for cellular imaging,there are a few limitations to their usage that should be addressed. A major drawback isthe size of the nanoparticles, which can range from approximately 5 to 30 nm in diame-ter, depending on the size of the QD itself, and the type of conjugation. This has raisedquestions about the effects of their size on some intracellular trafficking pathways [35].However, this issue is still in debate and more experiments must be completed to deter-mine what, if any, influence size has on the cell’s natural mechanisms. Furthermore, the“blinking” of quantum dots can cause difficulties in continuity of time-lapse imageseries. However, software programs exist to regain this continuity based on the locationand blinking pattern of the nanoparticle. Furthermore, groups are currently working tosynthesize QDs that do not “blink”[36].

In this chapter, we introduce the use of quantum dot conjugates to study theintracellular trafficking of biomolecules. Numerous conjugation techniques are avail-able to attach quantum dots to just about any biomolecule of interest [20, 37]. Covalentbioconjugation schemes have been successfully used for generating QDs that carry lig-ands, antibodies, affinity tags (such as biotin and hemmaglutinin (HA))[38–40] as well asother chemical handles such as azides [39, 41–43].

Currently the biomolecules most often studied in live cell imaging experimentsusing QDs are proteins with an extracellular domain, such as receptors, that are capableof extracellular QD-conjugation through ligands [44, 45], antibodies [46], or affinitytags [47]. However, microinjection and liposomal delivery are possible ways to intro-duce quantum dots to cytosolic or nuclear proteins for subsequent imaging [48–50].Although the intracellular study of nucleic acids [51, 52] and lipids is within capabilitiesof QD tracking techniques, they are underrepresented and hold promise for futureresearch.

Here, we outline a technique for studying the single-particle trafficking of mem-brane-expressed receptors inside live neurons. This is illustrated using a protocol forQD-nerve growth factor bioconjugates (QD-NGFs) to image different modes of move-ment that surface-expressed NGF receptors undergo after activation and internalization(Figure 5.2) [53]. These examples illustrate the application of this technique and themethodology below is offered as a starting point for customization to user-specificapplications.

5.2 Materials

5.2.1 Reagents

1. Streptavidin 655 QDots (Invitrogen, Carlsbad, CA) Store at 4°C. Do not freeze.

2. Nerve Growth Factor (β-NGF), (R&D Systems, Minneapolis, MN).

3. NHS-PEO4-biotin (Pierce, Rockford IL).

4. D-MEM (4500 mg/L glucose, 862 mg/L glutamine, and 110 mg/L sodium pyruvate)(Invitrogen, Carlsbad, CA).

5. Cell imaging solution: Add 10 μL B-27 serum-free supplement (Invitrogen, Carlsbad,CA) to 490 μl Hibernate E (Brain Bits, Springfield, IL). This provides enough for 5samples. Store at 4°C.


78

5.2.2 Imaging Equipment

1. Fluorescence microscope equipped with a high magnification (x63 or x100)objective.

2. Appropriate fluorescence filter set for QDs (available from Chroma or Semrock).

3. Sensitive digital CCD camera.

4. Appropriate computing software for acquiring digital images and analysis, such asImageJ.

5.3 Methods

5.3.1 Forming QD Bioconjugates

A number of chemical cross-linkers (i.e., EDC) or biochemical affinity tags (i.e., biotin,HA peptide) can be used to conjugate QDs to biomolecules. Here we form QD-nervegrowth factor (QD-NGFs) bioconjugates using biotin-streptavidin conjugation.1. Prepare biotinylated NGF: Add a 30-fold molar excess of NHS-PEO4-biotin to β-NGF

(200 ug/mL). Allow reaction to proceed for 1 hour at room temperature. To purifythe conjugates, dialyze the solution (7 kDa MWCO Slide-Alyzer, Pierce, Rockford, IL)against 500 mL of PBS (pH 7.2) for 3 hours.

2. To form QD-NGF bioconjugates, add streptavidin-QDs to biotinylated NGF at a 1:1molar ratio (typically 1 nM, 100 μL streptavidin-QD: 1 nM, 100 μL biotinylated NGF)in PBS at 4°C for 1 hour. Store at 4°C and use within 24 hours to minimizeaggregation.

5.3.2 Treating Cells with QD Bioconjugates

1. Plate neurons on a No. 1 glass coverslip at a density of about 360 cells/mm2. Allowcells to culture for about 1 week before QD treatment.

2. Prior to QD treatment, wash neurons twice with 1 mL D-MEM using a 3-cc syringe.Wash gently to minimize detachment of cells from the culture dish.

3. Treat cells with QD-NGF: Incubate with 10–200 pM QD-NGF at 37°C, 5% CO2 for 15minutes. Note: QD-NGF concentrations will need to be optimized. Use lowconcentrations to simplify single particle tracking and to minimize multiple QDinteractions.

4. Remove unbound QD-NGF from cells: Gently wash five times with 1 mL D-MEMusing a 3-cc syringe.

5. Add imaging media to cells.

5.4 Data Acquisition, Anticipated Results, andInterpretation

5.4.1 Imaging QD-Bound Complexes in Cells

1. Place cells on the stage of an inverted fluorescent microscope. A heating stage can beused to keep cells at 37°C. Cover the culture dish or imaging chamber with a glasscoverslip to prevent evaporation of media.

5.3 Methods

79

2. Use a high magnification objective (such as 63- or 100-x) to image QD-NGFcomplexes on cells. Select cells with an optimal number of QDs (typically 10–20QDs/field of view). Optimize exposure times to obtain the fastest capture rates withlowest amount of background. Capture a time-series stack of the quantum dots ofthe cells using a digital CCD camera. Note: If QD blinking is observed, this is a goodindication that single or small groups of QDs are present. Bright QD clusters and veryslow blinking rates may indicate QD aggregation. If this occurs, check the QDbioconjugates for blinking prior to introduction to cells.

5.4.2 Analysis of the Real-Time QD Dynamics

1. To obtain quantitative and detailed movement information from the time series,single particle tracking can be used to outline the trajectory of QD-bound cellularcomplexes. Software such as the ImageJ particle tracking plug-in can be used [54, 55].If a large number of QDs are in view, a single field of view may be segmented andprocessed in quadrants to increase processing speed. The following parameters canserve as starting points for this plug-in:

• Particle radius w [pixel]:3

• Intensity r [%]:0.05

• Cutoff score Ts [-]:0.0

• Maximum step length L [pixels]:1.0

• Link range R [frames]:1 or 10.

2. After running the automated QD tracking program, confirm that the tracked QDtrajectories are accurate. Compare, frame by frame, the movement of each QD withits assigned trajectory. Discard trajectories that have incorrect position assignments.These artifacts may occur due to multiple QD interactions, disappearance of the QDfrom the plane of focus, or QD blinking. Blinking may cause some trajectories to loseQDs. This can be minimized by increasing the image capture rate and/or extendingthe link range in the analysis program. This optimization allows QD blinking to be auseful feature of indicating single or low numbers of QDs in a complex while stillretaining accurate trajectory information.

3. Graph the 2-D trajectory of each QD complex. These 2-D trajectoriescontain qualitative features that can be used to estimate the trafficking mechanismsof the QDs. For example, linear displacements may suggest active transport,whereas diffusive movement in confined locations may suggest containment ofQD-complexes in vesicular compartments [53]. Further experiments usingpharmacological compounds (e.g., nocodazole to disrupt microtubule-basedtransport) or immunochemistry techniques can be used verify transportmechanisms [53].

4. The text file containing positional information for each QD trajectory can beimported into graphing/analysis software such as MATLAB or Excel to obtainquantitative positional/temporal information. Quantitative parameters such asaverage velocity, length of active motor steps, and mean square displacements canbe computed from this positional information.


80


Quantum dots are very bright nanoparticles that can overcome the drawbacks ofphotostability inherent in organic dyes. A wide variety of conjugation techniques existto tag biomolecules to these nanocrystals. Both chemical covalent conjugation and bio-tin-streptavidin binding are widely used as QD conjugation techniques. Because of theirbrightness and blinking pattern, individual QDs can be used to record the cellular loca-tion and distribution of their biomolecule conjugates. This type of information is verydifficult, if not impossible to gain from the use of traditional organic fluorophores, andopens up a tremendous opportunity to increase the knowledge in many areas of cellbiology. Many analysis programs, such as ImageJ, contain software to analyze the trajec-tory of QDs. These trajectories may allow the observation of very distinctive movementthat may indicate the transport mechanism. This type of single particle analysis allowsthe researcher to gain fine details about the locomotion of individual biomoleculesinside living cells.


81

33.5

44.5

55.5

66.5

0 5Time(s)

Yp

ositi

on(

m)

μ

10 15 201

1.52

2.53

3.54

4.5

0 5Time(s)

Xp

ositi

on(

m)

μ

10 15 20

(e) (f)

(c) (d)

(a) (b)

Figure 5.2 Molecular dynamics of single QD molecules in cortical neurons. (a) DIC image of a corti-cal neuron (5 days in vitro). Scale bar: 10 μm (b) Corresponding fluorescence image of the cortical neu-ron in (a) containing bound QD-NGFs. Scale bar: 10 μm (c) Single particle tracking reveals the motionof a QD-NGF complex undergoing linearly-directed active transport. Scale bar: 100 nm. (d) Singleparticle tracking reveals the motion of a QD-NGF complex undergoing restricted diffusive movement.Scale bar: 100 nm. (e) Position plot of the QD-NGF complex in (c) shows linear translation on theorder of a few micrometers. (f) Position plot of the QD-NGF complex in (e), shows restricted move-ment on the order of ~0.5 μm.


Problem Solution

Poor cell health Ensure that cells are in good health before QD incubation.Use an imaging media containing antioxidants to minimize phototoxicity.Use QDs that have a biocompatible coating such as PEG. Streptavidin andamine-functionalized Qdots, available from Invitrogen, contain PEG derivatives.Excess conjugation reagents may cause toxicity. Ensure that these reagents areremoved via dialysis or other purification.

Lack of QDs bound to cells Image conjugates separately in solution to check for aggregation. The presence ofvery large or bright particles with very slow blinking rates may indicate an aggre-gation problem.Increase the concentration of QD-conjugates or duration of QD incubation.Check the ligand to ensure bioactivity. For instance, free NGF should causeincreased growth of processes and differentiation.Ensure the ligand is bound to the QD. A coimmunoprecipitation assay may be usedto determine this.

Aggregation of conjugates Do not freeze QDs or QD-conjugates. Freezing may cause QDs to aggregate.Store QD stock solutions and ligand solutions only at high concentrations (100 nMor higher).Dilute QD conjugates into solutions containing 10% BSA.Use 0.1M borate buffer to store QD solutions for extended periods of time.

Insufficient trajectory data Decrease the number of QDs in the field of view by incubating with a lower con-centration of QDs. If QDs are allowed to come near each other, the trajectory datamay not be accurate.Use a heated stage and/or objective heater to help stabilize the focal plane andkeep QDs from going into and out of focus.Adjust the linkage rate in the ImageJ particle tracking software to account forblinking of QDs.

No QD movement Use a heated stage and objective heater to bring the system to physiological tem-perature.Ensure cells are in good health.Ensure QDs are not nonspecifically bound to the substrate.Check for specific binding of the ligand. An immunochemistry experiment canbe used to determine that QD-ligand-conjugates are binding specifically to thereceptor.

References

[1] Groc, L., Lafourcade, M., Heine, M., Renner, M., Racine, V., Sibarita, J.-B., Lounis, B., Choquet, D.,Cognet, L., “Surface Trafficking of Neurotransmitter Receptor: Comparison between Single-Mole-cule/Quantum Dot Strategies,” Journal of Neuroscience, Vol. 27, No. 46 2007, pp. 12433–12437.

[2] Levi, V., Gratton, E., “Exploring Dynamics in Living Cells by Tracking Single Particles,” Cell Bio-chemistry and Biophysics, Vol. 48, No. 1 2007, pp. 1–15.

[3] Moerner, W.E., “New Directions in Single-Molecule Imaging and Analysis,” Proc. Natl. Acad. Sci.USA, Vol. 104, No. 31 2007, pp. 12596–12602.

[4] Greenleaf, W.J., Woodside, M. T., Block, S. M., “High-Resolution Single-Molecule Measurements ofBiomolecular Motion,” Annu. Rev. Biophys, Biomol. Struct., Vol. 36, No. 2007, pp. 171–190.

[5] Xie, X.S., Trautman, J. K., “Optical Studies of Single Molecules at Room Temperature,” Annu. Rev.Phys. Chem., Vol. 49, No. 1998, pp. 441–480.

[6] Babcock, H.P., Zhuang, X., “Using Single Particle Tracking to Study Nuclear Trafficking of ViralGenes,” Biophysical Journal, Vol. 87, No. 4 2004, pp. 2749–2758.

[7] Suh, J., Dawson, M., Hanes, J., “Real-Time Multiple-Particle Tracking: Applications to Drug andGene Delivery,” Adv. Drug Del. Rev., Vol. 57, No. 1 2005, pp. 63–78.

[8] Wang, Y.-L., Taylor, D. L., Fluorescence Microscopy of Living Cells in Culture, Part A, San Diego: Aca-demic Press, 1989, p. 328.

[9] Wang, Y.-L., Taylor, D. L., Fluorescence Microscopy of Living Cells in Culture, Part. B, San Diego: Aca-demic Press, 1989, p. 498.


82

[10] Berlier, J.E., Rothe, A., Buller, G., Bradford, J., Gray, D. R., Filanoski, B. J., Telford, W. G., Yue, S.,Liu, J., Cheung, C.-Y., Chang, W, Hirsch, J. D., Beechem, J. M., Haugland, R. P., Gaugland, R. P.,“Quantitative Comparison of Long-Wavelength Alexa Fluor Dyes to Cy Dyes: Fluorescence of theDyes and Their Bioconjugates,” Journal of Histochemistry and Cytochemistry, Vol. 51, No. 12 2003,pp. 1699–1712.

[11] Heim, R., Prasher, D. C., Tsien, R. Y., “Wavelength Mutations and Posttranslational Autoxidationof Green Fluorescent Protein,” Proc. Natl. Acad. Sci. USA, Vol. 91, No. 26 1994, pp. 12501–12504.

[12] Tsien, R.Y., “The Green Fluorescent Protein,” Annu. Rev. Biochem., Vol. 67, No. 1998, pp. 507–544.[13] Campbell, R.E., Tour, O., Palmer, A. E., Steinbach, P. A., Baird, G. S., Zacharias, D. A., Tsien, R. Y.,

“A Monomeric Red Fluorescent Protein,” Proc. Natl. Acad. Sci. USA, Vol. 99, No. 12 2002,pp. 7877–7882.

[14] Zhang, J., Campbell, R. E., Ting, A. Y., Tsien, R. Y., “Creating New Fluorescent Probes for Cell Biol-ogy,” Mol. Cell. Bio., Vol. 3, No. 2002, pp. 906–918.

[15] Day, R.N., Schaufele, F., “Fluorescent Protein Tools for Studying Protein Dynamics in Living Cells:A Review,” J. Biomed. Opt., Vol. 3, No. 031202 2008.

[16] Ejeskar, K., Fransson, S., Zaibak, F., Ioannou, P. A., “Method for Efficient Transfection of inVitro-Transcribed Mrna into Sk-N-as and Hek293 Cells: Difference in the Toxicity of Nuclear EgfpCompared to Cytoplasmic Egfp,” International Journal of Molecular Medicine, Vol. 17, No. 6 2006,pp. 1011–1016.

[17] Kusumi, A., Iino, R., “Single-Fluorophore Dynamic Imaging in Living Cells,” Journal of Fluorescence,Vol. 11, No. 3 2001, pp. 187–195.

[18] Hirashima, N., Nishio, M., Nakanishi, M., “Intracellular Dynamics of a High Affinity Ngf ReceptorTrka in Pc12 Cell,” Biol. Pharm. Bull., Vol. 23, No. 9 2000, pp. 1097–1099.

[19] Yao, J., Larson, D. R., Vishwasrao, H. D., Zipfel, W. R., Webb, W. W., “Blinking and NonradiantDark Fraction of Water-Soluble Quantum Dots in Aquesous Solution,” Proc. Natl. Acad. Sci. USA,Vol. 102, No. 40 2005, pp. 14284–14289.

[20] Michalet, X., Pinaud, F. F., Bentolila, L. A., Tsay, J. M., Doose, S., Li, J. J., Sundaresan, G., Wu, A. M.,Gambhir, S. S., Weiss, S., “Quantum Dots for Live Cells, in Vivo Imaging and Diagnositics,” Science,Vol. 307, No. 5709 2005, pp. 538–544.

[21] Dubertret, B., Skourides, P., Norris, D. J., Noireaux, V., Brivanlou, A. H., Libchaber, A., “In VivoImaging of Quantum Dots Encapsulated in Phospholipid Micelles,” Science, Vol. 298, No. 55992002, pp. 1759–1762.

[22] Ballou, B., Lagerholm, B. C., Ernst, L. A., Bruchez, M. P., Waggoner, A. S., “Noninvasive Imaging ofQuantum Dots in Mice,” Bioconjugate Chem., Vol. 15, No. 1 2004, pp. 79–86.

[23] Zhang, C.Y., Johnson, L. W., “Simple and Accurate Quantification of Quantum Dots Via Sin-gle-Particle Counting,” J. Am. Chem. Soc., Vol. 130, No. 12 2008, pp. 3750–3751.

[24] Alivisatos, A.P., W. Gu, and C. Larabell, “Quantum Dots as Cellular Probes,” Annual Review of Bio-medical Engineering, Vol. 7, No. 2005, pp. 55–76, 3 plates.

[25] Bruchez, M.J., Moronne, M., Gin, P., Weiss, S., Alivisatos, A.P., “Semiconductor Nanocrystals asFluorescent Biological Labels,” Science, Vol. 281, No. 5385 1998, pp. 2013–2016.

[26] Chan, W.C., et al., “Luminescent Quantum Dots for Multiplexed Biological Detection and Imag-ing,” Curr Opin Biotechnol, Vol. 13, No. 1 2002, pp. 40–46.

[27] Gao, X., Nie, S., Quantum Dot-Encoded Beads, in Methods Mol. Biol., S.J. Rosenthal, Wright, D. W.,Editor. 2005, Humana Press: Totowa, NJ. p. 61–71.

[28] Gao, X., Chan, W. C. W., Nie, S., “Quantum-Dot Nanocrystals for Ultrasensitive Biological Label-ing and Multicolor Optical Encoding,” J. Biomed. Opt., Vol. 7, No. 4 2002, pp. 532–537.

[29] Dahan, M., Laurence, T., Pinaud, F., Chemla, D. S., Alivisatos, A. P., Sauer, M., Weiss, S.,“Time-Gated Biological Imaging by Use of Colloidal Quantum Dots,” Optics Letters, Vol. 26, No. 112001, pp. 825–827.

[30] Jaiswal, J.K., Mattoussi, H., Mauro, M. M., Simon, S. M., “Long-Term Multiple Color Imaging ofLive Cells Using Quantum Dot Bioconjugates,” Nat. Biotech., Vol. 21, No. 1 2003, pp. 47–51.

[31] Jaiswal, J.K., Goldman, E. R., Mattoussi, H., Simon, S. M., “Use of Quantum Dots of Live Cell Imag-ing,” Nat. Methods., Vol. 1, No. 1 2004, pp. 73–78.

[32] Bannai, H., Levi, S., Schweizer, C., Dahan, M., Triller, A., “Imaging the Lateral Diffusion of Mem-brane Molecules with Quantum Dots,” Nat. Protoc., Vol. 1, No. 6 2006, pp. 2628–2634.

[33] Lidke, D.S., Nagy, P., Heinzmann, R., Arndt-Jovin, D. J., Post, J. N., Grecco, H. E., Jares-Erijman, E.A., Jovin, T. M., “Quantum Dot Ligands Provide New Insights into Erbb\Her Receptor-MediatedSignal Transduction,” Nat. Biotech., Vol. 22, No. 2 2004, pp. 198–203.

[34] Dahan, M., Levi, S., Luccardini, C., Rostaing, P., Riveau, B., Triller, A., “Diffusion Dynamics ofGlycine Receptors Revealed by Single-Quantum Dot Tracking,” Science, Vol. 302, No. 5644 2003,pp. 442–445.

References

83

[35] Tekle, C., van Deurs, B., Sandvig, K., Iversen, T.-G., “Cellular Trafficking of Quantum Dot-LiganBioconjugates and Their Induction of Changes in Normal Routing of Unconjugated Ligands,”Nano Letters, Vol. 8, No. 7 2008, pp. 1858–1865.

[36] Mahler, B., Spinicelli, P., Buil, S., Quelin, X., Hermier, J. P., Dubertret, B., “Towards Non-BlinkingColloidal Quantum Dots,” Nat. Mater., Vol. 7, No. 8 2008, pp. 659–664.

[37] Medintz, I.L., et al., “Quantum Dot Bioconjugates for Imaging, Labelling and Sensing,” Nat Mater,Vol. 4, No. 6 2005, pp. 435–446.

[38] Rosenthal, S.J., Tomlinson, I., Adkins, E. M., Schroeter, S., Adams, S., Swafford, L., McBride, J.,Wang, Y., DeFelice, L. J., Blakely, R. D., “Targeting Cell Surface Receptors with Ligand-ConjugatedNanocrystals,” J. Am. Chem. Soc., Vol. 124, No. 17 2002, pp. 4586–4594.

[39] Howarth, M., Takao, K., Hayashi, Y., Ting, A. Y., “Targeting Quantum Dots to Surface Proteins inLiving Cells with Biotin Ligase,” Proc. Natl. Acad. Sci. USA, Vol. 102, No. 21 2005, pp. 7583–7588.

[40] McCann, C.M., Bareyre, F. M., Lichtman, J. W., Sanes, J. R., “Peptide Tags for Labeling MembraneProteins in Live Cells with Multiple Fluorphores,” BioTechniques, Vol. 38, No. 6 2005, pp. 945–952.

[41] Giepmans, B.N., et al., “The Fluorescent Toolbox for Assessing Protein Location and Function,” Sci-ence, Vol. 312, No. 5771 2006, pp. 217–24.

[42] Voggu, R., Suguna, P., Chandrasekaran, S., Rao, C. N. R., “Assembling Covalently LinkedNanocrystals and Nanotubes through Click Chemistry,” Chem. Phys. Lett., Vol. 443, No. (1–3)2007, pp. 118–121.

[43] Goldman, E.R., et al., “Avidin: A Natural Bridge for Quantum Dot-Antibody Conjugates,” J AmChem Soc, Vol. 124, No. 22 2002, pp. 6378–82.

[44] Vu, T.Q., Maddipati, R., Blute, T. A., Nehilla, B. J., Nusblat, L., Desal, T. A., “Peptide-ConjugatedQuantum Dots Activate Neuronal Receptors and Initiate Downstream Signaling of NeuriteGrowth,” Nano Letters, Vol. 5, No. 4 2005, pp. 603–607.

[45] Rajan, S.S., Liu, H. Y., Vu, T. Q., “Ligand-Bound Quantum Dot Probes for Studying the MolecularScale Dynamics of Receptor Endocytic Trafficking in Live Cells,” Nano Letters, Vol. 2, No. 6 2008,pp. 1153–1166.

[46] Rajan, S.S., Vu, T. Q., “Quantum Dots Monitor Trka Receptor Dynamics in the Interior of NeuralPc12 Cells,” Nano Letters, Vol. 6, No. 9 2006, pp. 2049–2059.

[47] Haggie, P.M., Kim, J. K., Lukacs, G. L., Verkman, A. S., “Tracking of Quantum Dot Labeled CftrShows near Immobilization by C-Terminal,” Molecular Biology of the Cell, Vol. 17, No. 12 2006, pp.4937–4945.

[48] Dudu, V., Ramcharan, M., Gilchrist, M. L., Holland, E. C., Vazquez, M., “Liposome Delivery ofQuantum Dots to the Cytosol of Live Cells,” J. Nanosci. Nanotechnol., Vol. 8, No. 5 2008, pp.2293–2300.

[49] Akerman, M.E., et al., “Nanocrystal Targeting in Vivo,” Proc Natl Acad Sci U S A, Vol. 99, No. 202002, pp. 12617–12621 (epub Sept. 16, 2002).

[50] Medintz, I.L., et al., “Intracellular Delivery of Quantum Dot-Protein Cargos Mediated by Cell Pene-trating Peptides,” Bioconjugate Chemistry, Vol. (epub ahead of print), No. 2008.

[51] Srinivasan, C., Lee, J., Papadimitrakopoulous, F., Silbart, L. K., Zhao, M., Burgess, D. J., “Labelingand Intracellular Tracking of Functionally Active Plasmid DNA with Semiconductor QuantumDots,” Mol. Ther., Vol. 14, No. 2 2006, pp. 192–201.

[52] Xiao, Y., Barker, P. E., “Semiconductor Nanocrystal Probes for Human Metaphase Chromosomes,”Nucleic Acids Research, Vol. 32, No. 3 2004.

[53] Sundara Rajan, S., H.Y. Liu, and T.Q. Vu, “Ligand-Bound Quantum Dot Probes for Studiyng theMolecular Scale Dynamics of Receptor Endocytic Trafficking in Live Cells,” ACS Nano, Vol. 2, No. 62008, pp. 1153–1166.

[54] Rasband, W.S., Image J. 1997–2007, U. S. National Institutes of Health: Bethesda, MD.[55] Sbalzarini, I.F., Koumoutsakos, P., “Feature Point Tracking and Trajectory Analysis for Video Imag-

ing in Cell Biology,” J. Struct. Biol., Vol. 151, No. 2 2005, pp. 182–195.


84

C H A P T E R

6Nanoparticles as Biodynamic Substrates forEngineering Cell Fates

María Pía Rossi1, 2, Ram I. Sharma2†, Emily Pawelski3, Jean E. Schwarzbauer,4 andPrabhas V. Moghe2, 3*1New Jersey Center for Biomaterials, 2Department of Chemical and Biochemical Engineering and 3Depart-ment of Biomedical Engineering, Rutgers University, Piscataway, NJ 08854, 4Department of MolecularBiology, Princeton University, Princeton, NJ 08544

*Corresponding Author: Professor Prabhas V. Moghe, Director, Rutgers NSF IGERT on Integrated Science &Engineering of Stem Cells, Department of Chemical and Biochemical Engineering, Rutgers University,599 Taylor Road, Piscataway, NJ 08854, Phone: 732-445-4500 x 6315, Fax: 732-445-3753, e-mail:[email protected]

†Currently at Orthopedic Biomechanics Laboratory, Universität Zürich, Zürich, Switzerland

85

Abstract

Cell behavior traditionally has been manipulated via biochemical cues. The useof nanoscale biointerfaces is particularly attractive because these could be usedto manipulate cell functions at their natural scale, and induce cell behaviorsthat had not been possible through “bulk” presentation of pharmaceutical orbiological factors. One of the advantages that nanomaterials can provide is tomimic the presentation of ligands and peptides in a clustered fashion viananoparticles. In this work, we utilized albumin nanoparticles functionalizedwith extracellular matrix ligands to activate and alter cell behavior. Focusingon the effect of biofunctional nanoparticles on skin cells such as keratinocytesand fibroblasts, we show the enhanced migration and matrix assembly by cells.Additionally, we show spatial guidance of cell processes by nanoparticles.Finally, the presentation of functionalized nanoparticles on three-dimensionalstructures is discussed.

Key terms biodegradable nanoparticles, cell migration,cell patterning, extracellular matrix assembly,ligand clustering, nanotechnology

6.1 Introduction

The extracellular matrix, commonly abbreviated as the ECM, is vital for many cellfunctions, as it provides not only biochemical cues that direct cell behavior but also thestructural support to cells [1]. The ECM is composed of a variety of proteins and polysac-charides that are secreted locally and assembled into an organized meshwork in closeassociation with the surface of the cells [2]. Once believed to be an inert framework forbolstering the physical conformation of tissues, it is now understood that the ECM playsa large role in different cell functions such as apoptosis, locomotion, morphogenesis anddifferentiation [3]. These functions are mediated by the interaction of integrins, cell-surface receptors, and ligands associated with the ECM [2].

Integrins are a large and widely studied family of cell surface receptors. Heterodimerscomprised of α and β subunits, these surface proteins have been found to bind to manydifferent ECM proteins, such as fibronectin, collagen, vitronectin, and laminin [4].When these ECM ligands, which commonly contain the Arg-Gly-Asp (or RGD) attach-ment site [5], bind to integrin receptors, important operations within the cell are trig-gered, such as the activation of second messenger cascades [6]. Studies have shown thatligand binding of integrin receptors leads to increased lateral mobility, which allows forintegrins to cluster and facilitates stronger adhesion at binding sites [6], influencingmorphogenesis, apoptosis, and proliferation. Researchers have proposed configurationsfor artificially clustering ligands to elicit changes in values associated with cell locomo-tion and adhesion [7]. For example, Maheshwari et al. explored whether the presenta-tion of integrin ligands in a clustered format affects cell adhesion and motility using amonomeric RGD peptide motif derived from the fibronectin integrin binding domain[7]. They presented the low-affinity RGD-derived ligand in a noncell-adhesive polyeth-ylene oxide (PEO) hydrogel background interspersed with polyethylene oxide moleculesconfigured in a star conformation. The ligand was bound to the PEO stars in clusterswith an average of 1, 5, or 9 ligands per star molecule. In addition to examining theeffects of different numbers of ligands bound to the star molecules, the researchers alsoexamined five different average ligand densities for each cluster size. Not only did thecells with clusters containing the highest number of ligands exhibit increased adhesion,but the clustered presentation also enhanced cell migration speeds.

One of the caveats associated with the PEO star model is the inability of PEOmacromolecules to allow for control of exact numbers of RGD peptides, relying insteadon averages. Additionally, it is difficult to calculate with any degree of precision thedistance between RGD peptides. This problem was overcome by the utilization ofAu-dot-containing micelles, which can be patterned with a high degree of precision via asubstrate-patterning strategy based on self-assembly of diblock copolymer micelles.These micelles are then treated with a gas plasma, which leaves only an extended, hexag-onal pattern of nanodots placed in nearly perfect regularity on a noncell-adhesive poly-mer background. As one nanodot can only anchor one integrin molecule, the regularityof the pattern provided the ability to calculate an optimal density of nanodots for celladhesion and motility purposes [8]. Using osteoblasts, the researchers determined thatthe role of spatial distribution of single RGD peptides on cell adhesion and spreadingwas best observed when the nanodots are closest together, and do not appear to formfocal adhesions or be affected in a clustering fashion when distances exceed 58 nm. In asimilar study, the same group demonstrated that fibroblast integrin clustering wasaffected by the spacing of RGD-functionalized gold nanodots [9]. At higher nanodot

Nanoparticles as Biodynamic Substrates for Engineering Cell Fates

86

spacings, fibroblast spreading was considerably compromised, but motility wasenhanced.

Another promising study by Lipski et al. involved the effect of silica nanoparticles oncells [10]. The biggest advantage of using silica particles is the versatility it affords withrespect to functional group and biological moiety modification. Therefore, Lipski et al.hypothesized that silica nanoparticles could be used to show the effects of both textureand chemistry in a decoupled fashion by allowing for greater ease in manipulatingnanoroughness. Bolstering Calvalcanti-Adams et al.’s research with respect to a proxim-ity threshold [8, 9], the researchers found that surface features within 50 nm producedthe greatest effect on cell functions such as proliferation. Additionally, the researchersalso found an effect on focal adhesion complexes and F-actin fiber alignment that wasspecific to cell type, with nanoroughness decreasing endothelial cell points of contactwhile having the opposite effects on preosteoblasts. These new forays into the explora-tion of silica nanoparticle topographies hold promise in affording researchers an easier,more versatile model with which to explore the effects of substrates on cell function.

One methodology that has been explored recently to manipulate cell behaviorthrough integrin-ligand binding involves the use of functionalized magnetic nano-particles. In the beginning of 2008, Mannix, Kumar et al. reported on the use ofsuperparamagnetic beads of 30 nm in diameter that were surface conjugatedwith N1-2,4-dinitrophenyl-L-lysine:L-lysine (DNP-Lys) to target binding of cell surfaceIgE–Fc1RI receptor complexes [11]. Upon application of an electromagnetic field, thebeads were attracted to each other, forcing the integrins to cluster in response. Further-more, upon removal of the electromagnetic field, bead:bead attraction was eliminated,and integrin clustering was reversed. Upon this induced clustering of the integrins, anincrease in calcium signaling by the cells was measured; removal of the field, andreversal of the clustering, resulted in calcium signaling to cease, demonstrating thereversibility of this technique.

Aside from using ligand-conjugated nanoparticles to target cell functions such asmigration, adhesion and spreading, these can also be used for targeted drug delivery.For example, in a recent study, Murphy et al. produced organic nanoparticles thatwere functionalized with a cyclic RGD peptide [12]. Furthermore, doxorubicin, a drugcommonly used in cancer therapy, was encapsulated into the nanoparticles. Thenanoparticles targeted the αv β3 integrin commonly found in tumor vasculature, andselective apoptosis was observed in the ávâ3-expressing sections of the vasculature. Thetreatment also demonstrated anti-metastatic activity, and no weight loss was observedas a result, indicating that functionalized nanoparticles are a viable technique for tar-geted drug delivery with minimal side effects.

It is clear that the manipulation of the interaction of cells with ECM ligands couldprove to be crucial to control cell behavior for applications in drug delivery, tissue andbioengineering. Nanotechnology could provide the materials necessary to promotethe appropriate presentation of ligands and activate or accelerate a diversity of cellfunctions. The use of nanoparticles is of particular interest for this purpose, as theyprovide an efficient way to present ligands in a clustered fashion and promote integrinclustering.

Many studies so far have involved the use of inorganic materials such as gold or ironoxide, which can provide interesting properties, such as electrical and magnetic conduc-tivity and easy functionalizability. However, the main shortcoming of these materials is

6.1 Introduction

87

that they are not biodegradable, and could have severe cytotoxic effects in vivo. For thisreason, some efforts have recently revolved around the use of organic and natural mate-rials to create functionalizable nanoparticles. Albumin nanoparticles are particularlyattractive because they are not cytotoxic or antigenic, are biodegradable and can be fab-ricated by a variety of techniques [13–15]. Therefore, while albumin nanoparticles can-not be intrinsically electrically or magnetically manipulated, they show great potentialfor use in vivo due to their biocompatibility. While albumin-derived nanoparticles haveshown great potential for the delivery of drugs, DNA and other macromolecules[13, 16, 17], in this work, we have concentrated on surface functionalization of thenanoparticles for the presentation of ligands to cells. In our work, we have observedenhanced migration of keratinocytes, protein assembly by fibroblasts and spatialguidance of cell attachment.

6.2 Experimental Design

Cell signaling can be regulated by promoting integrin clustering through the presenta-tion of extracellular matrix ligands. We hypothesized that the presentation of anextracellular matrix ligand on nanoparticles could be used as a tool to modify the dis-play, conformation, and/or overall organization of the ligand and engineer ligand clus-tering at the nanoscale to elicit differential cellular responses. The enhanced mobility ofthe nanoparticles could promote ligand availability, membrane based ligand/integrintranslocation, integrin mobilization and ligand internalization, activating cell signalingcascades to promote or guide cell functions. We selected albumin nanoparticles due totheir biocompatibility, biodegradability, and low cytotoxicity, as well as their ability tobe derivatized/encapsulated to achieve diverse biological functionality.

Throughout our experiments, we used extracellular matrix ligands and proteins aspositive controls, specifically, GST-FNIII9-10, the ligand we used to functionalize thenanoparticles, and whole length fibronectin, the extracellular matrix protein fromwhich the ligand was derived. We also used unfunctionalized nanoparticles and sub-strates blocked with bovine serum albumin or calcein as our negative controls. Samplesvaried nanoparticle size and ligand concentration, and experiments were always done intriplicate and repeated three times to ensure reproducibility and repeatability.

6.3 Materials

6.3.1 Cell Culture, Fixing, Staining, and Analysis Reagents

Human fibroblasts were cultured in McCoy’s 5A medium (Invitrogen, Chicago, IL)supplemented with 10% fetal bovine serum, 1% penicillin/streptopmycin(Biowhittaker, Walkersville, MD) and 1% L-glutamine (Invitrogen, Chicago, IL). Humankeratinocytes were cultured in serum-free keratinocyte growth medium (KGM)(Clonetics, San Diego, CA) containing 0.1 ng/ml epidermal growth factor (EGF), 5 μg/mlinsulin, 0.5 μg/ml hydrocortisone, 50 μg/ml gentamicin, 50 ng/ml amphotericin-B,0.15 mm calcium, and 30 μg/ml bovine pituitary extract (BPE). All cell culture reagentswere maintained at 4°C until use except the L-glutamine, which was maintained at

−20°C.


88

Formaldehyde and Triton X-100 were maintained at room temperature. Rhodamineand Texas Red Phalloidin, for actin staining, were maintained at –20oC. Monoclonalantifibronectin antibody produced in mouse, clone IST-4 for fibronectin assembly stain-ing was maintained at –20oC. Bovine serum albumin for blocking was maintained at 4oC,and calcein from fat-free milk for blocking was maintained at room temperature.B-nitrophenyl N-acetyl b-D glucosaminide was maintained at –20oC and glycine/5mMEDTA was maintained at room temperature; both were used for the cell attachmentassay. All these reagents were obtained from Sigma, St. Louis. MO. Secondary antibodies(fluorescein (FITC) and Texas Red AffiniPure Donkey anti-mouse (IgG) were maintainedat –80oC obtained from Jackson Immunolabs, Suffolk, U.K.

6.3.2 Nanoparticle Fabrication and Functionalization

Human serum albumin for nanoparticle synthesis (30% w/v, Sigma, St. Louis, MO) wasmaintained at 4oC. Iodoacetamide (Sigma, St. Louis, MO) was maintained at 4oC. HCland NaOH (Sigma, St. Louis, MO) were both maintained at room temperature. TheBCA protein assay (Pierce, Rockford, IL) was maintained at room temperature, andN-succinimidyl 3-(2-pyridyldithio) propionate (Sigma, St. Louis, MO) was maintained at–20oC. The recombinant fibronectin fragment GST-FNIII9-10 was expressed in E. coli, puri-fied, and stored at –20°C. The reagents for the alkaline phosphatase ELISA were obtainedfrom Sigma and maintained at 4oC.

6.3.3 Microscale Plasma Initiated Patterning

A Sylgard 184 silicone elastomer kit was employed to make micropatterning stamps, andpoly(DTE-co-8% PEG1K carbonate) was a courtesy of Prof. J. Kohn (Rutgers University).

6.4 Methods

There are several different ways of preparing albumin nanoparticles. Techniques involv-ing emulsification, controlled desolvation, and thermal denaturation all have their indi-vidual benefits and shortcomings. In this section, our most commonly used method foralbumin nanoparticle fabrication is described. This method was used to fabricate ANPsof sub-100 nm sizes, and was modified based on the original method by Takeoka andcolleagues [18, 19].

6.4.1 Albumin Nanoparticle Fabrication

Albumin-derived nanoparticles were synthesized by denaturing albumin monomersand stirring the suspension to generate nanoparticles by self-assembly processes, as dia-grammed in Figure 6.1. Specifically, the albumin nanoparticles (ANPs) were synthesizedby denaturing filtered (0.22 μm filter, Fisher) human serum albumin (30% w/v, Sigma,St. Louis, MO) diluted to 1% (v/v) in phosphate buffer saline (PBS) in a 250-mL glassbeaker through an increase in pH to ~10.6 by the drop-wise addition of 0.1N NaOH. Sub-sequently, temperature was slowly increased to 80°C with the use of a hot plate. Thetemperature was maintained on the hot plate at 80°C for 10 minutes and the solutionwas then rapidly cooled to 25°C by placing the glass beaker in an ice bath.

6.4 Methods

89

The glass beaker was then removed, and the temperature was maintained at 25°C for10 minutes. The pH was decreased to ~5.9 by the drop-wise 0.1N HCl and the tempera-ture was increased to 37°C slowly using a hot plate. Upon reaching the temperature,the solution was stirred gently using a magnetic stirrer on the hot plate in order toinduce self-assembly of the denatured albumin into the nanoparticles. The solution wasallowed to stir to allow for nanoparticles to aggregate and then incubated with 0.1%(w/v) iodoacetamide (Sigma, St. Louis, MO) gently shaking at room temperature for 1hour, covered with aluminum foil to prevent deactivation of the iodoacetamide, to stopthe aggregation reaction. The nanoparticle solution was placed in dialysis tubing(MWCO 100 kDa) and dialyzed at 4°C overnight to remove any unreacted monomericalbumin and filtered again (0.22 μm filter, Fisher Scientific, Pittsburgh, PA) to removelarge aggregates. Nanoparticle sizes ranged from ~30–200 nm.

This process exploits the use of alkaline conditions to expose the 17 pairs of disulfidebonds and one thiol group at 34 Cys [20] within albumin to the aqueous phase and con-vert the albumin from the more compact N-form to the B-form. By lowering the pH ofthe solution, the electrostatic repulsion among the negatively charged groups on theB-form albumin decreases, resulting in aggregation. Albumin was chosen to create afamily of various sized nanoparticles because it is (a) easily functionalized with theligand of interest, (b) it is biodegradable in vivo, (c) it has a high level of bio-compatibility, and (d) it allows effective exposure of adhesion ligands against a relativelyinert background.

To ensure elimination of unaggregated species, ANP preparations were filtered toremove particulates greater than 200 nm and dialyzed to remove monomeric albumin.


90

Figure 6.1 Schematic illustrating the synthesis and functionalization process of albuminnanoparticles (ANPs). (a) ANPs are synthesized by denaturing human serum albumin through anincrease in pH and temperature and aggregating the albumin into nanoparticles through a decrease intemperature and pH and a final increase in temperature and stirring. The ANPs are then reacted withSPDP for functionalization. (b) Amine-terminated ligands (in the case of this work, GST-FNIII9-10) arereacted with SPDP and then with DTT to make the ligands reactive for functionalization. (c) The reac-tive nanoparticles and ligands are incubated together at room temperature for 4 to 6 hours to induceconjugation, yielding functionalized ANPs.

SDS PAGE showed ANP preparations were significantly purified from monomeric albu-min following dialysis [21]. By changing the stirring time, it was also possible to producenanoparticles of different diameters (30–200 nm), which were confirmed with dynamiclight scattering (DLS) (data not shown). For DLS, samples were placed in the cuvettes rec-ommended by the manufacturer either in stock or at dilutions of up to 1:50 in PBS andanalyzed.

Scanning electron microscopy (SEM) confirmed the formation of nanoparticles fol-lowing the fabrication procedure (data not shown). For SEM analysis, nanoparticles wereadsorbed onto an aluminum stub by incubating at 4oC and washing two times with PBSand twice with water to remove unbound nanoparticles and remove salts from the PBSthat may dry and obscure imaging. The remaining solution was allowed to dry and thensputter-coated with gold-palladium to prevent charging during imaging.

Using a BCA assay kit (Pierce, Rockford, IL), we estimated the protein yield of ANPs inthe suspension post-dialysis and filtration (data not shown). The BCA assay was per-formed according to instructions in the kit; briefly, standards were prepared by dilutinga 2-mg/mL stock albumin solution at 2x dilutions and loaded onto a 96 well plate in trip-licate in the first three columns. Samples were loaded at different dilutions in triplicateonto the 96 well plate. Working reagent was prepared according to the instructions ofthe BCA assay kit and loaded into all wells. The plate was shaken gently for 30 secondsand incubated at 37oC for 30 minutes. Color changes in the plate were read using a platereader at an absorbance of 490 nm.

6.4.2 Albumin Nanoparticle Functionalization

ANPs were functionalized with a truncated fragment of fibronectin that consists of the9th and 10th type III domains of the protein (GST-FNIII9-10), as shown in Figure 6.1.Fibronectin, a dimeric glycoprotein, is involved in cellular processes such as adhesion,spreading and migration, and can help regulate tissue processes such as wound healing[22]. Both the 9th type III and the 10th type III domains within the selected fibronectinfragment associate with integrin cell surface receptors, and trigger intracellular signalingrelated to cell spreading, growth, and migration [23–25]. The GST-FNIII9-10 was producedas previously described, by cloning human fibronectin cDNA into a pGEX vector forexpression as a glutathione-S-transferase fusion protein [21]. Escherichia coli cells weretransformed with the expression plasmid and GST fusion proteins were separated frombacterial lysates by glutathione-sepharose affinity chromatography (GE Healthcare,Piscataway, NJ).

The ANPs were functionalized with the GST-FNIII9-10 ligand using bioconjugationand peptide chemistry techniques [21, 26]. Specifically, both GST-FNIII9-10 and ANPconcentrations were measured using a BCA protein assay kit (Pierce, Rockford, IL).N-succinimidyl 3-(2-pyridyldithio)propionate (SPDP, Sigma, St. Louis, MO), a hetero-bifunctional cross-linking agent, can react with the amine groups in the proteins to forman amide linkage at one end while the 2-pyridyldithiol group at the other end can reactwith sulfhydryl residues to form a disulfide bond. The GST-FNIII9-10 and ANPs were sepa-rately reacted with the SPDP for 30 minutes at room temperature at a concentration of500 μM. The GST-FNIII9-10 was then reacted with dithiothreitol (DTT) for 30 minutes atroom temperature at a concentration of 0.5 mg DTT per mg of GST-FNIII9-10 to form a freesulfhydryl group. The reacted protein and nanoparticles were then dialyzed (MWCO

6.4 Methods

91

6kDa) overnight at 4oC and the final concentration of each was again measured by theBCA protein assay. ANP-SPDP and GST-FNIII9-10-SPDP-DTT were then reacted togetherin a conical tube and shaken lightly for 4 to 6 hours at room temperature forfunctionalization and dialyzed (MWCO 100kDa) overnight at 4°C to remove anyunreacted species.

The efficiency of GST-FNIII9-10 conjugation to ANPs was examined using two types ofenzyme-linked immunosorbent assays (ELISAs), one specific for the cell binding domainon the recombinant fibronectin fragment and one for the GST tag in the GST-FNIII9-10;additionally, ELISAs for albumin in the nanoparticles were also performed.

For the albumin ELISA, albumin standards were loaded in triplicate onto a 96well plate starting at a concentration of 100 μg/mL and diluting 10x to 0 μg/mL.Functionalized nanoparticle samples were loaded in triplicate at varying concentrationsto avoid saturation. For the ligand cell binding domain ELISA, GST-FNIII9-10 standardswere loaded onto another 96 well plate at a starting concentration of 10 μg/mL anddiluted in 2x dilutions. Functionalized nanoparticle samples were loaded in triplicate atvarying concentrations to avoid saturation.

The plates were incubated at 4oC overnight and washed five times in DPBS with Ca2+

and Mg2+. Plates were then blocked using 13% casein (from fat free milk) for 1 hour at37°C and washed again. Plates were incubated with primary antibody (monoclonalantialbumin produced in mouse (Sigma, St. Louis, MO)) for albumin at a 1:10,000 dilu-tion in PBS; anti-fibronectin frag, cell attachment fragment, clone 3E3 (Millipore,Billerica, MA) for the ligand at a 1:1,000 dilution in PBS) for 1 hr at 37oC and washed. Thetwo plates were incubated with secondary antibody (anti-mouse IgG-alkalinephosphatase antibody (Sigma, St. Louis, MO)) at a 1:20,000 dilution in PBS), incubatedfor 1 hour at 37°C and washed. Both plates were then incubated with alkalinephosphatase yellow liquid substrate system (Sigma, St. Louis, MO) until color developed(about 45 minutes at room temperature) and read on a plate reader at 405 nm. To stopthe reaction, 3N NaOH can be added.

The ELISA for the GST tag in the ligand was done also by incubating standards andsamples overnight at 4°C, washing and blocking with casein. The primary antibody usedwas anti-glutathione-S-transferase antibody produced in rabbit (Sigma, St. Louis, MO) ata 1:2000 dilution in PBS for 1 hour at 37oC. After washing, the plate was incubated withanti-rabbit IgG (whole molecule)–peroxidase antibody produced in goat (Sigma, St.Louis, MO) at a 1:23,000 dilution in PBS for 1 hour at 37°C and washed. Plates were thenincubated with Sigma-FAST Fast Red TR/Naphthol AS-MX Tablets (Sigma, St. Louis, MO)for approximately 30 minutes and read at 450 nm. The solution can be stoppedwith H2SO4.

Increasing the amount of ligand in the conjugation reaction resulted in a propor-tionate increase in the levels of ligand conjugated to the surface of the nanoparticles,and nanoparticles were not saturated with ligand at lower loadings. When we examinedligand density for differentially sized nanoparticles, we found that, for a given initialmass of ligand reacted in the conjugation reaction, the extent of conjugation did not sig-nificantly differ for nanoparticles of different sizes, and nanoparticle size did not influ-ence ligand density [27]. By establishing adsorption isotherms, we could differentiallycontrol the presentation of the ligand on the ANPs by determining the bulk concentra-tions of ligand and GST-FNIII9-10-ANPs required to have equivalent net concentrations.ELISAs were performed in parallel in order to confirm that equivalent albumin amounts


92

were adsorbed to the substrate irrespective of nanoparticle size and ligand density. Also,the cell binding domain exposure of the GST-FNIII9-10 was determined and normalized byimmunosorbance assays [27]. By presenting the ligand on the ANPs, instead of byadsorption directly onto the substrate, the cell binding domain exposure was found tobe higher, as can be seen in Figure 6.2.

6.4.3 Albumin Nanoparticle Pattern Creation—Microscale Plasma InitiatedPatterning ( PIP)

Poly(DTE-co-8% PEG1K carbonate) was selected for patterning studies not only becauseof its biocompatibility but also because it inhibits both protein and cell attachment [28].The polymer, in powder form, was diluted in a 98.5% v/v methylene chloride/1.5% v/vmethanol solution at 1%w/v. The solutions were then spin-coated at 4,000 RPM ontoclean glass coverslips to form thin films of polymer on the glass. An elastomericpoly(dimethylsiloxane) (PDMS) stamp with parallel grooves 10 to 400 μm in width andopen at both ends was then utilized to selectively expose areas of the polymer surface tooxygen plasma. These sizes were specifically chosen to guide cell processes, which occurat the microscale, and confirm the functionality of the nanoparticles.

The stamp was fabricated by pouring a Sylgard 184 silicone elastomer kit at a baseweight to cross-linker weight ratio of 10:1 over lithographically created masters [29].Therefore, while some of the substrate is protected by the PDMS stamp, the area underthe grooves is exposed to the oxygen plasma. The polymer was treated at 50W for 60 to120 seconds to ensure sufficient functionalization. After plasma treatment, nanoparticlesolutions were incubated on the polymer surface overnight at 4oC to ensure binding andadsorption of the nanoparticles onto the substrate.

Fibroblasts were seeded at 10,000-20,000 cells/cm2 on the cover slips and incubatedat 37oC for 5 to 24 hours. Cells were then fixed and stained for actin as described below.

6.4 Methods

93

Figure 6.2 Cell binding domain exposure from GST-FNIII9-10, measured by ELISA using mouseantifibronectin cell binding domain (Clone 3E3), which recognizes the cell binding domain in humanFN. Conjugation of the ligand to ANPs increases exposure of the cell binding domain in comparison toadsorbing the ligand on a substrate directly, most likely through changes in conformation duringadsorption and functionalization. Values are the average of 3 experiments performed in triplicates.Error bars represent standard error around the mean.

6.4.4 Cell Culture

Human fibroblasts up to passage 32 were used for experiments. Fibroblasts weresupplemented with serum-free media during and at least 16 hours prior to experimenta-tion. Human keratinocyte passages 2–3 were utilized for all experimental studies.Keratinocytes were supplemented with KGM without BPE and EGF at least 16h prior tothe experiment and during the experiment.

For all cell experiments, nanoparticles were adsorbed on substrates at 4oC overnightor at 37oC for 1 hour. Unbound nanoparticles were then washed three times with PBSand blocked with 3% bovine serum albumin or casein at 37oC for 1 hour and washedthree times with PBS. Cells were then seeded on the substrates.

6.4.5 Keratinocyte Morphology and Migration

To evaluate cytoskeletal organization and morphology, keratinocytes were seeded at adensity of 8,400 cells/cm2 and fixed at 5 to 24 hours after seeding. For fixing, cells werewashed three times with DPBS with Ca2+ and Mg2+ and fixed with 3.7% formaldehyde inPBS (Sigma, St. Louis, MO) at room temperature for 15 minutes and washed. Cells werethen permeabilized with 0.5% Triton X-100 (Sigma, St. Louis, MO) for 15 minutesat room temperature and washed. Keratinocytes were then stained with rhodaminephalloidin (Sigma, St. Louis, MO) at a 1:200 dilution for 30 minutes at room temperaturein DPBS with Ca2+ and Mg2+ for visualization of actin.

Cell motility kinetics were investigated by seeding isolated keratinocytes at a concen-tration of 2,800 cells/cm2 on wells coated with either ligand-ANP, ANP, or ligand over-night at 4oC. Wells were then washed three times with PBS and blocked with BSA for 1hour at 37oC. Cells were incubated in the wells and then transferred to the microscopefor motility studies. Four nonoverlapping viewing fields containing single cells wereidentified in each of the wells and continually imaged at 20x magnification under trans-mitted light for a total of 10 hours at 10-minute intervals. Images were then analyzedwith Image Pro Plus (Media Cybernetics, Silver Springs, MD). For each image, the x and ylocation of the cell centroid was noted throughout each sequence of images and themean square displacement of the cell tracks was computed for each time interval:

( ) ( )( )( ) ( )[ ] ( )( ) ( )[ ][ ]d t n t

N nx n i t x i t y n i t y i t

i

N n2 2 2

0

11

− =− +

+ − + + −=

−

∑Δ Δ Δ Δ Δ

Cell motility was quantified by modeling the cell motility behavior as a persistentrandom walk in an isotropic environment [30]. Briefly, the mean-squared displacement,

given by d S P t P et

P2 22 1= − −−

[ ( )], is a function of time, with two major single cell motil-

ity parameters, root mean squared cell speed, S, and directional persistence time, P. Therandom motility coefficient was determined by ( )d t t P e t P2 4 1= − − −μ[ ( )]. Experimental

data was used to fit the above equations and regress the best estimates for S and P.

6.4.6 Fibroblast Extracellular Matrix Assembly

For extracellular fibronectin assembly, fibroblasts were seeded at a density 35,000cells/cm2 to ensure enough cell:cell contacts for fibroblasts to produce and assemble


94

extracellular matrix. Cells were maintained on the nanoparticle-adsorbed substrates for24 to 48 hours and fixed. For fixing, cells were washed three times with DPBS with Ca2+

and Mg2+ and fixed with 1% to 2% formaldehyde in water (Sigma, St. Louis, MO) at roomtemperature for 9 minutes and washed. Cells were then permeabilized with 0.5% TritonX-100 (Sigma, St. Louis, MO) for 15 minutes at room temperature and washed. To stainfor fibronectin in the extracellular matrix, samples were then stained with monoclonalanti-fibronectin antibody produced in mouse, clone IST-4 (Sigma, St. Louis, MO) at a1:100 dilution at 4oC overnight and washed three times with DPBS with Ca2+ and Mg2+.Samples were then stained with Fluorescein (FITC) AffiniPure donkey anti mouse IgG(Jackson Immunolabs, Suffolk, U.K.) at a 1:200 dilution for 2 hours at room temperatureand washed. Finally, cells were stained with Texas Red phalloidin (Sigma, St. Louis, MO)at a 1:200 dilution for 30 minutes at room temperature for visualization of actin.

6.4.7 Cell Attachment Assay

To test the degree of cell attachment as a function of GST-FNIII9-10 loading on thenanoparticles, 96 well nontissue culture dishes were coated overnight at 4ºC with eitherGST- FNIII9-10 at 2.5 to 10 mg/mL or nanoparticles conjugated with increasing levels ofGST- FNIII9-10. Wells were washed three times with PBS to remove unbound ligand andblocked with 1% bovine serum albumin (Sigma, St. Louis, MO) for 1 hour at 37ºC. Sub-strates were washed three times with PBS and cells added at 35,000 cells/well for 90 min-utes at 37ºC. Wells were washed twice with PBS and the number of cells adhered tosurface determined using the hexosaminidase assay [31]. Briefly, 60 μl of substrate (com-posed of equal volumes of 0.5% Triton X-100 (Sigma, St. Louis, MO) and 7.5 mMb-nitrophenyl N-acetyl b-D glucosaminide (Sigma, St. Louis, MO) in 0.1 M citrate buffer,pH 5.0) was added to the cells and incubated for 90 minutes at 37ºC. After terminatingthe reaction by the addition of 90 μl per well of 50 mM glycine/5mM EDTA (Sigma, St.Louis, MO), pH 10.4, the absorbance was read at 405 nm on a plate reader.

6.5 Results

In this section, activated response from cell interaction with functionalized albuminnanoparticles is outlined. As discussed in Section 6.3, it has been observed that cell bind-ing domain exposure of the ligand, GST-FNIII9-10, is increased by presenting the ligandon ANPs in comparison to adsorption on a 2-D substrate. Therefore, certain cellresponses important during wound healing events have been activated with the use ofGST-FNIII9-10-ANPs, including keratinocyte migration, extracellular matrix assembly byfibroblasts, and spatially guided attachment of fibroblasts and human mesenchymalstem cells.

6.5.1 Enhanced Cell Migration

Keratinocyte migration occurs early after the onset of a wound in the skin, in order toclose the wound and begin healing events. Previous studies showed that ligands pre-sented on a nanoscale system could lead to integrin clustering and enhanced migration[7], while other studies showed that ligands on dynamic, internalizable submicron parti-cles resulted in enhanced cell migration [32, 33]. Conjugating ligands on biodegradable

6.5 Results

95

nanoparticles presented the opportunity to create a nanoscale interface that is dynamicand would allow cells to interact with the matrix via integrins and promoting themto actively bind, sequester, and possibly internalize the ligand-functionalized nano-particles via specific receptor-mediated processes.

Incubation of keratinocytes with GST-FNIII9-10 altered cytoskeletal organization ofthe cells [21]. The effect on cytoskeletal morphology due to ligand presentation wasexamined by staining keratinocytes for F-actin, as shown in Figure 6.3. Cells culturedon GST-FNIII9-10 adsorbed substrates appear to have well defined stress fibers, indicat-ing strong attachment to the substrate; however, when the cells were cultured onGST-FNIII9-10-ANPs, they exhibited more filopodial extensions but fewer stress fibers,indicating a more motile phenotype [21]. Keratinocytes cultured on ANP-adsorbed sub-strates appeared more rounded and exhibited fewer filopodia and a less organizedcytoskeletal morphology. Staining for molecular markers of cell adhesion showed anenhanced localization of phosphorylated focal adhesion kinase and paxillin, both com-ponents of the focal adhesion complex, in cells seeded on GST-FNIII9-10-adsorbed sub-strates. Significantly lower expression was seen in cells seeded on GST-FNIII9-10-ANPs,while minimal levels are detected on unfunctionalized ANPs.

The process of attachment reflects the earliest response of cells to a surface. By vary-ing the density of the ligand presented to keratinocytes, it was possible to determinewhether the presentation of ligand via the ANPs modulated attachment of the cells.Keratinocytes were seeded on surfaces coated with the varying ligand densities (deter-mined by ELISA) either displayed on ANPs or directly adsorbed nontissue culture poly-styrene. Equal cell seeding densities were used, and cell attachment was determined bythe hexosaminidase assay.

While minimal cell attachment was detected for the unfunctionalized ANP condi-tions, at each ligand density a significant increase in cell attachment was observed whendisplaying the ligand on the ANPs in comparison to the ligand adsorbed to the substrate[21]. Previous reports of cell adhesion behavior on RGD-containing ligands displayedfrom surface configurations that induced ligand clustering [7, 34, 35] indicate enhancedcell attachment and adhesion strength. Our system differs, however, in that the cellsadhered to the GST-FNIII9-10-ANPs lacked dominant stress fibers and exhibited morefilopodial extensions as well as phosphorylated focal adhesion kinase and paxillin,


96

Figure 6.3 Fluorescent confocal microscopy images of keratinocytes incubated at 8,400 cells/cm2 onLab-Tek chamber slides with #1 glass coverslip bottoms coated with 10 μg/ml of (a) GST-FNIII9-10, (b)ligand-conjugated ANPs, and (c) unconjugated ANPs. After 5 hours, cells were fixed and stained withfluorescein-phalloidin to visualize the actin cytoskeleton.

which are components associated with stable focal adhesion [36, 37], whereaskeratinocytes on the GST-FNII9-10 controls showed a stationary phenotype andupregulated the expression of focal adhesion proteins. This supports the observationthat ligand presentation from the nanoparticles promotes availability of the ligand tocells and improves cell attachment while decreasing cell adhesion strength.

The increased filopodia observed on keratinocytes seeded on GST-FNIII9-10-ANPs indi-cates a more motile phenotype of these cells. To evaluate cellular migration with the useof functionalized nanoparticles, wells were adsorbed with either equivalent concentra-tions of GST-FNIII9-10, GST-FNIII9-10-ANPs and unfunctionalized ANPs. After 10 hours, themean squared displacement was larger for cells seeded on GST-FNIII9-10-ANPs comparedto GST-FNIII9-10 alone or unfunctionalized ANPs (Figure 6.4) [21]. Other studies previ-ously reported that keratinocyte migration can be governed by the availability of cellbinding domains (i.e., type III repeat domains 9 and 10 of fibronectin) [38, 39]. In ourstudies, when equivalent levels of ligand were presented either conjugated to thenanoparticles or adsorbed to the substrate, increased cell binding domain availability ofthe ligand by presentation on the ANPs was detected via immunoabsorbance assay andshown in Figure 6.2. Due to potential differences in surface energetics of the GST-FNIII9-10

fragment at the ANP surface [40–43], it is possible that conformational changes in theligand occurred upon functionalization to the ANPs [44].

6.5.2 Enhanced Extracellular Matrix Assembly

We also applied our ANP system to dermal fibroblast fibronectin matrix assembly. Therigidity of a substrate influences the organization of the actin cytoskeleton and changesfibroblast contractility, which has been shown to play a role in matrix assembly. In thepresent work, we explored the use of ANPs to alter the rigidity of the substrate at thenanoscale and regulate fibronectin matrix assembly. We hypothesized that, while the

6.5 Results

97

Figure 6.4 Single cell migration was examined on substrates with ligand, nanoparticles, andligand-conjugated nanoparticles. Keratinocytes were seeded at 2,800 cells/cm2 for 4 hours prior toimage acquisition. Images were taken over 10 hours. Cells were tracked and data was fit to modelscharacterizing cell migration for single-cell migration experiments. Error bars represent standard erroraround the mean. For each experiment, n = 60. Inset: Random motility coefficients were calculated forcells on various ligand-adsorbed substrates.

presentation of the extracellular ligand on the nanoparticles would modify the displayand overall organization of the ligand, varying the size of the nanoparticles would resultin different levels of cytoskeletal tension in fibroblasts, which would lead to differentdegrees of matrix assembly [27, 45].

To isolate the influence of the GST-FNIII9-10 presentation from ANPs on matrix assem-bly, experiments were conducted in a serum-free media that supported comparable lev-els of cell viability as examined by ethidium homodimer labeling [46] (data not shown).Serum-supplemented media may contain lysophosphatidic acids, which would promotematrix assembly by inducing contraction, and soluble fibronectin [47, 48]. In theserum-free environment, cells rely on the clustering of their integrins to the substrate toinduce spreading and adhesion by activating focal adhesion kinase and other small RhoGTPases [49].

Cell attachment, adhesion, and cytoskeletal organization are necessary during extra-cellular matrix assembly. To evaluate the role of ligand presentation in this process,immunofluorescence analysis was performed on fibroblasts seeded on substrates withdifferent densities of GST-FNIII9-10-ANPs [27, 45]. Using particles with the highest liganddensity, we observed a correlation between increasing numbers of assembled fibronectinfibrils in the extracellular matrix and size of the nanoparticles [27]. The largest GST-FNIII9-10-ANPs supported fibril formation, which was detectable by 24 hours (Figure 6.5)with more prominent fibronectin matrix fibrils forming after 48 hours. On smallernanoparticles, or on GST-FNIII9-10 alone adsorbed to the substrate, occasional short fibrilswere detected, but most of the fibronectin staining appeared to be intracellular.

At lower ligand densities, there was no detectable assembly of fibronectin matrix onGST-FNIII9-10-ANPs or GST-FNIII9-10 adsorbed on the substrate, and negligible amounts offibronectin were assembled by fibroblasts cultured on unfunctionalized nanoparticles.Using immunochemistry techniques, it was possibly to quantifying fibronectin matrixassembly by fibroblasts after 24 hours in culture at the highest ligand loading comparedto unfunctionalized nanoparticles. These results indicated a distinguishable differencein fibronectin assembled between different sized nanoparticles at the highest ligandloading (Figure 6.5). A 22% increase of assembled fibronectin matrix was observed withlarger nanoparticles, and quantification of fibril densities showed greater than 10-foldhigher number on 100- and 125-nm nanoparticles compared to 30- to 50-nmGST-FNIII9-10-ANPs. These results demonstrate that nanoparticle size is an important fac-tor of fibronectin matrix assembly, and suggest that cell binding events and subsequentcell function can be modulated not just by the nanoscale presentation of the ligand andligand density, but nanoparticle size as well.

To investigate the effect of nanoscale presentation of ligand on initial cell bindingand attachment events, equal numbers of fibroblasts were seeded in parallel on sub-strates of three different ligand densities either adsorbed on the surface or functionalizedonto the ANPs. Cell attachment increases with increasing ligand density [27]. For aspecific ligand density, the highest cell attachment was seen on the largest sizednanoparticles (~125 nm); significant attachment, but to a lesser degree, was alsoobserved on the 100-nm sized nanoparticles and on the ligand-only substrate, both dis-tinguishable from each other when analyzed by ANOVA. On 30- and 50-nm sizednanoparticles, attachment was low. Reports by our group and others affirm that cellattachment to ligands presented on substrates that promote integrin clustering also


98

enhances cell attachment and adhesion strength [7, 21, 45]. When adhered, these cellsproduce and deposit fibronectin into the extracellular matrix [50].

The effect of ligand presentation on cytoskeletal organization were examined bystaining for F-actin at different time points [27]. Cells on larger-sized GST-FNIII9-10-ANPsbegan to exhibit stress fibers as early as 1 hour postseeding. Cells cultured on ligand-adsorbed substrates had some filopodial projections after 1 hour in culture, whilethose on smaller-sized, functionalized nanoparticles showed restricted spreading andremained rounded. At 5 hours of culture, cells on all substrates appeared well-spreadwith well defined stress fibers, although cells on ligand-functionalized nanoparticlesappeared more elongated [27]. These results show that cell attachment and cytoskeletalorganization occur more rapidly (i.e., within 1 hour) on larger nanoparticles than onsmaller ones. This observation is congruent with previous studies that reported thatfibroblast matrix assembly can be governed by the dimensionality of the substrate[51, 52].

6.5 Results

99

Figure 6.5 (a) Increased ligand concentration and ANP size promote assembly of fibronectin matrix. Humanforeskin fibroblasts were serum-starved overnight and seeded on substrates with ligand at 2.2 μg/cm2 for 24hours (left column) or 48 hours (right column). Cells were fixed, permeabilized, and processed forimmunofluorescence. Matrix fibrils were visualized using a monoclonal mouse anti-human fibronectin epitopelocated within domain 5 of the type III repeats, followed by FITC-conjugated secondary antibody. Cells were alsostained for F-actin with Texas Red phalloidin. Increased culture time allowed cells seeded on smaller-sizednanoparticles to elongate and organize actin into filaments, yet matrix assembly did not commence, while cellson larger-sized nanoparticles not only developed a more organized cytoskeleton but also assembled moreextracellular fibronectin. Images were acquired at 63x, zoom 1. (b) Extent of matrix assembly was quantifiedusing ELISA techniques. Cells were cultured on substrates and lysed to leave behind the assembled matrix. Sub-strates were then blocked and incubated with anti-human fibronectin for domain 5 of the type III repeats, fol-lowed by enzyme linked secondary antibody. Values of ELISA absorbance were derived by back-calculating theconcentration based on the standard curve of whole length fibronectin. The star (*) represents statistical signifi-cance via ANOVA analysis when experiments were conducted in duplicate three times (p<0.05). In summary,greater levels of fibronectin matrix was assembled on substrates with larger sized nanoparticles.

6.6 Discussion of Pitfalls

6.6.1 Spatial Guidance of Cell Attachment—Microscale Plasma InitiatedPatterning

One of the shortcomings with nanoscale presentation of ligands via nanoparticles is thedifficulty to control the arrangement of nanoparticles, resulting also in a difficulty tospatially control cell attachment and tissue formation. The arrangement of living cells iscrucial for the functionality of tissues during development and regeneration [53, 54].Therefore, spatial guidance of cells can serve to provide clues on functions such asattachment and spreading [55], migration [56], and cell-matrix interactions [57], whichcould then provide crucial information on tissue morphogenesis and networking [55].However, most studies to date have been performed with whole proteins such asfibronectin, making it difficult to understand the role of smaller ligands and peptideson these cell processes. Patterning with these small biomolecules, however, can bechallenging, since conformational changes due to substrate binding can render thebiomolecules inactive.

We explored the use of microscale plasma initiated patterning (μPIP) [29], a novel,efficient, and widely applicable approach to direct the patterning of GST-FNIII9-10-ANPson nonconductive, biodegradable polymeric substrates that served as templates to elicitadhesion and spreading of human mesenchymal stem cells (hMSC’s) and fibroblastsinto arrays with superior ordering over similarly patterned ligands. During this process,selective regions of Poly(DTE-co-8% PEG1K carbonate) were exposed to oxygen plasmafor selective nanoparticle deposition (Figure 6.6(a) and (b)). Nanoparticle patterning wasconfirmed with atomic force microscopy and scanning electron microscopy (data notshown) and confirmed that while nanoparticles form a monolayer on plasma-exposedregions of the polymer, they minimally adsorb to the unexposed areas of the polymer.

Cell patterning with GST-FNIII9-10-ANPs was confirmed with the use of humanfibroblasts (Figure 6.6) and human mesenchymal stem cells (hMSC’s) (data not shown).Cell patterning is particularly interesting because, by controlling functions such asattachment and spreading, a control over differentiation and function can also beobtained [57, 58]. When using GST-FNIII9-10 alone, fibroblasts formed small patternedareas, but they did not spread evenly within the plasma-exposed areas (data not shown).Patterns appeared patchy, only forming in small areas scattered throughout the sample.Most cells spread on the substrate in a random, disorganized way, with cells thinand elongated and not spread inside the plasma-treated area of 40 μm. Furthermore,fibroblasts appear to be confined to the edges of the plasma-exposed polymer stripes,and the few patterns that were observed with the ligand alone may have resulted fromthe combination of the presence of protein and these plasma-exposed polymer edgesrather than from patterned ligand [59]. In Figure 6.6(c) through (d), patterns formed byseeding fibroblasts on GST-FNIII9-10-ANPs. In this case, distinct patterns covering theentire stamped area can be obtained by using the ligand-functionalized nanoparticles.Cells remained highly confined to the plasma-exposed areas yet spread along theplasma-exposed area, and edge effects were not observed even when using the smalleststamp size of 10 by 10 μm.

Adsorption of the ANPs on untreated poly(DTE-co-8% PEG1K carbonate) was con-firmed, by ELISA, to be lower than on poly(DTE-co-8% PEG1K carbonate) plasma-treatedfor 60 and 120 seconds (data not shown). The lower nanoparticle adsorption on the


100

untreated polymer is caused by the presence of poly(ethylene glycol), or PEG, whichinhibits protein, and subsequently nanoparticle, adsorption. Plasma treatment, how-ever, increases the surface energy and negative charge on the surface of the polymer. Thehigher surface energy of poly(DTE-co-8% PEG1K carbonate) is indicated by the com-pletely wetting contact angle upon plasma treatment, in comparison to the contactangle of the untreated polymer of 69±2o.

The ELISA data also showed that plasma-treatment of the polymer for 60 secondsinduces greater levels of ANP adsorption than plasma-treatment of the polymer for 120seconds. It is possible that while the 60-second exposure renders the polymer morehydrophilic than the untreated polymer, 120-second exposure renders the polymerexcessively hydrophilic and excessively increases the net negative charge (data notshown), inhibiting protein nanoparticle adsorption in comparison to 60-second plasmaexposure.

Adsorption of GST-FNIII9-10 on untreated poly(DTE-co-8% PEG1K carbonate) is statisti-cally similar to poly(DTE-co-8% PEG1K carbonate) treated for 60 seconds and onlyslightly higher than poly(DTE-co-8% PEG1K carbonate) treated for 120 seconds (data notshown). This data likely indicates that any difference in cell binding exposure from theligand is not due to differences in adsorption or binding of the ligand onto untreatedand plasma-treated surface, and the ligand is more uniformly exposed throughout thesubstrate.

6.6.2 Three-Dimensional Presentation of Albumin Nanoparticles

The surface presentation of ligands at the nanoscale has been successfully used to mimicligand clustering and induce integrin clustering [7, 9, 11, 60]. Nanoscale ligand presenta-


101

Figure 6.6 (a) Schematic illustrating the microscale plasma initiated patterning process; briefly, aPDMS stamp is placed on the surface of interest and plasma treated in oxygen gas. Areas exposed tothe plasma undergo surface functionalization via the formation of end groups (such as COO-, COOH)by interaction with the free radicals and ions in the oxygen gas. (b) Biofunctional nanoparticles thenpreferentially adsorb to the exposed area of the material. (c) Fluorescent microscopy image offibroblasts patterned with functionalized ANPs using a 40- by 40-μm stamp, showing that cells spreadacross the pattern, adapting to the topography of the stripe. (With ligand alone, patterning was spo-radic, and found only in small areas.) (Green=actin; blue= DAPI). (d) Edge effects are not observed evenby patterning using the 10- by 10-μm stamp with the nanoparticles, despite the confined area.

tion has induced cellular processes such as migration [7, 21] and extracellular matrixprotein secretion and assembly [27, 45]. No doubt that the use of nanoparticles in 2-Dcell studies has yielded interesting results and furthered the understanding of the invitro behavior of cells.

Nevertheless, the effect of nanoparticles in 3-D and in vivo models has been signifi-cantly unexplored, and their application in animal and human models is majorlyunknown. Nanoparticle toxicity is currently a hotly debated topic, and the scientificcommunity is just beginning to understand how nanoparticle geometry, size and chem-istry may affect their behavior in vivo.

We have begun exploring these possibilities by adsorbing functionalized albuminnanoparticles onto electrospun polymer scaffolds. These scaffolds, fabricated withPoly(DTE carbonate) (courtesy of Prof. J. Kohn, Rutgers University) were chosen due totheir fibrous architecture, which can serve as a model that resembles the fibrousextracellular matrix architecture of skin tissue. Human fibroblasts, the producers ofextracellular matrix in skin, were chosen for these preliminary experiments.

Figure 6.7(a) shows a scanning electron microscopy image of the electrospunscaffolds after fabrication. Fiber size was of 3.0 ± 0.7 μm, while porosity was of 11 ± 2μm. Figure 6.7(b) shows another scanning electron microscopy image confirming theadsorption of ANPs on the surface of the electrospun scaffold fibers. Figure 6.7(c) illus-trates the behavior of human fibroblasts when cultured on the electrospun fibrous scaf-fold with GST-FNIII9-10-ANPs, which induces cells to spread amongs the fibers. This couldindicate that the functionalized nanoparticles present multiple mobile anchor pointsalong the fibers and promote multiple attachment processes by the cells. Fibroblastsmake contact with several fibers, potentially showing improved cell:cell contacts thatare necessary for wound healing processes.

Based on studies currently underway, nanoparticles presenting biochemicalcues may also be capable of altering cell behavior in vivo, with applications to matrixregeneration and tissue repair processes. However, issues such as nanoparticle toxicity,long-term nanoparticle fates, and treatment effectiveness must be addressed in order forthese efforts to be successfully translated to practice.

6.7 Summary Points

In this work, we explored the use of biodegradable, biocompatible albumin nano-particles that were functionalized with a truncated fragment of fibronectin that encom-


102

Figure 6.7 Scanning electron microscopy image of (a) electrospun poly(DTE carbonate) fibrous scaf-folds and (b) ANPs adsorbed on the fibers of the electrospun fibrous scaffolds. (c) Confocal microscopyimages of fibroblast morphology on electrospun fibrous scaffolds with GST-FNIII9-10-ANPs.

passes the synergy sequence and the RGD-containing sequence of fibronectin. Thefunctionalized nanoparticles were shown to:

• Enhance the presentation of the cell binding domain of the ligand to cells;

• Enhance cell attachment onto 2-D substrates;

• Promote a more motile phenotype in keratinocytes;

• Increase the production and assembly of fibronectin matrix by fibroblasts;

• Promote the spatial attachment of cells into patterned templates;

• Enhance spreading of fibroblasts onto 3-D, electrospun fibrous scaffolds.

Acknowledgments

The authors would like to thank Prof. J. Kohn for production of the Poly(DTE-co-8%PEG1K carbonate) and Prof. D. Shreiber for helpful discussion. The authors would alsolike to thank Matthew Treiser, Rebecca Moore, Jing Xu, and Vanesa Figueroa for helpfuldiscussion. This work was funded by the NSF NIRT Grant Number 0609000. The projectdescribed was supported by Grant Number T32EB005583 from the National Institute ofBiomedical Imaging and Bioengineering. The content is solely the responsibility of theauthors and does not necessarily represent the official views of the National Institute ofBiomedical Imaging and Bioengineering or the National Institutes of Health.

References

[1] Gumbiner, B.M., “Cell adhesion the molecular basis of tissue architecture and morphogenesis,”Cell, Vol. 84, No. 3 1996, pp. 345–357.

[2] Alberts, B., et al., Molecular Biology of the Cell, 4th Edition. 2002, New York: Garland Publishing.[3] Brassard, D.L., et al., “Armstrong, Integrin alpha(v)beta(3)-mediated activation of apoptosis,” Exp

Cell Res, Vol. 251, No. 1 1999, pp. 33–45.[4] Giancotti, F.G. and E. Ruoslahti, “Integrin signaling,” Science, Vol. 285, No. 5430 1999,

pp. 1028–1032.[5] Ruoslahti, E., “RGD and other recognition sequences for integrins,” Annu. Rev. Cell Dev. Biol., Vol.

12, No. 1996, pp. 697–715.[6] Yauch, R.L., et al., “Mutational evidence for control of cell adhesion through integrin diffu-

sion/clustering, independent of ligand binding,” J Exp Med, Vol. 186, No. 8 1997, pp. 1347–1355.[7] Maheshwari, G., et al., “Cell adhesion and motility depend on nanoscale RGD clustering,” J Cell

Sci, Vol. 113 ( Pt 10), No. 2000, pp. 1677–1686.[8] Arnold, M., et al., “Activation of integrin function by nanopatterned adhesive interfaces,”

Chemphyschem, Vol. 5, No. 3 2004, pp. 383–388.[9] Cavalcanti-Adam, E.A., et al., “Cell spreading and focal adhesion dynamics are regulated by spac-

ing of integrin ligands,” Biophys J, Vol. 92, No. 8 2007, pp. 2964–2974.[10] Lipski, A.M., et al., “The effect of silica nanoparticle-modified surfaces on cell morphology,

cytoskeletal organization and function,” Biomaterials, Vol. No. 2008.[11] Mannix, R.J., et al., “Nanomagnetic actuation of receptor-mediated signal transduction,” Nature

Nanotechnology, Vol. 3, No. 1 2008, pp. 36–40.[12] Murphy, E.A., et al., “Nanoparticle-mediated drug delivery to tumor vasculature suppresses metas-

tasis,” Proc Natl Acad Sci U S A, Vol. 105, No. 27 2008, pp. 9343–9348.[13] Arnedo, A., S. Espuelas, and J.M. Irache, “Albumin nanoparticles as carriers for a phosphodiester

oligonucleotide,” International Journal of Pharmaceutics, Vol. 244, No. 1-2 2002, pp. 59–72.[14] Arshady, R., “Albumin Microspheres and Microcapsules—Methodology of Manufacturing Tech-

niques,” Journal of Controlled Release, Vol. 14, No. 2 1990, pp. 111–131.[15] Rhodes, B.A., et al., “Radioactive albumin microspheres for studies of the pulmonary circulation,”

Radiology, Vol. 92, No. 7 1969, pp. 1453–1460.

Acknowledgments

103

[16] Michaelis, K., et al., “Covalent linkage of apolipoprotein E to albumin nanoparticles stronglyenhances drug transport into the brain,” Journal of Pharmacology and Experimental Therapeutics,Vol. 317, No. 3 2006, pp. 1246–1253.

[17] Wartlick, H., et al., “Tumour cell delivery of antisense oligonuclceotides by human serum albuminnanoparticles,” J Control Release, Vol. 96, No. 3 2004, pp. 483–495.

[18] Takeoka, S., et al., “Conjugation of von Willebrand factor-binding domain of platelet glycoproteinIb alpha to size-controlled albumin microspheres,” Biomacromolecules, Vol. 1, No. 2 2000,pp. 290–295.

[19] Takeoka, S., et al., “Fibrinogen-conjugated albumin polymers and their interaction with plateletsunder flow conditions,” Biomacromolecules, Vol. 2, No. 4 2001, pp. 1192–1197.

[20] Albumin : structure, biosynthesis, function, I.S. Theodore Peters, ed., 1978, New York: PergamonPress.

[21] Sharma, R.I., et al., “Albumin-derived nanocarriers: substrates for enhanced cell adhesive liganddisplay and cell motility,” Biomaterials, Vol. 27, No. 19 2006, pp. 3589–3598.

[22] Petersen, T.E., et al., “Partial primary structure of bovine plasma fibronectin: three types of internalhomology,” Proc Natl Acad Sci U S A, Vol. 80, No. 1 1983, pp. 137–141.

[23] Altroff, H., L. Choulier, and H.J. Mardon, “Synergistic activity of the ninth and tenth FIII domainsof human fibronectin depends upon structural stability,” J Biol Chem, Vol. 278, No. 1 2003,pp. 491–497.

[24] Main, A.L., et al., “The three-dimensional structure of the tenth type III module of fibronectin: aninsight into RGD-mediated interactions,” Cell, Vol. 71, No. 4 1992, pp. 67167–8.

[25] Schwarzbauer, J.E. and J.L. Sechler, “Fibronectin fibrillogenesis: a paradigm for extracellular matrixassembly,” Curr Opin Cell Biol, Vol. 11, No. 5 1999, pp. 622–627.

[26] Hermanson, G., Bioconjugate Techniques, 1996, New York: Academic Press.[27] Sharma, R.I., D.I. Shreiber, and P.V. Moghe, “Nanoscale Variation of Bioadhesive Substrates as a

Tool for Engineering of Cell Matrix Assembly,” Tissue Engineering, Vol. 14, No. 7 2008,pp. 1237–1250.

[28] Tziampazis, E., J. Kohn, and P.V. Moghe, “PEG-variant biomaterials as selectively adhesive proteintemplates: model surfaces for controlled cell adhesion and migration,” Biomaterials, Vol. 21, No. 52000, pp. 511–520.

[29] Langowski, B.A. and K.E. Uhrich, “Microscale plasma-initiated patterning (mu PIP),” Langmuir,Vol. 21, No. 23 2005, pp. 10509–10514.

[30] Dimilla, P.A., et al., “Maximal Migration of Human Smooth-Muscle Cells on Fibronectin andType-Iv Collagen Occurs at an Intermediate Attachment Strength,” Journal of Cell Biology, Vol. 122,No. 3 1993, pp. 729–737.

[31] Landegren, U., “Measurement of Cell Numbers by Means of the Endogenous EnzymeHexosaminidase - Applications to Detection of Lymphokines and Cell-Surface Antigens,” Journal ofImmunological Methods, Vol. 67, No. 2 1984, pp. 379–388.

[32] Tjia, J.S. and P.V. Moghe, “Cell migration on cell-internalizable ligand microdepots: Aphenomenological model,” Annals of Biomedical Engineering, Vol. 30, No. 6 2002, pp. 851–866.

[33] Tjia, J.S. and P.V. Moghe, “Regulation of cell motility on polymer substrates via “dynamic,” cellinternalizable, ligand microinterfaces,” Tissue Engineering, Vol. 8, No. 2 2002, pp. 247–261.

[34] Danilov, Y.N. and R.L. Juliano, “(Arg-Gly-Asp)n-albumin conjugates as a model substratum forintegrin-mediated cell adhesion,” Exp Cell Res, Vol. 182, No. 1 1989, pp. 186–196.

[35] Koo, L.Y., et al., “Co-regulation of cell adhesion by nanoscale RGD organization and mechanicalstimulus,” J Cell Sci, Vol. 115, No. Pt 7 2002, pp. 1423–1433.

[36] Tran, H., et al., “Integrin clustering induces kinectin accumulation,” J Cell Sci, Vol. 115, No. Pt 102002, pp. 2031–2040.

[37] Turner, C.E., “Paxillin interactions,” J Cell Sci, Vol. 113 Pt 23, No. 2000, pp. 4139–4140.[38] Kim, J.P., et al., “Mechanism of human keratinocyte migration on fibronectin: unique roles of RGD

site and integrins,” J Cell Physiol, Vol. 151, No. 3 1992, pp. 443–450.[39] Kim, J.P., et al., “Integrin receptors and RGD sequences in human keratinocyte migration: unique

anti-migratory function of alpha 3 beta 1 epiligrin receptor,” J Invest Dermatol, Vol. 98, No. 5 1992,pp. 764–770.

[40] Latha, M.S. and A. Jayakrishnan, “A new method for the synthesis of smooth, round, hydrophilicprotein microspheres using low concentrations of polymeric dispersing agents,” J Microencapsul,Vol. 12, No. 1 1995, pp. 7–12.

[41] Longo, W.E. and E.P. Goldberg, “Hydrophilic albumin microspheres,” Methods Enzymol, Vol. 112,No. 1985, pp. 18–26.

[42] Toworfe, G.K., et al., “Fibronectin adsorption on surface-activated poly(dimethylsiloxane) and itseffect on cellular function,” J Biomed Mater Res A, Vol. 71, No. 3 2004, pp. 449–61.


104

[43] van der Walle, C.F., H. Altroff, and H.J. Mardon, “Novel mutant human fibronectin FIII9-10domain pair with increased conformational stability and biological activity,” Protein Eng, Vol. 15,No. 12 2002, pp. 1021–1024.

[44] Du, H., P. Chandaroy, and S.W. Hui, “Grafted poly-(ethylene glycol) on lipid surfaces inhibits pro-tein adsorption and cell adhesion,” Biochim Biophys Acta, Vol. 1326, No. 2 1997, pp. 236–248.

[45] Pereira, M., et al., “Engineered cell-adhesive nanoparticles nucleate extracellular matrix assembly,”Tissue Eng, Vol. 13, No. 3 2007, pp. 567–578.

[46] Imbert, D. and C. Cullander, “Assessment of cornea viability by confocal laser scanning micros-copy and MTT assay,” Cornea, Vol. 16, No. 6 1997, pp. 666–674.

[47] Zhang, Q., et al., “Modulation of cell surface fibronectin assembly sites by lysophosphatidic acid,” JCell Biol, Vol. 127, No. 5 1994, pp. 1447–1459.

[48] Zhang, Q., M.K. Magnusson, and D.F. Mosher, “Lysophosphatidic acid andmicrotubule-destabilizing agents stimulate fibronectin matrix assembly through Rho-dependentactin stress fiber formation and cell contraction,” Mol Biol Cell, Vol. 8, No. 8 1997, pp. 1415–1425.

[49] Lehnert, D., et al., “Cell behaviour on micropatterned substrata: limits of extracellular matrixgeometry for spreading and adhesion,” J Cell Sci, Vol. 117, No. Pt 1 2004, pp. 41–52.

[50] Sottile, J., D.C. Hocking, and P.J. Swiatek, “Fibronectin matrix assembly enhances adhe-sion-dependent cell growth,” J Cell Sci, Vol. 111 ( Pt 19), No. 1998, pp. 2933–2943.

[51] Beningo, K.A., M. Dembo, and Y.L. Wang, “Responses of fibroblasts to anchorage of dorsalextracellular matrix receptors,” Proc Natl Acad Sci U S A, Vol. 101, No. 52 2004, pp. 18024–18029.

[52] Mao, Y. and J.E. Schwarzbauer, “Stimulatory effects of a three-dimensional microenvironment oncell-mediated fibronectin fibrillogenesis,” J Cell Sci, Vol. 118, No. Pt 19 2005, pp. 4427–4436.

[53] Bao, S. and R. Cagan, “Preferential adhesion mediated by Hibris and Roughest regulatesmorphogenesis and patterning in the Drosophila eye,” Dev Cell, Vol. 8, No. 6 2005, pp. 925–935.

[54] Hayashi, T. and R.W. Carthew, “Surface mechanics mediate pattern formation in the developingretina,” Nature, Vol. 431, No. 7009 2004, pp. 647–652.

[55] Chen, C.S., et al., “Micropatterned surfaces for control of cell shape, position, and function,”Biotechnol Prog, Vol. 14, No. 3 1998, pp. 356–363.

[56] Li, S., et al., “Effects of morphological patterning on endothelial cell migration,” Biorheology,Vol. 38, No. 2-3 2001, pp. 101–108.

[57] Goffin, J.M., et al., “Focal adhesion size controls tension-dependent recruitment of alpha-smoothmuscle actin to stress fibers,” J Cell Biol, Vol. 172, No. 2 2006, pp. 259–268.

[58] McBeath, R., et al., “Cell shape, cytoskeletal tension, and RhoA regulate stem cell lineage commit-ment,” Developmental Cell, Vol. 6, No. 2004, pp. 483–495.

[59] Rossi, M.P., et al., “Enhanced Cell-Interactive Display of Biofunctionalized Nanoparticles viaPlasma-Initiated Patterning,” Small, Vol. In Review, No. 2008.

[60] Cavalcanti-Adam, E.A., et al., “Lateral spacing of integrin ligands influences cell spreading andfocal adhesion assembly,” Eur J Cell Biol, Vol. 85, No. 3-4 2006, pp. 219–224.

References

105

C H A P T E R

7Magnetic Cell Separation to Enrich forRare Cells

Ying Xiong1, Mei Shao1, Maciej Zborowski2, William G. Lowrie1, andJeffrey J. Chalmers1,3, *

1Department of Chemical and Biomolecular Engineering, The Ohio State University, 125 Koffolt Laborato-ries, 140 West 19th Avenue, Columbus, OH 432102Department of Biomedical Engineering, Cleveland Clinic, 9500 Euclid Avenue, Cleveland, OH 44195,Phone: 216-445-9342, Fax: 216-444-9198, e-mail: [email protected]

*Corresponding Author: 3Director, University Cell Analysis and Sorting Core,Phone: 216-292-2727, Fax: 216-292-3769, e-mail: [email protected]

107

Key terms magnetic cell separationmagnetic nanoparticlescirculating tumor cellscell staining

Abstract

Magnetic cell separation has become a basic cell preparation tool used in a largenumber of research, and to less extent, clinical laboratories. A potential clinicalapplication of magnetic cell separation is the enrichment of circulating tumorcells, CTC. In this chapter, a methodology will be presented in which a purelynegative depletion of human blood cells from cancer patients is used to enrichfor rare circulating tumor cells. It is suggested that such a separation approachis more general than positive selection and allows for a variety of rare, unde-fined cells to be isolated.

7.1 Introduction

Some of the earliest reports of magnetic cell separation include the application ofmagnetic fields to remove iron-loaded cells from heterogeneous cell suspension; forexample, kupffer cells in rat liver [1] and erythrocytes in blood [2, 3]. The theoreticalbasis of magnetically separating cells containing iron was originally proposed by Paulingand Coryell [4].

As might be expected, as the concept of magnetically separating cells containingiron was being demonstrated, the idea of magnetically separating cells labeled with anti-bodies conjugated to particles containing iron was proposed and developed [5]. Sincethen, magnetic cell separation has become increasingly popular and has expandedbeyond cells to viruses and proteins. The popularity is probably most evident by thesignificant number of commercial system on the market (i.e., the Miltenyi Biotech fam-ily of MACS products, Dynabeads, Easy Sep by Stem Cell Technologies, and the Veridexsystem).

Generally, the magnetic particles employed in magnetic cell separation consist ofiron oxides cores (such as magnetite, Fe3O4) surrounded by a polymeric layer [6, 7]. Thesepolymers include natural dextran, various polysaccharides, polyacrylamide-agarose [8],polyglutaraldehyde [9], polyacrolein [10], and even proteins [11]. The size of these mag-netite core and polymeric shell combinations can vary from 10 nm for ferritin, to 4.5microns for Dynabeads. A summary of a number of commercially available, magneticparticle conjugated antibodies is shown in Table 7.1.

Obviously, in addition to the reagents, a magnetic system and labeling methodologyis also required to separate the cells. From a methodology perspective, cell separationcan be considered either as positive selection: targeting of the desired cell (i.e.,immunomagnetically labeling the targeted cell) and indirectly removing all other celltypes, or negative depletion: targeting of the undesirable cell (i.e., immunomagneticallylabeling all of the unwanted cells). The actual magnetic separation systems reported inthe literature can be classified as (1) batch systems (i.e., collection of the magneticallylabeled cells on the walls of a test tube, unlabeled cells settling to the bottom of a tubeor remaining in suspension; EasySep), (2) continuous flow through systems (a heteroge-neous mixture of labeled and unlabeled cells enter the separator, and labeled andunlabeled cells flow out (i.e. Quaduropole magnetic cell separation system [12]), or (3) ahybrid of the two in which the magnetically labeled cells are retained, either on the wallof a column, or within a “packed column” containing small steel spheres or wire (i.e.,MACS columns or annular tubes within a magnetic field [13]). Figure 7.1 presents pho-tos and or diagrams representative of commercial representatives of two of these types ofsystems and Table 7.2 lists a number of commercial representatives.

Of all of the commercial system available, only one, the CellSearch system byVeredex, LLC, is approved by FDA in the United States as a diagnostic instrument todetect circulating tumor cells in patients. With respect to the actual magnetic separationof targeted cells, with a subsequent reinfusion into a patient, two systems have been orare current in clinical trails: CliniMACS cell separation system from Miltenyi Biotec andISOLEX 300i magnetic cell, originally manufactured by Baxter Healthcare Corporationand now also sold by Miltenyi Biotec.

No matter which system or magnetic particles employed, fundamentally, thetargeted cells must be imparted with a sufficient magnetic susceptibility such that it canbe easily separated from the nontargeted cells. Fortunately, except for a few specific

Magnetic Cell Separation to Enrich for Rare Cells

108

exceptions, such as deoxygenated red blood cells and specific spores of bacillus and sev-eral strains of bacteria, cells are intrinsic diamagnetic [14–16]. This is in contrast to fluo-rescent-based separations (i.e., FACS) where cellular autofluorescence effectively reducesthe sensitivity of a number of fluorescent labels.

7.1 Introduction

109

Table 7.1 Summary of a Number of Commercially Available, Magnetic Particle Conjugated Antibodies

Company Bead Size Antibody Application

Ademtech 300 nm Anti-IgG, Anti-IgM, Human CD4,CD8, CD14Bangs Laboratories, Inc. ~3 μm Anti-human leukocyte

Anti-mouse leukocyteAnti-IgG, Anti-IgM

BD Bioscience ~200 nm Anti-human cellAnti-mouse cellAnti-rat cellSecondary antibody

Bioclone Inc. 1–2 μm StreptavidinEMD 300 nm Anti IgG, streptavidinInvitrogen 1.0 μm (MyOneTM)

2.8 μm (M-280)4.5 μm (M-450)

Anti-human cellAnti-mouse B cell, dendritic cell, T cellSecondary antibody

Cortex Biochem 1–10 μm (MagaCell) Anti-IgGG. Kisker GbR 0.5–10 μm Anti-IgG,SreptavidinImmunicon Nanometers Anti-epithelial cellIndicia Biotechnology 0.3–1 μm Anti-IgG, anti-IgM, streptavidinMagSense Life Science 0.5,1 μm Goat anti-Mouse IgG, streptavidinMiltenyi Biotec <200 μm Anti-human cell

Anti-mouse cellAnti-rat cellAnti-nonhuman primate cellSecondary antibody

R & D systems ~150 μm Anti-human B cell, T cellAnti-mouse B cell, T cellAnti-rat T cell

(a) (b)Figure 7.1 (a–b) Representative photos of commercial batch, 1A (EasySep) or hybrid, flow throughmagnetic cell separation systems, 1B (MACS system).

7.1.1 Principle

7.1.1.1 Magnetic Force on an Immunomagnetically Labeled Cell

Fundamental to any magnetic cell separation approach is the magnetic forces that areapplied to the targeted cells as a result of the bound antibody-magnetic particle conju-gate. To obtain a high level of cell separation performance, optimization of the magneticforces acting on the targeted cells is usually needed. The magnetic force, on a per mag-netic particle basis, is mathematically expressed as:

FB

Sm part p p m, = ∇ =φμ

φ02

02(7.1)

where p is the field interaction parameter, B0 is the applied magnetic field induction,and Sm is the magnetic energy gradient [17]. The field interaction parameter is definedby:

( )φ χ χp p f p satV B B= − <0 (7.2)

or

φμ

χpsat

f p sat

MB

V B B= −⎛⎝⎜

⎞⎠⎟ ≥0

00 (7.3)

where p is the magnetic susceptibility of the magnetic particle, f is the magnetic suscep-tibility of the suspending buffer, Vp is the volume of the magnetic particle, Bsat is magni-tude of the magnetic field induction above which the magnetic particle is saturated, Msat

is the value of the saturated magnetization of the magnetic particle, and 0 is the mag-netic permeability of free space [17]. Here the functional dependence of particle magne-tization, M, on the applied field, B0, has been simplified to a linear function for low fields(7.2) and a constant for high fields (7.3).


110

Table 7.2 Commercially Available Magnetic Separators

System Type Company Product

Batch Veredex, LLC CellSearchAdemtech Adem-Mag MV, Adem-Mag HVBangs Laboratories, Inc. LS001, MS002, MS003, MS004BD Biosciences BD ImagnetInvitrogen DynaMag, Magna SepCortex Biochem Magnetic separator (2,10,20 position), magnetic block,

96 well plate magnetic separatorImmunocon MagNest deviceR & D systems Magcellect magnetStemcell Technologies EasySep magnetStemcell Technologies StemSep magnet

Partial flow-through Miltenyi Biotec MiniMACS, OctoMACS, MidiMACS, QuadroMACS,VarioMACS, SuperMACS II, antoMACS

Theoretically, the magnetic force acting on a magnetically labeled cell is propor-tional to the number of magnetic particles conjugated to the cell. The total force operat-ing on an immunomagnetically labeled cell can then be written as:

( )F n F Fm m part m cell= ⋅ ⋅ +1θλ β , , (7.4)

where n1 is the number of antigen molecules expressed on each cell, λ is the fraction ofantigen molecules bound by the antibody, and θ is the valence of primary antibodybinding, which depends on the specificity of antigen-antibody binding. The product ofthese three terms is also referred to as the antibody binding capacity, ABC, for a specificantigen expressed on the cell. β is the magnification factor and corresponds to the num-ber of magnetic nanoparticles conjugated per antibody, and Fm,cell is the intrinsic mag-netic force operating on a cell. Summarizing, the total magnetic force, Fm, operating on aimmunomagnetically labeled cell can be expressed as:

F ABC F Fm m part m cell= ⋅ ⋅ +β , , (7.5)

Similar to magnetic particles, the magnetic force acting on the cell without labeling canbe expressed as:

( ) ( )F VB

V Sm cell cell f cell cell f cell m, = − ∇ = −χ χμ

χ χ02

02(7.6)

where χcell and Vcell are the magnetic susceptibility and volume of the cell, respectively.Introducing (7.1) and (7.6) into (7.5),

( )F ABC S V Sm p m cell f cell m= ⋅ ⋅ + −β φ χ χ (7.7)

Therefore, the more magnetic particles that can bind to the targeted cell (ABC ⋅ β), thehigher the magnetic force operating on the cells. Similarly, increasing the magneticenergy gradient, Sm, also increases the magnetic force operating on the cell [18]. Further,using particles with a higher field interaction parameter, φp, a higher magnetic force isobtained [17].

7.1.1.2 Interaction Between Magnetic Particles and the Targeted Cell

The binding between magnetic nano- or microparticles and the targeted cell is typicallythrough an antibody-antigen interaction. In general, there are three ways to magneti-cally label a cell (Figure 7.2). The first method is a one-step labeling in which a magneticparticle (from 50 nm to over 1 micron in diameter) is conjugated to the Ab targeting thecell surface marker. The second method is a two-step labeling, employing a primary Abspecific to the cell surface antigen and a secondary Ab targeting the primary Ab to whicha magnetic particle is conjugated. The secondary Ab either targets the primary Ab, or amolecule bound to the primary antibody (FITC, PE, etc.). A common alternative to anti-body-antigen interactions for the primary-secondary interaction is streptavidin-biotinlabeling. The third method is a combination of a one-step and a two-step approach: atetrameric Ab is used that simultaneously targets a marker on the cell surface and a mag-

7.1 Introduction

111

netic particle, which is added as a second step after the cell has been labeled withthe Ab [19].

However, practically, when an antibody-conjugate binds to a cell, at least five scenar-ios can occur: (1) monovalent binding, (2) homogeneous bivalent binding, (3) multipleantibodies binding to a single antigen, (4) heterogeneous bivalent binding, and (5)cross-linked binding [20]. Figure 7.3 presents examples of each of these five cases. Thecomplexity of antibody-antigen interaction has made the accurate prediction of thecell-magnetic particle conjugation difficult. Yet effort has been made to describe thebinding affinity between antigen-antibody-conjugate in the first three scenarios ofFigure 7.3, which gave rise to the following equation [20]:

[ ]( ) [ ]( ) [ ]θ

λ

α α=

+ + ′ − + + ′ − ′

′

′

Ab K C Ab K C Ab C

CTotal D Total D Total1 1

24

2

CABC C

NA

= ⋅(7.8)

where θ corresponds to the fraction of the total surface antigen sites bound with anti-body, [Ab]Total represents the total concentration of antibodies, KD1 is the dissociation con-stant for scenario A of Figure 7.3, C is the concentration of cells, and NA is Acogadro’snumber. λ in (7.8) is the valence of the antibody binding, the value of which is intro-duced for scenario B:

[ ] [ ][ ] [ ]

λ =⋅ + ⋅ ⋅

⋅ + ⋅ ⋅Ag Ab Ag Ab Ag

Ag Ab Ag Ab Ag2(7.9)

and for scenario C:


112

Figure 7.2 (a–c) Example of a one step, two step, and a modified two step labeling process.

[ ] [ ][ ] [ ]

λ =⋅ + ⋅ ⋅⋅ + ⋅ ⋅

Ag Ab Ag Ab Ag

Ag Ab Ag Ab Ag

2(7.10)

If all the antibodies bind to the cell in a monovalent nature as in Scenario A, λ would beequal to 1. If all the antibody binding is of the homogeneous, bivalent nature (scenarioB), λ is 0.5. If all of the antigen binding is bivalent as in scenario C, λ is 2.

Equation (7.8) indicates that the saturation of the antibody binding sites, θ, are afunction of four primary variables: the concentrations antibody in suspension, [Ab]total,the equilibrium dissociation constant of the antibody with the antigen, αKD1, the ABC ofthe cells, and the total concentration of the cells. Zhang et al. [20] further demonstratedthat the conjugation of molecules and magnetic particles to antibodies can significantly

7.1 Introduction

113

Figure 7.3 Scenarios A–E.

change the value of αKD1 such that the amount of antibody needed to achieve an equiva-lent saturation of the antigen binding sites can increase by over an order of magnitudewhen a 200-nm magnetic particle is conjugated to an antibody. Such negative effectscan have significant financial implications when a separation process is scaled up to aclinical scale.

7.1.1.3 Quantification of Magnetic Cell Separation Performance

As with any separation technology, a complete analysis of the system performancerequires an accurate characterization of the heterogeneous mixture before and after theseparation. In addition, depending on the desired outcome the performance evaluationcan be presented in a number of ways. In addition to cell mixture quantification, totalviable cell number before and after separation is a further key factor. While such quanti-fication factors at first thought can appear to be straightforward, the sheer number ofcells, complexity in accurately quantifying cell subtypes (i.e., differentiating betweenT-cell subtypes based on low expressing cell surface markers), and even measuringcell viability can be very challenging when higher performance cell separations aredesirable [21].

Traditional determination of cell number usually is based on the cell size and shape,which can be realized either by manually counting on a hemacytometer or automati-cally counting using an electronic cell counting system, such as a Coulter coun-ter. Recently software has been developed that allows electronic counting using ahemacytometer, a digital camera, and computer, such as Cellometer Cell Counter fromNexcelom Bioscience. A further choice is a flow cytometer, with the help of TruCountbeads from BD Bioscience. While the automatic methods at first thought appear to bemore accurate, it has been our experience that manual counting with a hemacytometeris still the most accurate.

The cell viability can be determined with the traditional trypan blue exclusion tech-nique, or more recent flow cytometry methods that include 7-AAD and propidiumiodide (PI) stains. However, these techniques are still a “snapshot” of the current state ofthe cells, and as has been reported, inability or delay growth of cells after a separationcan still occur despite the cells appearing to be healthy [22].

Once the viable cell number as well as the specific cell fraction is determined, theanalysis of separation performance is typically based on the various definitions of purityand recovery. In some cases, the depletion of an unwanted cell types is also presented,and if a high level of depletion is desired, it is presented using a logarithm scale, typicallyreferred to as log depletion. The formula of several commonly used measures ofperformance is:

purity PN

N Ntt

t nt

= =+

(7.11)

recovery = =⋅⋅

N

N

N P

Nt final

t inital

final t final

inital

,

,

,

Pt inital,

⎛

⎝⎜⎜

⎞

⎠⎟⎟ (7.12)

log log ,

,10 10depletion

N P

N Pinital nt inital

final nt f

=⋅⋅ inal

⎛

⎝⎜⎜

⎞

⎠⎟⎟ (7.13)


114

where N represents total cell number, the subscript t and nt refer to the targeted cell typeand the nontargeted cell type, respectively, and Final and Initial stand for after andbefore separation, respectively.

7.1.2 Examples of Cell Magnetic Separation Applications

7.1.2.1 Isolation of Human Stem Cells

The term stem cell has been used to refer to cells that originate from a variety ofsources including bone marrow, peripheral blood, cord blood, embryos, and variousadult organs. While these cells are, obviously, not all the same, with respect tocell separation, a number of characteristics are common, including distinct surfacemarker(s) which define the cells that are otherwise indistinguishable from the othercells in the cell source; they are rare; and their value is such that maximizing theirrecovery is crucial.

With respect to the stem cell for hemopoetic purposes, which is one of themost commonly reported stem cell separations, a majority of the reports involve apositive selection for the stem cell, typically using an antibody targeting the CD34-cellsurface marker [23, 24]. Although some recent studies also used antiCD133 as theantibody [25] the final results analysis would still employ CD34 as stem cell indicatorin flow cytometry. In addition, ISHAGE gating strategy described by Sutherland et al.is used by most of the researchers in the field to define the presence of the stem cell [26].

For somatic stem cells, which are currently less well defined, fewer separation reportsexist. While a few markers have been reported for the cells derived from certain organs[27], for others no specific surface antigen can be used to define the cells directly. Thepurification of these stem cells correspondingly employ negative separation in which allthe other cells in the mixture are labeled with magnetic beads, thereby depleting theunwanted cells.

7.1.2.2 T Cell Depletion

The main application of T cell depletion (TCD) is in bone marrow or hematopoietic stemcell transplantation, to reduce post-transplant graft-versus-host disease (GVHD) thatcould lead to patient death. Early studies of TCD between the 1980s and mid-1990s didnot use magnetic separation and the depletion results were subsatisfying. To date it isgenerally believed that >3.5 log10 depletion is needed for an effective treatment [28],which could be achieved either by immunomagnetic positive CD34 selection as men-tioned above or direct T cell depletion using markers such as CD3 [29–32].

7.1.2.3 Rare Cancer Cell Detection

Over the last decade, significant interest has developed in the potential for the detectionand quantification of circulating tumor cells (CTCs) in the peripheral blood of cancerpatients. In fact, one system, the CellSearch by Verdex LLC is FDA-approved as a prog-nostic test for breast cancer based on the number of CTCs detected. The concentration ofcancer cells in the peripheral blood of cancer patients has been reported to range fromless than one per ml of blood to over 1,000 [33]. Considering a typical milliliter of blood

7.1 Introduction

115

contains on the order of 5 × 109 cells (both red blood cells and nucleated cells), easily andreliably finding one cell in 109 is a significant undertaking.

Usually rare cancer cell detection includes several steps. First, most of the red bloodcells and platelets will be removed either by centrifugation or lysis. Then magnetic parti-cle conjugated antibodies are added to the cell mixture targeting white blood cells. Atthis stage, either cancer cells are targeted (a positive selection approach) or normal bloodcells are targeted (a negative depletion approach) [33, 34]. The most commonly reportedapproach is to target epithelial markers on the surface of the circulating tumor cells;however, it has been recently suggested that not all tumor cells express epithelial mark-ers, thereby biasing the positive selection approach [34].

7.1.2.4 Bacteria

Magnetotactic bacteria are a group of bacteria capable of synthesizing magnetosomes[35] or contain magnetically susceptible elements such as iron or manganese, showingobservable intrinsic magnetic susceptibility in a strong magnetic field [36]. Given therelatively small number of bacteria that are intrinsically magnetic, few studies exist onthe applications of these organisms. However, the level of magnetism of these few mag-netic bacteria is such that significant applications are possible.

A recent technology that allows for enrichment of circulating tumor cells in cancerpatients by removing normal cells [19] will presented next.

7.2 Materials and Methods

Blood samples were obtained from patients who presented with squamous cell carci-noma of the head and neck (SCCHN) that were undergoing surgery. Operators wereblinded to clinical correlative information during the cell suspension processing andanalysis. From 10- to 18.5-ml peripheral blood was taken from each SCCHN patient,who was undergoing surgical resection for squamous cell carcinoma of the oral cavity,oropharynx, hypopharynx, or larynx and that had not been previously treated for thisdisease. Fresh blood samples from cancer patients or healthy donors were collected ingreen-top BD Vacutainer blood collection tubes containing sodium heparin (Cat#367874, BD). System performance was measured by spiking known amounts of cancercells from two cancer cells, Detroit-562 (a SCCHN line) and F-01 (a melanoma cell line)into buffy coats purchased from the American Red Cross.

7.2.1 Enrichment Process

Figure 7.4 presents a diagram of the overall process. Blood samples were either processedimmediately, or stored at 4°C overnight and processed early the next day. If more thanone Vacutainer tube was used for collection, the blood samples were pooled and sub-jected to a red cell lysis step.


116

7.2.2 Red Cell Lysis Step

Red blood cells were removed by applying a lysis buffer (154 mM NH4Cl, 10 mM KHCO3,0.1 mM EDTA), at a ratio of 25-ml lysis buffer to 1 ml of blood and incubating at roomtemperature for 5 minutes. After 5-minute centrifugation at 300 xg, the cell pellet waswashed and then resuspended in labeling buffer (PBS supplemented with 2-mM EDTAand 0.5% bovine serum albumin).

7.2.3 Immunomagnetic Labeling

For immunomagnetic labeling, a tetrameric antibody complex (TAC) structure was usedconsisting of bifunctional antibodies [37] purchased from StemCell Technologies (Van-couver, BC). The TAC complex used targets both CD45 cell surface receptors and dex-tran-coated, nano- or micromagnetic particles. Figure 7.2(c) presents this bifunctionalcharacteristic. The cell sample was prepared by suspending in a labeling buffer (PBS sup-plemented with 2 mM EDTA and 0.5 % bovine serum albumin) to which 0.5 ul of theTAC complex was added per million cells and the cell suspension was incubated for 30minute at room temperature in a shaker. Without washing the cell suspension, 1 μl ofthe magnetic nano- or micromagnetic particle suspension per million cells was thenadded and incubated for15 minutes at room temperature. The cells were then washedwith labeling buffer, centrifuged, resuspended in labeling buffer, and incubated for15 minutes.

7.2.4 Magnetic Cell Separation Step

To conduct a purely negative depletion of unwanted cells, a flow-through of theunlabeled cells, with a retention of magnetically labeled cells on the wall of the channelwithin the magnet was used and is presented in Figure 7.5. As Figure 7.4 indicates, oncethe unlabeled cells flowed through the system, they were collected and subjected toeither immunocytochemistry or RT-PCR.

7.3 Data Acquisition, Results, and Interpretation

Since the number of circulating tumor cells in a cancer patient are not known, spikingstudies were conducted in which a known number of cancer cells, from a cancer cell line,was spiked into blood samples from the American Red Cross (buffy coats). Table 7.3 pres-ents the results of these spiking studies using the protocol listed above, except when


117

Figure 7.4 Overview of enrichment process for circulating tumor cells (from Ying et al. 2009).

buffy coats were used, all of the RBCs had been previously removed. As can be observed,an average log10 enrichment of 2.9 was obtained for the spiked cancer cells, with an aver-age recovery of 83.5%. Cell concentrations were determined visually and volumes were


118

Figure 7.5 Flow through system enrichment for circulating tumor cells.

Table 7.3 Enrichment Performance of Human, Buffy-Coat Spiked with Cancer Cells

Run 1 2 3 Avg 4 5 6 Ave

Total number ofPBL used

8.0 × 107 8.0 × 107 8.0 × 107 8.0 × 107 8.0 × 107 8.0 × 107 8.0 × 107 8.0 × 107

Total number ofcancer1 cells added

800 800 800 800 800 800 800 800

Initial cancer cellconcentration(cancer/total cells)

1 × 10–5 1 × 10–5 1 × 10–5 1 × 10–5 1 × 10–5 1 × 10–5 1 × 10–5 1 × 10–5

Number of PBLrecovered afterenrichment2

1.4 × 105 1.5 × 105 1.6 × 105 1.5 × 105 3.7 × 104 7.4 × 104 1.1 × 105 7.3 × 104

Number of cancercells recovered

690 670 630 663 690 680 650 670

Final purity(cancer cells/totalcells)

4.9 × 10–3 4.6 × 10–3 4.0 × 10–3 4.5 × 10–3 1.9 × 10–2 9.2 × 10–3 5.9 × 10–3 1.1 × 10–2

Percent recovery ofcancer cells

86% 84% 79% 83% 86% 85% 81% 84%

Log10 enrichment 2.8 2.7 2.7 2.7 3.3 3.0 2.9 3.11 Runs 1, 2, and 3 used SCC-4 squamous cell carcinoma cell line while runs 4, 5, and 6 use F-01, which is a melanoma cell line.2 Runs 1, 2, and 3 used Stem Cell Technologies nanoparticles, while runs 4, 5 and 6 used Stem Cell Technologies microparticles.

measured analytically. Purity, recovery, and log10 depletion determined as indicated in(7.11), (7.12), and (7.13).

In addition to the spiking studies presented in Table 7.3, 25 enrichments of periph-eral blood samples from patients with squamous cell carcinoma of the head and neckusing the protocol outlined in Figure 7.3, were conducted. The average log10 enrichmentof nucleated cells was 2.64 and the overall enrichment, including the lysis of red bloodcells, was 5.3. Figure 7.6 is an example set of photographs of a cytospin and subsequentimmunocyotchemistry staining of one of these enriched samples. Figure 7.6(a) is abrightfield photograph, Figure 7.6(b) is a photograph of a fluorescent image with filtersset for fluorescein isothiocyanate, FITC, Figure 7.6(c) is photograph of a fluorescentimage with filters set for 4’,6-diamidino-2-phenylindole, DAPI, and Figure 7.6(d) is acomputer created fusion of Figure 7.6(b) and (c). The dye FITC was conjugated to anantibody that targets cytokeratins and DAPI targets all cell nuclei. It is generally acceptedthat a circulating tumor cell is positive for cytokeratins and the cell must be intact andcontain a nuclei (DAPI positive) [38, 39].

As can be imaged, the final purity in these 25 samples varies widely, since in somecases, no CTCs were found, while in other cases over 1,000 CTCs per ml of original bloodsample was detected. For the samples with a high number of CTCs, up to 75, cells pres-ent on the cytospin are CTC.


119

(a) (b)

(c) (d)

Figure 7.6 Photographs of microscopic images of a cytospin of one of the enriched peripheral bloodsamples from a cancer patient. (a) is a brightfield image; (b) is an image filtered for cytokeratin stain-ing (yellow-green); (c) is an image filtered for nuclei (DAPI) staining; (d) is an electronic superpositionof (b) and (c). Original magnification ×200.


From a basic, or fundamental perspective, magnetic cell separation has significantadvantages compared to many other cell separation technologies. With the significant

drop (by a factor of at 10×) in prices of very high power magnets, and continuousflow-through designs with no working parts [12], the potential of very high (> 106

cells/s) magnetic cell separation throughput exist [40] with high levels of recovery [41].While complexities are introduced with the use of antibodies, including binding affini-ties, selectivity, specificities, and cost, antibody challenges are similar no matter whatseparation technologies is used (i.e., FACS, affinity columns, and panning).

However, as with most positive attributes of any technology, significant limitationsexists, probably most notably the single parameter separation, in contrast with fluores-cence-based technology in which greater than 10 simultaneous parameters can be evalu-ated and used to separate a cell. While it is possible to separate cells based on the numberof antibody-magnetic particles conjugated to the cell [42], only one cell surface markeris targeted at a time. An alternative approach is to sequentially perform magnetic cellseparations by removing the magnetic particles between separations.



Red blood cells fail to lysis Old or ineffective lysis buffer Make up fresh bufferFailure to achieve high log10 depletion Binding affinity of magnetic

reagents to antigen to lowUse tetrameric antibody complex

Yield of no magnetically targeted cellslow

Nonspecific binding of magneticreagents to nontargeted cells

Use magnetic reagents that havelarger magnetic particles and use aflow-through separation system thatcan wash the nonspecifically boundcells out of column

7.5 Summary Points to Obtain High-Performance,Magnetic Cell Separations

1. Use magnetic reagents, which have high specificity for targeted cells and lownonspecific binding;

2. Have sufficient flow rates through the system, if it is a flow-through magneticallyseparator, to reduce nonspecific losses in the system;

3. Minimize the number of handling and process steps to minimize nonspecific losses.

Acknowledgments

The authors wish to acknowledge the financial support of National Science Foundation(BES-0124897 to J.J.C.), the National Cancer Institute (R01 CA62349 to M.Z., R01CA97391-01A1 to J.J.C.), and the State of Ohio Third Frontier Program (ODOD26140000: TECH 07-001).


120

References

[1] Berg, T., and Boman, D., “Distribution of Lysosomal Enzymes between Parenchymal and KupfferCells of Rat Liver,” Biochimica et Biophysica Acta, Enzymology, Vol. 321, No. 2, 1973, pp. 585–596.

[2] Melville, D., Paul F., and Roath S. “High gradient magnetic separation of red cells from wholeblood,” Transactions on Magnetics, Vol. 6 1975, pp. 1701–1704.

[3] Plyavin, Y., and Blum E. “Magnetic parameters of blood cells and high gradient paramagnetic anddiamagnetic phoresis,” Magnetohydrodynamics, Vol. 19 1983, pp. 349–359.

[4] Pauling, L. and Coryell C. D., “The magnetic properties and structure of hemoglobin,oxyhemoglobin, and carbonmonoxyhemoglobin,” Proc. Natl. Acad. Sci. USA, Vol. 22 1936,pp. 210–216.

[5] Giaever I., Magnetic Separation of Biological Particles, U.S. Patent 3,970,518, filed July 25, 1975, andissued July 20, 1976.

[6] Jovin M.T., Arndt-Jovin D.J., “Cell separation,” Trends in Biochemical Sciences, Vol. 5 1980,pp. 214–219.

[7] Safarik, I.; Safarikova, M., J Chromatogr B Biomed Sci Appl, 1999, 722, 33–53.[8] Antonio, J.C., Ternvnck T., Rodrigot M., and Avrameas, S. “Lymphoid cell fractionation on

magnetic polyacryamided-agarose beads,” Immunochemistry, Vol. 15, No. 7 1978, pp. 443–452.[9] Margel, S., and Rembaum A., “Polyglutaraldehyde microspheres: new reagent for cell labeling and

cell separation,” Polymer Preprints, Vol. 20, No. 1 1979, pp. 589–593.[10] Margel, S., Beitler, U., and Ofarim M., “A novel synthesis of polyacrolein microspheres and their

application for cell labeling and cell separation,” Immunological Communications, Vol. 10, No. 71981, pp. 567–575.

[11] Zborowski, M., Fuh, C.B., Green, R., Sun, L., and Chalmers, J.J., “Analytical Magnetopheresis ofFerritin-labeled Lymphocytes,” Analytical Chemistry, Vol. 67. No. 20 1995, pp. 3702–3712.

[12] Sun, L., Zborowski, M., Moore, L., and Chalmers, J.J., “Continuous, Flow-ThroughImmunomagnetic Cell Separation in a Quadrupole Field,” Cytometry. Vol. 33 1998, pp. 469–475.

[13] Tong, X., Xiong Y., Zborowski, M., Farag S.S., and Chalmers J.J., “A novel high throughputimmunomagnetic cell sorting system for potential clinical scale depletion of T cells for allogeneicstem cell transplantation,” Experimental Hematology, Vol. 35 2007, pp. 1613–1622.

[14] Zborowski, M., Ostera, G.R., Moore, L.R., Milliron, S., Chalmers, J.J., and Schechter, A.N. “RedBlood Cell Magnetophoresis, “Biophysics Journal, Vol. 84 2003, pp. 2638–2645.

[15] Melnik, K., Sun, J., Fleischman, A., Roy, S., Zborowski, M., Chalmers, J.J. “Quantification of mag-netic susceptibility in several strains of bacillus spores: implications for separation and detection.”Biotechnology and Bioengineering Vol. 98 2007, pp. 186–192.

[16] Schuller, D., “Genetics and cell biology of magnetosome formation in magnetotactic bacteria”FEBS Microbiol Rev. Vol. 32 No. 4 2008, pp. 654–672.

[17] Zhang, H., Moore, L.R., Zborowski, M., and Chalmers J.J., “Establishment and implications of acharacterization method for magnetic nanoparticle using cell tracking velocimetry and magneticsusceptibility modified solutions,” The Analyst, Vol. 130, 2005, pp. 514–527.

[18] McCloskey, K., Chalmers, J.J., and Zborowski, M., “Magnetic Cell Separation: Characterization ofMagnetophoretic Mobility,” Analytical Chemistry, Vol. 75, No. 4, 2003, pp. 6868–6874.

[19] Yang, L., Lang, J.C., Balasubramanian, P., Jantan, K.R., Schuller, D., Agrawal, A., Zborowski, M.,and Chalmers, J.J., “Optimization of an Enrichment process for Circulating tumor cells from theblood of Head and Neck Cancer patients through depletion of normal cells,” Biotechnol. Bioeng,Vol. 102, No. 2 2009, pp. 521–534.

[20] Zhang, H., Williams, P.S., Zborowski, M., and Chalmers J.J. “Binding affinities/avidities of anti-body-antigen interactions: quantification and scale-up implications,” Biotechnology and Bioengi-neering, Vol. 9, No. 2 2006, pp. 812–829.

[21] Tong et al., 2007.[22] Mollet, M., Godoy-Silva, R., Berdugo, C., and Chalmers, J.J. “Computer Simulations of the Energy

Dissipation Rate in a Fluorescence Activated Cell Sorter: Implications to Cells,” Biotechnol. Bioeng.,Vol. 100, 2008, pp.260–272.

[23] Hoppe, B., Mohr, M., Roots,-Weiss, A., Kienast, J., and Berdel, W.E. “Improvement of tumor celldepletion by combining immunomagnetic positive selection of CD34-positive hematopoieticstem cells and negative selection (purging) of tumor cells,” Bone Marrow Transplant., Vol. 23, 1999,pp. 809–807.

[24] Martín-Henao, G.A., and Picón M., “CD34+ cell selection: Combined positive and negative cellselection from allogeneic peripheral blood progenitor cells (PBPC) by use of immunomagneticmethods,” Bone Marrow Transplantation, Vol. 27 2001, pp. 683–687.

References

121

[25] Bitan, M., and Shapira M.Y. “Successful transplantation of haploidentically mismatched peripheralblood stem cells using CD133+ purified stem cells,” Experimental Hematology, Vol. 33, 2005,pp. 713–718.

[26] Sutherland, D.R., Anderson, L., Keeney, M., Nayar, R., and Chin-Yee, I., “The ISHAGE guidelinesfor CD34+ cell determination by flow cytometry,” Journal of Hematotherapy, Vol. 5, 1996,pp. 213–226.

[27] Romagnani, P., Lasagni, L., Mazzinghi, B., Lazzeri, E., and Romagnani, S., “Pharmacological modu-lation of stem cell function,” Curr Med Chem, Vol. 14, No 10 2007, pp. 1129–1139.

[28] Koh, L.P., Rizzieri, D.A., and Chao, N.J. “Allogeneic hematopoietic stem cell transplant using mis-matched/haploidentical donors,” Biology of Blood and Marrow Transplantation, Vol. 13, No. 11 2007,pp. 1249–1267.

[29] Barfield, R.C., Otto, M., and Houston, J., “A one-step large-scale method for T- and B-cell depletionof mobilized PBSC for Allogeneic transplantation,” Cytotherapy, Vol. 6, 2004, pp. 1–6.

[30] Gordon, P. R., Leimig, T., Mueller, I., “A large-scale method for T cell depletion: towards graft engi-neering of mobilized peripheral blood stem cells,” Bone Marrow Transplantation. Vol. 30, 2002,pp. 69–74.

[31] Schumm, M., Handgretinger, R., Pfeiffer, M., “Determination of residual T- and B-cell content afterimmunomagentic depletion: proposal for flow cytometric analysis and results from 103 separa-tions,” Cytotherapy, Vol. 8, 2006, pp. 465–472.

[32] Tong et al., ibid.[33] Cristofanilli, M., Budd, G.T., Ellis, M.J. “Circulating Tumor Cells, Disease Progression, and Survival

in Metastatic Breast Cancer,” N EnGL. J Med, Vol. 351, No. 8 2004, pp. 781–791.[34] Yang et al., 2009.[35] Hergt, R., Hiergeist, R., Zeisberger, M., Schuler, D., Heyen, U. Hilger, I., and Kaiser, W.A., “Magnetic

properties of bacterial magnetosomes as potential diagnostic and therapeutic tools,” Journal ofMagnetism and Magnetic Materials Vol. 293, 2005, pp. 80–86.

[36] Melnik et al., 2007.[37] Lansdorp, P.M., Aalberse, R.C., Bos, R., Schutter, W.G., Van Bruggen, E.F., “Cyclic tetramolecular

complexes of monoclonal antibodies: a new type of cross-linking reagent.” Eur. J. Immunol,. Vol.16, No. 6, 1986, pp. 679–683.

[38] Partridge, M., Brkenhoff, R., and Phillips E, “Detection of Rare Disseminated Tumor Cells IdentifiesHead and Neck Cancer Patients at Risk of Treatment Failure,” Clinical Cancer Research, Vol. 9, 2003.pp. 5287–5294.

[39] Riethdorf, S., Fritsche, H., and Muller, V., “Detection of Circulating Tumor Cells in PeripheralBlood of Patients with Metastatic Breast Cancer: A Validation Study of the CellSearch System,”Clin. Cancer Res,. Vol. 13, No. 3 2007, pp. 920–928.

[40] Williams, P.S., Zborowski, M., and Chalmers, J.J., “Flow Rate for the Quadrupole Magnetic CellSorter,” Analytical Chemistry, Vol. 71, 1999, pp. 3799–3807.

[41] Tong et al., 2007.[42] Moore, L.R., Zborowski, M., Sun, L. and Chalmers, J.J., “Lymphocyte Fractionation Using

Immunomagnetic Colloid and Dipole Magnet Flow Cell Sorter,” J. Biochemical and BiophysicalMethod, Vol. 37, 1998, pp. 11–33.


122

C H A P T E R

8Magnetic Nanoparticles for Drug Delivery

Susan P. Foy, Andrew Stine, Tapan K. Jain, and Vinod Labhasetwar*

Department of Biomedical Engineering, Lerner Research Institute, Cleveland Clinic, Cleveland, OH 44195

*Corresponding author: Vinod Labhasetwar, Ph.D., Department of Biomedical Engineering/ND-20, Cleveland Clinic, 9500Euclid Avenue, Cleveland, OH 44195, Phone: 216-445-9364, Fax: 216-444-9198, e-mail: [email protected]

123

Abstract

Magnetic nanoparticles (MNPs) are a multifunctional system capable of beingimaged, loaded with drug, and targeted to a particular region of interest thoughan externally applied magnetic field (MF). The use of an oleic acid (OA) coatingbetween the iron-oxide core and Pluronic in this method allows hydrophobicdrugs to be loaded into the MNPs alone or in combination for drug delivery.With a small size of around 200 nm (hydrodynamic diameter), MNPs may dif-fuse easily across the cell membrane, and their uptake and drug delivery can befurther increased by an external MF. All of these aspects help ensure optimaldosing, reducing toxicity to other organs, and increasing drug delivery to a tar-geted area.

Key terms cellular uptakedoxorubicindrug deliveryiron-oxidemagnetic nanoparticleswater-insoluble drugs

8.1 IntroductionThe magnetic properties of MNPs allow them to be imaged via magnetic resonanceimaging (MRI) and targeted to a particular region by an externally applied MF. TheOA-Pluronic MNPs detailed below have an iron-oxide core surrounded by OA, whichencapsulates a hydrophobic drug. The hydrophobic portion (PPO) of a Pluronic coatinganchors onto the OA, while the hydrophilic portion of the Pluronic (PEO) forms acorona, allowing aqueous dispersity of the MNPs.

Once loaded with drug, MNPs can be targeted to a region of interest through anexternally applied MF, and the drug released over a period of weeks [1]. This helps toachieve optimal dosing by reducing the systemic toxicity of the drug, and decreases thelikelihood of drug resistance that would result from insufficient drug present [2].Through MRI imaging, the biodistribution of MNPs and indirectly the drug concentra-tion may be determined.

Usually, MNPs use dextran or starch conjugated as an outer layer of the MNP toachieve aqueous dispersity in water [3]. Such methods use complex chemistry, and canonly conjugate a limited amount of drugs. The conjugation methods also lead to thedrug being released within a few hours [4]. The method described here takes advantageof the OA shell, with the potential to easily incorporate many hydrophobic drugs, aloneor in combination. Hydrophilic drugs may also be incorporated into the MNPs after con-version to their hydrophobic form, as detailed in the example below.

There are several different steps involved in the synthesis and characterization of thedrug-loaded MNPs (Figure 8.1). First, the MNPs are synthesized and coated with OA andPluronic. The size and charge are determined by a zeta potential/particle sizing system.Separately, the anticancer drug doxorubicin (DOX) used in the example is converted toits hydrophobic form, and loaded into the MNPs through stirring. If a drug is hydropho-bic and can be dispersed in acetone or ethanol, this conversion is not necessary. Oncethe drug is incorporated in the OA shell, the MNPs are collected by magnetic separationand the unentrapped drug washed away. The amount of drug loaded is determined byextracting the drug from the MNPs in a methanol-chloroform mixture, and quantifiedusing a suitable analytical method. The kinetics of drug release are determined using adouble diffusion cell, in which one side of the cell contains the drug-loaded MNPs andthe other a phosphate buffered saline (PBS)-Tween-80 (0.1%w/v) mixture acting as asink for the drug. The drug released from the MNPs can be collected at varying timeintervals over a period of weeks. A practical application of drug delivery by MNPs is dem-onstrated in an in vitro antiproliferative activity experiment using the breast cancercell-line (MCF-7) as an example. In addition, a magnet can be placed below the cells toattract the MNPs into the cells faster than they would be internalized by diffusion.

MNPs are used in drug delivery and targeting for a single drug in this method, butthey may also be used for delivery of multiple water-insoluble agents. Certain combina-tions of drugs, such as doxorubicin and paclitaxel, show synergistic anticancer activity.Thus the ability to load multiple drugs in the MNPs could improve therapeutic out-comes when using MNPs as a drug delivery system.

8.2 Experimental DesignOne of the unique properties of the Pluronic-OA MNP formulation is its ability toachieve sustained release of a drug over a period of several weeks. A custom-designed

Magnetic Nanoparticles for Drug Delivery

124

double diffusion cell is used to determine the kinetics of drug release (Figure 8.2(a)). Thedouble diffusion cell has two chambers, separating the drug-loaded MNPs from thereleased drug, which freely diffuses across a PVDF membrane. The released drug may beremoved and analyzed to determine the rate of drug release (Figure 8.2(b)).

An in vitro method is detailed to test the hypothesis that in the presence of a MF,more MNPs and thus drug will be taken up by a cancer cell-line than those entering bydiffusion alone. Several controls are necessary in this experiment, including DOX insolution, equivalent amounts of plain MNPs to DOX-MNPs, and the presence or absenceof a MF for both the plain and drug-loaded MNPs. In the presence of a MF, the treat-


125

5-8 DaysDOX-MNPs

treatment andMTS analysis

Uptake of DOXMNPs under MF

DOX-MNP uptake

5-8 DaysDOX-MNPs

treatment andMTS analysis

Uptake of DOXMNPs under MF

DOX-MNP uptakeUptake of DOX-MNPs under MF

DOX-MNP uptake

48-72 HoursMagnetic

NanoparticleSynthesis

Iron(II) + Iron(III)Ammonium Hydroxide

+ Oleic Acid

+ Pluronic

24 HoursDrug Incorporation

and DOX-MNPsseparation

24 HoursConversion ofDOX·HCl to

Hydrophobic DOX

Iron-Oxide Core

Oleic AcidPluronic

Magnet

MNP

PluronicDrug

PEO PPO PEO

CellsMCF-7 Cancer

+ DOXMNPs- MF

+ DOX-MNPs+ MF

Figure 8.1 Schematic for the synthesis and incorporation of drug in MNPs, and delivery of the MNPsto a breast cancer cell-line.

ments at varying concentrations are delivered to the cells in a 24 well plate, and eachtreatment re-seeded in 6 wells in a 96 well plate. Reseeding the cells after incubating withMNPs ensures that the effects of the drug are due to actual uptake of the drug-loadedMNPs and not due to drug being released from the MNPs and then entering the cells.

8.3 Materials

8.3.1 Reagents

• Ammonium hydroxide (5 M; Fisher Scientific).

• Cancer cell-line (MCF-7 breast cancer cell-line, American Type Culture CollectionATCC, Manassas, VA).

• Chloroform (HPLC Grade, Fisher Scientific).

• Doxorubicin hydrochloride (DOX·HCl, Dabur Research Foundation, Ghaziabad,India). Doxorubicin (DOX) is light-sensitive; keep protected from light and store at-20ºC when not in use.

• Fetal bovine serum (FBS, Invitrogen, Grand Island, NY).

• Hydrochloric acid (HCl, trace metal grade, Fisher Scientific).


126

PVDF membrane

Drug loaded MNPs

Receiver cell

PVDF membrane

Drug loaded MNPs

Receiver cell

0 5 10 15 20 25 300

20

40

60

80

100

Time (Day)

Dru

gre

leas

ed(%

)Donor cell

(b)

(a)

Figure 8.2 The double diffusion cell (a) allows drug released from the MNPs to flow freely across a0.1-μm PVDF membrane, where it can be collected and quantified to determine the drug release (b).(Figure 8.2(b) is reprinted in part with permission from [5]. Copyright 2005 American ChemicalSociety.)

• Iron(II) chloride tetrahydrate (FeCl2•4H2O, 99+%, Fisher Scientific).

• Iron(III) chloride hexahydrate (FeCl3•6H2O, 99% pure granulated, FisherScientific).

• Medium for cells (depending on cell-line).

• MEM supplemented with 10% v/v FBS, 100 mg/mL Penicillin-streptomycin, 1%v/v minimum essential amino acids, and 1% v/v sodium pyruvate for MCF-7cell-line.

• Methanol (HPLC Grade, 99.9%, Acros, New Jersey).

• MTS assay (CellTiter 96 AQueous, Promega, Madison, WI).

• Nitrogen-purged distilled (DI) water.

• Oleic acid (OA, Fisher Scientific).

• Phosphate-buffered saline (PBS, pH 7.4).

• Pluronic (F127, BASF Corporation, Mt. Olive, NJ).

• Triethylamine (> = 99.5%, Sigma-Aldrich).

• Trypsin.

• Tween-80 (Sigma-Aldrich).

8.3.2 Facilities and Equipment

• Centrifuge tubes (15 mL, 50 mL).

• Cuvettes (Brookhaven Instruments Corporation).

• Double diffusion cell.

• Environ-Shaker (Max Q 4000, Barnstead|Lab-Line).

• Fine-tip transfer pipette (Samco Scientific Corporation, San Fernando, CA).

• Fluorescence spectrophotometer (LS55, PerkinElmer, Waltham, MA).

• Inorganic membrane syringe-driven filter (0.02 μm, Anatop 25, Whatman Interna-tional Ltd, Maidstone, England).

• Lyophilizer (FreeZone 4.5, Labconco, Kansas City, MO).

• Magnetic block (4”× 6”; Dura Magnetics, Sylvania, OH).

• Magnetic stir bars.

• Magnetic stirring hot plate (PC-420D, Corning).

• Microcentrifuge tubes (1.5 mL, Fisher Scientific).

• Neodymium iron boron magnets (12,200 G, Edmund Scientific, Tonawanda, NY).

• Plates (24 and 96 wells, Becton Dickinson Labware, Franklin Lakes, NJ).

• Plate reader (BT 2000 Microkinetics Reader; BioTek Instruments, Inc., Winooski,VT).

• PVDF membrane (0.1 μm, VVLP, Durapore Millipore, Billerica, MA).

• Syringe (10 mL, HSW/Norm-Ject, Germany).

• Thermometer (Quartz digi-thermo, Fisher Scientific).

• Vials (20 mL, Sigma-Aldrich, 40 mL, Fisher Scientific).

• Water-bath sonicator (FS-30, Fisher Scientific).

• Zeta Potential/Particle Sizer (NICOMP 380 ZLS, Particle Sizing Systems, SantaBarbara, CA).

8.3 Materials

127

8.4 Methods

8.4.1 Synthesis of Magnetic Nanoparticles

Bubble 1L of DI water with nitrogen for 15 minutes the day of use. Use in all the stepsinvolved in the synthesis and formulation of MNPs. Nitrogen is used to prevent oxida-tion of the MNPs.

1. Prepare a 30-mL aqueous solution of 0.1 M Fe (III), and 15 mL of 0.1 M Fe (II) inwater. Combine in a 150-mL beaker, add a magnetic stir bar, and cover withparafilm. Stir the solution at ~400 rpm for 5 minutes on magnetic stirring hot platein a fume hood.

2. Add 3 mL of 5M ammonium hydroxide drop-wise over 1 minute to coprecipitatemagnetite particles. Continue stirring for 20 minutes.

3. Add 100 mg OA (~10 drops with a fine tip transfer pipet), and heat to 80ºC for 30minutes to evaporate the excess ammonia. Check the temperature of the solutionevery 5 minutes with a thermometer. Do not let the mixture boil.

4. Remove from heat and allow the solution to cool to room temperature. Separate theMNPs from excess OA by placing a magnet beneath the beaker until the MNPs settle.Pour off the supernatant while holding one magnet on the bottom and an additionalmagnet on the side just below the spout. Resuspend in 30-mL water. Repeat thiswash cycle two more times, adding 45 mL water after the final wash.

5. Add 100-mg Pluronic and stir overnight, with a parafilm cover to prevent oxidationof the MNPs.

6. Remove from stirring and remove the magnetic stir bar by attracting it to the side ofthe beaker with a magnet on the outside. Rinse the stir bar with solution to allow theexcess MNPs to fall back into the solution before removing the magnet completely.

7. Divide the MNPs into two 40-mL vials and tape two neodymium iron boron magnetswith opposite polarity on either side of each vial. Allow the MNPs to separate for 4hours at 4ºC or 7 hours at room temperature. Discard the supernatant and resuspendthe MNPs in 20 mL of sterile filtered nitrogen-purged water. Repeat the wash cycletwo more times. Recombine the MNPs after the final wash in a known volume ofwater (~10 mL), transfer to a 15-mL tube and sonicate for 5 minutes in a water-bathsonicator. (Sterile filtered nitrogen-purged water is used in the above step so that theMNPs are not contaminated by any other large particles, and to avoid contaminationin cell culture experiments.)

8. Centrifuge for 10 minutes at 1,000 rpm at 4ºC and carefully transfer the supernatantinto a new 15-mL tube without disturbing the pellet. The smaller MNPs will remainsuspended while larger MNPs will be left behind in the pellet and may be discarded.The MNPs can be stored for 3 months at 4ºC under a nitrogen gas atmosphere.

9. Determine the nanoparticle yield by suspending the MNPs by sonication for 10minutes, freezing a 1-mL aliquot at -70ºC in a tube of known mass, lyophilizing thesample for 2 days and weighing the dry particles.

10. Clean the beakers and stir bars by rinsing and sonication to remove loose MNPs,then swirling with a small amount of HCl in the fume hood to dissolve any MNPsthat remain. After dissolving the excess MNPs, add water in excess to dilute anyremaining HCl and discard in the sink.


128

8.4.2 Physical Characterization of Magnetic Nanoparticles

1. Sonicate the MNPs for 1 minute and suspend a sample at 2-μg/mL in water. (A 2–5mL suspension is required to carry out the size and zeta potential measurements ofthe sample.)

2. In a cuvette, sonicate the suspension for 1 minute in a water-bath sonicator.

3. Measure both the size and zeta potential of the sample.

8.4.3 Conversion of DOX HCl

8.4.3.1 Convert DOX•HCl into Water-Insoluble Doxorubicin

1. Weigh out 49 mg DOX•HCl in a small beaker, add 14 mL of 12.5% v/v methanol inchloroform, and sonicate briefly to disperse.

2. Add 60-μl triethylamine and stir for 2–3 hours. (The solution becomes clearer onaddition of triethylamine.)

3. Filter the solution into a 20-mL vial (of known mass) with a 10-mL syringe and0.02-μm filter, then filter an extra 1 mL of methanol-chloroform into the vial towash any DOX remaining in the filter disc.

4. Cover with aluminum foil with holes in the top and leave in a fume hood to beginevaporation. (Nitrogen gas may be flushed over the surface of the DOX solution inmethanol-chloroform to speed up the evaporation process if necessary. Keep the vialin a room temperature water bath if using this method to prevent the mixture fromgetting cold, which will slow the evaporation process.)

5. Lyophilize the sample to remove residual chloroform and determine its dry weight.

Store protected from light at –20ºC for up to 1 year.

8.4.3.2 Doxorubicin in Solution

1. Prepare a concentrated hydrophilic DOX solution by dissolving DOX•HCl into a66% v/v solution of ethanol in sterile water.

2. To prepare a 4-mg/mL solution, add 1.25 mL of 66% v/v ethanol in water to 5.0 mg

of DOX•HCl and vortex until dissolved.

Store protected from light at –20ºC for up to 1 year.

8.4.4 Drug Loading and Release Kinetics

8.4.4.1 Drug Loading

1. Suspend the hydrophobic DOX at 5 mg/mL in ethanol and sonicate briefly. Add 600μL of the DOX solution while stirring to 7 mL of MNPs (4.28 mg/mL) in a 20-mL vial.Continue stirring overnight. (The drug will become incorporated into the OA shellsurrounding the MNPs.)

2. Separate the MNPs from the unentrapped drug by placing magnets on either side ofthe vial, and pouring off the solution when they separate out. Wash the MNPs threetimes by resuspending the particles in water and separating them out from solution

8.4 Methods

129

with the magnets. Save the first wash to analyze how much drug was not entrappedin the MNPs. Resuspend in a known volume of water (~5 mL) after the final wash.

8.4.4.2 Determine Drug Loading

1. Take a 1-mL aliquot of the DOX-loaded MNPs (DOX-MNPs) in a tube of known mass,freeze at –70ºC, lyophilize, and determine the mass of the dried sample.

2. Add 2 mL of 12.5% v/v methanol in chloroform to the dried sample and leave it toshake for 24 hours at room temperature. (This combination of solvents will extractthe drug from the MNPs, with greater solubility than either solvent alone.Twenty-four hours is sufficient time for drug extraction.)

3. Divide the sample into microcentrifuge tubes, centrifuge for 10 minutes at 14,000gin an Eppendorf microcentrifuge, and collect the supernatant.

4. Make two dilutions of the supernatant, one twice as dilute as the first. Forexample, dilute a 100-μL aliquot of the supernatant to 5 mL in 12.5% v/vmethanol-chloroform mixture, and a 100-μL sample to 2.5 mL. (Two dilutions aremade to ensure that the DOX measured is in the linear portion of the calibrationcurve. If the samples are too concentrated, the fluorescence will be quenched and themore dilute sample will increase in fluorescence intensity.)

5. Prepare standards of DOX from 0–10 μg in 12.5% v/v methanol in chloroform.6. Determine the drug concentration using a fluorescence spectrophotometer at λex =

485 nm and λem = 591 nm. Calculate the amount of drug loaded in the MNPs bycomparing the measured value with the standard plot. (To check whether the sampleis in the linear portion of the calibration curve, the sample can be diluted and thefluorescence value should decrease proportionally. If the fluorescence intensitydoes not decrease proportionally, quenching is occurring and the samples need to befurther diluted.)

8.4.5 Kinetics of DOX Release from Magnetic Nanoparticles

1. Suspend the DOX-MNPs at 2 mg/mL in PBS buffer containing 0.1% w/v Tween-80.(Tween-80 is used to maintain sink condition so that the drug is released freely fromMNPs.)

2. In a double diffusion cell with a 0.1-μm porosity PVDF membrane, fill the donorchamber with 2.5-mL DOX-MNPs and the receiver chamber with 2.5-mLPBS-Tween-80. (The drug released from MNPs will diffuse freely across themembrane but the MNPs will not.)

3. Leave the suspension to shake on rotating shaker at 110 rpm at 37ºC.

4. Completely remove the buffer from the receiving chamber at different time intervalsand replace with fresh PBS-Tween-80 buffer.

5. Freeze and lyophilize the collected sample and dissolve in 12.5% v/v methanol inchloroform.

6. Prepare a standard plot (0–100-μg/mL DOX) under identical conditions bydissolving the drug in PBS-Tween-80, freezing at -70ºC, lyophilizing the sample andresuspending it in 12.5% v/v methanol in chloroform.

7. Measure the fluorescence intensity at λex = 485 nm and λem = 591 nm.


130

8.4.6 Antiproliferative Activity of Doxorubicin Loaded MagneticNanoparticles on MCF-7 Cells

1. Seed MCF-7 cells in a 96 well plate at 3,000 cells/well (100-μL/well) and allow thecells to attach for 24 hours. (Fill the perimeter wells in the 96 well plate with mediaonly—cells are not seeded in these wells because the media will evaporate.)

2. Suspend the desired concentrations of DOX-MNPs, control MNPs, and DOXsolution in supplemented MEM. Prepare the different concentrations with respect tothe DOX content.

3. Remove media from the 96 well plates and add 100 μL of treatment to each of 6 wells,leaving some wells with plain media as a control. This is considered day 0 of theexperiment for the MTS assay.

4. Replace the old media with fresh supplemented media on days 2, 4, and 5 withoutany additional treatment. For the MTS assay on day 5, add 20 μL of MTS reagent toeach well after the media change, and incubate for 90 minutes. Place 90 μL of mediafrom each well in a fresh 96 well plate. Measure the absorbance at 490 nm on amicroplate reader.

5. Determine the effect of drug on cell-proliferation by calculating the percentdifference in intensity of the treated cells compared to the untreated controls.

8.4.7 Antiproliferative Activity of Doxorubicin Loaded MagneticNanoparticles on MCF-7 Cells in the Presence of a Magnetic Field

1. Seed MCF-7 cells in a 24 well plate at 100,000 cells/well (1 mL/well).

2. When cells reach confluency (~2 days after seeding), suspend desired concentrationsof DOX-MNPs, control MNPs, and DOX solution in supplemented MEM.

3. Remove media from the 24 well plates and add 1 mL of treatment to each well,treating some wells with plain media as a control. Stack the 24 well plate with cells

on an empty 24 well plate on a 4”×6” magnet and return the plates to the incubatorfor 2 hours. (A 24 well plate is placed between the plate with the cells and the magnetto allow uniform attraction of the MNPs over the cells on the surface of the plate.)

4. Remove the magnet, wash the cells two times with PBS, add 50 μL of trypsin to eachwell and return the plate to the incubator until cells have detached (2–3 minutes).Add 1 mL of supplemented media to each well to neutralize the trypsin and transferthe contents of each well to separate 15-mL tubes. Centrifuge the cells at 1,000 rpmfor 10 minutes at 4ºC and resuspend them in supplemented MEM at 30,000 cells/mL.(The cell count can be determined from one control well and the same volume ofmedia added to all of the 15-mL tubes.)

5. In a 96 well plate, add 100 μL of the cells (30,000 cells/mL) to each of 6 wells and fillthe wells in the perimeter with 100-μL media. Prepare two identical plates if runningthe MTS assay on days 2 and 5. Add the magnet below the 96 well plate (with 24 wellplate in between) and return to incubator. This is considered day 0 for the MTS assay.(The perimeter wells will lose media due to evaporation, so cells are not seeded inthese wells.)

6. On the second day after seeding the cells, aspirate off the media and add 100 μL offresh media to each well. For the MTS assay, add 20 μL of MTS reagent to each welland incubate for 90 minutes. Place 90 μL of media from each well in a fresh 96 wellplate. Measure the absorbance at 490 nm on a microplate reader.

8.4 Methods

131

7. Determine the effect of drug on cell-proliferation by calculating the percentdifference in intensity of the treated cells compared to the untreated controls.

8. For a 5-day MTS assay, remove and discard the old media from each 96 well plate andadd 100 μL of fresh media on days 2 and 4. On day 5, change the media and run theMTS assay (as described above).


The approximate MNP yield in one batch is 90 mg. The size of the Pluronic F127 MNPs isabout 200 nm as determined by the particle sizing system (Figure 8.3).

After the MTS assay, the percent growth can be calculated for each treatment con-centration according to the following formula:

%Growthmeanmean

TreatedCells

ControlCells

= ×100

A curve can be fit to the data to determine the IC50, or the drug concentration neededto inhibit 50% of the cell growth, using the following equation:

( )y

A A

x xA

op= −

++1 2

21

Where y = % Growth at drug concentration x, A1 = maximum % Growth, A2 = mini-mum % Growth, xo = inflection point of the curve, and p = largest absolute value of theslope of the curve. The IC50 for MCF-7 cells varies with the drug used. As an example,drug in solution and in MNPs for two different anticancer drugs, DOX and paclitaxelwere tested in MCF-7 cells and their IC50 determined (Table 8.1).


132

Intensity-Wt gaussian distribution

Diameter (nm)

Rela

tive

inte

nsity

50 100 200 500 1000

20

0

40

60

80

100

Figure 8.3 Particle sizing system output with an average MNP size of about 200 nm.


The OA shell for drug loading allows multiple hydrophobic drugs to be loaded alone orin combination in the MNP formulation, targeted by MF, and released over severalweeks. In the synthesis of the MNPs, prolonged exposure to an oxygen environmentmay cause oxidation of the MNPs and decrease their overall magnetic properties. Severalsmall steps can be taken to decrease this risk. All water used in the synthesis, formula-tion, and washing of MNPs is purged with nitrogen gas to minimize the dissolved oxy-gen in the aqueous phase and prevent MNP oxidation. The rpm for the stir bars has beensuggested, but it is most important to allow the solution to mix at the highest rpm possi-ble without causing violent stirring, which would also introduce oxygen into the solu-tion. In addition, when the MNPs are stored for an extended period of time, flushingnitrogen gas over the solution before covering and storing it will decrease the risk of oxi-dation and loss of the magnetic properties. The OA coating on MNPs will further protectthe iron-oxide core from oxidation.

The Pluronic coating used to disperse the MNPs in aqueous solution comes in severalformulations with varying lengths in the hydrophilic (PEO) and hydrophobic (PPO)chains. The Pluronic used can alter several properties in the MNPs, including the particlesize, surface characteristics like hydrophilicity and zeta potential, and the percentage ofdrug loading. Pluronic F127 is used in the method described, but there are severalPluronics that show increased circulation time, including Pluronic L64, Pluronic F68,and Pluronic F108. After the Pluronic has been added to the MNP formulation, improperhandling of the MNPs may also lead to aggregation. For example, allowing the MNPs toremain in suspension for a long time before washing off excess Pluronic may increasetheir aggregation and size. Placing a magnet on the vial to attract the MNPs removes theMNPs from free Pluronic in solution and is one step that may prevent aggregation.Washing the MNPs several times removes this free Pluronic, though some may remainin solution regardless of the number of washes. This free Pluronic may also form micellesin solution, and if the critical micelle concentration (CMC) is reached, when the MNPsare loaded with drug, the free Pluronic may solubilize some of the drug and decrease theoverall drug loading in the MNPs.

After lyophilizing a sample of the MNPs, static charge may develop causing theapparent yield to be negative. This has been observed in particular while wearing latexgloves. Nitrile gloves produce less static charge, though handling the vial with thelyophilized sample with forceps may be the best solution in overcoming this problem.

Different magnets used for targeting of the MNPs in the in vitro experiments greatlyaffect the uniformity of MNP uptake in the cells. The most uniform MF is achieved using

a 4”×6” magnetic block from Dura Magnetics. The BioMag 96 well plate separator from


133

Table 8.1 IC50 of DOX and Paclitaxel in MCF-7 with Drug in Solution andLoaded in MNPs

Doxorubicin (ng/mL) Paclitaxel (ng/mL)

MNPs1 Solution MNPs2 SolutionIC50 795.5 102.9 10.6 9.8

1 Doxorubicin loading in MNPs: 8.2% w/w2 Paclitaxel loading in MNPs: 9.5% w/wIC50= Drug concentration required to kill 50% of cells

Polysciences, Inc. uses 24 small square magnets imbedded in a protective case, whichcauses great variability in MF to each well and thus MNP uptake by cells. The 1-cm2 neo-dymium iron boron magnets used in MNP separation can be placed under eight wellchamber slides with 1-cm2 wells. However, this method is much more tedious andrequires many chamber slides to achieve the same sample size of just one 24 well plate

on a 4”×6” Dura Magnetics magnet.



MNPs did not suspend. MNPs may have boiled during synthesis. During synthesis, do not allow the solu-tion to boil.

Smaller, hydrophobic particlesdon’t disperse with Pluronic.

Too much OA added during synthesis. Decrease the amount of OA added duringMNP synthesis.

MNPs won’t resuspend. Final suspension may have been frozen. Do not freeze or lyophilize MNPs aftersynthesis.

Uniform suspension lost. Excess sonication. Avoid hand held or high-poweredsonication. If loaded with drug, sonicateonly briefly in a water bath sonicator.

Negative MNP yield afterlyophilization.

Vials used for lyophilization may havedeveloped static charge.

Avoid holding the lyophilized sample withnitrile or latex gloves; consider handlingwith forceps.

The spectrophotometer readingfor DOX is higher when theamount of drug measured isdiluted.

The DOX sample measured may be tooconcentrated; the spectrophotometer maybe reading in the fluorescence quenchingportion, with a nonlinear or negativeslope.

Dilute the samples until the values mea-sured are in the linear region (i.e., thesample with twice as much drug willhave twice the fluorescence).


MNPs have wide application and can be used in drug loading and targeted drug delivery,as a contrast agent in MRI imaging, and to induce hyperthermia with an alternating MF.

8.8 Summary Points

• MNPs with high drug-loading capacity and sustained release properties aredeveloped [1].

• Any hydrophobic drug or substance should be able to incorporate into the OA por-tion of the MNPs formulated, alone or in combination. If possible, hydrophilicdrugs may be converted to a hydrophobic form and incorporated into the MNPs.

• Drug loading into an OA shell allows sustained drug delivery over a periodof weeks [1].

• The MNPs may be targeted by a MF, increasing cellular uptake and drug delivery ascompared to that achieved by diffusion.


134

Acknowledgments

The study reported here is funded by grant R01 EB005822 (to VL) from the NationalInstitute of Biomedical Imaging and Bioengineering of the National Institutes of Health.

References

[1] Jain, T. K., M. A. Morales, S. K. Sahoo, D. L. Leslie-Pelecky, and V. Labhasetwar, “Iron OxideNanoparticles for Sustained Delivery of Anticancer Agents,” Molecular Pharmaceutics, Vol. 2, No. 3,2005, pp. 194–205.

[2] Bezwoda, W. R., “High-Dose Chemotherapy with Hematopoietic Rescue in Breast Cancer: fromTheory to Practice,” Cancer Chemotherapy and Pharmacology, Vol. 40, 1997, pp. S79–S87.

[3] LaConte, L., N. Nitin, and G. Bao, “Magnetic Nanoparticle Probes,” Materials Today, Vol. 8, No. 5,2005, pp. 32–38.

[4] Alexiou, C., W. Arnold, R. J. Klein, and F. G Parak, et al., “Locoregional Cancer Treatment withMagnetic Drug Targeting,” Cancer Research, Vol. 60, No. 23, 2000, pp. 6641–6648.

[5] Jain, T. K., M. A. Morales, and S. K. Sahoo, et al., “Iron Oxide Nanoparticles for Sustained Deliveryof Anticancer Agents,” Molecular Pharmaceutics, Vol. 2, No. 3, 2005, pp. 194–205.

Acknowledgments

135

C H A P T E R

9Imaging and Therapy of AtheroscleroticLesions with Theranostic Nanoparticles

Jason R. McCarthy1, Farouc A. Jaffer2, and Ralph Weissleder1

1 Center for Systems Biology, Harvard Medical School and Massachusetts General Hospital, 149 13th St.,Rm 5406, Charlestown, MA 021292 Cardiovascular Research Center, Cardiology Division, Harvard Medical School and Massachusetts Gen-eral Hospital, 149 13th St., 4th Floor, Charlestown, MA 02129

Corresponding author: Jason R. McCarthy, Center for Systems Biology, Harvard Medical School andMassachusetts General Hospital, 149 13th St., Rm 5406, Charlestown, MA, 02129, Phone: 617-726-5788,Fax: 617-726-5708, e-mail: [email protected]

137

Abstract

Theranostic nanomaterials, or those bearing both therapeutic and diagnosticentities, are capable of simultaneously imaging and treating disease. In thismethod, we synthesize a novel atherosclerosis-targeted theranostic nanoagentbased upon crosslinked iron oxide nanoparticles (CLIO) bearing fluorophoresfor near infrared fluorescence imaging, and near infrared light activated thera-peutic (NILAT) agents for therapy. These macrophage-targeted nanoparticlesare applied to the detection and localized therapy of atherosclerotic lesions inapolipoprotein E deficient mice. Intravital fluorescence microscopy enablesthe longitudinal examination of nanoparticle uptake before and after therapy,thus allowing for an in vivo determination of therapeutic efficacy. Whiletheranostic nanoagents have unique strengths, including the concomitantassessment of the diagnosis and therapy of disease, the field is still in itsinfancy. This method provides for further study of these capabilities.

Key terms theranostic nanoagentintravital fluorescence microscopyiron oxidenanoparticlesmolecular imaginglight-activated therapy

9.1 Introduction

The combination of diagnostic and therapeutic entities onto one nanoscaffold enablesthe simultaneous diagnosis and treatment of disease. These integrated “theranostic”materials offer several potential advantages over conventional therapeutic agents,including feedback mechanisms for the determination of the localization, and therapeu-tic efficacy of treatments. The success of these agents in this burgeoning field is not fullyrealized at present, partly due to mismatches between the diagnostic and therapeuticportions, including dosing, which is often significantly higher for treatment. While thisfield is still in its infancy, it is clear that theranostics offer unique capabilities and theirapplications require further study.

Atherosclerosis is a leading cause of death worldwide, and new treatments areurgently needed to limit myocardial infarction, stroke, and death. An intriguingtreatment strategy is localized therapy of inflamed atherosclerotic lesions, as researchover the past decade demonstrates that inflammation and the innate immuneresponse participate critically in the initiation and progression of atherosclerosis[1–3]. In particular, macrophages contribute crucially to all stages of atherogenesis,from foam cell and fatty streak formation to the coordination of the inflammatoryresponse leading to plaque rupture and thrombosis in advanced atherosclerotic lesions.Histopathological studies of clinical atheromata further link macrophage content,apoptosis, and macrophage-derived proteinases to rupture-prone plaques [4–8].Macrophages thus represent an important cellular target for atherosclerosis therapies[1–3, 9–12].

In this method, we investigate the use of theranostic nanoagents in the localizationand treatment of atherosclerosis, via the focal ablation of inflammatory macrophages.This is enabled by the affinity of macrophages for polysaccharide-coated iron oxidenanoparticles. Dextran-coated monocrystalline iron oxide nanoparticles (MION) havebeen utilized clinically to better delineate primary tumors [13], image angiogenesis [14],and detect metasteses [15, 16]. Additionally, these particles have been used to image theinflammatory cells, predominantly macrophages, of human carotid atheroscleroticlesions [17–19].

One of the greatly enabling modifications made to MION has been the crosslinkingof the dextran and its amination [20]. The resulting particle, CLIO (cross-linked ironoxide) allows for facile functionalization via amide bond formation. It also offers superbstability under harsh conditions without a change in size, blood half-life, or loss of itsdextran coat. Due to the similarities between dextran coated MION and CLIO, it is notsurprising that it is also readily internalized by plaque inflammatory cells. In fact, 65% ofthe cells in experimental atherosclerotic lesions that contain CLIO are macrophages,with plaque smooth muscle and endothelial cells demonstrating modest uptake [21].Thus, CLIO appears to be a promising scaffold for the development of theranosticnanoagents targeted to inflammatory macrophages in atherosclerosis.

While any number of therapeutic moieties can be utilized to bring about a therapeu-tic effect, many of the options are intrinsically toxic. As these theranostic agents are tobe administered systemically, complications may arise, such as extraneous tissue dam-age. In order to circumvent this, agents that are activate only at the site of interest suchas prodrugs or photosensitizers, become attractive options. Near infrared light-activated

Imaging and Therapy of Atherosclerotic Lesions with Theranostic Nanoparticles

138

therapeutic (NILAT) agents generate cytotoxic singlet oxygen upon illumination withthe appropriate wavelength of light. Thus, the action of these agents is focal, beinglimited only to areas receiving laser irradiation.

The combination of highly phototoxic NILAT agents with plaque-targeted opticaland magnetic resonance imaging agents may yield theranostic nanoparticles capable oflocating and treating inflamed atherosclerotic lesions. In this method we outline thesteps necessary to synthesize a macrophage-targeted theranostic nanoparticle with theabove capabilities. We begin with the synthesis of CLIO via epichlorohydrin-inducedcrosslinking of dextran coated MION. The particles are then made optically activeby the conjugation of 5-(4-carboxyphenyl)-10,15,20-triphenyl-2,3-dihydroxychlorin(TPC) [22, 23], a potent NILAT agent, and Alexa Fluor 750 (AF750), a near infraredfluorophore, to the particle surface. A second control nanoparticle not bearing a NILATagent is also synthesized, and is utilized as a nontherapeutic control agent in allexperiments.

The resulting agents are then injected into atherosclerosis laden apolipoprotein Edeficient (apoE-/-) mice and the surgically exposed carotid atheromata are imaged byintravital fluorescence microscopy (IVFM). Importantly, the particles are given 24 hoursto localize within the lesions, as they are long circulating, and can accumulate over timevia the enhanced permeability and retention effect. Following the survival imaging ses-sion, the exposed lesions are irradiated with a 650-nm laser in order to bring about thetherapeutic effect of the NILAT agent. The surgical incisions are then sutured and themice are allowed to recover. At the designated time point, at either 1 or 3 weeks aftertherapy, the mice are re-injected with the respective active or control agents, which aregiven 24 hours to localize. Next the surgical incision is reopened, and the mice arereimaged. One of the main advantages of this procedure is that it allows for longitudinalstudies of nanoparticle uptake within the atheromata.

While this method primarily focuses on the application of theranostic nano-materials to the diagnosis and therapy of inflamed atherosclerotic lesions, it can easilybe applied to any number of diseases, such as cancers and autoimmune diseases.Nanoparticulate scaffolds can be targeted to many different cell or tissue types via conju-gation of the appropriate targeting ligands to the particle surface. As well, the therapeu-tic functionality of the particles can also be chosen in order to elicit the requiredtherapeutic effect. Most importantly, the ablation of specific cell types within the targettissues and resulting therapeutic efficacy can be readily monitored by IVFM.


In this method we examine the utility of theranostic nanoagents in the focal ablationof macrophages within atherosclerotic lesions. Therapeutic efficacy is determined bylongitudinal examination of IVFM data. While the initial uptake and fluorescence of thesynthesized theranostic nanoagent in the atherosclerotic lesions is expected to be high,reinjection and imaging of the agent localization at time points 1 and 3 weeks aftertherapy is expected to reveal minimal uptake of the agents due to the ablation ofthe inflammatory cells (predominantly macrophages) of treated lesions. The resultsobtained for the theranostic nanoparticle are compared with those obtained with a con-trol nanoparticle. This control particle does not contain the therapeutic portion, but is


139

otherwise identical to the theranostic agent. While not included in this method,another possible control is the injection of saline into the atherosclerotic mice instead ofthe nanoparticles.

As this method involves in vivo experimentation, it is important to include an ade-quate number of animals in each cohort to allow for inherent variability. We recom-mend a minimum of five animals per cohort. In addition there is inherent variabilitybetween nanoparticle preparations, thus it is important to synthesize an appropriateamount of each agent. The initial and followup injections must be from the same batchof particles, as the fluorescence of the agents is matched, and the data is obtained as acomparison of the pre- and post-treatment fluorescence intensity.

9.3 Materials

9.3.1 Reagents

• MION-47 (Center for Molecular Imaging Research, Massachusetts GeneralHospital);

• 5M NaOH (Fisher, cat. no. S256-500);

• Epichlorohydrin (Fluka, cat. no. 45340);

• 30% (wt/vol) ammonium hydroxide (Aldrich, cat. no. 221228);

• Citrate buffer: 0.025M sodium citrate pH 8 (Fisher, cat. no. S279-500);

• 6N Hydrochloric acid (Fisher, cat. no. SA56-1);

• Hydrogen peroxide 3% (wt/vol) (Aldrich, cat. no. 323381);

• Iron atomic spectroscopy standard concentrate, 1.00g Fe (Fluka, cat. no. 02679);

• Phosphate buffered saline without calcium and magnesium, 10x solution (Fisher,cat. no. BP399-500);

• Phosphate buffered saline without calcium and magnesium, 1x solution (Fisher,cat. no. BP2438-4);

• Alexa Fluor 750 (Invitrogen, cat. no. A-20011);

• Dimethylsulfoxide (Fisher, cat. no. D128-500);

• Succinimidyl ester of 5-(4-carboxyphenyl)-10,15,20-triphenyl-2,3-dihydroxychlorin (TPC) [23];

• Fluorescein (Acros, cat. no. AC11924-0250 );

• Fluorescein isothiocyanate-dextran (Sigma, cat. no. FD2000S).

9.3.2 Facilities/Equipment

• QuixStand Benchtop System (A/G Technology) with cartridge UFP-100-E-5A (100kDa NMWC);

• Amicon Ultra 15 (Fisher, cat. No. UFC9 100 08, 100 kDa NMWC);

• Cary 50 UV-visible spectrophotometer (Varian);

• Cary Eclipse fluorescence spectrophotometer (Varian);

• Sephadex G-25 (Aldrich, cat no. G25150);


140

• Prototypical laser scanning fluorescence microscope (Olympus Corporation,Japan);

• 650-nm diode laser, 250 mW (B&W Tech, BWF1).

9.3.3 Animal Model

All animal studies should be performed in accordance with relevant guidelines and regu-lations. Apolipoprotein E deficient (apoE–/–) mice are employed as a well-characterizedexperimental model of atherosclerosis. Female apoE–/– mice (Jackson Laboratory, BarHarbor, Maine) are placed on an atherogenic diet (21% fat, 0.15% cholesterol, HarlanTeklad, Madison, Wisconsin) from 10 weeks until 28 weeks of age. After the initialimaging session, animals are placed on a regular chow diet to limit the formation ofnew lesions adjacent to those undergoing treatment.

9.3.4 Alternate Reagents and Equipment

MION-47. This reagent is available to the scientific community through ourlaboratory (http://cmir.mgh.harvard.edu). A number of alternate, dextran-coatediron oxide nanoparticles can be considered but have not been comparativelytested by the authors.

AF750. While AF750 has been utilized by the authors for this method, a numberof other fluorophores that absorb and emit in the same regime can be considered,such as Cy 7 (GE Healthcare), VivoTag 750 (VisEn Medical), and IRDye 800(LI-COR). Fluorescent dyes, such as Cy 5.5 cannot be utilized, as they overlap withthe absorption spectra of the near-infrared light activated therapeutic moieties.

TPC. TPC is utilized in this method due to its availability to the authors, as it issynthesized by our laboratory, and its utility as a highly phototoxic agent. TheUV-visible absorption profile is also optimized for use in tandem with longerwavelength fluorescent dyes, such as AF750, for the development of theranosticnanoagents. Other photosensitizers, such as chlorin e6 can be used, although theirefficacy has not been tested by us.

Prototypical laser scanning fluorescence microscope. The prototypical laser scanningfluorescence microscope used in this method was developed by Olympus.Quantitative imaging can also be done on a number of other fluorescence basedsystems, including those used in fluorescence molecular tomography andmultiphoton microscopy.

9.4 Methods

9.4.1 Synthesis of Theranostic Nanoparticles

1. Aminated crosslinked iron oxide nanoparticles (CLIO-NH2) are synthesized bythe epichlorohydrin-mediated crosslinking of dextran coated monocrystalline ironoxide nanoparticles (MION). To MION-47 is added aqueous sodium hydroxide (5M)in a ratio of 5 volumes NaOH to 3 volumes MION-47 over the course of 15 minuteswhile stirring at room temperature. Two volumes of epichlorohydrin are slowly

9.4 Methods

141

added and stirred vigorously for an additional 8 hours, at which time 3 volumes of30% ammonium hydroxide are added and stirred for 10 hours.

2. Excess epichlorohydrin and ammonia are removed by diafiltration (QuixstandBenchtop System) using citrate buffer. The resulting nanomaterial is concentrated toapproximately 10 mg Fe/mL using centrifugal filtration (Amicon Ultra-15, 100 kDanominal molecular weight cutoff). CLIO-NH2 is stable and can be stored in Nalgenebottles at 4°C.

3. Determine CLIO-NH2 concentration spectrophotometrically. Mix 10 μl of CLIO-NH2

with 1-ml 6M hydrochloric acid and 10 μl of 3% hydrogen peroxide, let sit for 1 hourat room temperature, and measure optical density at 410 nm. During this time,prepare standard solutions containing 0.1–4.0 mg of Fe per ml of iron atomicspectroscopy standard concentrate in 6M HCl.

4. CLIO-NH2 is fluorescently labeled with amine reactive fluorophores. Succinimidylesters of fluorescent dyes are preferred, although isothiocyanates demonstrate equalutility. In this instance, we are using Alexa Fluor 750 (AF750), which absorbs at 752nm and emits at 779 nm. This dye is visualized in the Cy 7 channel of all opticalimaging systems. To 20 mg CLIO-NH2 (2 mL in citrate buffer) is added 10x PBS (222μL), followed by 1 mg of AF750 dissolved in 200 μL DMSO. The resulting solution isshaken for 4 hours at room temperature, and then purified by size exclusionchromatography (Sephadex G-25) according to the manufacturer’s instructionsusing PBS as the eluent to yield the magnetofluorescent nanoparticle (MFNP). Thismaterial is stable at 4°C. A portion (10 mg) of the AF750-labeled particles is set asidefor use as the control (CLIO-AF750, nontherapeutic) agent.

5. The dye labeled particles are then labeled with near-infrared light activatedtherapeutic (NILAT) moieties. The NILAT agent used in this method, thesuccinimidyl ester of 5-(4-carboxyphenyl)-10,15,20-triphenyl-2,3-dihydroxychlorin(TPC), is reacted with the VT-680-labeled nanoparticle in a ratio of 1-mg NILAT to10-mg Fe. The NILAT agent is dissolved in enough DMSO prior to addition to thenanoparticle suspension that it is 20% of the solution by volume. The reaction isallowed to proceed for 4 hours while shaking, at which time it is purified by sizeexclusion chromatography (Sephadex G-25) according to the manufacturer’sinstructions using PBS as the eluent to yield the theranostic nanoparticle (TNP). Thismaterial is also stable at 4°C.

6. The concentration of the dye-labeled particles and chromophores are determinedspectrophotometrically (Figure 9.1). The UV-visible absorption of a standardsolution of CLIO-NH2 (10 μL of CLIO-NH2 in 3 mL PBS, calculated in step 3) isdetermined at 300 nm. Similarly, 10 µL of the MFNP is diluted to 3 mL with PBS andits absorption is also determined at 300 nm. The concentration of the particle insuspension is then calculated comparatively. The concentration of thechromophores is determined from the absorption of the dyes at their maxima in thesame diluted sample and its extinction coefficient using Beer’s law. Allmeasurements are performed in triplicate. Ideally, particle concentrations will begreater than 1-mg Fe/mL in order to decrease the volume injected into the mice inthe later steps. If the concentration is found to be less than 1-mg Fe/mL, thesuspension can be concentrated by centrifugal filtration, as described above.

7. The fluorescence emission of the particles is also determined by diluting 10 μL of theparticle solution in 3 mL of PBS. A standard solution of AF750 in PBS is also created


142

with its optical density (OD) matched to that of the AF750 on the nanoparticle. Thesolutions are then excited at 730 nm, and observed from 750 to 850 nm, and the areaunder the curve is integrated. Fluorescence quenching is determined by the relativeintegrated area.

9.4.2 Intravital Fluorescence Microscopy

1. Twenty four hours prior to imaging, mice are injected with the TNP or the controlagent at a dose of 10-mg Fe/kg weight via the tail vein. This allows for nanoparticlelocalization and blood clearance prior to imaging.

2. On the day of imaging, the mice are anesthetized by inhalation anesthesia (2%isoflurane, 1 L/min O2) using an isoflurane vaporizer. The distal right commoncarotid artery is carefully exposed with removal of the periadventitial tissues and theatherosclerotic plaques are visually identified (Figure 9.2). Animals are placed on awarmed glass plate and maintained on inhalation anesthesia during the imagingsession.

9.4 Methods

143

Figure 9.1 (a) UV-vis absorption spectrum of TNP. (b) Normalized UV-vis absorption (—) and fluo-rescence emission spectra (- - -) of TNP. Fluorescence emission excited at 730 nm.

Figure 9.2 Field of view through a dissecting microscope following exposure of the right carotidartery. The atherosclerotic lesions is circled.

3. Multichannel intravital fluorescence imaging is performed with a prototypical laserscanning fluorescence microscope (Olympus Corporation, Japan) after carotid arteryisolation. Two excitation wavelengths, 488 and 748 nm, are used. Image acquisitionis 1 second. Software (FluoView 300, Olympus) is used to control the fluorescence

microscope. The images collected are 512 × 512 pixels with a pixel size of about 5.4

μm/pixel. The total image size is approximately 2.75 mm × 2.8 mm. Acquired magesare stored as 16-bit multilayer Tagged Image File Format (TIFF) files. Images in theFITC channel and Cy7 channel, with a 505- to 525-nm bandpass and a 770-nmlong-pass filter, respectively, are collected simultaneously. A dry objective (4x) with afield of view of 3.25 mm and a theoretical lateral resolution of about 2.6 µm at 680nm is used. The detectors for both visible light and near infrared signals are widespectral response photomultiplier tubes (model R928P, Hamamatsu, Japan).

4. Once the atheroma is located within the field of view, fluorescein-labeled dextran(FITC-dextran, Sigma) is injected into the mice in order to better delineate thevasculature and luminal-encroaching plaques. FITC-dextran is a long circulatingagent utilized for fluorescent angiograms, and is injected at a dose of 5 mg/kg weightat the time of imaging. Often filling defects become evident in areas of high plaqueburden. These plaque-induced defects often correspond to areas of increased uptakeof the TNP.

9.4.3 Light-Based Therapy

1. Following the imaging session, the atheroma is treated with wavelength-specificlight in order to elicit a therapeutic effect. The exposed carotid artery is illuminatedwith a 650-nm diode laser (150 mW, 3 min, total fluence = 11 J/cm2) utilizingan optical fiber and collimator in order to ensure homogenous light distribution(Figure 9.3).


144

Figure 9.3 Illumination of the right carotid artery with a 650-nm diode laser after the initial imagingsession in order to elicit a therapeutic response.

2. After illumination, the incision is sutured and the mice are allowed to recover.

3. The mice are divided into cohorts with two different endpoints: 1 and 3 weeks. Oneday prior to the designated endpoint, the mice are reinjected with the respectiveagent, which is given 24 hours to localize, followed by exposure of the carotid arteryand IVFM, as described above.


9.5.1 Characterization of Theranostic Nanoparticles

Once synthesized, it is important to determine the concentrations of the particles insolution, given as mg iron/mL, and the concentrations of each of the chromophores inthe nanoparticle suspension. The concentration of the TNP is determined using thevalue of the concentration of the CLIO-NH2 determined in Section 9.4.1, step 3. Tenmicroliters of CLIO-NH2 is diluted to 3 mL, and its optical density (OD) is determined at300 nm. (Note: The OD of the particles is highly variable between particle preparations,thus it is important to determine the OD each time a new preparation of nanoparticlesis synthesized.) Similar to an extinction coefficient, the ratio of concentration toOD should be constant, and can be used in the determination of the concentrationof the TNP.

The concentration of the TNP is determined ratiometrically after dilution of theproduct with PBS. For example, if the initial concentration of CLIO-NH2 was 8.43-mgFe mL–1, and it was diluted as described (10 μL in 3 mL total volume), the final concentra-

tion would be 2.81 × 10–2 mg Fe mL–1. This solution would have an optical density of 1.67AU at 300 nm. If 10 μL of the TNP solution is diluted to 3 mL and the OD was measuredto be 0.297 AU, the concentration of the TNP would be calculated to be 1.5-mg Fe mL-1

from the equation below:

[ ] [ ][ ] [ ]ST OD TNP OD

TNP TNPdil st dil TNP

dil

=

= × D

Where [ST]dil is the concentration of the diluted standard. ODst is the optical density ofthe diluted standard, [TNP]dil is the concentration of the diluted TNP, ODTNP is the opticaldensity of the diluted TNP, [TNP] is the concentration of TNP in the original sample, andD is the dilution factor. Thus:

( ) [ ][ ]

2 18 10 167 0297

499 10

2 1

3

. . .

.

× =

= ×

− −

−

mg Fe mL U TNP U

TNP mg F

Α Α

e mL mL mL mg Fe mL− − − −× × =1 1 2 13 1 10 150.

The concentration of the AF750 is calculated from the optical density of the dye at itsmaxima in the diluted solution and the extinction coefficient for the dye times the dilu-tion factor using Beer’s law as follows:

A Cd D= ε

9.5 Data Acquisition, Anticipated Results, and Interpretation

145

Where A is the optical density of the diluted sample, ε is the extinction coefficient ofthe dye, c is the concentration of the dye in solution, and d is the pathlength of thecuvette. As before, D is the dilution factor. Thus, for the diluted solution above with anOD of 0.0848 AU at 760 nm, the concentration of AF750 is:

( )( )( ) ( )0 0848 2 10 1 3 1 10

127 10

5 1 1 2

4

.

.

AU = × ×

= ×

− − −

−

L C

C

mol cm cm mL mL

M

The calculation of the concentration of the TPC is complicated by its overlap at 648 nmwith the broad absorption of the AF750. It is thus necessary to determine the absorptionof AF750 relative to its absorption at 760 nm. This can be accomplished using theCLIO-AF750 that was synthesized for use as the control, nontherapeutic nanoparticle.Dilution of 10 μL of CLIO-AF750 to 3 mL followed by acquisition of its UV-visibleabsorption spectra gives ratio of the OD of AF 750 at 760 and 648 nm, which is approxi-mately 0.45. Thus the concentration of the TPC is calculated from the followingequations:

ODTPC = OD648 − 0.45OD760

ODTPC = 0.0611 AU − (0.45 × 0.0848 AU) = 0.0229 AU

Where OD648 and OD760 are the optical densities at the respective wavelengths in thedilute TNP solution. The concentration is then calculated from the OD of the TPC at648, using Beer’s law and the dilution factor.

( )( )( ) ( )0 0229 3 10 1 3 1 10

2 29 10

4 1 1 2

4

.

.

AU = × ×

= ×

− − −

−

L C

C

mol cm cm mL mL

M

All measurements should be conducted in triplicate with the average value used inall subsequent experimentation.

9.5.2 Animal Experimentation

For all animal experimentation, the minimal number of animals per cohort should be 5to allow for a more accurate determination of the results. For the experiments detailedabove, there are four total cohorts of mice, due to the number of endpoints, as well as theuse of therapeutic and control nanoparticle preparations. Thus, the minimum numberof animals used in this method will be 20, although more can be added in order toincrease the significance of the results.

9.5.3 Intravital Fluorescence Microscopy

The fluorescent signal was determined as integrated signal intensities (SI) from manuallydrawn regions of interest (ROI) on areas of plaque using ImageJ software (NationalInstitutes of Health, Bethesda, Maryland). The plaque target-to-background ratio (TBR)was calculated as follows: TBR = [SI(plaque) / SI(blood)]. SI is equal to the integrated sig-nal density divided by the area of the ROI. For the treated plaque depicted in Figure 9.4,below, the initial TBR, before laser illumination was 4.08, which is derived from the SI ofthe yellow ROI divided by the SI of the green ROI (45.6 A.U./11.1 A.U.). One week after


146

treatment the TBR decreases significantly to 0.6 (7.42 A.U./11.9 A.U., Figure 9.5). For thecontrol nanoparticles, the TBR remains increases, with 5.2 (34.7 A.U./6.02 A.U.) beforelaser irradiation and 13.0 (37.1 A.U./2.8 A.U.) after.

9.5.4 Statistical Analyses

All results should be reported as mean ± standard deviation. For differences betweenmultiple groups, a one-way ANOVA followed by a posthoc Tukey’s test for multiple com-parisons should be used. A p-value of < 0.05 is considered significant.

9.5 Data Acquisition, Anticipated Results, and Interpretation

147

Pretreatment 1 week post-treatment

Figure 9.4 Carotid atheroma before and after therapy with TNP. The signal from the Cy7 channel ofthe IVFM (red) decreases due to the focal ablation of inflammatory macrophages (4.1 pretreatment vs.0.6 post-treatment). TBRs are calculated as a ratio of the integrated signal intensity for a specific ROI ofthe plaques (yellow circles) versus an ROI for the blood (green circle). The bottom images (blue) are theangiogram acquired in the FITC channel after administration of FITC-dextran.

Pretreatment 1 week post-treatment

Figure 9.5 Carotid atheroma before and after therapy with the control agent, CLIO-AF750. The sig-nal from the Cy7 channel of the IVFM (red) increases (5.2 pretreatment vs. 13.0 post-treatment). Thebottom images (blue) are the angiogram acquired in the FITC channel after administration ofFITC-dextran.

9.5.5 Anticipated Results

Upon completion of the experiments and relevant calculations, trends should beobserved in both the treated and untreated cohorts at all time points. The TBR for alltreated animals should decrease between the pre- and post-treated imaging sessions,while those animals in the control groups should show increases in TBR. This is indica-tive of a decrease in the number of phagocytic cells contained within the atheromata.This data can be further confirmed by correlative histology examining the localizationof the agent on the microscopic level, as well as the relative macrophage content of thelesions after therapy.


The utility of theranostic nanoagents in vivo has, thus far, received little attention.While there are a number of publications detailing the synthesis of this class of materi-als, methods must be developed in order to determine their efficacy. Although thismethod is very specific in its scope, it can be readily applied to investigate the role offocal cell ablation in any number of diseases.

The synthesis of the light-activated theranostic nanoparticles is based upon a stan-dard protocol. CLIO is formed by the epichlorohydrin induced crosslinking of the dex-tran coating material of MION, followed by its amination. This reaction is usually doneon a large scale to limit the batch-to-batch variability inherent in the synthesis, espe-cially the amination. Following the purification and concentration of CLIO-NH2, theAF750 and TPC are conjugated to the particle by reaction of the succinimidyl esterfunctionalized dyes with the amines on the nanoparticle surface, thus forming amidebonds. The control nanoparticle, bearing only AF750, is taken as an aliquot from theproduct formed after reaction with AF750 (Section 9.4.1, step 4). This allows for the fluo-rescence of the control particle to be matched to the fluorescence of the TNP. At thispoint, the optical properties of the products are quantified, including the concentrationof the particle (in mg Fe mL-1), and the concentration of the dyes (in M). The fluores-cence emission of the particle is also qualitatively examined, as compared to anequimolar solution of AF750. If too many dye molecules are present on the nanoparticlesurface, dye-dye quenching can occur, resulting in a particle with minimal fluorescenceemission. Unfortunately, if the fluorescence is quenched, the particles are no longerviable, and should be discarded and resynthesized using decreased amounts of thedye starting materials.

The in vivo efficacy of the TNP is examined in aged atherosclerosis-laden apoE-/-

mice. These mice are put on a high-cholesterol diet at about 10 weeks of age, and are kepton that diet until the beginning of the study, in order to induce atherosclerosis withhigh levels of inflammation. One day prior to the initial imaging session, the mice areinjected with the respective agents. This time allows for the maximum localization ofthe nanoparticles to the lesions of interest. On the day of imaging, the mice are anesthe-tized and the carotid artery is surgically exposed, and examined visually for the presenceof atheromata. Occasionally, the artery that is exposed contains no lesions. If this doesoccur, the contralateral carotid artery can be used after surgical exposure. The lesions arethen located in the Cy7 channel of the IVFM. At this point, FITC-dextran is injected inorder to better delineate the vasculature, and Z-stack images are acquired in both the


148

FITC and Cy7 channels of the IVFM. Following imaging, the exposed carotid artery isilluminated with a 650-nm laser in order to elicit a therapeutic response, the surgicalincision is sutured, and the mice are allowed to recover. The mice are also returned to anormal diet in order to prevent the formation of new lesions.

At the requisite time point after therapy (1 or 3 weeks), the animals are reinjectedwith the respective agents, which are given 24 hours to localize. The surgical incision isreopened and the mice are reimaged. Following the imaging sessions, the Z-stack imagesare summed for each mouse, and the TBR for each mouse is determined before and aftertherapy using hand-drawn ROIs. The TBR is the ratio of the signal intensity from thelesion to the signal intensity from the adjacent blood. As is illustrated above, all treatedmice should exhibit a significant decrease in TBR, indicative of macrophage ablation,while the control group should show no change, or a slight increase in TBR. Althoughunlikely, it is also possible for these mice to show a decrease in signal intensity, due tothe withdrawal of the high-cholesterol diet.

While this method enables the longitudinal in vivo study of theranostic nanoagentlocalization and therapeutic efficacy, there are several other techniques that can beutilized ex vivo, such as flow cytometry and immunohistochemistry. Digestion ofcarotid arteries in a collagenase cocktail, followed by fluorescent antibody labeling andflow cytometric analysis enables the identification of the cell types containing thenanoagent, as well as the relative proportion of each cell type [24]. Similarly, the carotidarteries can be embedded and sectioned for histological identification of nanoparticlelocalization, and the relative content of each cell type within the lesion. The main draw-back of these techniques is that they require the sacrifice of the animal, and as such, cannot be utilized to examine the therapeutic response over the course of the study.



Section 9.4.1, step 7The material is minimally ornonfluorescent.

The conjugation of the chromophoresto the particle was inefficient.

The conjugation of the chromophores to theparticle was too efficient, causing quenching.

1. Repeat UV-vis absorption measurements.2. Remove any impurities by dialysis.3. Repeat reaction with chromophores.1. Repeat UV-vis absorption measurements.2. Material is of no utility. Start again fromCLIO-NH2.

Section 9.4.1, step 5Material precipitates.

The presence of divalent cations can causeprecipitation.

Ensure that PBS does not contain calcium andmagnesium.

Section 9.4.2, step 2No lesions present.

Lesion formation is variable in the apoE-/-model.

Expose contralateral carotid artery and inspectfor lesions.

Section 9.4.2, step 3No signal from lesion.

Lesion uptake of particle is poor, or nanoagenthas degraded.

Ensure that mice have visible carotid plaques.Repeat UV-vis and fluorescence measurementsto ensure nanoparticle composition.

9.7 Summary Points

Theranostic nanomaterials comprised of crosslinked dextran coated iron oxidenanoparticles, NILAT agents, and fluorophores, are capable of imaging atheroscleroticlesions.

9.7 Summary Points

149

• Focal macrophage ablation results in a decrease in signal from the theranosticnanoparticle upon reinjection and imaging.

• Theranostic nanoparticles can be functionalized for use in numerous diseases, asthey can be targeted to specific cell or tissue types.

• The therapeutic portion of the theranostic nanoagent must be optimized for spe-cific applications, with special regard paid to the intrinsic toxicity of the therapeu-tic ligand.

Acknowledgments

We would like to thank Drs. Ethan Korngold, Jose-Luiz Figueiredo, and Rainer Kohler,and Purvish Patel for their assistance in developing this method. This work wassupported in part by NIH grants U01-HL080731 (RW), U54-CA119349 (RW), andU54-CA126515 (RW).

References

[1] Hansson, G. K., “Inflammation, Atherosclerosis, and Coronary Artery Disease,” N. Engl. J. Med.,Vol. 352, No. 16 2005, pp. 1685–1695.

[2] Hansson, G. K., and P. Libby, “The Immune Response in Atherosclerosis: A Double-Edged Sword,”Nat Rev Immunol, Vol. 6, No. 7, 2006, pp. 508–519.

[3] Libby, P., “Inflammation in Atherosclerosis,” Nature, Vol. 420, No. 6917 2002, pp. 868–874.[4] Galis, Z. S., G. K. Sukhova, and R. Kranzhofer, et al., “Macrophage Foam Cells from Experimental

Atheroma Constitutively Produce Matrix-Degrading Proteinases,” Proc. Natl. Acad. Sci. U S A, Vol.92, No. 2, 1995, pp. 402–406.

[5] Lendon, C. L., M. J. Davies, G. V. Born, and P. D. Richardson, “Atherosclerotic Plaque Caps AreLocally Weakened When Macrophages Density Is Increased,” Atherosclerosis, Vol. 87, No. 1, 1991,pp. 87–90.

[6] Moreno, P. R., E. Falk, and I. F. Palacios, et al., “Macrophage Infiltration in Acute Coronary Syn-dromes. Implications for Plaque Rupture,” Circulation, Vol. 90, No. 2, 1994, pp. 775–778.

[7] Redgrave, J. N., J. K. Lovett, P. J. Gallagher, and P. M. Rothwell, “Histological Assessment of 526Symptomatic Carotid Plaques in Relation to the Nature and Timing of Ischemic Symptoms: TheOxford Plaque Study,” Circulation, Vol. 113, No. 19, 2006, pp. 2320–2328.

[8] van der Wal, A. C., A. E. Becker, and C.M. van der Loos, et al., “Fibrous and Lipid-RichAtherosclerotic Plaques are Part of Interchangeable Morphologies Related to Inflammation: A Con-cept,” Coron Artery Dis, Vol. 5, No. 6, 1994, pp. 463–469.

[9] Cuchel, M., and D. J. Rader, “Macrophage Reverse Cholesterol Transport: Key to the Regression ofAtherosclerosis?,” Circulation, Vol. 113, No. 21, 2006, pp. 2548–2555.

[10] Li, A. C., and C. K. Glass, “The Macrophage Foam Cell as a Target for Therapeutic Intervention,”Nat. Med., Vol. 8, No. 11, 2002, pp. 1235–1242.

[11] Liang, C. P., S. Han, T. Senokuchi, and A. R. Tall, “The Macrophage at the Crossroads of InsulinResistance and Atherosclerosis,” Circ. Res., Vol. 100, No. 11, 2007, pp. 1546–1555.

[12] Naghavi, M., P. Libby, and E. Falk, et al., “From Vulnerable Plaque to Vulnerable Patient: A Call forNew Definitions and Risk Assessment Strategies: Part II,” Circulation, Vol. 108, No. 15, 2003,pp. 1772–1778.

[13] Enochs, W. S., G. Harsh, F. Hochberg, and R. Weissleder, “Improved Delineation of Human BrainTumors on MR Images Using a Long-Circulating, Superparamagnetic Iron Oxide Agent,” J MagnReson Imaging, Vol. 9, No. 2, 1999, pp. 228–232.

[14] Rydland, J., A. Bjornerud, and O. Haugen, et al., “New Intravascular Contrast Agent Applied toDynamic Contrast Enhanced MR Imaging of Human Breast Cancer,” Acta Radiol, Vol. 44, No. 3,2003, pp. 275–283.

[15] Harisinghani, M. G., S. Saini, and R. Weissleder, et al., “Splenic Imaging with UltrasmallSuperparamagnetic Iron Oxide Ferumoxtran-10 (AMI-7227): Preliminary Observations,” J ComputAssist Tomogr, Vol. 25, No. 5, 2001, pp. 770–776.


150

[16] Saini, S., R. Sharma, and R. L. Baron, et al., “Multicentre Dose-Ranging Study on the Efficacy ofUSPIO Ferumoxtran-10 for Liver MR Imaging,” Clin Radiol, Vol. 55, No. 9, 2000, pp. 690–695.

[17] Kooi, M. E., V. C. Cappendijk, and K. B. Cleutjens, et al., “Accumulation of UltrasmallSuperparamagnetic Particles of Iron Oxide in Human Atherosclerotic Plaques Can Be Detected byIn Vivo Magnetic Resonance Imaging,” Circulation, Vol. 107, No. 19, 2003, pp. 2453–2458.

[18] Trivedi, R. A., C. Mallawarachi, and J. M. U-King-Im, et al., “Identifying Inflamed Carotid PlaquesUsing In Vivo USPIO-Enhanced MR Imaging to Label Plaque Macrophages,” Arterioscler ThrombVasc Biol, Vol. 26, No. 7, 2006, pp. 1601–1606.

[19] Trivedi, R. A., J. M. U-King-Im, and M. J. Graves, et al., “In Vivo Detection of Macrophages inHuman Carotid Atheroma: Temporal Dependence of Ultrasmall Superparamagnetic Particles ofIron Oxide-Enhanced MRI,” Stroke, Vol. 35, No. 7, 2004, pp. 1631–1635.

[20] Josephson, L., C. H. Tung, A. Moore, and R. Weissleder, “High-Efficiency Intracellular MagneticLabeling with Novel Superparamagnetic-Tat Peptide Conjugates,” Bioconjug Chem, Vol. 10, No. 2,1999, pp. 186–191.

[21] Jaffer, F. A., M. Nahrendorf, and D. Sosnovik, et al., “Cellular imaging of inflammation in athero-sclerosis using magnetofluorescent nanomaterials,” Mol Imaging, Vol. 5, No. 2, 2006, pp. 85–92.

[22] Choi, Y., J. R. McCarthy, R. Weissleder, and C. H. Tung, “Conjugation of a Photosensitizer to anOligoarginine-Based Cell-Penetrating Peptide Increases the Efficacy of Photodynamic Therapy,”ChemMedChem, Vol. 1, No. 4, 2006, pp. 458–463.

[23] McCarthy, J. R., F. A. Jaffer, and R. Weissleder, “A Macrophage-Targeted Theranostic Nanoparticlefor Biomedical Applications,” Small, Vol. 2, No. 8–9, 2006, pp. 983–987.

[24] Swirski, F. K., P. Libby, E. Aikawa, et al., “Ly-6Chi Monocytes Dominate Hypercholesterolemia-Associated Monocytosis and Give Rise to Macrophages in Atheromata,” J. Clin. Invest., Vol. 117,No. 1, 2007, pp. 195–205.

References

151

C H A P T E R

1 0Biomedical Applications of Metal Nanoshells

André M. Gobin

University of Louisville, Louisville, KY

153

Key terms nanoshellsnear infraredphotothermal therapylaser therapyplasmon resonance

Abstract

This chapter details the methods associated with producing near infrared(NIR) resonant composite nanoparticles called nanoshells. Engineered nano-structures called nanoshells were first designed and fabricated at Rice Univer-sity and consist of a dielectric core of silica and a metal shell, generally gold.Gold nanoshells are particularly useful for biomedical applications due tobiocompatibility of gold and the ability to tune the resonance of these particlesto match virtually any wavelength. This chapter addresses the methods of pro-ducing gold nanoshells, passivating the surface for in vivo studies and conju-gating biomolecules to the surface followed by testing the concentration ofbound antibodies on the surface. These techniques allows one to producenanoshells with specific NIR resonance and allow targeting to a variety of celltypes via antibodies or ligands or to other targets and could be used to extendthe use of nanoshells beyond therapeutic applications.

10.1 Introduction

Nanoshells are a relatively new class of nanoparticles consisting of an ultrathin metalshell (generally gold) surrounding a dielectric core such as silica. Gold nanoparticleshave low toxicity, thus gold-coated nanoshells displaying a gold surface has the samedegree of biocompatibility as solid gold nanoparticles used in a variety of applicationstoday. With their facile optical tunability, nanoshells are ideal for biological applica-tions in the near infrared (NIR). The NIR window is defined as a region where energy oflight having wavelengths between 650 to 900 nm can penetrate through tissue relativelyunimpeded by hemoglobin or water. This allows one to tune the nanoshells to matcha laser wavelength in this region and create a pair of nanoparticle + laser that caneffectively be used for therapeutic purposes.

This chapter examines the methods used in making nanoshells, conjugation ofbiomolecules to the surface, quantification of attached antibodies on nanoshells as wellas in vitro and in vivo testing of nanoshells to determine efficacy as a therapeutic agent.

Nanoshells can be designed to either strongly absorb or scatter light in the NIR basedon the dimensions of the core and shell and overall size, permitting applications forheating or optical contrast as discussed in detail in this chapter. The gold surface allowsfor easy conjugation of proteins through the use of a PEG linker that contains a disulfideor thiol moiety. Due to their unique properties and biocompatibility, gold nanoshellshave been investigated for a variety of biomedical applications. These include beingused as a mechanism to provide heating for photothermally modulated drug deliverysystems, fast antigen detection systems with whole blood, use in imaging applications,as an exogenous NIR absorber for tissue welding or bonding, cancer therapy by nonspe-cific accumulation in tumors and for targeted cell ablation using antibody targetingmechanisms. In this chapter we study the binding of an antibody to nanoshells asa method to target prostate adenocarcinoma using the prostate specific membraneantigen (PSMA). The conjugation technique to the polymer linker and to the nano-shell is highlighted in the methods section and results of measurements of antibodyconcentrations are shown in the results section.

10.1.1 Biomedical Applications of Metal Nanoshells

Since the development of gold nanoshells in 1997 by the Halas group, numerous poten-tial applications have been explored. Of particular interest to this discussion is theirapplication to biomedicine due to the inert and biocompatible nature of the gold coat-ing, the flexibility of the chemistry that can be performed on gold surfaces, and the easewith which the optical properties of metal nanoshells can be manipulated. Gold/silicananoshells have already proven very effective for photothermal cancer therapy in vivoby taking advantage of its ability to absorb NIR energy and create heat. In work by Hirschand O’Neal, the nanoshells used were primarily absorbing, at about 85% absorbing effi-ciency and provided up to 100% regression of tumors in mice after treatment [1]; how-ever the scattering properties of the nanoshell has also been exploited for in vitroimaging [2] as well as for combined imaging and photothermal ablation in vitro [3]. Inthe imaging studies it was demonstrated that the nanoshell could be used to provideadequate scattering for imaging contrast and retain NIR absorbing properties sufficientto allow photothermal ablation in vitro [3]. This chapter details the successful in vivodemonstration of the use of near infrared resonant gold nanoshells, to first increase opti-

Biomedical Applications of Metal Nanoshells

154

cal contrast in tumors for optical coherence tomography (OCT) imaging for diagnosticsand second, to subsequently treat the tumors by absorption of near infrared (NIR)light for photothermal ablation. The approach discussed in this chapter uses a singlenanoparticle formulation that has been designed to have both absorption and scatteringin the NIR to accomplish diagnostic imaging and therapeutic benefits simultaneously[4]. Nanotherapeutics like these can allow the development of “see and treat”applications that is expected to reduce patient care costs and allow wider delivery oftreatments.

Nanoshell mediated cancer therapy has many benefits compared to traditional che-motherapy or radiotherapy methods, particularly in its potential to reduce side effects.Whereas, the side effects of the drugs typically used in chemotherapy or the radiationused leaves various uncomfortable side effects, the gold nanoparticles by themselves arenot known to cause any side effects. Nanoshell mediated cancer therapy begins by pref-erential accumulation of nanoshells into the tumors due to the leaky vasculature that ischaracteristic of fast growing tumors. Tumor vasculature have pore sizes that are hun-dreds of nanometers in diameter compared to normal vessels that have pore sizes on theorder of tens of nanometers; this allows easy extravasation of nanoparticles into tumors.Permeability of particles up to 400 nm has been shown in human colon carcinoma, sug-gesting pore sizes up to 600 nm. Thus, tumors become laden with nanoshells while otherhealthy tissues with normal tight endothelial junctions in the vasculature have minimalaccumulation of nanoparticles. The application of NIR light causes heating only in thenanoshell-laden tumor, leaving the healthy tissue unaffected. NIR light energy isabsorbed by the nanoshell creating heat. Data shows that the heating of nanoshellsupon exposure to NIR light disrupts the integrity of the cell membrane, causing death ofthe cells [5]. Heating of the tumor cells by this mechanism causes irreversible thermaldamage thus allowing the tumor to be destroyed. It has been shown that irreversiblethermal damage occurs and is evident at temperatures between 55°C to 59°C manifest-ing as edema, whitening and eventually tissue necrosis in the region. Since the NIR lightis minimally absorbed by normal tissue components, there is minimal temperatureincrease in the absence of nanoshells and no detectable damage in surrounding tissue.The delivery of PEGylated nanoshells to tumors followed by therapeutic administrationof NIR light showed up to 100% regression of tumors in a murine model [1].

10.1.2 Nanoshells for Combined Optical Contrast and TherapeuticApplication

Extinction of light on a nanoshell is due to scattering and absorbing events. The scatter-ing property of a nanoshell can be exploited for imaging applications just as theabsorbing property can be exploited for therapeutic benefit. This was demonstrated byLoo et al. in vitro using antibody targeted nanoshells to specifically bind to HER2overexpressing tumors [3]. In their study, imaging was performed using darkfieldmicroscopy; this allows imaging by illuminating the sample with light at an angle andcollecting light scattered from the objects to create an image. Given these advantages,nanoshells can be used as contrast agents for enhanced OCT imaging based ontheir backscattering properties, as well as a cancer therapeutic, due to their absorbingproperties.

10.1 Introduction

155


To evaluate targeting of nanoshells to cells of a particular type will depend on the needsof the experimenter and the final outcome desired, whether it is in vitro targeting stud-ies or ultimately in vivo targeting to tumors of a particular type. In this chapter we focuson targeting nanoshells to prostate cancer by the use of antibodies specific to receptorson this cell type which is over-expressed in greater amounts (100–1000x) more than innormal cell types. This is the first fundamental issue with being able to target, one has toevaluate the availability of a target on the intended cell type and evaluate the relativeabundance of that marker on the cell of interest in comparison to other cells.

For antibody targeting the assay should be run in triplicate as a minimum to ensureconsistency of results. For this assay it is necessary to have a negative control ofnanoshells with PEG only. This allows one to determine the background amount ofsecondary antibodies that may become entrapped with the nanoshells during thecentrifugation process.

For animal studies, the minimum number of animals required is dependent on thesignificance levels desired for a certain percentage change in the result. To this end thereare many sources including software, books, and articles for determination of thesenumbers. For our studies we chose 10 to 12 animals as the minimum to show signifi-cance greater than 95% between treated and untreated groups when using the formula-tion of nanoshells for imaging and therapeutic application.

10.3 Materials

High-purity chemicals are essential for producing good nanoparticles. Except wherenoted, chemical were obtained from Sigma (Milwaukee, WI). The chemicals required forthe many processes are grouped and listed below.

10.3.1 Nanoparticle Production

Tetraethyl orthosilicate (TEOS, 99.999%) used for producing silica cores; ammoniumhydroxide, 14-15N was used as the base catalyst in silica core nanoparticle production;and (3-aminopropyl) triethoxysilane (APTES, 99%) used to provide amine groups on thesurface of silica nanoparticles. Tetrakis (hydroxymethyl) phosphonium chloride (THPC,80%) and 1M NaOH were used to produce gold colloid via the Duff process. Gold in theform of hydrogen tetrachloroaurate (III) trihydrate (chloroauric acid) 99.99% purity waspurchased from Alfa Aesar (Ward Hill, MA) and used for all procedures requiring goldsolutions. Polyethylene glycol–SH (PEG-SH, 5000 MW) was used for blocking and passi-vating nanoshells surfaces.

10.3.2 Protein Conjugation to Nanoshells Surface

Bifunctional PEG: orthopyridyl–disulfide–poly(ethylene glycol)–N– hydroxy-succinimide ester (OPSS-PEG-NHS, 2000MW) for conjugating proteins to nano- shellssurfaces was obtained from Nektar (Birmingham, AL). PEG-SH (MW 5000; Nektar, Bir-mingham, AL) was used to block exposed gold surfaces on nanoshells to resist proteinadsorption and allow better circulation in vivo. A monoclonal anti-PSMA in the form of


156

mouse- anti-HuPSMA, clone Y/PSMA1 (M20454M) was obtained from Biodesign Inter-national (Meridian Life Sciences, Saco, ME) for PSMA targeting. Recombinant fusionproteins of mouse ephrin-A1/Fc chimera (R & D Systems, Minneapolis, MN) wereobtained for EphA2 targeting.

10.3.3 Cell Culture

Media was obtained from ATCC including: Ham’s F12K, RPMI-1640, and DMEM supple-mented with 4 mM l-glutamine, 1% penicillin, 1% streptomycin (GPS) and 10% fetalbovine serum (FBS).

10.3.4 In Vitro Assays

Quantification of antibody concentration on nanoshell surfaces through horseradishperoxidase (HRP) activity was measured with 3, 3’, 5, 5’–tetramethylbenzidine (TMB)assay (Sigma, Milwaukee, WI).

10.4 Methods

10.4.1 Fabrication of Gold/Silica Core Nanoshells

Gold nanoshell synthesis has been previously described by others [6]. First, silica coreswere grown using the Stöber process, the basic reduction of tetraethyl orthosilicate(TEOS). Next, 45 ml of 200-proof ethanol was used with 3.0 to 5.5 ml in 0.5 ml incre-ments of 14.8 N NH4OH to make six batches at different ammonia volumes. Then, 1.5 mlTEOS was added to each batch and allowed to react a minimum of 8 hours. Higher vol-ume of ammonia produces larger silica nanoparticles. Silica precipitates were centri-fuged and washed with 200-proof ethanol twice to remove any remaining NH4OH (200–3500g (size-dependent), 20 ml for 20 minutes in each step). The resultant silicananoparticles were sized using scanning electron microscopy (SEM; Philips FEI XL30).Average diameters of different batches ranged between 98 and 112 nm. Only batcheswith a polydispersity of less than 10% were used in subsequent steps. Reaction of the sil-ica core nanoparticles with 200 μl of (3-aminopropyl) triethoxysilane (APTES) per batchprovided amine groups on the surface of the cores to allow for adsorption of gold colloidin the subsequent step. Aminated silica cores were boiled for 2 hours with addition of200° ethanol to maintain volume, then cooled and washed twice by centrifugation. Thesilica core suspensions were measured to determine the weight percent of solids andadjusted to 4 wt% for storage by addition of ethanol.

For colloid production, a 1wt % gold salt solution was prepared with 99 grams 18.2

MΩ−cm H2O and 1 gram hydrogen tetrachloroaurate (III) trihydrate (chloroauric acid)99.99% purity (HAuCl4) purchased from Alfa Aesar (Ward Hill, MA) and stored in amberbottles for use in various steps requiring gold. 400 μl of (hydroxymethyl) phosphoniumchloride (THPC, 80%) was mixed with 33-ml DI water as a stock solution. To produce thecolloidal gold particles the following were mixed together: 180-ml DI water, 1.2-ml 1MNaOH, 4-ml THPC stock solution, and 6.75 ml of 1 wt % gold solution. This gold colloidmade through the Duff process has a size of 2 to 4 nm after aging for 2 to 3 weeks at 4°C.After aging the colloid was then concentrated ~20X through rotary evaporation and

10.4 Methods

157

mixed with the ammine coated silica particles at a volume of 10 ml concentrated colloid:300 μl stored silica core suspension, thus allowing small gold colloid to attach to thelarger silica nanoparticle surface to act as nucleation sites in the subsequent reductionstep. This resulted in the seed particles from which nanoshells are grown by reduction ofadditional gold using formaldehyde as the reducing agent.

Finally, the gold shell was then grown by reduction of gold using 0.4 mM HAuCl4

solution (plating solution) in the presence of formaldehyde. The plating solution ismade with 50-mM potassium carbonate and gold salt from the 1% solution to a finalconcentration of 0.4 mM HAuCl4. To produce particles with varying shell thicknesses wevaried the concentration of seed particles while using the same amount of plating solu-tion. The spectra of each set of samples were examined for optimal conditions to pro-duce desired NIR absorbing nanoshells. NIR absorption characteristics of the nanoshellswere determined using a UV-Vis spectrophotometer (Carey 5000 Varian, Walnut Creek,CA). Samples with the appropriate NIR peak resonance (~ 800 nm) were scaled up lin-early to provide nanoshells for the experiment.

10.4.2 Nanoshells for Combined Imaging and Therapy In Vivo

10.4.2.1 In Vivo Model

BALBc mice inoculated with 150,000 murine colon carcinoma cells (CT-26; ATCC) in25 μl of PBS. Tumors were allowed to grow to a cross-sectional area of 20 to25 mm2 andno more that 4 to 5 mm in any one dimension before treatment. Then, 150 μl ofPEGylated nanoshells at a concentration of 1.5 x 1010 nanoshells/ml were injected intothe tail vein of the animals 20 hours prior to imaging and laser irradiation. A total of 36animals were inoculated with the cancer cells. Animals were randomly divided intothree groups; Group 1: Nanoshell + Laser, Group 2: PBS + Laser, and Group 3: UntreatedControl.

10.4.2.2 OCT Imaging

This study used a commercially available OCT imaging system, Niris Imaging System,(Imalux; 1300 nm, Cleveland, OH). The axial and transverse resolutions were approxi-mately 10 and 15 μm, respectively. OCT images were collected for nanoshell-injectedand control mice 20 hours following injection (to allow time for passive accumulation ofnanoshells) and analyzed to assess the increase in contrast provided by the nanoshells intumor tissue compared to normal tissue. OCT images of the tumor and normal tissuewere taken after 20 hours of circulation. The animals were not anesthetized during theinjection or circulation period, only during imaging and treatment by the NIR laser. Thetumors were imaged using the Niris OCT imaging device by applying glycerol on theshaved tumor site for index matching and placing the probe in contact with the skindirectly above the tumor. Images were captured at several locations on each tumorthrough the integrated computer and image analysis system. Normal tissue images weretaken at a location at least 2 cm distant to the tumor on the same animal. For statisticalanalysis, images were analyzed to first quantify the contrast levels using standardthresholding for image analysis then intensity data were analyzed using an unpaired stu-dent t-test assuming equal variance with a confidence interval of 95%, p < 0.05 of thetwo populations of images from PBS-treated and nanoshell-treated mice.


158

10.4.2.3 Therapeutic Laser Irradiation

After imaging, the tumors were irradiated with a NIR laser. In vivo irradiation wasaccomplished using an Integrated Fiber Array Packet, FAP-I System, with a wavelength of808 nm (Coherent, Santa Clara, CA) at a power density of 4 W/cm2, 5-mm diameterspot for 3 minutes. Animal survival was monitored for 7 weeks after imaging and treat-ment. Following treatment, survival data analysis was performed using the standardKaplan-Meier analysis using MedCalc software to determine statistical significance aftertherapy. Analysis of the tumor regression was performed using the average measure-ments of the tumor size of the surviving populations at the times shown and comparedusing an unpaired student t-test assuming equal variance with a confidence interval of95%, p< 0.05.

10.4.2.4 Nanoshell Accumulation in Tissue

Three animals from each group were sacrificed following treatment to examine thetumors for the presence of nanoshells using silver enhancement staining. One half ofthe frozen tumors were sectioned to 8 μm, and silver staining was performed using theSigma Silver Enhancement solutions (Sigma, Milwaukee, WI). Images of each sectionwere taken at 64x magnification to look for the presence of nanoshells; silver stainingallows the nanoshells to act as nucleation sites for deposition of silver to grow largeenough to allow for visualization under light microscopy. The second half of the tumorwas sent to Texas A&M University for nuclear activation analysis (NAA). Tissue samplesfor NAA were lyophilized and weighed; blanks and the dried tumor sample were irradi-ated along with precise calibration standards at the Texas A&M University’s Nuclear Sci-ence Center 1 MW TRIGA research reactor for 14 hours. The irradiation position used inthis study has an average neutron flux of approximately 1 x 1013 sec-1cm-2. High-puritygermanium detectors with nominal resolutions (FWHM) of 1.74 keV or better and effi-ciencies of 25-47 % by industry standard relative measurement were used to quantify the412 keV gamma line from 198Au. The Canberra Industries OpenVMS alpha processor-based Genie-ESP software was used for acquisition and computation of goldconcentrations.

10.4.3 Passivation of Nanoshells with PEG

Nanoshells were surface-coated with poly (ethylene glycol) PEG to enhance circulationtimes and reduce immune response in vivo. PEGylation was accomplished by adding100 μl of 5-μM PEG-SH, molecular weight 5 kDa (Nektar, Huntsville, AL) to 20 ml of ananoshell suspension with an optical density (OD) of 2.0 (~6 x 108 particles/ml) in DIwater for a minimum of 8 hr at 4°C. PEG-modified nanoshells were sterilized by filtra-tion using a 0.22-μm filter and subsequently centrifuged to increase concentration. Tofacilitate injection in vivo, nanoshells were resuspended in sterile phosphate-bufferedsaline (PBS), pH = 7.4 at physiological salt concentration, to an OD =50 (~1.5x1010 parti-cles/ml). For concentration of the sample, a force of 1500g was used to spin down thesample to a pellet and the supernatant was removed. The sample was then diluted withsterile PBS and measured and adjusted as necessary to ensure final concentration ofOD = 50. At this point the suspension of nanoshells is ready for in vivo use.

10.4 Methods

159

10.4.4 Conjugation of Biomolecules to Nanoshells

A bifunctional PEG polymer, Orthopyridyl-disulfide-poly(ethylene glycol)-N-hydroxysuccinimide ester (OPSS-PEG-NHS, 2000MW) was obtained from Nektar (Bir-mingham, AL). Protein of interest is dissolved or diluted to known concentrations with100-mM sodium bicarbonate at pH 8.5. In the case of anti-PSMA targeted nanoshells weused a monoclonal anti-PSMA in the form of mouse-anti-HuPSMA, clone Y/PSMA1(M20454M) obtained from Biodesign International (Meridian Life Sciences, Saco, ME).The polymer was reacted at a mole ratio of 2:1 with anti-PSMA for four hours at 4 °C. TheOPSS-PEG-NHS molecule binds when the NHS group cleaves in aqueous environmentleaving an activated carboxylic terminus that can bind to a free primary amine group onthe antibody or other protein forming a peptide bond and covalently linking the PEG tothe antibody to form OPSS-PEG-antibody. After reaction the conjugated PEG-Ab solu-tion is incubated with nanoshells at calculated to be ~2000 Ab fragments per nanoshellparticle. The mixture is allowed to incubate for 1 hour after which PEG-SH at the con-centrations discussed above is added to complete passivation of the rest of the gold sur-face. Binding of the antibody to the surface prior to blocking allows for a higherconcentration of Ab on the nanoshell surface and maximizes the use of the antibodysolution. This reaction scheme is shown in Figure 10.1.

10.4.5 Quantification of Antibodies on Nanoshells

Suspensions of conjugated nanoshells prepared as described above were centrifuged at1000g to separate unbound antibodies from the particles. Nanoshells were blocked ina 3% bovine serum albumin (BSA) for 1 hour. The washing step was repeated for a total


160

-PEG -SS

H2O

+N

H

H

NR N

O

O

O C

O

(CH2CH2O)n-NH-CH2CH2C-S-S-

O

CH2CH2

N

H+

N

H

H

NR -O C

O


O

CH2CH2

N

NHS Cleaves in water

Peptide bond forms at activated carboxylic terminus

NR C

O


O

CH2CH2

NH -PEG -SS-PEG -SS

H2O

+N

H

H

NR N

H

H

NR N

O

O

O C

O


O

CH2CH2

NN

O

O

O C

O

N

O

O

O C

O


OO

CH2CH2

NN

H+

N

H

H

NR N

H

H

NR -O C

O


O

CH2CH2

N

-O C

O


OO

CH2CH2

NN

NR C

O


O

CH2CH2

NHNR C

O


OO

CH2CH2

NNH

Antibody ( ) has PEG with a disulfide covalently attachedR

Step 1

Step 2

Step 3

Antibody ( )with availableprimary amine

R OPSS-PEG-N-hydroxysuccinimide

Figure 10.1 Representation of antibody binding to bifunctional PEG for subsequent conjugation tonanoshells. At the end of the reaction the PEG is covalently attached to the antibody and the disulfideis able to bind to the gold surface of the nanoshell after the protecting group leaves.

of two times. Suspensions were then incubated with an anti-mouse IgG antibody conju-gated to horseradish peroxidase (HRP), Sigma, A-4416.Then, 450 μl of nanoshell suspen-sion was added to 50 μl of A-4416 diluted to100 μg/ml and incubated for 1 hour.Suspensions were washed by centrifugation and resuspension twice to remove unboundsecondary antibody; after a third centrifugation, supernatant and nanoshells wereretained for HRP quantification. HRP standards were made up at concentrations rangingfrom 2 ng/ml to 100 ng/ml and nanoshells as well as supernatant were assayed for HRPusing 3,3’,5,5’–tetramethylbenzidine (TMB). The reaction is generally stopped after 5-7minutes by use of H2SO4 and read using a plate reader at 450 nm (model ELX800; BioTekInstruments, Winooski, VT). It may be useful to run the assay once to determine thereaction rate during development relative to the standards to align the assay so that thebest standard curve can be obtained for the range of the concentration of antibody inthe nanoshell suspension. After determination of the concentration of antibody one candetermine the number of antibody molecules on the nanoshells’ surface by using theconcentration of nanoshells as determined by spectroscopic measurements. It is essen-tial to use nanoshells that contain PEG only as a control to assure the backgroundamount of secondary antibody is accounted for during incubation step and subsequentwashing.

10.5 Results

10.5.1 Gold/Silica Nanoshells Allow Both Imaging Contrast Increase andTherapeutic Benefit

10.5.1.1 OCT Image Analysis

PEG-modified nanoshells were injected intravenously in tumor-bearing mice andallowed to passively accumulate in the tumor tissue due to the leakiness of the tumorvasculature. The significant accumulation of particles within the tumor tissue dramati-cally increased the NIR scattering within the tumor, enhancing the OCT contrast.Figure 10.2 shows representative OCT images of tumors of mice prior to irradiation withthe 808-nm laser. OCT images of normal and tumor tissue of mice treated injected withsaline are shown in Figure 10.2(a) and (c). Figure 10.2(b) and (d) are of mice injectedwith nanoshells. Note the enhanced contrast in the image (d) indicates that the goldnanoshells can be visualized with OCT system and shows higher contrast within the tis-sue of either the normal tissue area or the tumor treated with saline. Figure 10.3 showsthe quantification of the image intensity of normal tissue (n = 3) and tumor tissue (n = 6)with PBS injection and nanoshell injections. OCT images were analyzed to quantitatethe contrast and analyzed using a student’s t-test of the two populations of images fromPBS treated and nanoshell (NS) treated mice. The data shows a significant increase in thecontrast of tumor compared to normal tissue when nanoshells are used. No statisticaldifference is observed in the contrast of images of normal tissue whether nanoshells orsaline are used.

10.5.1.2 Histological Analysis

Histological examination of tumors using silver staining confirmed that OCT signalswere the result of scattering from nanoshells within the tumor. Figure 10.4 shows the sil-

10.5 Results

161

ver staining of representative areas of tumors from mice treated with nanoshells (a) orwith PBS (b) showing a marked increase in darkening of the tissue in (a), indicating thepresence of nanoshells within the tumor. Additionally, neutron activation analysis(NAA) verified nanoshells present in the tumor shown in Figure 10.4(a) at 12.5 ppmcompared to 0 ppm for tumors of mice injected with just PBS Figure 10.4(b).

10.5.1.3 Survival Following Imaging and Therapy

Tumor regression and survival of the mice were followed for 7 weeks after treatment.Figure 10.5(a) shows the tumor sizes on the day of treatment and 12 days aftertreatment; tumors on nanoshell-treated mice were completely regressed except forone mouse. Figure 10.5(b) shows the survival of the mice during the studyperiod. Kaplan-Meier statistical analysis shows a median survival of 14 days for thePBS + Laser group and 10 days for the Untreated Control group. By day 21 the survivalof the Nanoshell + Laser group was significantly greater than either control groups,p < 0.001.


162

Min Max

(C) Tumor tissue + PBS (D) Tumor tissue + Nanoshells

Muscle

Skin

Glass

Muscle

Skin

Glass

Min MaxMin Max

(B) Normal tissue + Nanoshells(A) Normal tissue + PBS

200 mμ

Figure 10.2 Representative OCT images from normal skin and muscle tissue areas of mice systemi-cally injected with nanoshells (a) or with PBS (b). Representative OCT images from tumors of mice sys-temically injected with nanoshells (c) or with PBS (d). Analysis of all images shows a significantincrease in contrast intensity after nanoshell injection in the tumors of mice treated with nanoshellswhile no increase in intensity is observed in the normal tissue. The glass of the probe is 200-μm thickand shows as a dark nonscattering layer [4].

10.5.2 Evaluation of Antibody Concentration per Nanoshell

Anti-PSMA concentration on nanoshells were measured on several different days and onseveral different batches and consistently shows an average surface concentration of~170 antibody molecules per nanoshell, or 0.3 ± 0.02 pmol/cm2 (Figure 10.6). This mea-surement is consistent with values for other antibodies previously measured in our labincluding the concentration of anti-HER2 on nanoshells for breast cancer targeting.


The goal of this research is to develop minimally invasive systems for cancer therapy anddiagnosis. The use of NIR light and gold nanoparticle aids in the achievement of thisgoal as gold is relatively bioinert and tissue has few absorbers of NIR light. This combina-tion may provide the best opportunity for advancement of a nanotherapeutic particleplatform. Combined with the ability to extract diagnostic information from the same


163

0.00

0.25

0.50

0.75

1.00

PBS injected NS injected

*

Perc

ent

incr

ease

inco

ntra

st:

Tum

ors

vers

usno

rmal

tissu

e

**

Figure 10.3 Quantification of OCT images shows a significant increase in intensity of images oftumors from mice with systemic nanoshell injection. For PBS-only injection there is a 16% increase innormal tissue compared to tumor tissue scattered intensity, while for nanoshell-injected mice the dif-ference in normal compared to tumor tissue was an increase of 56% (*p<0.00002).

10 mμ10 mμ 10 mμ

(A) (B)

Figure 10.4 Silver enhancement staining of tissue shows heterogeneous staining of the tumor tissuefrom mice injected with nanoshells (a), indicating the presence of nanoshells. In contrast, there is littlesilver enhancement of sections taken from mice with PBS injection (b).

particle, the development of see-and-treat modalities can reduce lag in treatment andenhance outcome for many forms of cancer. Still there are a few areas where particularattention must be paid in order to avoid potential problems with this mode of therapySize and surface properties of the nanoparticle is an important consideration in develop-ment of treatment options. The larger the particles are to the upper limit of vasculardefects in the tumor bed, the lower the likelihood of having uniform distribution ofnanoparticles within the tumor; this may be significant in allowing for complete treat-ment of the tumor. Differences in tumor growth rate and heterogeneity in vascular leaksof tumors from one patient to another may require different accumulation times fortreatment. Nanotherapeutic options will need to be tailored specifically to patients as iscurrently being done with many types of cancer drug therapies. To accomplish moreuniform treatment, accurate imaging and quantification of nanoparticles in vivo willneed to be addressed.


164

(a)

Time post-treatment [days]0 12

0

20

40

60

80

Tum

orar

ea,m

m2

100

(b)

Time post-treatment [weeks]

0 1 2 3 4 5 6 70%

Perc

ent

surv

ivin

g

25%

50%

75%

100%

Figure 10.5 (a) Tumor size before irradiation and 12 days post-irradiation of mice treated withNanoshell + NIR laser irradiation (green bar); PBS Sham + NIR laser treatment (blue bar) or untreatedcontrol (red bar); values are average ±SEM. (b) Kaplan-Meier survival data for the treatment groups postirradiation; Nanoshell + NIR laser irradiation (solid green line); PBS Sham + NIR laser treatment(dashed blue line) or untreated control (red line); survival was followed for 7 weeks post-treatment.After 21 days, the nanoshell therapy group survival rate was significantly higher than either controlgroup; p<0.001.

The use of NIR light to image large sections of the body is increasing though its reso-lution has not yet achieved that of magnetic resonance imaging (MRI) or X-ray comput-erized tomography (CT). High-resolution optical systems such as OCT has penetrationlimits of 2 to 3 mm reducing the ability to locate deeply embed tumors with this imagingmode. This could be compensated by introducing MR contrast agents on the surface ofnanoshells or within the core itself. Nanoshells can be developed with magnetic cores orwith gadolinium doped polymer coatings that could allow MRI imaging capabilities. Inaddition to imaging with NIR light, the ability to get therapeutic doses of NIR lightenergy to heat nanoshells within deeply situated tumors is more difficult due to scatter-ing of the light by structures within the tissue. This could be compensated for by the useof diffuse fiber optic probes to penetrate the tumor to deliver the light at the cost ofincreasing complexity and invasiveness of the procedure.

10.7 Statistical Analysis

Various statistical analysis methods were used in different portions of testing to deter-mine differences. OCT Images were analyzed to first quantify the contrast levels usingstandard thresholding for image analysis then intensity data were analyzed using anunpaired student t-test assuming equal variance with a confidence interval of 95%,

α<0.05 of the two populations of images from PBS-treated and nanoshell-treated mice.Analysis of the tumor regression (Figure 10.5(a)) was performed using the average mea-surements of the tumor size of the surviving populations at the times shown and com-pared using an unpaired student t-test assuming equal variance with a confidence

interval of 95%, α <0.05. Kaplan-Meier statistical analysis was used to analyze survivaldata after treatment of nanoshell in the imaging and treatment study. A median survivalof 14 days for the PBS + Laser group and 10 days for the Untreated Control group wasobserved. By day 21, the survival of the Nanoshell + Laser group was significantly greaterthan either the control or sham groups, p<0.001 and this then continued for the dura-

10.7 Statistical Analysis

165

First measurement, t=0 Second measurement, t=3 weeks

Batch A Batch B

−

50

100

150

200

Mea

sure

dC

onc

ofA

bp

erN

S

Figure 10.6 Antibody concentration on two batches of nanoshells measured on different dates.

tion of the study. Median survival time could not be calculated for this group as the longterm survival was 83%.

Quantification of the image intensity of OCT images of normal tissue (n = 3) andtumor tissue (n = 6) (Figure 10.3) with PBS injections and nanoshell injections wasperformed with NIH imageJ software and analyzed using a student’s t-test of the twopopulations of images from PBS-treated and nanoshell-treated mice. The data showed asignificant increase in the optical contrast of tumor compared to normal tissue whennanoshells are used; p<0.00002. No statistical difference is observed in the intensity ofthe optical contrast of images of normal tissue whether nanoshells are used or PBS.


Problem Explanation Additional Indications Potential Solutions

Nanoshells peak resonance islower than desired

Shell too thick Seeds have good coverage ofgold colloid

Add less gold for reductionstep

Colloidal particles on seeds tolarge

Gold colloid was aged toolong, age for shorterperiod/and at lower tempera-ture

Colloidal particles on seeds tosparse

Incubate amine coated coreswith higher concentration ofgold colloid or increase lengthof incubation

Nanoshells not stable in salinesolution

Inadequate PEG coverage Surface charge measurementsindicate negative charges, notneutral

Add more PEG during incuba-tion with nanoshells

Nanoshells too old, repeatwith fresh batch of nanoshells

Antibody coverage low Antibody not bound tonanoshell

OPSS-PEG-NHS not reacting toantibody

Use fresh OPSS-PEG-NHS

PEG-Ab conjugate not bindingto nanoshells

PEG-SH added too early or intoo large a concentration,reduce incubation period orconcentration

PEG only nanoshells Abconcentration too high

The control is not adequatelycovered with PEG

Use new control samples

No accumulation in tumor Nanoshells not stable in vivo Check stability in salinesolution

See above

Acknowledgments

The author wishes to acknowledge his sincerest thanks to Dr Jennifer West for herencouragement in compiling this chapter.

References

[1] O’Neal, D. P., et al., “Photo-Thermal Tumor Ablation in Mice Using Near Infrared-AbsorbingNanoparticles,” Cancer Lett.,. 209(2): 2004, p. 171–176.

[2] Loo, C., et al., “Gold Nanoshell Bioconjugates for Molecular Imaging in Living Cells,” Opt. Lett.,30(9): 2005, pp. 1012–1014.

[3] Loo, C., et al., “Immunotargeted Nanoshells for Integrated Cancer Imaging and Therapy,” NanoLett., 5(4): 2005, pp. 709–711.


166

[4] Gobin, A. M., et al., “Near-Infrared Resonant Nanoshells for Combined Optical Imaging andPhotothermal Cancer Therapy,” Nano Lett, 7(7): 2007, pp. 1929–1934.

[5] Hirsch, L. R., et al., “Nanoshell-Mediated Near-Infrared Thermal Therapy of Tumors under Mag-netic Resonance Guidance,” Proc. Natl. Acad. Sci. U. S. A., 100(23): 2003, pp. 13549–13554.

[6] Oldenburg, S. J., et al., “Nanoengineering of Optical Resonances,” Ch. Phys. Lett., 288(2): 1998, pp.243–247.

Acknowledgments

167

C H A P T E R

1 1Environmentally Responsive MultifunctionalLiposomes

Amit A. Kale and Vladimir P. Torchilin*

*Corresponding author: Vladimir P. TorchilinCenter for Pharmaceutical Biotechnology and Nanomedicine, Northeastern University312 Mugar Hall, 360 Huntington Avenue, Boston, MA 02125e-mail: [email protected], Phone: 617-373-3206, Fax: 617-373-8886

169

Abstract

The liposomal drug carriers capable of spontaneous accumulation in pathologi-cal “acidic” sites via the enhanced permeability and retention (EPR) effectand further penetration and drug delivery inside the target cells via the actionof the cell-penetrating peptide (CPP), have been prepared in such a way thatliposomes simultaneously bear on their surface CPP (TATp) moieties and pro-tective PEG chains. PEG chains were incorporated into the liposome membranevia the PEG-attached phosphatidylethanolamine (PE) residue with PEGand PE being conjugated with the lowered pH-degradable hydrazone bond(PEG-HZ-PE). Under normal conditions, liposome-grafted PEG shieldedliposome-attached TATp moieties since the PEG spacer for TATp attachment(PEG(1000)) was shorter than protective PEG(2000). PEGylated liposomes areexpected to accumulate in targets via the EPR effect, but inside the acidifiedtumor or ischemic tissues lose their PEG coating due to the lowered pH-inducedhydrolysis of HZ and penetrate inside cells via the now-exposed TATp moieties.This concept is shown here to work in cell cultures in vitro as well as in tumorsin experimental mice in vivo. These nanocarriers also showed enhanced pGFPtransfection efficiency upon intratumoral administration in mice, compared tocontrol pH nonsensitive counterpart. These results can be considered as animportant step in the development of tumor-specific stimuli-sensitive drug andgene delivery systems.

Key terms pH-sensitive liposomes, cell penetrating peptide, TATp,hydrazone, PEG-PE, enhanced permeability and retention(EPR)

11.1 Introduction

The development of an optimal drug delivery systems has the utmost importance incontemporary medicine [1]. The ultimate goal in drug delivery is to achieve therapeuticconcentrations of the drug at the target site while drug concentrations at other tissuesare kept in safe levels. This has a particular importance in case of cancer treatment,where the challenge is to selectively destroy the tumor without damaging normal tis-sues. Disease (tumor) site-specific targeting of drugs and drug carriers may, at least par-tially, solve the problem. However, the issue remains how to achieve fast and effectivedrug release from the pharmaceutical carrier when it has already reached its target, suchas tumor [2–4]. This issue is equally important when long-circulating PEGylated drugdelivery systems are used [5–7], since PEG prevents normal interaction of the carrierwith cells and other destabilizing factors, or when drug carrier is intended for theintracellular penetration [8, 9] and a properly scheduled cytoplasmic release of the activedrug is expected to prevent its degradation in lysosomes [10].

There are several approaches to this problem including the use of stimuli-sensitivepharmaceutical nanocarriers, which is based on the fact that many pathological sitesincluding tumors demonstrate hyperthermia or acidification [11–13]. In general, envi-ronmentally sensitive carriers exhibit dramatic changes in their swelling behavior, net-work structure, permeability, or stability in response to changes in the pH or ionicstrength of the surrounding fluid or temperature [14].

Researchers working in the area of development of environment-responsivedrug delivery systems have architectured numerous carriers or conjugate systems toselectively deliver actives to pathological sites. Kataoka’s group has prepareddoxorubicin-physically loaded poly(beta-benzyl-L-aspartate) copolymer micelles andevaluated their pharmaceutical properties and biological significance [15]. AcceleratedDOX release was observed after lowering the surrounding pH from 7.4 to 5.0, suggestinga pH-sensitive release of DOX from the micelles. DOX loaded in the micelle showed aconsiderably higher antitumor activity compared to free DOX against mouse C26 tumorby i.v. injection, indicating a promising feature for PEG-PBLA pH-sensitive micelle as along-circulating carrier system useful in modulated drug delivery.

Hydrophobically-modified copolymers of N-isopropylacrylamide bearing a pH-sen-sitive moiety were investigated for the preparation of pH-responsive liposomes andpolymeric micelles [16]. The copolymers having the hydrophobic anchor randomly dis-tributed within the polymeric chain were found to more efficiently destabilize eggphosphatidylcholine (EPC)/cholesterol liposomes than the alkyl terminated polymers.Release of both a highly water soluble fluorescent contents marker, pyranine, andan amphipathic cytotoxic anticancer drug, doxorubicin, from copolymer-modifiedliposomes was shown to be dependent on pH. Also, polymeric micelles were studied as adelivery system for the photosensitizer aluminum chloride phthalocyanine, (AlClPc),currently evaluated in photodynamic therapy. pH-responsive polymeric micelles loadedwith AlClPc were found to exhibit increased cytotoxicity against EMT-6 mouse mam-mary cells in vitro than the control Cremophor EL formulation [17, 18]. Cremophor ELis a solubilizer used for solubilization of poorly soluble active-AlClPc. Drug carriers con-taining weak acids or bases can promote cytosolic delivery of macromolecules byexploiting the acidic pH of the endosome. Asokan et al. have prepared two pH-sensitivemono-stearoyl derivatives of morpholine, one with a (2-hydroxy) propylene (ML1)linker and the other, an ethylene (ML2) linker. The pK(a) values of lipids ML1 and ML2,

Environmentally Responsive Multifunctional Liposomes

170

when incorporated into liposomes, are 6.12 and 5.91, respectively. Both lipids disrupthuman erythrocytes at pH equal to or below their pK(a) but show no such activity atpH 7.4. This group has also synthesized two Gemini surfactants or bis-detergentsby cross-linking the headgroups of single-tailed, tertiary amine detergents throughoxyethylene (BD1) or acid-labile acetal (BD2) moieties [19]. As evidenced by thin-layerchromatography, BD2 was hydrolyzed under acidic conditions (pH 5.0) with an approx-imate half-life of 3 hours at 37°C, while BD1 remained stable. Low pH-induced collapseof liposomes containing acid-labile BD2 into micelles was more facile than that of BD1.With BD1, the process appeared to be reversible in that aggregation of micelles wasobserved at basic pH. The irreversible lamellar-to-micellar transition observed withBD2-containing liposomes can possibly be attributed to acid-catalyzed hydrolysis ofthe acetal cross-linker, which generates two detergent monomers within the bilayer.Liposomes composed of 75 mol % bis-detergent and 25 mol % phosphatidylcholinewere readily prepared and could entrap macromolecules such as polyanionic dextran ofMW 40 kDa with moderate efficiency. The ability of BD2-containing liposomes to pro-mote efficient cytosolic delivery of antisense oligonucleotides was confirmed by theirdiffuse intracellular distribution seen in fluorescence micrographs, and the upregulationof luciferase in an antisense functional assay. Bae et al. formulated pH-sensitive poly-meric mixed micelles composed of poly(L-histidine) (polyHis; M(w) 5000)/PEG (M(n)2000) and poly(L-lactic acid) (PLLA) (M(n) 3000)/PEG (M(n) 2000) block copolymerswith or without folate conjugation [20, 21]. The polyHis/PEG micelles showed acceler-ated adriamycin release as the pH decreased from 8.0. In order to tailor the triggering pHof the polymeric micelles to the more acidic extracellular pH of tumors, while improvingthe micelle stability at pH 7.4, the PLLA/PEG block copolymer was blended withpolyHis/PEG to form mixed micelles. Blending shifted the triggering pH to a lowervalue. Depending on the amount of PLLA/PEG, the mixed micelles were destabilized inthe pH range of 7.2 to 6.6 (triggering pH for adriamycin release). When the mixedmicelles were conjugated with folic acid, the in vitro results demonstrated that themicelles were more effective in tumor cell kill due to accelerated drug release and folatereceptor-mediated tumor uptake. In addition, after internalization polyHis was found tobe effective for cytosolic ADR delivery by virtue of fusogenic activity.

Certain pH-sensitive linkages have been popularly used to allow the drug release,protective “coat” removal, or new function appearance because of their fast degradationin acidified pathological sites [22–24]. These include cis-aconityls [25, 26], electron-richtrityls [27], polyketals [28], acetals [29, 30], vinyl ethers [31, 32], hydrazones [33–35],poly(ortho-esters) [36], and thiopropionates [37]. Such constructs may turn out to beuseful for the site-specific delivery of drugs at the tumor sites [12], infarcts [38], inflam-mation zones [39] or cell cytoplasm or endosomes [40], since at these “acidic” sites, pHdrops from the normal physiologic value of pH 7.4 to pH 6.0 and below.

11.1.1 Cis-Aconityl Linkage

A pH-sensitive cis-aconityl linkage has been used to make immunoconjugates ofdaunorubicin by Shen et al. [41] and Diener et al. [42] while doxorubicin was conjugatedto murine monoclonal antibodies (MoAb) raised against human breast tumor cells[43] or murine monoclonal antibody (MAb) developed against human pulmonaryadenocarcinoma [44].

11.1 Introduction

171

11.1.2 Trityl Linkage

The trityl group has been used in organic chemistry as an acid-cleavable protectinggroup for amino and hydroxyl groups. Patel group at Lilly Research Laboratories haveestablished structure-stability relationship of different trityl-nucleoside derivatives byusing NMR-spectroscopy [45]. In general, the acid-sensitivity of these compoundsincreases with the electron-donating effects of the substituents (e.g., methoxy groups)that stabilize the intermediatory formed carbocation in the hydrolysis step. In vitroactivity in a human colon carcinoma cell line showed that the antibody conjugates withthe most pronounced acid lability exhibited the strongest inhibitory effects. However,the most stable conjugates were 20 to 30 times less active than the free nucleosideantimetabolite [46, 47]. These structure-activity relationships also confirmed in animalexperiments [45].

11.1.3 Acetal Linkage

Acetals have the potential to be used as linkages for a range of alcohol functionalities,because their hydrolysis is generally first-order relative to the hydronium ion, makingthe expected rate of hydrolysis 10 times faster with each unit of pH decrease and by alter-ing their chemical structure, it is possible to tune their hydrolysis rate. In addition,acetals can be formed using a variety of types of hydroxyl groups including primary, sec-ondary, tertiary and syn-1,2- and -1,3-diols, and the rate of hydrolysis can be tuned byvarying the structure of the acetal. Gillies et al. synthesized a four different acetal-basedconjugates using model drugs and PEO polymer [48]. The hydrolysis kinetics of the con-jugates had half-lives ranging from less than 1 minute to several days at pH 5.0, withslower hydrolysis at pH 7.4 in all cases. Encrypted polymers containing pH-sensitiveacetal linkage between either oligonucleotide or macromolecule and PEG showed directvesicular escape and efficiently deliver oligonucleotides and macromolecules into thecytoplasm of hepatocytes [49]. Acetal-based acid-degradable protein-loaded microgelsalso have showed promising results for delivery of protein-based vaccines [50].

11.1.4 Polyketal Linkage

Murthy group has introduced an acid-sensitive hydrophobic nanoparticle based on anew polymer, poly(1,4-phenyleneacetone dimethylene ketal) (PPADK), which comple-ments existing biodegradable nanoparticle technologies [51]. This polymer has ketallinkages in its backbone and degrades via acid-catalyzed hydrolysis into low molecularweight compounds that can be easily excreted. PPADK forms micro- and nanoparticles,via an emulsion procedure, and can be used for the delivery of hydrophobic drugs andpotentially proteins [52].

11.1.5 Vinyl Ether Linkage

Acid-labile polyethylene glycol (PEG) conjugated vinyl ether lipids were synthesizedand used at low molar ratios to stabilize the nonlamellar, highly fusogeniclipid, dioleoylphosphatidyl ethanolamine, as unilamellar liposomes [32]. Acid-catalyzedhydrolysis of the vinyl ether bond destabilized these liposomes by removal ofthe sterically-stabilizing PEG layer, thereby promoting contents release on the hours


172

timescale at pH < 5. pH-Sensitive amphiphilic hydrogels were synthesized by radiationcopolymerization of ethylene glycol vinyl ether (EGVE), butyl vinyl ether (BVE) andacrylic acid (AA) in the presence of the crosslinking agent, diethylene glycol divinylether (DEGDVE) [53, 54]. The results of the swelling experiments indicated that thehydrogel, which has 60:40:5 comonomer ratio (mol% of EGVE:BVE:AA in monomericmixture), is pH-sensitive. While the hydrogel is in a fully hydrated form at pH>6, itextensively dehydrates below pH 6. A two-stage volume phase transition was observedin the range of pH 6.0 to 7.0 and 7.5 to 8.0.

11.1.6 Hydrazone Linkage

In 1980, Hurwitz and coworkers reported for the first time that hydrazone-based poly-mer-daunorubicin conjugates have substantial cytotoxicity than the analogues contain-ing noncleavable linkers between those conjugates that appeared to be completelyinactive [55]. In 1989, the Lilly labs reported the use of hydrazone linkages to targetmonoclonal antibodies to potent cytotoxic DAVLB hydrazide [56]. In vivo studies ofantitumor activity showed that the efficiency and safety of the conjugate was increasedover that of the unconjugated. The Kratz group has prepared trasnferin and albumin ascarriers for targeting of chlorambusil, an anticancer active [57, 58]. In vitro studies withboth conjugates demonstrated them to be as active or more active than the free drug,whereas they had reduced toxicities.

11.1.7 Poly(Ortho-Esters)

Toncheva et al. have prepared amphiphilic AB and ABA block copolymers from poly(ortho-esters) and poly (ethylene glycol). The micelles formed by these coblock poly-mers were stable in PBS at pH 7.4 and 37°C for 3 days and in a citrate buffer at pH 5.5 and37°C for 2 hours [36].

11.1.8 Thiopropionates

The remarkably enhanced gene silencing in hepatoma cells was achieved by assemblinglactosylated-PEG-siRNA conjugates bearing acid-labile beta-thiopropionate linkagesinto polyion complex (PIC) micelles through the mixing with poly(l-lysine) [59]. ThePIC micelles with clustered lactose moieties on the periphery were successfully trans-ported into hepatoma cells in a receptor-mediated manner, releasing hundreds ofactive siRNA molecules into the cellular interior responding to the pH decrease in theendosomal compartment. Eventually, almost 100 times enhancement in gene silencingactivity compared to that of the free conjugate was achieved for the micelle system, facil-itating the practical utility of siRNA therapeutics. Kataoka’s group [60] also architecturedthree types of newly engineered block copolymers forming polyplex micelles useful foroligonucleotides and siRNA delivery: (1) PEG-polycation diblock copolymers possessingdiamine side-chain with distinctive pKa for siRNA encapsulation into polyplex micelleswith high endosomal escaping ability, (2) lactosylated PEG-(oligonucleotide or siRNA)conjugate through acid-labile beta-thiopropionate linkage to construct pH-sensitive PICmicelles, and (3) PEG-poly(methacrylic acid) block copolymer for the construction oforganic/inorganic hybrid nanoparticles encapsulating siRNA.

11.1 Introduction

173

Recently, NEBI linkers were introduced as potential pH-sensitive linkages. Kineticanalysis of eight derivatives of N-ethoxybenzylimidazoles (NEBIs) showed that theirrates of hydrolysis are accelerated in mild aqueous acidic solutions compared to in solu-tions at normal, physiological pH. A derivative of NEBI carrying doxorubicin, a widelyused anticancer agent, also showed an increased rate of hydrolysis under mild acid com-pared to that at normal physiological pH. The doxorubicin analogue resulting fromhydrolysis from the NEBI exhibited good cytotoxic activity when exposed to humanovarian cancer cells [61].

We have demonstrated the utility of highly pH-sensitive hydrazone bond-basedPEG-PE conjugates in preparing double-targeted stimuli-sensitive pharmaceuticalnanocarriers [62, 63]. Two important temporal characteristics of such carriers includetheir sufficiently long lifetime under normal physiological conditions and theirsufficiently fast destabilization within the acidic target. Since real practical tasks mayrequire different times for such carriers to stay in the blood and to release theircontents (or “develop” an additional function) inside the target, we have synthesized aseries of PEG-HZ-PE conjugates with different substituents at the hydrazone bondand evaluated their hydrolytic stability at normal and slightly acidic pH values. Theseconjugates differed from each other with respect to the exact structure of groupsforming the hydrazone linkage between phospholipid and PEG. The characterizationof the in vitro behavior of these conjugates has provided important information use-ful for future design and development of pH-sensitive nanocarriers with controlledproperties.

11.2 Materials

11.2.1 Chemicals

1,2-dioleoyl-sn-glycero-3-phosphoethanolamine, DOPE; 1,2-dipalmitoyl-sn-glycero-3-phosphothioethanolamine (sodium salt), DPPE-SH; 1,2-dimyristoyl-sn-glycero-3-phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (ammonium salt), (Rh-PE),Egg phosphatidylcholine (egg PC), cholesterol (Ch), mPEG2000-DSPE and1,2-dioleoyl-3-trimethylammonium-propane (DOTAP) were purchased from AvantiPolar Lipids, (Alabaster, AL); (N-e-maleimidocaproic acid) hydrazide, EMCH;4-(4-N-maleimidophenyl) butyric acid hydrazide hydrochloride, MPBH;N-(k-maleimidoundecanoic acid) hydrazide, KMUH; succinimidyl4-(N-maleimidomethyl) cyclohexane-1-carboxylate, SMCC were purchased from PierceBiotechnology Inc. (Rockford, IL). 2-acetamido-4-mecrcapto butanoic acid hydrazide,AMBH was purchased from Molecular Probes (Invitrogen, Carlsbad, CA); methoxypoly(ethylene) glycol butyraldehyde (MW 2000), and mPEG-SH (MW 2000) were pur-chased from Nektar Therapeutics (Huntsville, AL). Triethylamine was purchased fromAldrich Chemicals. 4-succinimidyl formylbenzoate (SFB) was purchased from Molbio(Boulder, CO). Maleimide-PEG1000-NHS was purchased from Quanta Biodesign (Powell,OH); TATp-cysteine from Research Genetics (Huntsville, AL). Succinimidyl4-(N-maleimidomethyl)cyclohexane-1-carboxylate hydrazide (SMCCHz) was purchasedfrom Molecular Biosciences (Boulder, CO). 4-acetyl phenyl maleimide, Sephadex G25m,


174

and Sepharose CL4B were purchased from Sigma-Aldrich. All solvents were purchasedfrom Fisher Scientific (HPLC grade) and used without further purification.

Lewis Lung Carcinoma (LLC) cell line was purchased from ATCC (Rockville, MD).Delbecco’s minimal essential medium, complete serum free medium and fetal bovineserum were purchased from Cellgro (Kansas City, MO).

11.2.2 Syntheses

All reactions were monitored by TLC using 0.25 mm × 7.5 cm silica plates with UV-indicator (Merck 60F-254), and mobile phase chloroform:methanol (80:20% v/v).Phospholipid and PEG alone or their conjugates were visualized by phosphomolybdicacid and Dragendorff spray reagents. Silica gel (240–360 μm) and size exclusion media,Sepharose CL4B (40–165 μm) and Sephadex G25m (Sigma-Aldrich) were used for silicacolumn chromatography and size exclusion chromatography, respectively.

11.2.3 Preparation of the TATp-Bearing, Rhodamine-Labeled LiposomalFormulations

The pH-sensitive or pH-insensitive, Rh-labeled, TATp-bearing liposomes were preparedby the lipid film hydration method. A mixture of PC:Chol (7:3), TATp-PEG1000-PE, Rh-PEand either mPEG2000-HZ-PE (pH-sensitive) or mPEG2000-DSPE (pH-insensitive) at molarratio 10:0.25:0.1:15 was evaporated under reduced pressure. The dry lipid formed washydrated with phosphate buffer saline, pH 7.4. The liposomal suspension was filteredthrough 0.2-µm polycarbonate filters and stored at 4°C until use. The liposome particlemean size and size distribution were observed using a Coulter N4 Plus submicronparticle analyzer.

11.2.4 Preparation of the TAtp-Bearing, Rhodamine Labeled, pGFPComplexed Liposomal Formulations

The pH-sensitive or pH-insensitive, TATp-bearing pGFP-complexed liposomes wereprepared by the spontaneous vesicle formation (SVF) method adopted from [64] withfew modifications. A plasmid solution was prepared by combining pGFP and 10-mMTris EDTA (TE) buffer, pH 7.4. A lipid solution in ethanol was prepared by dissolvingegg PC:Chol (7:3) in anhydrous ethanol, and then adding DOTAP, TATp-PEG1000-PE,and either mPEG2000-HZ-PE (22, pH-sensitive) or mPEG2000-DSPE (pH-insensitive) at10:0.25:15 molar ratio. The charge (+/-) ratio was 10:1. The lipid and plasmid solutionswere preheated to 37°C before mixing together. After mixing these solutions for 10minutes, ethanol was evaporated under the reduced pressure. The samples were filteredthrough 0.2-μm polycarbonate filters and stored at 3C until use. The liposomalformulations were subjected to the agarose gel electrophoresis to test for the quantita-tive presence and intactness of the plasmid within the liposomes [65]. In a typicalcase, the pGFP concentration was 3.22-μg/mg of total lipid. The liposome particlemean size and size distribution were observed using a Coulter N4 Plus submicronparticle analyzer.

11.2 Materials

175

11.3 Methods

11.3.1 Synthesis of Hydrazone-Based mPEG-HZ-PE Conjugates [63, 66]

11.3.1.1 Synthesis of Aliphatic Aldehyde-Derived Hydrazone-Based mPEG-HZ-PEConjugates

Step 1: Synthesis of Hydrazide-Activated Phospholipids22 μmoles of phosphatidylthioethanolamine, 2, were mixed with 1.5 molar excess ofeach acyl hydrazide linker (Table 11.1) in 3-mL anhydrous methanol containing 5 molarexcess of triethylamine over lipid (Scheme 11.1). The reaction was performed at 25°Cunder argon for 8 hours. Solvent was removed under reduced pressure, and the residuewas dissolved in chloroform and applied to a 5-mL silica gel column that had been acti-vated (150°C overnight) and prewashed with 20 mL of chloroform. The column wasequilibrated with an additional 15 mL of chloroform followed by 5 mL of each of thefollowing chloroform:methanol mixtures 4:0.25, 4:0.5, 4:0.75, 4:1, 4:2, and finallywith 6 mL of 4:3 v/v. The phosphate-containing fractions eluting in 4:l, 4:2, and 4:3chloroform:methanol (v/v) were pooled and concentrated under reduced pressure. Theproduct was stored in glass ampoules as chloroform solution under argon at –80°C.

For the activation of phospholipid with AMBH, a maleimide derivative ofphosphatidylethanolamine, 7, was prepared using SMCC (Scheme 11.3). In brief,phosphatidylethanolamine, 6, in chloroform was reacted with 1.5 molar excess ofSMCC, 5, over lipid in presence of 5 molar excess of TEA under argon for 5 hours. Themaleimide-derivative was separated from excess SMCC on silica gel column using chlo-roform:methanol (4:0.2 v/v) mobile phase. The elution fractions containingNinhydrin-negative and phosphorus-positive fractions were pooled and concentratedunder reduced pressure. DOPE-maleimide was further used to synthesize AMBH-acti-vated derivative of phospholipid, 8, by reacting with 1.5 molar excess of AMBH usingTEA as catalyst (Scheme 11.4).

Step 2: Synthesis of mPEG-HZ-PE ConjugatesTwenty-one μmoles of mPEG2000-butyraldehyde were reacted with 14 µmoles of linker-activated phospholipid in 2-ml chloroform at 25°C in a tightly closed reaction vessel


176

Table 11.1 List of Acyl Hydrazide Cross-Linkers

Linker Used Mol. Wt. Length of Spacer Arm

AMBH2-acetamido-4-mercaptobutanoic acid hydrazide

191.25 —

EMCH(N-e-maleimidocaproic acid)hydrazide

225.24 11.8 A

MPBH4-(4-N-maleimidophenyl)butyric acid hydrazide

309.5 17.9 A

KMUHN-(k-maleimido undecanoic acid)hydrazide

295.8 19.0 A

SMCCHSuccinimidyl 4-(N-maleimidomethyl)cyclohexane-1-carboxylate hydrazide

365.31 —

(Schemes 11.2 and 11.5). After an overnight stirring, chloroform was evaporated undervacuum in rotary evaporator. The excess mPEG2000-butyraldehyde was separated fromPEG-HZ-PE conjugates using gel filtration chromatography. The gel filtration chroma-tography was performed using sepharose-CL4B equilibrated overnight in pH 9–10degassed ultra pure water (elution medium) in 1.5 × 30 cm glass column. The thin filmformed in a round-bottom flask after evaporating chloroform was hydrated with the elu-tion medium and applied to the column. The micelles formed by PEG-HZ-PE conjugatewere the first to elute from the column. Micelle containing fractions were identified byDragendorff spray reagent and pooled together, kept in freezer at -80°C overnight beforesubjecting to freeze-drying. The freeze-dried PEG-HZ-PE conjugates were weighed andstored at –80°C as chloroform solution.

11.3.1.2 Synthesis of Aromatic Aldehyde-Derived Hydrazone-Based mPEG-HZ-PEConjugates

Step 1: Synthesis of Hydrazide-Activated PEG DerivativesForty μmoles of mPEG-SH in chloroform were mixed with two molar excess of acylhydrazide cross-linkers: EMCH (10a), MPBH (10b), KMUH (10c) presence of 5 molar

11.3 Methods

177

Scheme 1: Synthesis of acyl hydrazide-activated phospholipids

OO

OPO

OO

C15 H31

O

C15 H31

O

HS

O

ON

X

CNHH2N

O

ON

X

C

O

NHH2N +

X =

(3a)CH2 4

3

(3c)

(3b)

Na

1 2

O

OOPOOH

OC15 H31

O

C15 H31

O

S

CH2 9

CH2 2

O

Anhydrous methanol/TEA

KMUH

MPBH

EMCH

Scheme 11.1


178

Scheme 3: Maleimide activation of phosphatidylethanolamine

O

O

O

O

N

O

O

N

OO

O

P

O

OHO

C17 H33

O

C17 H33

O

H2N

+

OO

O

N

OO

O

P

O

OHO

C17 H33

O

C17 H33

O

NH

5 6

7

Anhydrous methanol/TEANHS

Scheme 11.3

Scheme 2: Synthesis of aliphatic aldehyde-based hydrazone-derived mPEG-HZ-PE

OO

On

H+

H2O

O

ON

X

CNHH2N

X =

(4a)CH2 4EMCH

3

(4c)

(4b)MPBH

KMUH

O

OOPOOH

OC15 H31

O

C15 H31

O

S

CH2 9

CH2 2

O

OO

n

O

ON

X

CNHN

O

OO

POOH

OC15 H31

O

C15 H31

O

S

O

4

Scheme 11.2

excess of triethylamine over lipid. The excess EMCH was separated from the product bysize exclusion chromatography using Sephadex G25m media. The acyl hydrazide deriva-tives of PEG, (11a), (11b), (11c) were freeze-dried and stored as chloroform solution at–80°C (see Scheme 11.6).

Step 2: Synthesis of Aromatic Aldehyde-Activated PhospholipidThirty-five μmoles of phosphatidylethanolamine, DOPE-NH2, 12, in chloroform weremixed with 2 molar excess of 4-succinimidylformyl benzoate, SFB, 13, in presence of

11.3 Methods

179

Scheme 5. Synthesis of PEG-HZ-PE conjugate using AMBH-activated phospholipid

O

O

O

n

H+

H2O

O NH

N

S

O

HN

OO

O

N

OO

O

P

O

OHO

C17 H33

O

C17 H33

O

NH

OO

n

O NH

NH2

S

O

HN

OO

O

N

OO

O

P

O

OHO

C17 H33

O

C17 H33

O

NH

8

9

Scheme 11.5

Scheme 4. AMBH-derivatized phospholipid via sulfhydryl-maleimide addition reaction

OO

O

N

O

O

O

P

O

OHO

C17 H33

O

C17 H33

O

NH

7

O NH

NH2

SH

O

HN

+

O NH

NH2

S

O

HN

OO

O

N

O

O

O

P

O

OHO

C17 H33

O

C17 H33

O

NH

8

Anhydrous methanol/TEA

Scheme 11.4

3 molar excess triethylamine over lipid (See Scheme 11.7). After stirring for 3 hours, sol-vent was evaporated, residue was redissolved in chloroform, and product was separatedon silica gel column using acetonitrile:methanol mobile phases: 4:0, 4:0.25, 4:0.5,4:0.75, and 4: 1 v/v. The fractions containing product were identified by TLC analysis,pooled, and concentrated. The product was stored as chloroform solution at –80°C.

Step 3: Synthesis of mPEG-HZ-PE ConjugatesA 1.5-molar excess of SFB activated phospholipid, 14, was reacted with acyl hydrazidederivatized PEGs, 11a, 11b, and 11c, respectively, in chloroform at room temperature(see Scheme 11.8). After overnight stirring, chloroform was evaporated under reducedpressure. The PEG-HZ-PE conjugate was purified using size exclusion chromatographyusing Sepharose CL4B as described before.

11.3.1.3 Synthesis of Aromatic Ketone-Derived Hydrazone-Based Mpeg-HZ-PEConjugates

Step 1: Synthesis of Hydrazide Derivative of PEGmPEG-SH (MW 2000), 16, was reacted with 2 molar excess of SMCCHz, 17, in presenceof triethylamine for 8 hours in dry chloroform (see Scheme 11.9). Chloroform was evap-


180

Scheme 6: Synthesis of acyl hydrazide activated PEG

OO n

SH O

NH

NH2

X

O

O

N+

OO n

S

X =

O

NH

NH2

X

O

O

N

(11a)CH2 5EMCH

(11c)

(11b)CH2 3

CH2 10

MPBH

KMUH

10

11

EMCH: 10aMPBH:10bKMUH: 10c

Chloroform/TEA, RT, Overnight

Mol wt ~2000

Scheme 11.6

11.3 Methods

181

Scheme 7: SFB activation of phosphatidylethanolamine

O

OO

POHO

ONH2

H33C17

O

C17H33O

+

O

O

O

O

O

N

12

13

O

OO

POHO

ONH

H33C17

O

C17HOC

O

O

14

Chloroform/TEA, RT, overnight

NHS

33

Scheme 11.7

Scheme 8: Synthesis of PEG-HZ-PE conjugate

OO n

SO

HNNH2

X

O

O

N+

H2O

OO n

SO

NH

N X

O

O

N

X =

CH2 5 EMCH (15a)

CH2 3

CH2 9

MPBH (15b)

KMUH (15c)

O

OO

POHO

ONH

H33C17

O

C17H33OC

O

O

O

OO

POHO

ONH

H33C17

O

C17H33OC

O

14(11a), (11b), (11c)

15

Scheme 11.8

orated, and the residue was dissolved in water. The PEG-hydrazide derivative, 18, wasseparated and purified by the size exclusion gel chromatography using Sephadex G25mmedia. The product was freeze-dried and stored as chloroform solution at –80°C.

Step 2: Activation of Phospholipid with 4-Acetyl Phenyl MaleimideForty μmoles of 4-acetyl phenyl maleimide, 19, were reacted with 27 μmoles ofphosphatidylthioethanol (DPPE-SH), 20, in presence of triethylamine overnight withconstant stirring under inert atmosphere of argon (see Scheme 11.9). The product, 21,was separated on a silica gel column using chloroform:methanol mobile phase (4:1 v/v).The fractions containing product were identified by TLC analysis, pooled, and concen-trated. Aromatic ketone-activated phospholipid was stored as chloroform solution at–80°C.


182

Scheme 9: Synthesis of aromatic ketone-derived hydrazone based mPEG-HZ-PE

N

O

O NH

NH2

O O n SH

Mol wt. ~ 2000

+

N

O

O NHNH2

O O n S

O

O

18

O

O

OO

OPO

OHO

C15 H31

O

C15 H31

O

SO

N

O

O

ON

O

OO

O

OPO

OHO

C15 H31

O

C15 H31

O

HS

+

21

+

N

O

O NH

N

O O nS

OO

OPO

OHO

C15 H31

O

C15 H31O

SN

O

O

H2O

O

16 17

19 20

2118

22

Step 3

Step 2

Step 1

Scheme 11.9

Step 3: Synthesis of mPEG-HZ-PE ConjugateHydrazide-activated PEG derivative, 18, was reacted with 1.5 molar excess of the aro-matic ketone-derivatized phospholipid, 21, overnight under the constant stirring atroom temperature (see Scheme 11.9). The PEG-HZ-PE conjugate, 22, was separated andpurified by size exclusion gel chromatography using Sepharose-CL4B media.

11.3.2 Synthesis of PE-PEG1000-TATp Conjugate [66]

Step 1: Synthesis of PE-PEG1000-MaleimideA 1.5 molar excess of DOPE-NH2, 23, was reacted with NHS-PEG1000-maleimide, 24, inchloroform under argon at room temperature in presence of 3 molar excesstriethylamine overnight with stirring (See Scheme 11.10). The productPE-PEG1000-maleimide, 25, was separated on the Sephadex G25m column equilibratedovernight with the degassed double deionized water. The product was freeze-dried andstored under chloroform at –80°C.

11.3 Methods

183

Scheme 10: Synthesis of PE-PEG-TATp conjugate

O

O

O

HN

O

NO O

N

O

OnO

OO

PO

HO

O

NH2

H33C17

O

C17H33O

+

O

OO

PO

HO

O NH

H33C17

O

C17H33O

O

HN

O

NO O

n

O

2324

25

TATp-SH

O

OO

PO

HO

O NH

H33C17

O

C17H33O

O

HN

O

NO O

n

O

S TATp

26

NHS

Chloroform/TEA

Scheme 11.10

Step 2: Synthesis of PE-PEG1000-TATpA 2-fold molar excess of TATp-SH was mixed with PE-PEG1000-maleimide, 25, in chloro-form under inert atmosphere with gentle shaking for 8 hours (See Scheme 11.10). Theexcess TATp-SH was separated from the product, 26, by gel filtration chromatographyusing Sephadex G25m media. The freeze-dried product was stored under chloroform at–80°C until further use.

11.3.3 In Vitro pH-Dependant Degradation of PEG-HZ-PE Conjugates

The time-dependant degradation of PEG-HZ-PE micelles incubated in buffer solutions(phosphate buffer saline, pH 7.4 and pH 5.0) maintained at 37°C was followed by HPLCusing Shodex KW-804 size exclusion column. The elution buffer used was pH 7.0, Phos-phate buffer (100 mM phosphate, 150 mM sodium sulfate), run at 1.0 ml/min. For fluo-rescent detection (Ex 550 nm/Em 590 nm) of micelle peak, Rh-PE (1 mol % of PEG-PE)was added to the PEG-PE conjugate in chloroform. A film was prepared by evaporatingthe chloroform under argon stream and hydrated with the phosphate buffer saline, pH7.4 or 5.0 (adjusted by precalculated quantity of 1N HCl). A peak that represents micellepopulation appeared at the retention time between 9 to 10 minutes. The degradationkinetics of micelles was assessed by following the area under micelle curve that repre-sents intact micellar population. Half-lives were calculated by noting the time at whichhalf of the initial (t = 0) micellar population existed [63].

11.3.4 Avidin-Biotin Affinity Chromatography

To check the pH-sensitivity, biotin containing micelles were formulated by mixingmPEG2000-HZ-PE (60 % mol), PEG750-PE (37 % mol), Rhodamine-PE (0.5 % mol, fluores-cent marker), biotin-PE (2.5 % mol, biotin component) in chloroform. Chloroform wasevaporated and a thin film was formed using rotary evaporator. To test the binding ofbiotin-bearing Rh-PE-labeled, TATp-bearing liposomes before and after incubation atlowered pH values, the corresponding samples were kept for 3 hours at pH 7.4 or pH 5.0and then applied onto the Immobilized NeutrAvidin protein column. The degree of theretention of the corresponding preparation on the column was estimated following thedecrease in the sample rhodamine fluorescence at 550/590 nm after passing through theNeutrAvidin column [67].

11.3.5 In Vitro Cell-Culture Study

H9C2 rat embryonic cardiomyocytes in 10% fetal bovine serum DMEM were grown oncoverslips in 6-well plates, then treated with various Rh-PE-labeled liposome samples(with and without preincubation for 3 hours at pH 5.0) in serum-free medium(2 mL/well, 30 mg total lipid/mL). After a 1-hour incubation period, the media wereremoved and the plates washed with serum-free medium three times. Individualcoverslips were mounted cell-side down onto fresh glass slides with PBS. Cells wereviewed with a Nikon Eclipse E400 microscope under bright light or underepifluorescence with rhodamine/TRITC filter [67]. The images were analyzed usingImageJ 1.34I software (NIH) for integrated density comparison of red fluorescencebetween two groups.


184

11.3.6 In Vivo Study

LLC tumors were grown in nu/nu mice (Charles River Breeding Laboratories, MA) by thes.c. injection of 8 x 104 LLC cells per mouse into the left flank (protocol # 05-1233R,approved by the Institutional Animal Care and Use Committee at Northeastern Univer-sity, Boston). When tumor reached 5 to 10 mm in diameter, they were injected at four tofive different spots with 150 µl of Rh-labeled, TATp-bearing pH-sensitive or pH-insensi-tive liposomes in phosphate-buffered saline, pH 7.4. Mice were killed 6 hours later bycervical dislocation, and excised tumors were cryo-fixed as described above. Microtomecut sections were washed thoroughly with phosphate buffer saline (pH 7.4), dried andfixed on slides using Fluor Mounting medium. These sections were observed under fluo-rescence microscopy using TRITC filter [68]. Further, the images were analyzed usingImageJ 1.34I software (NIH) for integrated density comparison of red fluorescencebetween pH-sensitive and pH-insensitive groups.

11.3.7 In Vivo Transfection with pGFP

LLC tumors were grown as described above. When tumor reached 5 to 10 mm in diame-ter, they were injected at four to five different spots with 150 μl of pGFP-loaded,TATp-bearing pH-sensitive or pH-insensitive liposomes in phosphate-buffered saline,pH 7.4. Mice were killed 72 hours later by cervical dislocation, and excised tumors werefixed in a 4% buffered paraformaldehyde overnight at 4°C, blotted dry of excessparaformaldehyde and kept in 20% sucrose in PBS overnight at 4°C. Cryofixation wasdone by the immersion of tissues in ice-cold isopentane for 3 minutes followed by freez-ing at –80°C. Fixed, frozen tumors were mounted in Tissue-Tek OCT 4583 compound(Sakura Finetek, Torrance, CA) and sectioned on a Microtome Plus (TBS). Sections weremounted on slides and analyzed by the fluorescence microscopy using FITC filter andwith hematoxylin-eosin staining. The images were analyzed using ImageJ 1.34I software(NIH) for integrated density comparison of green fluorescence between pH-sensitive andnon-pH-sensitive groups.


11.4.1 Synthesis of Hydrazone-Based mPEG-HZ-PE Conjugates

The success of hydrazones as pH-sensitive linkages derives from the fact that theirhydrolytic stability is governed by the nature of hydrazone bond formed. Hydrazonesare much more stable than imines as a result of the delocalization of the π–electrons inthe former. In fact, parent hydrazones are too stable for the application in drug deliverysystems, and an electron withdrawing group has to be introduced to moderate the stabil-ity by somewhat disfavoring electron delocalization throughout the molecule as com-pared to the parent hydrazone. Hydrazones can be prepared from aldehydes or ketonesand hydrazides under very mild conditions including aqueous solutions. Hydrazonebond formation can take place even in vivo from separate fragments that self-assembleunder physiological conditions [69].

We have applied different synthetic methods based on the use of various aldehydesthat can produce the hydrazone linkage between PEG and PE [63]. Synthesis of aliphaticaldehyde-derived hydrazone containing PEG-PE conjugate was pursued in two steps.


185

First, phospholipid was activated with four different acyl hydrazides. The sulfhydrylreactive group of phosphatidylthioethanolamine was reacted with maleimide end ofmaleimido acyl hydrazides through Michael addition, thus providing acyl hydrazideactivated PE. mPEG-butyraldehyde, an aliphatic aldehyde, was then reacted with acylhydrazide activated PE to get hydrazone based PEG-PE conjugate. To synthesizearomatic aldehyde-derived hydrazone, an aromatic aldehyde moiety was introducedinto the phospholipid by reacting succinimidyl 4-formylbenzoate (SFB) withphosphatidylethanolamine under mild alkaline conditions. The acyl hydrazide-PEGderivatives were synthesized using mPEG-SH and maleimido acyl hydrazides (EMCH,MPBH, and KMUH). The SFB-activated phospholipid was then reacted with acylhydrazide derivatized PEG. Aromatic ketone-derived hydrazone-based PEG-PE conju-gates were synthesized by reacting aromatic ketone-activated phospholipids with acylhydrazide-activated PEG [66].

11.4.2 Synthesis of PE-PEG1000-TATp Conjugate

TATp-SH was attached to the heterobifunctional PEG via the two step synthesis asshown in Scheme 11.10. First, Mal-PEG-PE conjugate was synthesized by reactingDOPE-NH2 with the NHS end of heterobifunctional PEG derivative, 23,NHS-PEG1000-maleimide. PE-PEG1000-maleimide was then reacted with TATp-SH to formPE-PEG1000-TATp conjugate. The conjugate was separated by gel chromatography usingthe Sephadex G25m media.

The time required for each synthetic scheme and corresponding practical yield areshown in Table 11.2.

11.4.3 In Vitro pH-Dependant Degradation of PEG-HZ-PE Conjugates

All PEG-HZ-PE derivatives spontaneously form micelle in aqueous surroundings [70].The stability of hydrazone-based PEG-PE conjugates incubated at physiological pH 7.4and acidic pH 5.0 in buffer solutions maintained at 37°C was investigated by HPLC. Forthis purpose, the area under the micelle peak of PEG-HZ-PE (Rt 9–10 min) was observedover a period of time. PEG-HZ-PE conjugates derived from an aliphatic aldehyde and dif-ferent acyl hydrazides were found to be highly unstable under acidic conditions, withthe micelle peak was completely disappearing within 2 minutes of incubation at pH 5.0.At the same time, these conjugates were relatively stable at physiological pH: thePEG-HZ-PE conjugate, 9, with AMBH as cross-linker showed the half-life of 150 minutesfollowed by EMCH, 4a, (120 min), MPBH, 4b, (90 min), and KMUH, 4c, (20min) (Table 11.2). The rate of hydrolysis among the aliphatic aldehyde-derivedhydrazone-based PEG-PE conjugates (4a, 4b, 4c, and 9) at pH 7.4 seems to be dependanton carbon chain length of acyl hydrazide. The increase in number of carbon atoms inacyl hydrazide led to increase in rate of hydrolysis (PEG-PE conjugate 4c, acyl hydrazidewith 10-C atoms > 4a, acyl hydrazide with 5-C atoms > 9, acyl hydrazide with 3-Catoms). Introducing an aromatic character within carbon chain of acyl hydrazide led toincrease in hydrolysis as observed in case of 4b and 4a. (rate of hydrolysis of 4b > 4a).

Alternatively, the PEG-HZ-PE conjugates derived from an aromatic aldehyde andacyl hydrazides were found to be highly stable at pH 7.4 and 5.0 (Table 11.3). Thehalf-life values were not attained at either of those pH values even at the end of incuba-tion period of 72 hours in pH 7.4 and 48 hours in pH 5.0 buffer solutions maintained at


186

37°C. The resistance to hydrolysis exhibited by hydrazones derived from aromatic alde-hydes can be attributed to the conjugation of the π bonds of –C=N- bond of thehydrazone with the π bonding benzene ring. Thus, it supports the finding that hydra-


187

Table 11.2 Time Required and % Practical Yield of Synthetic Schemes

Scheme # and Description Product % YieldApprox. Total TimeRequired for Each Step

Scheme 1: Synthesis of acylhydrazide-activated phospholipids

3a3b3c

65.771.968.0

12 hr

Scheme 2: Synthesis of aliphaticaldehyde-based hydrazone-derivedmPEG-HZ-PE

4a4b4c

55.157.453.5

24 hr

Scheme 3: Maleimide activation ofphosphatidylethanolamine

7 65.7 12 hr

Scheme 4: AMBH-derivatizedphospholipid viasulfhydryl-maleimide addition reac-tion

8 70.1 12 hr

Scheme 5: Synthesis of PEG-HZ-PEconjugate using AMBH-activatedphospholipid

9 61.0 24 hr

Scheme 6: Synthesis of acylhydrazide activated PEG

11a11b11c

80.380.584.0

24 hr

Scheme 7: SFB activation ofphosphatidylethanolamine

14 73.0 12 hr

Scheme 8: Synthesis of PEG-HZ-PEconjugate

15a15b15c

57.864.562.0

24 hr

Scheme 9: Synthesis of aromaticketone-derived hydrazone basedmPEG-HZ-PE

182122

78.565.556.2

36 hr

Scheme 10: Synthesis ofPE-PEG-TATp conjugate

2526

62.551.8

48 hr

Table 11.3 Half-lives of Different Hydrazone-Based mPEG-HZ-PE Conju-gates Incubated in Phosphate Buffered Saline, pH 7.4 and pH 5.0 at 37°Cover a Period of Time,

mPEG-HZ-PEConjugate

Half-Life ( )pH 7.4 pH 5.0

4a 2 < 0.034b 1.5 < 0.034c 0.33 < 0.039 2.5 < 0.0315a > 72 > 4815b > 72 > 4815c > 72 > 4822 40 2.0

zones formed from aromatic aldehydes are more stable to acidic hydrolysis than thoseformed from aliphatic ones [71, 72]. The hydrazone hydrolysis involves the protonationof the –C=N nitrogen followed by the nucleophilic attack of water and cleavage of C-Nbond of tetrahedran intermediate [73]. Any of these steps is determining and dependanton the pH. The substituents on the carbonyl reaction partner influence the rate ofhydrolysis through altering the pKa of the hydrazone with electron donating substitu-ents facilitating protonation of the –C=N nitrogen [74].

This would support the fact that PEG-HZ-PE conjugates containing hydrazone bondderived from the aliphatic aldehyde are more prone to hydrolytic degradation. Aromaticaldehyde-derived hydrazone bond is too stable for the purpose of pH-triggered drugrelease. Careful selection of an aldehyde and an acyl hydrazide would be necessary forthe application of the hydrazone-based chemistry for the development of pH-sensitivepharmaceutical nanocarriers.

As Scheme 11.9 shows, an aromatic ketone-derived hydrazone bond was introducedbetween PEG and PE. The presence of a methyl group (electron donating) on the car-bonyl functional group would provide a sufficient lability of the hydrazone bond undermildly acidic conditions while an immediate aromatic ring (electron withdrawing) nextto the hydrazone bond would offer the stability under acidic and neutral conditions.mPEG-HZ-PE conjugate, wherein the hydrazone bond is derived from an aromaticketone, exhibited the half-lives of 2-to-3h at slightly acidic pH values, and much higherstability (up to 40 h) at the physiological pH (Table 11.3).

11.4.4 Avidin-Biotin Affinity Chromatography

To determine the pH-sensitivity of mPEG-HZ-PE conjugates, biotin-embedded micellesshielded by cleavable mPEG2000-HZ-PE, were eluted through avidin immobilized gelmedia columns. The control micelle formulation (incubated at pH 7.4 at 37°C for 3h)showed only a minimal biotin binding against 69% biotin binding of test micelle formu-lation (incubated at pH 5.0 at 37°C for 3 h), Figure 11.1. This proves shielding effect ofmPEG2000-HZ-PE conjugate under physiological pH condition and deshielding afterexposure to acidic environment.

11.4.5 In Vitro Cell Culture Study

To study shielding/de-shielding effect of mPEG-HZ-PE under the influence of acidic pH,internalization of Rh-labeled, TATp-bearing, mPEG-HZ-PE shielded liposomes pre-incu-bated at pH 7.4 and pH 5.0 was followed using H9C2 cells. As seen in Figure 11.2(a) and(b), Rh-labeled TATp-bearing, pH-sensitive liposomes incubated at pH 5.0 showed 2.5times (ImageJ 1.34I data) more internalization than when incubated at pH 7.4 becauseof better accessibility of TATp for its action after detachment of pH-sensitive PEG coronafrom liposomal surface under the influence of acidic pH.

11.4.6 In Vivo Study

Trying to cover different physiological conditions, we attempted intratumoral injec-tions of Rh-labeled, TATp-bearing pH-sensitive or pH-insensitive liposomes into LLCtumor bearing mice. An acidic pH at the tumor site is a well-known fact that is of interestwhile developing physiology-based targeted delivery systems. Under the fluorescence


188

microscope with TRITC filter, samples prepared 6 hours post-injection from tumorsinjected with TATp-bearing, Rh-labeled, pH-sensitive liposomes demonstrated intensiveand bright red fluorescence which was four times (as per ImageJ 1.34I data) more thanthat observed in the samples obtained from the tumors injected with TATp-bearing,Rh-labeled, pH-insensitive liposomes (Figure 11.3(a) and (b)).

11.4.7 In Vivo pGFP Transfection Experiment

We attempted a localized transfection of tumor cells by the direct intratumoral adminis-tration of sterically shielded with pH-sensitive (containing mPEG-HZ-PE, 25) orpH-insensitive (containing mPEG-DSPE) conjugates TATp-liposome-pGFP complexesinto the tumor tissue by the intratumoral injections. Histologically,hematoxylin/eosin-stained tumor slices in animals injected with both preparations


189

0

10

20

30

40

50

60

70

80

pH 5.0 pH 7.4

Perc

ent

Biot

inbo

und

Figure 11.1 Binding of pH-sensitive biotin-micelles to NeutrAvidin columns after incubation atroom temperature at pH 5.0 and 7.4

(a) (b)

Figure 11.2 Fluorescence microscopy showing internalization of Rh-PE-labeled/TATp/pH-sensitiveliposomes by H9C2 cells after incubation at pH 7.4 (12.2a) and pH 5.0 (12.2b)

showed the identical typical pattern of poorly differentiated carcinoma (polymorphiccells with basophilic nuclei forming nests and sheets and containing multiple sites ofneoangiogenesis; Figure 11.4(a) and (b)). However, under the fluorescence microscopewith FITC filter, samples prepared 72 hours postinjection from tumors injected withpH-sensitive PEG-TATp-liposome-pGFP complexes demonstrated intensive and brightgreen fluorescence compared to only minimal GFP fluorescence observed in the samplesobtained from the tumors injected with pH-insensitive PEG-TATp-liposome-pGFPcomplexes (Figure 11.5(a) and (b)).

The enhanced pGFP transfection by using pH-sensitive PEG-TATp-liposome-pGFPcomplexes is an ultimate result of the removal of mPEG-HZ-PE coat under the decreasedpH of the tumor tissue, and better accessibility of deshielded TATp moieties inTATp-liposome-pGFP complexes for internalization by the cancer cells allowing for theincreased interactions of pGFP with cancer cell nuclei.

Owing to their physicochemical properties, the long-circulating (PEGylated)liposomal carriers have the ability to accumulate inside the tumor tissue via the EPReffect, without further escape into undesired nontarget sites. The pH at tumor sites isacidic [12, 13]. Therefore, when TATp-pGFP-liposomes with an additional pH-sensitive


190

(a) (b)

Figure 11.3 TRITC image of frozen tissue section treated with intratumoral injection ofRh-labeled/TAT/pH-nonsensitive liposome (a) or Rh-labeled/TATp/pH-sensitive liposome (b) into LLCtumor bearing mice.

(a) (b)

Figure 11.4 Histology of tumor tissue after the hematoxylin/eosin staining under bright-field lightmicroscopy. Untreated tumor (a), and treated tumor (b).

PEG coating accumulate in the tumor tissue, the lowered pH-mediated removal of theprotective PEG coat takes place, and TATp moieties become exposed and accessible forthe interaction with cells. This leads to rapid pGFP pay-load delivery into the cancer cellsas result of the extensive TATp-mediated internalization of liposomes, and therebyenhanced transfection. The ImageJ analysis indicated a three times less transfection inthe case of PEG-TATp-pGFP-pH-insensitive liposomes as non-detachable PEG coat inter-feres and sterically hinders the interactions between TATp and target cancer cells.

11.5 Conclusion

pH-sensitive mPEG-HZ-PE conjugates based on hydrazone bond chemistry were synthe-sized. The pH-dependant hydrolytic kinetics could be tuned using appropriate aldehydeor ketone and acyl hydrazide. These conjugates have immense applications in targeteddrug delivery systems (e.g., the development of the targeted drug carriers carrying a tem-porarily hidden function such as cell penetrating peptide, TATp), and a detachablePEG-HZ-PE, which, in addition to prolonging circulation half-life of carriers, can exposeTATp function only under the action of certain local stimuli (such as lowered pH), repre-sent a significant step on the way toward “smart” multifunctional pharmaceuticalnanocarriers for target accumulation by EPR effect and intracellular penetration in acontrolled fashion.

11.5 Conclusion

191

(a) (b)

Figure 11.5 Fluorescence microscopy images of the LLC tumor sections fom the tumors injectedwith pGFP-loaded TATp-bearing liposomes with the pH-cleavable PEG coat (a) and with thepH-nonclevable PEG coat (b).



Presence of impurities in the finalproduct while separation on silica gelcolumn.

Impurities of unreacted starting mate-rials or byproducts show up in thefinal product due to many reasons.

Optimize the mobile phase compo-nents, and composition taking intoaccount sample loading and dimen-sions of the column.

Difficulty in identification of PEG orPEG-lipid conjugates on TLC plates.

PEG or PEG-lipid components showsimilar Rf values.

Use specific visualizing agents such asDragendorff for PEG andPhosphomolybdnum spray reagent forlipids.

Difficulty in growing tumors in somemice injected with tumor cells by s.c.route.

Some animals show delayed growthof tumors after s.c. injection of tumorcell.

Wait until tumor grows to desired size(allow some more time).

11.6 Summary Points

1. Hydrazone-based pH-sensitive linkages were introduced between polyethyleneglycol and lipid moieties to synthesize pH-sensitive PEG-PE conjugates. Thehydrolytic kinetics of such linkages was monitored using size exclusionchromatographic method.

2. The pH-dependant hydrolytic stability of hydrazone-based linkages is influenced bynature of carbonyl function and substitutions on acyl hydrazide and carbonyl part ofthe linkage.

3. The in vitro biotin-avidin binding, internalization of fluorescently labelednanocarriers in the in vitro cell culture using H9C2 cells clearly indicated pHsensitivity of designed environmentally sensitive nanocarriers.

4. A cell penetrating peptide, TATp, was successfully anchored on the surface ofenvironmentally sensitive nanocarriers.

5. Rh-labeled or pGFP complexed, TATp bearing pH-sensitive nanocarriers showedincreased accumulation or enhanced transfection, respectively, in tumor bearingmice after intratumoral injections of these prototypes compared to pH-nonsensitivecounterpart.

Acknowledgments

This work was supported by the NIH grants RO1 HL55519 and RO1 CA121838 to VPT.

References

[1] Mainardes, R. M., and Silva, L. P., “Drug delivery systems: past, present, and future,” Curr. Drug Tar-gets 5, 2004, 449–55.

[2] Nishiyama, N., Okazaki, S., Cabral, H., Miyamoto, M., Kato, Y., Sugiyama, Y., Nishio, K.,Matsumura, Y., and Kataoka, K. “Novel cisplatin-incorporated polymeric micelles can eradicatesolid tumors in mice.” Cancer Res. 63, 2003, 8977–8983.

[3] Kreuter, J. “Drug targeting with nanoparticles.” Eur. J. Drug Metab. Pharmacokinet. 19, 1994,253–256.

[4] Storm, G., and Crommelin, D. J. “Colloidal systems for tumor targeting.” Hybridoma 16, 1997,119–125.


192

[5] Francis, G. E., Delgado, C., Fisher, D., Malik, F., and Agrawal, A. K. “Polyethylene glycol modifica-tion: relevance of improved methodology to tumour targeting.” J. Drug Target. 3, 1996, 321–340.

[6] Simard, P., Hoarau, D., Khalid, M. N., Roux, E., and Leroux, J. C. “Preparation and in vivo evalua-tion of PEGylated spherulite formulations.” Biochim. Biophys. Acta 1715, 2005, 37–48.

[7] Maruyama, K. “In vivo targeting by liposomes.” Biol. Pharm. Bull. 23, 2000, 791–799.[8] Xiong, X. B., Huang, Y., Lu, W. L., Zhang, X., Zhang, H., Nagai, T., and Zhang, Q. “Intracellular

delivery of doxorubicin with RGD-modified sterically stabilized liposomes for an improvedantitumor efficacy: in vitro and in vivo.” J. Pharm. Sci. 94, 2005, 1782–1793.

[9] Torchilin, V. P. “TAT peptide-modified liposomes for intracellular delivery of drugs and DNA.” Cel-lular & Molecular Biology Letters 7, 2002, 265–267.

[10] Nori, A., and Kopecek, J. “Intracellular targeting of polymer-bound drugs for cancer chemother-apy.” Adv. Drug Deliv. Rev. 57, 2005, 609–636.

[11] Jayasundar, R., and Singh, V. P. “In vivo temperature measurements in brain tumors using protonMR spectroscopy.” Neurol. India 50, 2002, 436–9.

[12] Engin, K., Leeper, D. B., Cater, J. R., Thistlethwaite, A. J., Tupchong, L., and McFarlane, J. D.“Extracellular pH distribution in human tumours.” Int. J. Hyperthermia 11, 1995, 211–216.

[13] Ojugo, A. S., McSheehy, P. M., McIntyre, D. J., McCoy, C., Stubbs, M., Leach, M. O., Judson, I. R.,and Griffiths, J. R. “Measurement of the extracellular pH of solid tumours in mice by magnetic reso-nance spectroscopy: a comparison of exogenous (19)F and (31)P probes.” NMR Biomed. 12, 1999,495–504.

[14] Khare, A. R., and Peppas, N.A. “Release behavior of bioactive agents from pH-sensitive hydrogels.”J. Biomater. Sci. Polym. Ed. 4, 1993, 275–289.

[15] Kataoka, K., Matsumoto, T., Yokoyama, M., Okano, T., Sakurai, Y., Fukushima, S., Okamoto, K.,and Kwon, G. S. “Doxorubicin-loaded poly(ethylene glycol)-poly(beta-benzyl-L-aspartate) copoly-mer micelles: their pharmaceutical characteristics and biological significance.” J Control Release 64,2000, 143–153.

[16] Leroux, J., Roux, E., Le Garrec, D., Hong, K., and Drummond, D. C. “N-isopropylacrylamide copol-ymers for the preparation of pH-sensitive liposomes and polymeric micelles.” J Control Release 72,2001, 71–84.

[17] Le Garrec, D., Taillefer, J., Van Lier, J. E., Lenaerts, V., and Leroux, J. C. “Optimizing pH-responsivepolymeric micelles for drug delivery in a cancer photodynamic therapy model.” J Drug Target 10,2002, 429–437.

[18] Taillefer, J., Brasseur, N., van Lier, J. E., Lenaerts, V., Le Garrec, D., and Leroux, J. C. “In-vitro andin-vivo evaluation of pH-responsive polymeric micelles in a photodynamic cancer therapymodel.” J Pharm Pharmacol 53, 2001, 155–166.

[19] Asokan, A., and Cho, M. J. “Cytosolic delivery of macromolecules. 3. Synthesis and characteriza-tion of acid-sensitive bis-detergents.” Bioconjug Chem 15, 2004, 1166–1173.

[20] Lee, E. S., Na, K., and Bae, Y. H. “Polymeric micelle for tumor pH and folate-mediated targeting.” JControl Release 91, 2003, 103–113.

[21] Lee, E. S., Shin, H. J., Na, K., and Bae, Y. H. “Poly(L-histidine)-PEG block copolymer micelles andpH-induced destabilization.” J Control Release 90, 2003, 363–74.

[22] Braslawsky, G. R., Kadow, K., Knipe, J., McGoff, K., Edson, M., Kaneko, T., and Greenfield, R. S.“Adriamycin(hydrazone)-antibody conjugates require internalization and intracellular acidhydrolysis for antitumor activity.” Cancer Immunol. Immunother. 33, 1991, 367–374.

[23] Yoo, H. S., Lee, E. A., and Park, T. G. “Doxorubicin-conjugated biodegradable polymeric micelleshaving acid-cleavable linkages.” Journal of Controlled Release 82, 2002, 17–27.

[24] Lee, E. S., Na, K. and Bae, Y. H. “Super pH-sensitive multifunctional polymeric micelle.” Nano Lett.5, 2005, 325–329.

[25] Shen, W. C., and Ryser, H. J. “cis-Aconityl spacer between daunomycin and macromolecular carri-ers: a model of pH-sensitive linkage releasing drug from a lysosomotropic conjugate.” Biochem.Biophys. Res. Commun. 102, 1981, 1048–1054.

[26] Ogden, J. R., Leung, K., Kunda, S. A., Telander, M. W., Avner, B. P., Liao, S. K., Thurman, G. B., andOldham, R. K. “Immunoconjugates of doxorubicin and murine antihuman breast carcinomamonoclonal antibodies prepared via an N-hydroxysuccinimide active ester intermediate ofcis-aconityl-doxorubicin: preparation and in vitro cytotoxicity.” Mol. Biother. 1, 1989, 170–174.

[27] Patel, V. F., Hardin, J. N., Mastro, J. M., Law, K. L., Zimmermann, J. L., Ehlhardt, W. J., Woodland, J.M., and Starling, J. J. “Novel acid labile COL1 trityl-linked difluoronucleoside immunoconjugates:synthesis, characterization, and biological activity.” Bioconjugate Chem. 7, 1996, 497–510.

[28] Heffernan, M. J., and Murthy, N. “Polyketal nanoparticles: a new pH-sensitive biodegradable drugdelivery vehicle.” Bioconjugate Chem. 16, 2005, 1340–1342.

[29] Gillies, E. R., and Frechet, J. M. “pH-Responsive copolymer assemblies for controlled release ofdoxorubicin.” Bioconjugate Chem. 16, 2005, 361–368.

References

193

[30] Gillies, E. R., Jonsson, T. B., and Frechet, J. M. “Stimuli-responsive supramolecular assemblies oflinear-dendritic copolymers.” J. Am. Chem. Soc. 126, 2004, 11936–11943.

[31] Gumusderelioglu, M., and Kesgin, D. “Release kinetics of bovine serum albumin from pH-sensitivepoly(vinyl ether) based hydrogels.” Int. J. Pharm. 288, 2005, 273–279.

[32] Shin, J., Shum, P., and Thompson, D. H. “Acid-triggered release via dePEGylation of DOPEliposomes containing acid-labile vinyl ether PEG-lipids.” Journal of Controlled Release 91, 2003,187–200.

[33] Kratz, F., Beyer, U., Roth, T., Schutte, M. T., Unold, A., Fiebig, H. H., and Unger, C. “Albumin conju-gates of the anticancer drug chlorambucil: synthesis, characterization, and in vitro efficacy.” Arch.Pharm. (Weinheim) 331, 1998, 47–53.

[34] Beyer, U., Roth, T., Schumacher, P., Maier, G., Unold, A., Frahm, A. W., Fiebig, H. H., Unger, C., andKratz, F. “Synthesis and in vitro efficacy of transferrin conjugates of the anticancer drugchlorambucil.” J. Med. Chem. 41, 1998, 2701–2708.

[35] Di Stefano, G., Lanza, M., Kratz, F., Merina, L., and Fiume, L. “A novel method for couplingdoxorubicin to lactosaminated human albumin by an acid sensitive hydrazone bond: Synthesis,characterization and preliminary biological properties of the conjugate.” Eur. J. Pharm. Sci. 23,2004, 393–397.

[36] Toncheva, V., Schacht, E., Ng, S. Y., Barr, J., and Heller, J. “Use of block copolymers of poly(orthoesters) and poly (ethylene glycol) micellar carriers as potential tumour targeting systems.” J. DrugTarget. 11, 2003, 345–353.

[37] Oishi, M., Nagasaki, Y., Itaka, K., Nishiyama, N., and Kataoka, K. “Lactosylated poly(ethylene gly-col)-siRNA conjugate through acid-labile beta-thiopropionate linkage to construct pH-sensitivepolyion complex micelles achieving enhanced gene silencing in hepatoma cells.” J. Am. Chem. Soc.127, 2005, 1624–1625.

[38] Steenbergen, C., Deleeuw, G., Rich, T., and Williamson, J. R. “Effects of acidosis and ischemia oncontractility and intracellular pH of rat heart.” Circ. Res. 41, 1977, 849–858.

[39] Frunder, H. “The pH changes of living tissue during activity and inflammation.” Pharmazie 4,1949, 345–355.

[40] Mellman, I., Fuchs, R., and Helenius, A. “Acidification of the endocytic and exocytic pathways.”Annu. Rev. Biochem. 55, 1986, 663–700.

[41] Shen, W. C., and Ryser, H. J. “cis-Aconityl spacer between daunomycin and macromolecular carri-ers: a model of pH-sensitive linkage releasing drug from a lysosomotropic conjugate.” BiochemBiophys Res Commun 102, 1981, 1048–1054.

[42] Diener, E., Diner, U. E., Sinha, A., Xie, S., and Vergidis, R. “Specific immunosuppression byimmunotoxins containing daunomycin.” Science 231, 1986, 148–150.

[43] Ogden, J. R., Leung, K., Kunda, S. A., Telander, M. W., Avner, B. P., Liao, S. K., Thurman, G. B., andOldham, R. K. “Immunoconjugates of doxorubicin and murine antihuman breast carcinomamonoclonal antibodies prepared via an N-hydroxysuccinimide active ester intermediate ofcis-aconityl-doxorubicin: preparation and in vitro cytotoxicity.” Mol Biother 1, 1989, 170–174.

[44] Sinkule, J. A., Rosen, S. T., and Radosevich, J. A. “Monoclonal antibody 44-3A6 doxorubicinimmunoconjugates: comparative in vitro anti-tumor efficacy of different conjugation methods.”Tumour Biol 12, 1991, 198–206.

[45] Patel, V. F., Hardin, J. N., Mastro, J. M., Law, K. L., Zimmermann, J. L., Ehlhardt, W. J., Woodland, J.M., and Starling, J. J. “Novel acid labile COL1 trityl-linked difluoronucleoside immunoconjugates:synthesis, characterization, and biological activity.” Bioconjug Chem 7, 1996, 497–510.

[46] Patel, V. F., Hardin, J. N., Starling, J. J., and Mastro, J. M. “Novel trityl linked drugimmunoconjugates for cancer therapy.” Bioorganic & Medicinal Chemistry Letters 5, 1995, 507–512.

[47] Patel, V. F., Hardin, J. N., Grindey, G. B., and Schultz, R. M. “Tritylated oncolytics as prodrugs.”Bioorganic & Medicinal Chemistry Letters 5, 1995, 513–518.

[48] Gillies, E. R., Goodwin, A. P., and Frechet, J. M. “Acetals as pH-sensitive linkages for drug delivery.”Bioconjug Chem 15, 2004, 1254–1263.

[49] Murthy, N., Campbell, J., Fausto, N., Hoffman, A. S., and Stayton, P. S. “Design and synthesis ofpH-responsive polymeric carriers that target uptake and enhance the intracellular delivery ofoligonucleotides.” J Control Release 89, 2003, 365–374.

[50] Murthy, N., Xu, M., Schuck, S., Kunisawa, J., Shastri, N., and Frechet, J. M. “A macromoleculardelivery vehicle for protein-based vaccines: acid-degradable protein-loaded microgels.” Proc NatlAcad Sci U S A 100, 2003, 4995–5000.

[51] Heffernan, M. J., and Murthy, N. “Polyketal nanoparticles: a new pH-sensitive biodegradable drugdelivery vehicle.” Bioconjug Chem 16, 2005, 1340–1342.

[52] Lee, S., Yang, S. C., Heffernan, M. J., Taylor, W. R., and Murthy, N. “Polyketal microparticles: A newdelivery vehicle for superoxide dismutase.” Bioconjug Chem 18, 2007, 4–7.

[53] Gümüsderelioglu, M., and Kesgin, D. “Release kinetics of bovine serum albumin from pH-sensitivepoly(vinyl ether) based hydrogels.” International Journal of Pharmaceutics 288, 2005, 273–279.


194

[54] Gümüsderelioglu, M., and Topal, I. U. “Vinyl ether/acrylic acid terpolymer hydrogels synthesizedby [gamma]-radiation: characterization, thermosensitivity and pH-sensitivity.” Radiation Physicsand Chemistry 73, 2005, 272–279.

[55] Hurwitz, E., Wilchek, M., and Pitha, J. “Soluble macromolecules as carriers for daunorubicin.” J.Appl. Biochem. 2, 1980, 25–35.

[56] Laguzza, B. C., Nichols, C. L., Briggs, S. L., Cullinan, G. J., Johnson, D. A., Starling, J. J., Baker, A. L.,Bumol, T. F., and Corvalan, J. R. “New antitumor monoclonal antibody-vinca conjugatesLY203725 and related compounds: design, preparation, and representative in vivo activity.” J MedChem 32, 1989, 548–555.

[57] Beyer, U., Roth, T., Schumacher, P., Maier, G., Unold, A., Frahm, A. W., Fiebig, H. H., Unger, C., andKratz, F. “Synthesis and in vitro efficacy of transferrin conjugates of the anticancer drugchlorambucil.” J Med Chem 41, 1998, 2701–2708.

[58] Kratz, F., Beyer, U., Roth, T., Schutte, M. T., Unold, A., Fiebig, H. H., and Unger, C. “Albumin conju-gates of the anticancer drug chlorambucil: synthesis, characterization, and in vitro efficacy.” ArchPharm (Weinheim) 331, 1998, 47–53.

[59] Oishi, M., Nagasaki, Y., Itaka, K., Nishiyama, N., and Kataoka, K. “Lactosylated poly(ethylene gly-col)-siRNA conjugate through acid-labile beta-thiopropionate linkage to construct pH-sensitivepolyion complex micelles achieving enhanced gene silencing in hepatoma cells.” J Am Chem Soc127, 2005, 1624–1625.

[60] Kataoka, K., Itaka, K., Nishiyama, N., Yamasaki, Y., Oishi, M., and Nagasaki, Y. “Smart polymericmicelles as nanocarriers for oligonucleotides and siRNA delivery.” Nucleic Acids Symp Ser (Oxf), 49,2005, 17–18.

[61] Kong, S. D., Luong, A., Manorek, G., Howell, S. B., and Yang, J. “Acidic hydrolysis ofN-Ethoxybenzylimidazoles (NEBIs): Potential applications as pH-sensitive linkers for drug deliv-ery.” Bioconjug Chem 18, 2007, 293–296.

[62] Sawant, R. M., Hurley, J.P., Salmaso S., Kale, A. A., Tolcheva, E., Levchenko, T. and Torchilin, V. P. “‘Smart’ Drug Delivery Systems: Double-targeted pH-responsive pharmaceutical nanocarriers.”Bioconjugate Chem. 17, 2006, 943–949.

[63] Kale, A. A., and Torchilin, V. P. “Design, synthesis, and characterization of pH-sensitive PEG-PEconjugates for stimuli-sensitive pharmaceutical nanocarriers: the effect of substitutes at thehydrazone linkage on the ph stability of PEG-PE conjugates.” Bioconjug Chem 18, 2007, 363–370.

[64] Jeffs, L. B., Palmer, L. R., Ambegia, E. G., Giesbrecht, C., Ewanick, S., and MacLachlan, I. “A scal-able, extrusion-free method for efficient liposomal encapsulation of plasmid DNA.” Pharm Res 22,2005, 362–372.

[65] Torchilin, V. P., Levchenko, T. S., Rammohan, R., Volodina, N., Papahadjopoulos-Sternberg, B.,and D’Souza Gerard, G. M. “Cell transfection in vitro and in vivo with nontoxic TAT pep-tide-liposome-DNA complexes.” Proc. Natl. Acad. Sci. U. S. A. 100, 2003, 1972–1977.

[66] Kale, A. A., and Torchilin, V. P. “Enhanced transfection of tumor cells in vivo using “Smart”pH-sensitive TAT-modified pegylated liposomes.” J Drug Target 15, 2007, 538–545.

[67] Sawant, R. M., Hurley, J. P., Salmaso, S., Kale, A. A., Tolcheva, E., Levchenko, T., and Torchilin, V.P.“ ‘Smart’ Drug Delivery Systems: Double-targeted pH-responsive pharmaceutical nanocarriers.”Bioconjug Chem. 17, 2006, 943–949.

[68] Torchilin, V. P., Levchenko, T. S., Rammohan, R., Volodina, N., Papahadjopoulos-Sternberg, B. andD’Souza, G. G. M. “Cell transfection in vitro and in vivo with nontoxic TAT peptide-liposome-DNAcomplexes.” Proceedings of the National Academy of Sciences of the United States of America 100, 2003,1972–1977.

[69] Rideout, D. “Self-assembling drugs: a new approach to biochemical modulation in cancer chemo-therapy.” Cancer Invest. 12, 1994, 189-202; discussion 268–269.

[70] Lukyanov, A. N., Gao, Z. and Torchilin, V. P. “Micelles from polyethylene glycol/phosphatidylethanolamine conjugates for tumor drug delivery.” Journal of Controlled Release 91,2003, 97–102.

[71] Apelgren, L. D., Bailey, D. L., Briggs, S. L., et al. “Chemoimmunoconjugate development for ovar-ian carcinoma therapy: preclinical studies with vinca alkaloid-monoclonal antibody constructs.”Bioconjugate Chem 4, 1993, 121–126.

[72] Baker, M. A., Gray, B. D., Ohlsson-Wilhelm, B. M., Carpenter, D. C., and Muirhead, K. A.“Zyn-Linked colchicines: Controlled-release lipophilic prodrugs with enhanced antitumor effi-cacy.” Journal of Controlled Release 40, 1996, 89–100.

[73] Cordes, E. H., and Jencks, W.P. “The Mechanism of hydrolysis of schiff’s bases derived fromaliphatic amines.” J. Am. Chem. Soc. 85, 1963, 2843–2848.

[74] Harnsberger, H. F., Cochran, E.L., and Szmant, H.H. “The basicity of hydrazones.” J. Am. Chem. Soc.77, 1955, 5048–5050.

References

195

C H A P T E R

1 2Biodegradable, Targeted PolymericNanoparticle Drug Delivery Formulation forCancer Therapy

Eric M. Pridgen,1,2 Frank Alexis,2,3,4 Robert S. Langer,1,2,4 and Omid C. Farohkzad2,3*

1Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 021392MIT-Harvard Center for Cancer Nanotechnology Excellence, Cambridge, MA 021393Labortatory of Nanomedicine and Biomaterials, Departments of Anesthesiology, Brigham and Women’sHospital and Harvard Medical School, Boston, MA 021154Harvard-MIT Division of Health Sciences and Technology, Cambridge, MA 02139

*Corresponding author: Omid C. Farokhzad, M.D., Assistant Professor of Anesthesiology, Harvard MedicalSchool, Department of Anesthesiology, Brigham and Women’s Hospital, 75 Francis Street, Boston, MA02115, e-mail: [email protected], Phone: 617-732-6093, Fax: 617-730-2801

197

Abstract

Polymeric nanoparticle delivery systems have the potential to significantlyimpact the treatment of cancer. Nanoparticles offer the ability to design a deliv-ery vehicle that maximizes the therapeutic index of a drug by encapsulating thedrug, targeting it to cancerous tissue, and releasing it in a controlled manner foroptimal dosing. This chapter describes the complete technique for the prepara-tion and characterization of a polymeric nanoparticle delivery system. Thepreparation of the delivery system includes descriptions for the synthesis ofthe polymers, formation of nanoparticles that encapsulate chemotherapeuticdrugs, and surface functionalization with ligands for targeting to canceroustissue. The characterization of nanoparticle physicochemical properties isdescribed along with the evaluation of the delivery system in a cell-basedmodel for binding, uptake, and cytotoxicity. A discussion of methods to opti-mize the delivery system is included to provide a guide for the engineering of adelivery system for specific applications.

Key terms polymeric nanoparticles, biodegradable polymers, cancertherapy, surface functionalization, chemotherapeuticdrugs, controlled drug release, aptamers, targetednanoparticles, drug delivery

12.1 Introduction

Although research efforts over the past 30 years have led to improvements in patient sur-vival, cancer is currently the second-leading cause of death in the United States. Onepotential way to achieve dramatic improvements in the treatment of cancer is throughthe use of new technologies. Nanotechnology is an emerging field that the NationalCancer Institute (NCI) has recognized as having the potential to make paradigm-chang-ing impacts on the detection, treatment, and prevention of cancer [1].

Nanoparticle delivery systems have the potential to become a key technology in thetreatment of cancer. Nanoparticles have several advantages as delivery vehicles thatmake them useful for cancer therapy. They are typically on the order of 100 nm, compa-rable in size to many viruses, although these systems can be fabricated over a wide sizerange [2]. The small size allows nanoparticles to overcome many biological barriers,access tumor tissue through porous vasculature [3, 4], and achieve cellular uptake(Figure 12.1) [5]. The surface of nanoparticles can be engineered to increase blood circu-lation time and influence biodistribution [6], while targeting ligands attached to thesurface can result in enhanced uptake by target tissues [7]. Encapsulation of chemo-therapeutic drugs inside nanoparticles can increase the therapeutic index by deliveringan elevated dose directly to a tumor while limiting systemic toxicity [8]. Drug releasefrom nanoparticles can either be controlled over a period of time or triggered based onan environmental stimulus specific to the tumor tissue such as pH or temperature[9, 10]. Furthermore, the solubility and stability of chemotherapeutic drugs can beimproved through encapsulation, providing an opportunity to reevaluate potentialdrugs that were previously ignored based on poor pharmacokinetics or high toxicity

Biodegradable, Targeted Polymeric Nanoparticle Drug Delivery Formulation for Cancer Therapy

198

Nucleus

Binding

Uptake

Drug Release Nucleus

BindingUptake

Drug Release

Targeted,drug-loadednanoparticles

Endothelialcells

Normal cells

Malignant cells

Extravasation

Figure 12.1 The small size of nanoparticles allows them to extravasate into malignant tissue throughleaky tumor vasculature. Targeting ligands on the surface of nanoparticles are able to bind to receptorson malignant cells, causing uptake through receptor-mediated endocytosis. Encapsulated drug canthen be released from the nanoparticles in a controlled manner for a therapeutic response.

[11]. In addition to chemotherapeutic drugs, imaging agents can also be encapsulatedwithin or conjugated to the surface of nanoparticles to improve tumor detection[12, 13]. Finally, nanoparticles can be engineered to be multifunctional with theability to target cancerous tissue, carry imaging agents for detection, and deliver achemotherapeutic payload [14]. The flexibility in design of nanoparticle delivery sys-tems offers an opportunity to develop novel approaches to deliver drugs that may resultin alternative or complementary therapeutic options for patients with cancer.

Polymeric nanoparticle delivery systems consist of several components that can beengineered based on the desired application. These components are the core, corona,targeting ligand, and payload (Figure 12.2). Considerations of each component are nec-essary when designing a delivery system because each component affects the overall per-formance of the system. The core region affects drug encapsulation and release profiles.The corona region influences particle size, blood circulation half-life, and particle stabil-ity. Targeting ligands are used to enhance cellular uptake after accumulation in tumortissue through binding and endocytosis. The payload used is based on the application,but could consist of a chemotherapeutic drug for therapy or imaging agents fordetection and monitoring of a tumor.

The design criteria for a nanoparticle drug delivery system to treat cancer include thefollowing specifications [15]:

1. Small size (preferably between 10 and 200 nm);

2. High drug loading and encapsulation efficiency;

3. Low rate of aggregation (particle stability);

4. Optimized pharmacokinetics and biodistribution properties.

12.1 Introduction

199

Core

Targeting Ligand

Payload

· Affects particle size andstability

· Influences biodistributionand circulation half-life

Corona

· Increases cellular uptake afteraccumulation in tumor tissuethrough binding and endocytosis

· Ligands include peptides,antibodies, nucleic acids,carbohyrates, small molecules,and surface morphology

· Consists of biodegradablepolymer

· Properties affect drugencapsulation and release

· Includes chemotherapeuticdrugs and imaging agents

· Properties affect encapsulationand release

· Loaded by physical entrapmentor chemical conjugation

Figure 12.2 Components of a nanoparticle delivery system and the effects of each component onthe properties of the system.

In this chapter, the formulation of targeted, biodegradable polymeric nanoparticledrug delivery systems for cancer therapy will be described. Several different methods willbe discussed in order to provide the reader with the flexibility to design a nanoparticledelivery system for a desired application. The materials comprising the core and coronawill be biodegradable and biocompatible polymers approved by the U.S. Food and DrugAdministration (FDA) for clinical use. The use of approved biomaterials will facilitate thetranslation of the delivery system into clinical practice. The core will consist of a polyes-ter such as poly(D,L-lactic acid) (PLA) or poly(D,L-lactide-co-glycolic acid) (PLGA) [16]. Thesafety of these polymers in clinical use is well established, first as a biomaterial in Vicrylsutures [17] and later as excipients for sustained release of parenteral drugs [18]. The sur-face will be modified with poly(ethylene glycol) (PEG), a hydrophilic polymer that sig-nificantly reduces nonspecific interactions with proteins, resulting in increased bloodcirculation times [19–21]. Targeting ligands such as aptamers and antibodies will be con-jugated to the PEG corona through several different chemistries that are common forbioconjugation. Chemotherapeutic drugs will be encapsulated during nanoparticle for-mation using several different synthesis methods. In addition to a detailed protocol forthe formulation of a nanoparticle delivery system, this chapter will also describe how tocharacterize the physicochemical properties of the delivery system and evaluate the sys-tem’s performance in vitro using a cell model.

12.2 Materials

12.2.1 Polymer Synthesis of PLA-PEG and PLGA-PEG

12.2.1.1 Materials for Conjugation via carbodiimide Chemistry

• PLGA-COOH or PLA-COOH (Store under nitrogen at –20°C.)

• NH2-PEG-X, where X = –CH3, –OH, –MAL, or –COOH (Store at –20°C.)

• EDC [1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride] (Preparefresh before use.)

• Anhydrous dichloromethane (DCM) (Safety note: Avoid contact. Use proper gloveswhen handling and use in a hood only.)

• Cold methanol (Safety note: Avoid contact. Highly Flammable. Use proper gloves whenhandling and use in a hood only with sources of ignition removed.)

12.2.1.2 Materials for Conjugation via Ring Opening Polymerization

•D,L-Lactide (Store under nitrogen at –20°C)

• HO-PEG-X (where X = –COOH or –MAL) (Store at –20°C)

• Anhydrous toluene (Safety note: Avoid contact. Highly Flammable. Use proper gloveswhen handling and use in a hood only with sources of ignition removed.)

• Tin(II) 2-ethylhexanoate (Store under dry conditions.)

• Sodium sulfate

• Cold methanol (Safety note: Avoid contact. Highly Flammable. Use proper gloves whenhandling and use in a hood only with sources of ignition removed.)


200

• Acetonitrile (Safety note: Avoid contact. Highly Flammable. Use proper gloves when han-dling and use in a hood only with sources of ignition removed.)

• Chloroform (Safety note: Avoid contact. Use proper gloves when handling and use in ahood only.)

• Magnesium sulfate

• 47 mm PTFE filter membrane, 0.45 μm

12.2.2 Nanoparticle Formation

12.2.2.1 Materials

• PLGA-PEG or PLA-PEG polymer (from Section 12.3.1)

• Drug of interest


• Ultrapure water

• Millipore Amicon Ultra-4 or Ultra-15 centrifugal filter units (NMWL – 100 kDa)

• Dichloromethane (DCM) (Safety note: Avoid contact. Use proper gloves when handlingand use in a hood only.)

• Polyvinyl alcohol (PVA) (Molecular weight – 30 kDa). Prepare 1% (w/v) or 0.3%(w/v) PVA in water solution.

12.2.2.2 Facilities/Equipment

• Probe sonicator

12.2.3 Ligand Conjugation

12.2.3.1 Materials

• PLA-PEG-COOH or PLGA-PEG-COOH (carbodiimide chemistry)

• PLA-PEG-MAL or PLGA-PEG-MAL (maleimide-thiol chemistry)

• Ligand of interest

• Phosphate-buffered saline (PBS), pH 7.4

• Ultrapure water (RNase/DNase free depending on the targeting ligand)

• EDC [1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride] (Preparefresh before use.)

• NHS (N-hydroxysuccinimide) (Prepare fresh before use.)


• 2-Iminothiolane-HCl (Traut’s Reagent) or other reagent for introducing thiolgroups (Prepare fresh before use.)

12.2.4 Quantification of Drug Encapsulation

12.2.4.1 Materials

• Drug-encapsulating nanoparticles

12.2 Materials

201

• Water (HPLC grade)

• Acetonitrile (HPLC grade)


• HPLC system with UV detector

• Reversed-phase column (column specifications will be specific to drug used)

12.2.5 Release Experiments

12.2.5.1 Materials



• Dialysis units (Molecular weight cutoff will be dependent on drug molecularweight)



• HPLC system with UV detector

• Reversed-phase column (column specifications will be specific to drug used)

12.2.6 Postformulation Treatment

12.2.6.1 Materials


• 10% (w/v) sucrose in water solution

12.2.7 Cell Binding and Uptake Experiments

12.2.7.1 Materials

• 8-well microscope chamber slides

• 6-well tissue culture plates

• Cell growth medium

• Opti-MEM reduced-serum medium

• Fluorescent nanoparticles


• 4% (v/v) formaldehyde in ultrapure water

• 0.1% (v/v) Triton-X in PBS

• Rhodamine phalloidin (available from Invitrogen) (Dilute 20 μL of dye in 1 mLPBS)

• Mounting medium with or without DAPI

• Trypsin, 0.25% (1×) with EDTA


202


• Fluorescence microscope or confocal fluorescence microscope

• Flow cytometer with appropriate lasers and detectors

12.2.8 Cytotoxicity Experiments

12.2.8.1 Materials

• 48-well tissue culture plates

• Cell growth medium

• Opti-MEM reduced-serum medium


• CellTiter 96 AQueous One Solution Cell Proliferation Assay (MTS) (available fromPromega)



• Plate reader

12.3 Methods

A complete description of the techniques used for the formulation and characterizationof a nanoparticle delivery system for cancer therapy is provided in this section. Asummary of the steps and characterization parameters is provided in Figure 12.3.Section 12.3.1 describes the synthesis of poly(D,L-lactic acid)-block-poly(ethylene glycol)(PLA-PEG) and poly(D,L-lactide-co-glycolic acid)-block-poly(ethylene glycol) (PLGA-PEG)diblock copolymers using carbodiimide chemistry or ring opening polymerization(ROP). The resulting copolymers are then formed into nanoparticles in Section12.3.2 using several different methods, which also allow the encapsulation ofchemotherapeutic drugs. Section 12.3.3 details the conjugation of targeting ligands to

12.3 Methods

203

Polymer Synthesis

PLGA

PEG

+ +

Drug

Ligand

· Cell binding and uptake· Cytotoxicity

In vitro Evaluation

· Ligand density· Ligand orientation

Ligand ConjugationNanoparticle Formation

····

Particle size and shapeSurface chargeDrug encapsulationDrug release

·Chemical structureMolecular weight

·

PLGA-PEG

Figure 12.3 Overall procedure for formulation and characterization of polymeric, targetednanoparticles. The major steps are in bold and key characterization parameters are listed beneath eachstep.

the nanoparticle surface to complete the delivery system. In Sections 12.3.1 through12.3.3, relevant characterization tools are discussed. Further characterization of theencapsulation and release of the drug component are described in Sections 12.3.4 and12.3.5. Long-term storage of nanoparticles is detailed in Section 12.3.6. Finally, in vitroexperiments are described in Sections 12.3.7 and 12.3.8 to evaluate the delivery systemin the context of a cell-based disease model.

12.3.1 Polymer Synthesis of PLA-PEG and PLGA-PEG

In this section, the synthesis of PLGA-b-PEG and PLA-b-PEG is described using two dif-ferent methods. The first method is the conjugation of commercially available PLGA orPLA to PEG using carbodiimide chemistry. Carbodiimides are zero-length cross-linkersused to aid in the formation of amide linkages between carboxylate (–COOH) and amine(–NH2) functional groups. EDC [1-ethyl-3-(3-dimethylaminopropyl)-carbodiimidehydrochloride] is a popular carbodiimide. It can react with carboxylic acids to form ahighly reactive O-acylisourea intermediate, which can then react with a nucleophilesuch as a primary amine to form an amide bond [22]. Other potential reactions canoccur between the active intermediate and thiol groups or oxygen atoms such as thosein water.

Using EDC, PLGA-COOH or PLA-COOH is conjugated to heterobifunctionalNH2-PEG-X, where X represents possible functional groups such as methyl (–CH3),hydroxyl (–OH), maleimide (–MAL), or carboxylate (–COOH) (Figure 12.4(a)). In thecase of a carboxylic acid on the PEG, EDC is used to activate the carboxylic acid on PLGAor PLA and then separated from the polymer before addition of the NH2-PEG-COOH(Figure 12.4(b)). The only functional group that X cannot be is an amine because thePLGA or PLA could then conjugate to both ends of the PEG unless it is desired to usePLA-PEG-PLA or PLGA-PEG-PLGA to form nanoparticles. The choice of functional groupon the PEG depends on the active functional group of the targeting ligand that will beconjugated or the surface properties desired for the nanoparticle.

The second method for synthesizing PLA-b-PEG is to use ring opening polymeriza-tion (ROP) (Figure 12.4(c)). In this method, the PEG must have a hydroxyl functionalgroup (HO-PEG-X, where X = –COOH or –MAL) from which D,L-lactide can polymerizethrough ROP using tin(II) 2-ethylhexanoate as a catalyst. Similarly, PLyGzA copolymersare synthesized using a mixture of D,L-lactide (y molar) and D,L-glycolide (z molar). Reactiv-ity of D,L-glycolide is higher than D,L-lactide, so control of the random copolymerizationcontent should be optimized using different molar ratios of the two monomers.

Following synthesis, the polymers should be characterized with nuclear magneticresonance (NMR) for chemical structure and conjugation efficiency as well as gel perme-ation chromatography (GPC) for polymer molecular weight.

12.3.1.1 Protocol for Conjugation via Carbodiimide Chemistry (if X = –CH3, –OH, or–MAL)

1. Dissolve PLGA-COOH or PLA-COOH in anhydrous DCM at a concentration of 10mg/mL.

2. Dissolve NH2-PEG-X in anhydrous DCM at a concentration of 10 mg/mL in aseparate vial.


204

3. Dissolve EDC in anhydrous DCM at a concentration of 10 mg/mL in a separate vial.

4. Add EDC to PLGA or PLA using a 5× molar excess of EDC and vortex.

5. Add PEG to the PLGA/EDC or PLA/EDC solution using a 2× molar excess of PEG.

6. React overnight for 15–20 hours at room temperature while stirring. Cover thesolution to protect from light.

7. Precipitate the polymer in cold water or cold methanol.

8. Centrifuge the resulting solution for 30 minutes at 2,500 rpm.

9. Discard the supernatant and dry resulting pellet under vacuum until solvent isremoved.

10. Store polymer under nitrogen at –20°C.

12.3 Methods

205

(a)

(b)

(c)

PL G A-COOHx y

PL G A-COOHx y PL G A-PEG-Xx y

EDC

EDC

H N-PEG-X2

H N-PEG-COOH2

PL G A-PEG-COOHx y

Activated PL G Ax y Activated PL G Ax y

EDC

Remove excess EDC

+

HO-PEG-X

Sn(Oct) catalyst2

PLA-PEG-COOH

Lactide monomer

Figure 12.4 Schematic diagrams of diblock polymer synthesis via carbodiimide chemistry conjuga-tion using NH2-PEG-X with (a) X = –CH3, –OH, –MAL, and (b) X = –COOH or via (c) ring opening poly-merization using tin(II) 2-ethylhexanoate (Sn(Oct)2) as a catalyst.

12.3.1.2 Protocol for Conjugation via Carbodiimide Chemistry (if X = –COOH)

1. Dissolve PLGA-COOH or PLA-COOH in anhydrous DCM at a concentration of 10mg/mL.

2. Dissolve NH2-PEG-X in anhydrous DCM at a concentration of 10 mg/mL in aseparate vial.

3. Dissolve EDC in anhydrous DCM at a concentration of 10 mg/mL in a separate vial.

4. Mix EDC with PLGA or PLA using a 5× molar excess of EDC and allow reaction tooccur for 2 hours at room temperature while stirring. Cover the solution to protectfrom light.



7. Discard the supernatant. Repeat twice to remove all EDC.

8. After the third wash, dry the resulting pellet under vacuum until solvent is removed.

9. Dissolve the activated PLGA or PLA in anhydrous DCM at a concentration of 10mg/mL.

10. Add PEG to the PLGA solution using a 2× molar excess of PEG.

11. React overnight for 15–20 hours at room temperature while stirring. Cover thesolution to protect from light.



14. Discard the supernatant and dry resulting pellet under vacuum until solvent isremoved.


12.3.1.3 Protocol for Conjugation via Ring Opening Polymerization

1. Dissolve vacuum-dried D,L-Lactide (1.6 g, 11.1 mmol) and HO-PEG-X (0.289 g, 0.085mmol) in anhydrous toluene (10 mL) containing anhydrous Na2SO4 (200 mg, 1.4mmol) in a round-bottom flask (see Figure 12.5 for experimental setup).

2. Heat to a reflux temperature of 120°C.

3. Add tin (II) 2-ethylhexanoate (20 mg, 0.05 mmol) to initiate the polymerization. Stirfor 15 minutes to remove all water.

4. Stir for 12 hours with reflux.

5. Cool solution to room temperature.

6. Add cold water (10 mL) and stir vigorously at room temperature for 30 minutes tohydrolyze unreacted lactide monomer.

7. Transfer resulting mixture to a separation funnel containing chloroform (50 mL)and water (30 mL).

8. After layer separation, collect the organic layer (bottom layer) and dry by addinganhydrous magnesium sulfate (200 mg).

9. Filter the solution using 0.45 μm PTFE filter membrane and concentrate undervacuum.

10. Dissolve the dried material in acetonitrile.

11. Pour solution in cold methanol for precipitation.

12. Centrifuge resulting solution for 10 minutes at 4,000 rpm.


206

13. Remove the supernatant and dry resulting pellet under vacuum until solvent isremoved.


12.3.2 Nanoparticle Formation

This section describes three methods for the formation of nanoparticles using the poly-mers synthesized in Section 12.3.1. The three methods are nanoprecipitation [23],oil-in-water (o/w) emulsification-solvent evaporation (single emulsion) [24, 25], andwater-in-oil-in-water (w/o/w) emulsification-solvent evaporation (double emulsion) [6](Figure 12.6). The choice of method is usually dependent on the drug physicochemicalproperties along with the requirements for encapsulation and particle size.

Nanoprecipitation and single emulsion are methods typically used to encapsulatelipophilic drugs. The nanoprecipitation technique requires the drug to be soluble in awater-miscible organic solvent. Nanoparticles are formed instantaneously upon addi-tion of the organic phase to the aqueous phase due to rapid solvent displacement, result-ing in a reduced particle size without the need for sonication or homogenization [26].The single emulsion technique requires the drug to be soluble in a water-immiscibleorganic solvent. Oil-in-water emulsions are formed with the addition of surfactants aftersonication or homogenization. Solvent evaporation results in polymer precipitationinto nanoparticles.

The third method, double emulsion, is used to encapsulate hydrophilic drugs. In thistechnique, the drug is dissolved in the aqueous phase and emulsified with a surfactant ina water-immiscible organic solvent containing the polymer. This first emulsion is thenadded to a second aqueous phase with or without surfactant to form the second emul-sion, where polymer precipitation into nanoparticles occurs due to solvent evaporation.

12.3 Methods

207

Condenser

Oil bath Stir plate

Screw caps

Oil bathtemperaturecontrol

Water flowat roomtemperature

Round-bottom flaskwith stir bar

Figure 12.5 Experimental setup for ring opening polymerization reaction.

The double emulsion technique typically yields nanoparticles with larger sizes than inthe other two methods [27].

Once formed, the particle size and surface charge of the nanoparticles should becharacterized. Light scattering techniques can be used to determine particle size andpopulation uniformity, while electron microscopy (TEM) can be used to image the size,shape, and uniformity of the nanoparticle population. Light scattering instruments canalso be used to measure the zeta potential of the nanoparticles, providing a measure-ment of the surface charge of the nanoparticles. Encapsulation of the drug in thenanoparticles must also be characterized, and this process is described in Section 12.3.4.

12.3.2.1 Protocol for Nanoprecipitation Method

1. Dissolve 1 mg of polymer in 100 μL of acetonitrile and 100 μg of drug in 100 μL ofacetonitrile.

2. Add the polymer/drug solution (200 μL total volume) dropwise to 400 μL ofultrapure water under stirring.

3. Mix the resulting solution for at least 2 hours.

4. Wash the nanoparticle solution at least twice with ultrapure water using 100 kDaAmicon filters. The nanoparticles should be centrifuged at 3,000 rpm or less.

5. Resuspend the nanoparticles in the desired buffer.

12.3.2.2 Protocol for Single Emulsion Method

1. Dissolve 20 mg of polymer and 0.5 mg of drug in 1 mL of dichloromethane (DCM).

2. Add the resulting solution to 2 mL of PVA (1% w/v) in ultrapure water.


208

Nanoprecipitation Single Emulsion (o/w) Double Emulsion (w/o/w)

Drug/polymersolution addeddropwise toaqueous solution

Incubate toallow solventdisplacement

Sonicate orhomogenize

Incubate toallow solventevaporation

Sonicate orhomogenize

Form primaryemulsion (w/o)

Add aqueous phaseSonicate or homogenize

Form secondaryemulsion (w/o/w)

Incubate to allowsolvent evaporation

SurfactantPolymerDrugOrganic PhaseAqueous Phase

Figure 12.6 Nanoparticle formation using the nanoprecipitation, single emulsion, or double emul-sion method.

3. Sonicate with a probe sonicator for 15–30 seconds at 10W.

4. Stir moderately overnight in a hood to evaporate the solvent.



12.3.2.3 Protocol for Double Emulsion Method

1. Dissolve 1 mg of the drug in 2 mL of ultrapure water to prepare a drug stock solution.

2. Dissolve 50 mg of the polymer in 1 mL of DCM to prepare a polymer stock solution.

3. Add 50 μL of drug solution to 1 mL of the polymer solution and emulsify the mixtureusing a probe sonicator at 10W for 15–30 seconds.

4. Add 2 mL of 1% w/v PVA in water to the emulsion and sonicate for 15 seconds at10W using a probe sonicator.

5. Pour the resulting solution into 50 mL of aqueous PVA (0.3% w/v) with gentlestirring.

6. Stir the solution overnight to allow evaporation of the solvent.



12.3.3 Conjugation of Targeting Ligand

This section describes two different chemistries for the conjugation of targeting ligandsto the nanoparticle surface: carbodiimide and maleimide-thiol chemistry (Figure 12.7).Conjugation occurs through functional groups on the ligand and the end of the PEGcorona. Carbodiimide chemistry, which was used for polymer conjugation, forms a sta-ble amide linkage between carboxylate and amine functional groups. Maleimide-thiolchemistry forms a stable thioester linkage between maleimide and thiol (–SH) functionalgroups. Both chemistries result in covalent linkages that are favored over noncovalentlinkages for stability in the physiological environments (pH, high salt concentrations)of the body.

The choice of conjugation chemistry depends on the targeting ligand and thedesired surface properties of the delivery system. For instance, nucleic acid ligands canbe modified with thiol, carboxylate, or amine end groups for conjugation. If a negativesurface charge is desired, which has been shown to minimize interactions with proteinsin the blood [28], a carboxylate functional group at the end of the PEG would be favored,leading to the use of amine-modified nucleic acid ligands for conjugation. In addition,the negative charge may be used to prevent electrostatic interactions between the sur-face and the negatively charged nucleic acids, resulting in less physical adsorption of theligand. This is described later in the protocol for ligand conjugation using carbodiimidechemistry. In the case of peptide ligands, the addition of a cysteine amino acid with afree thiol group favors the use of maleimide-thiol chemistry for the conjugation [29]. Forprotein ligands such as antibodies, the maleimide-thiol conjugation chemistry is alsocommonly used [30–32]. The frequency of free thiol groups in proteins is usually lowcompared with groups such as carboxylates and amines [22]. Conjugation using thesegroups will therefore restrict the thioester linkage to a limited number of sites within the

12.3 Methods

209

protein targeting ligand. Free thiol groups can also be introduced into a proteinligand using reagents that modify amine groups. One commercially available reagent is2-iminothiolane (Traut’s Reagent) (Pierce). A description of this is included later in theprotocol for ligand conjugation using maleimide-thiol chemistry. Other reagents suchas N-succinimidyl S-acetylthioacetate (SATA) (Pierce) have spacers between the aminelinkage and the free thiol generated so that the thiol group is away from the surface ofthe protein, potentially improving the conjugation efficiency. One disadvantage ofusing the thiol chemistry is the potential for disulfide formation between thiol groupson different proteins, leading to cross-linking of the targeting ligand. This can bemitigated through the addition of chelating agents such as ethylenediaminetetraaceticacid (EDTA) or slightly acidic pH.

After conjugation of the targeting ligand to the nanoparticle delivery system, theconjugation should be confirmed and quantified. By attaching a fluorescent probe tothe ligand, the presence of the ligand on the nanoparticles can be qualitatively con-firmed using fluorescence microscopy, flow cytometry, or a fluorescence plate reader. Inaddition, gel electrophoresis can be used to separate nanoparticles with ligand from freeligand to confirm conjugation. For quantitative assessment of conjugation, assay kitssuch as Picogreen (DNA), Ribogreen (RNA), or BCA (protein) can be used to measure theamount of ligand on the surface of the nanoparticles.

12.3.3.1 Protocol for Ligand Conjugation via Carbodiimide Chemistry

1. Suspend PLA-PEG-COOH or PLGA-PEG-COOH nanoparticles in phosphate-bufferedsaline (PBS), pH 7.4, at 10 mg/mL.


210

NP NP

NP

NP

NP

Antibody

2-Iminothiolane HCI

Thiol-modifiedantibody

NHS

EDC

NH -modifiedaptamer

2

(b)

(a)

Figure 12.7 Schematic diagrams of ligand conjugation chemistries. (a) Conjugation of anamine-modified aptamer with a stable amide bond using carbodiimide chemistry. (b) Conjugation ofan antibody with a stable thioester bond using maleimide-thiol chemistry.

2. Add a 5× molar excess of EDC and a 10× molar excess of NHS to the nanoparticlesolution and incubate for 20 minutes at room temperature.

3. Rinse the nanoparticle solution three times with PBS using 100 kDa Amicon filters toremove excess EDC.

4. Add 1 mg/mL amine-modified nucleic acid ligand in a 1:1 molar ratio(ligand:polymer ratio) and incubate 1 hour at 37°C with gentle agitation or 4 hourson ice.

5. Rinse nanoparticle solution twice with ultrapure water using 100 kDa Amicon filtersto remove unconjugated ligand and suspend nanoparticles in desired buffer. Thenanoparticles should be centrifuged at 3,000 rpm or less. The targeted nanoparticledelivery system is now ready for characterization or use.

12.3.3.2 Protocol for Ligand Conjugation via Maleimide-Thiol Chemistry

1. Dissolve protein ligand in PBS, pH 7.4, at a concentration of 10 mg/mL. If usingprotein or peptide ligand with free thiol groups, skip to step 4.

2. Dissolve 2-iminothiolane-HCl (Traut’s Reagent) in PBS at a concentration of 5mg/mL.

3. Mix protein solution with Traut’s Reagent solution with a 40× (10×–50×) molar ratioof Traut’s Reagent to modify protein with a free thiol group and incubate for 1 hourat room temperature.

4. After incubation, add the resulting solution to 10 mg/mL PLA-PEG-MAL or PLGA-PEG-MAL nanoparticles in PBS. The ligand should be added in a 5% (1%–50%) molarratio (protein:polymer).

5. Incubate the resulting solution overnight at 4°C with gentle agitation.

6. Rinse nanoparticle solution twice with ultrapure water using 100 kDa Amicon filtersto remove unconjugated ligand and suspend nanoparticles in desired buffer. Thenanoparticles should be centrifuged at 3,000 rpm or less. The targeted nanoparticledelivery system is now ready for characterization or use.

When using aptamer ligands, it is necessary to heat the aptamers prior to conjuga-tion to expose the functional group. Aptamers can be heated at 90°C for 5 minutes or60°C for 15 minutes, and then incubated with the nanoparticles. Nanoparticles can alsobe heated, although heating will increase the drug release rate.

12.3.4 Quantification of Drug Encapsulation

This section describes the quantification of drug encapsulation within nanoparticles.Drug encapsulation is measured using either a direct or indirect method. For the directmethod, nanoparticles are dissolved in an organic solvent that the polymer is soluble into extract the drug. The extracted drug is then quantified using a convenient assay. Formany drugs, reversed-phase high performance liquid chromatography (RP-HPLC) withUV detection is used for quantification based on a calibration curve. In this case, a con-venient solvent to use for extraction is acetonitrile because it is present in the RP-HPLCmobile phase. The following protocol describes quantification using the direct method.For the indirect method, the drug present in the aqueous phase after encapsulationis measured to determine the amount of drug that was not encapsulated in the

12.3 Methods

211

nanoparticles. The flow-through during the nanoparticle wash steps must be collectedand the drug present is quantified using a convenient assay. From the quantification ofthe drug, the drug encapsulation efficiency and the drug loading can be calculatedaccording to the following equations.

( )Encapsulation EfficiencyMass of encapsulated dr

% = ugMass of initial drug

∗100 (12.1)

( )Drug loadingMass of encapsulated drug

Mass of po% =

lymer used for encapsulation∗100 (12.2)

12.3.4.1 Protocol for Quantification of Drug Encapsulation

1. Collect 500 μg of drug-encapsulating nanoparticles in 200 μL of water or PBSsolution.

2. Add 200 μL of acetonitrile to the nanoparticle solution, mix vigorously, and incubatefor 24 hours. If the incubation lasts longer than 24 hours, store sample at 4°C tominimize the evaporation of the acetonitrile.

3. Quantify the drug in the resulting solution using RP-HPLC.

12.3.5 Drug Release Studies

This section describes the measurement of drug release profiles for nanoparticle deliverysystems. There are two methods used to measure release rate, with the choice dependenton the solubility of the drug in the release medium. With either method, drug-encapsu-lating nanoparticles are contained within a dialysis unit that is incubated in releasemedium. The dialysis membrane must have a molecular weight cutoff that allows thedrug to diffuse through while retaining the nanoparticles. The release medium andconditions should mimic the physiological conditions under which the drug willbe released in the body. For instance, many release studies are conducted with nano-particles incubated at 37°C in PBS, as described later in the release protocols.

For drugs with low solubility in the release medium, a large reservoir of the releasemedium should be used to maintain the condition of infinite sink for the drug.Nanoparticles are collected from the dialysis unit at specified time points for measure-ment of the drug remaining in the nanoparticles using the direct method described inSection 12.3.4. For drugs with higher solubility in the release medium, the same methodcan be used. However, an alternative method is to use a small reservoir of releasemedium. Samples can be collected from the release medium to quantify the drugreleased using the indirect method described in Section 12.3.4. This method reducesthe amount of material required for the study. The frequency of sampling in eachmethod will depend on the release rate, with more time points required during thefaster release periods.

12.3.5.1 Protocol for Release Experiment with a Low-Solubility Drug

1. Prepare 15 mg of drug-encapsulating nanoparticles and resuspend in PBS, pH 7.4,with a final concentration of 2.5 mg nanoparticles/mL; 15 mg will be enough


212

material for 10 time points using 500 μg per sample and triplicate samples at eachtime point.

2. Split nanoparticles equally into 27 dialysis units (500 μg nanoparticles per unit, 200

μL of sample). The remaining three 500-μg samples should be used to measure thedrug mass at t = 0 minutes.

3. Incubate the dialysis units in 4L of PBS buffer at 37°C with gentle stirring.

4. At each time point, collect the nanoparticle samples from three dialysis units andkeep separate for triplicate measurements of the drug release. To evaluate possibleburst release from nanoparticles, early time points should be analyzed (15, 30, and60 minutes).

5. Add 200 μL of acetonitrile to each 200-μL nanoparticle sample, mix vigorously, andincubate for a minimum of 24 hours.

6. Quantify the drug mass.

The release medium should be changed frequently, such as every hour or at everytime point, to ensure that the infinite sink condition remains throughout the releasestudy. The drug release can be calculated using the drug mass (MD) measured at t = 0 min-utes and at a specified time point n as shown in (12.3).

( ) ( ) ( )( )

Drug Release t nM t M t n

M tD D

D

= == − =

=∗,% %

0

0100 (12.3)

12.3.5.2 Protocol for Release Experiment with High-Solubility Drug

1. Prepare 1.5 mg of drug-encapsulating nanoparticles and resuspend in PBS, pH 7.4,with a final concentration of 2.5 mg nanoparticles/mL.

2. Split nanoparticles equally into three dialysis units (500 μg nanoparticles per unit,

200 μL sample).

3. Incubate the dialysis units in 1 mL of PBS buffer at 37°C.

4. At each time point, collect 100 μL of dialysate from each of the three samples and

replace it with 100 μL of fresh PBS buffer.

5. Quantify the drug mass.

The drug release can be calculated using the drug mass (MD) measured at t = 0 minutesand the drug mass measured in the release medium at a specified time point n. However,the mass of drug removed (MD,R) from the release medium for sampling at each timepoint must be accounted for when calculating the total mass of drug released at eachtime point.

( ) ( )( )

Drug Release t nM t n M

M tD D R

D

= == +

=∗,% %,

0100 (12.4)

12.3.6 Postformulation Treatment

This section describes the treatment of nanoparticles post-formulation to improvestability if the nanoparticles will be stored instead of immediately used. Lyophilizationor freezing at –20°C are two methods used for the long-term storage of nanoparticles. Inboth methods, a lyoprotectant (or cryoprotectant) needs to be added to prevent

12.3 Methods

213

aggregation such as sucrose or trehalose [33]. To ensure that the nanoparticles recoveredafter storage are the same, they should be tested for changes in particle size, drug encap-sulation, and drug activity.

12.3.6.1 Method for Particle Storage with Sucrose Lyoprotection

1. Prepare drug-encapsulating nanoparticles.

2. After washing the nanoparticles with water using a 100-kDa filter, resuspend nano-particles in 10% (w/v) sucrose with a 4:1 mass ratio of sucrose to nanoparticles. Thefinal nanoparticle concentration should be 2 mg/mL.

3. Store nanoparticles either by freezing at –20°C or by lyophilization.

4. When ready for use, resuspend nanoparticles in desired medium and wash threetimes with a 100-kDa Amicon filter to remove all sucrose from the sample.

12.3.7 In Vitro Experiments: Cell Binding and Uptake Studies

This section describes in vitro experiments aimed at studying the cell binding anduptake of the nanoparticle delivery system. If using nontargeted nanoparticles, the pur-pose of the experiment is to demonstrate that the delivery system is taken up by the cellsof interest. For targeted nanoparticles, the purpose is to show that the targeted deliverysystem has enhanced selective binding and uptake by cells expressing the targeted recep-tor. To do these types of studies, a cell line is required that expresses the targeted receptoras well as a control cell line that does not express the receptor.

Fluorescence is a convenient tool for cell uptake experiments using fluorescentnanoparticles, although radioactivity is also used for these types of experiments. Thereare two ways to prepare fluorescent nanoparticles. One method is to encapsulate ahydrophobic fluorescent dye such as NBD cholesterol (22-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-23,24-bisnor-5-cholen-3β-ol) (Invitrogen) within nanoparticlesusing the same protocol used to encapsulate hydrophobic drugs [34]. While this methodis simple, the disadvantage is that the dye can escape from the nanoparticle during incu-bation with cells. A second method which avoids this issue is to conjugate a dye such asAlexaFluor (Invitrogen) to PLA or PLGA through an amine functional group using thecarbodiimide chemistry described earlier. This approach will slow the release of dye dur-ing the incubation. Using either approach, fluorescent nanoparticles with and withoutthe targeting ligand can be formulated and tested for uptake and specificity.

Binding and uptake by cells can be observed qualitatively using fluorescence micros-copy or quantitatively using flow cytometry. Microscopy allows the determination ofwhether nanoparticles are bound to the surface of a cell or internalized within thecell [35]. Colocalization studies can also be conducted to determine whether thenanoparticles end up in endosomes, lysosomes, or escape into the cytoplasm of the cell[36]. Flow cytometry can only quantify nanoparticle internalization by cells since thefollowing method uses trypsin to collect the cells.

12.3.7.1 Protocol for Fluorescence Microscopy Imaging

1. Grow adherent cells on 8-well microscope chamber slides in appropriate cell growthmedium until the cells are 70% confluent.


214

2. On the day of the experiment, remove cell growth medium and incubate the cellswith Opti-MEM medium prewarmed to 37°C for 30 minutes.

3. Fluorescent nanoparticles should be prepared in PBS and concentrated to 4 mg/mL.

4. Add 50 μg of fluorescent nanoparticles to the cells and incubate for 2 hours at 37°C.

5. Remove nanoparticles and gently wash the cells twice with 500 μL PBS.

6. Add 250 μL of 4% formaldehyde to fix the cells and incubate for 20 minutes.

7. Wash cells twice with 500 μL PBS.

8. Add 250 μL of 0.1% Triton-X and incubate for 3 minutes.


10. Add 250 μL of rhodamine phalloidin dye and incubate for 20 minutes. (This step isonly necessary if interested in staining the cytoskeleton.)


12. Remove all liquid and mount cells using mounting medium with DAPI if stainingthe nucleus. Otherwise, use the mounting medium without DAPI.

13. Image cells using a fluorescence microscope or confocal fluorescence microscope.

12.3.7.2 Protocol for Quantification of Internalization by Flow Cytometry

1. Grow adherent cells on 6-well plates in appropriate cell medium until the cells are70% confluent.

2. On the day of the experiment, remove cell growth medium and incubate the cellswith Opti-MEM media prewarmed to 37°C for 30 minutes.

3. Fluorescent nanoparticles should be prepared in PBS and concentrated to 4 mg/mL.

4. Add 100 μg of fluorescent nanoparticles to the cells and incubate for 2 hours at 37°C.

5. Remove nanoparticles and gently wash the cells twice with PBS.

6. Add 500 μL of trypsin and incubate until cells release from the plate surface.

7. Add 3 mL of media to the cells.

8. Collect cells and centrifuge for 1 minute at 1,000 ×g to recover the cells.

9. Remove the media and resuspend in a buffer such as PBS for flow cytometry analysis.

10. Analyze cells using flow cytometry.

12.3.8 In Vitro Experiments: Cytotoxicity Studies

This section describes in vitro experiments aimed at studying the cytotoxicity of thenanoparticle delivery system. The purpose of these experiments is to demonstrate theenhanced toxicity of the nanoparticles in cells expressing the targeted receptor com-pared with cells that do not express the receptor due to enhanced uptake. These experi-ments are similar to the uptake studies described in the previous section except that thenanoparticles contain a drug instead of a fluorescent dye. Toxicity is evaluated using acell proliferation assay such as the MTS assay (Promega).

12.3.8.1 Protocol for Cytotoxicity Study

1. Grow adherent cells on 48-well plates in appropriate cell growth medium untilthe cells are 70% confluent.

12.3 Methods

215

2. On the day of the experiment, remove cell growth medium and incubate the cellswith Opti-MEM medium prewarmed to 37°C for 30 minutes.

3. Drug-encapsulating nanoparticles should be prepared in PBS, washed just prior touse to remove any free drug, and concentrated to 4 mg/mL.

4. Add varying amounts of nanoparticles to the cells and incubate for 1 hour at 37°C.

5. Remove nanoparticles and gently wash the cells twice with PBS.

6. Incubate the cells for 72 hours in cell growth medium without changing the mediumto allow cells to proliferate.

7. Add MTS reagent to cells and quantify cell proliferation using a plate reader.

When seeding cells, the outer wells of the plates show greater variability, so only theinner wells should be used for more consistent results. The optimal nanoparticle con-centration and incubation times will need to be determined experimentally for eachdelivery system and each cell model. The incubation times should be kept short sincelonger times allow the drug to escape from the nanoparticles. The free drug could thenbe taken up by the cells and contribute to the toxicity for both targeted and nontargetednanoparticles.


In the methods described in the previous section, data was generated for three maingoals: characterization of the synthesized polymers, characterization of the nanoparticledelivery system, and evaluation of performance in an in vitro cell model. This sectionwill discuss the acquisition and interpretation of that data.

12.4.1 Polymer Characterization

Before using a polymer for nanoparticle formation, the polymer needs to be character-ized thoroughly since it has a significant influence on the properties of the nanoparticle.The critical parameters include the averaged molecular weight and polydispersity,which are characterized using gel permeation chromatography (GPC). The polymershould also be chemically characterized using nuclear magnetic resonance (NMR).

GPC is a technique used to separate polymers based on size. A set of standards withknown molecular weights are used to generate a standard curve of retention time versusmolecular weight from which the polymer analyzed can be compared. A molecularweight distribution can then be generated for the unknown polymer. Polystyrene orPEG is usually used as the standard. One limitation of this technique is that the correla-tion between molecular weight and hydrodynamic radius for the standard may be differ-ent than the polymer analyzed, leading to an error in the absolute value of the molecularweight for the analyzed polymer. If the polymers are purchased from a vendor, thenmolecular weight information would be provided. However, if using ring opening poly-merization, then the molecular weight would need to be determined. For the ring open-ing polymerization protocol described in Section 12.3.1.3, the average molecular weightshould be ~10.5 kDa. An alternative method to GPC is the use of viscosimetry, whichcan be used to determine molecular weight based on the concept that larger moleculeswill be a greater impediment to flow and result in higher solution viscosities.


216

NMR is a spectroscopic technique that allows different chemical groups in a mole-cule to be identified based on their chemical shifts. If the polymer structures are known,which is the case for PLA-PEG and PLGA-PEG, the conjugation efficiency can be esti-mated. By dissolving a polymer in a deuterated solvent such as deuterated chloroform ordeuterated dimethyl sulfoxide (DMSO), 1H NMR can be used to identify the differentchemical groups. An example 1H NMR spectrum for PLA-PEG is shown in Figure 12.8. Inthe figure, the different peaks correspond to the –CH, –CH2, and –CH3 groups in thepolymer. Since the –CH and –CH3 groups are only present in PLA monomer, the peakarea ratio of –CH3 to –CH should be approximately 3 to correspond with the ratio ofhydrogen atoms. The PEG polymer has two –CH2 groups per monomer. If the signalfrom –CH in PLA is compared with the signal from –CH2 divided by 4 to account for thetwo methylene groups per PEG and two hydrogen atoms per methylene group, theratio can be compared with the ratio of the expected molecular weights to determinewhether there is free PLA or free PEG remaining. If nonconjugated PEG is remaining,another separation can be performed to remove the remaining unreacted PEG usingprecipitation in cold water.

12.4.2 Nanoparticle Characterization

The nanoparticle delivery system requires significant characterization because thephysicochemical properties of the system determine its performance. The critical param-eters include particle size, surface charge, drug encapsulation, drug release, and ligandconjugation. Several analytical tools are available to analyze each parameter.


217

-CH peak

- pCH eak

- pCH eak

2

3

DMSO solvent peak

PLA-PEG

OHC N

H

O

y

H2C

CH2

O

z

CH3

5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 ppm

Figure 12.8 1H NMR characterization of PLA-PEG dissolved in deuterated DMSO synthesized usingring opening polymerization.

Particle size and polydispersity can be estimated in solution using light scattering ortreated for analysis using scanning electron microscopy (SEM) or transmission electronmicroscopy (TEM) (Figure 12.9). Light scattering is used to quantify the hydrodynamicradius and polydispersity of a nanoparticle population and is slightly affected by thesolution. TEM and SEM analyze dry particles, providing images of the nanoparticles thatcan be used to qualitatively observe particle size, shape, and polydispersity. The two ana-lytical tools are complementary and should both be used to fully characterize thenanoparticle size. However, because the samples are treated differently, the particle sizemeasurements will not correspond exactly with each other. The particle size will varygreatly depending on the polymer used and nanoparticle formation conditions.

In addition to particle size, light scattering instruments can be used to measure thezeta potential, which is an estimate of the surface charge of the nanoparticles. For zetapotential, the measurement is very sensitive to the ionic environment. Therefore, themeasurement will be most accurate in a solution that mimics physiological conditions,such as PBS. The zeta potential will vary depending on the functional end groups andthe presence of targeting ligands. For example, carboxyl end groups result in a negativecharge of approximately –50 mV [35].

Drug encapsulation can be quantified using the protocols described in Section12.3.4. Regardless of the assay used to quantify the drug, it is important to determinewhether the polymer or solution interferes with the drug quantification. Controlsusing nanoparticles without the drug should be prepared and treated the same as thedrug-encapsulating nanoparticles to determine interference. Encapsulation efficiencies


218

0

20

40

60

80

100

40 49 60 74 92 113 139 172 211

Hydrodynamic diameter (nm)

Freq

uenc

y

0

20

40

60

80

100

40 50 51 54 67 83 102 125 155

Hydrodynamic diameter (nm)

Freq

uenc

y

190

(c)

(b)(a)

100 nm

Figure 12.9 Particle size distributions of PLA-PEG nanoparticles prepared using nanoprecipitation (a)after washing (mean hydrodynamic diameter = 83 nm; polydispersity = 0.348) and (b) after filtrationusing a 0.1 μm filter (mean hydrodynamic diameter = 56 nm; polydispersity = 0.053). (c) Transmissionelectron micrograph (TEM) of PLA-PEG nanoparticles negatively stained with uranyl acetate solution.

and drug loads vary based on the physicochemical properties of the drug and polymer,formulation conditions, and initial drug load. For hydrophobic drugs such as docetaxelencapsulated in PLGA-PEG nanoparticles, a typical drug load is 1% (w/w) and encapsula-tion efficiency is approximately 10% [34]. Drug release rates will also vary dependingon several parameters discussed in Section 12.5.5 (Figure 12.10). For docetaxel encapsu-lated in PLGA-PEG nanoparticles with a PLGA molecular weight of approximately 10.5kDa, half of the drug is released in 12 hours [36]. The release profile is usually biphasic,with an initial burst release followed by a slower release over a few days.

Ligand conjugation can be quantified using two different methods. The first is thedirect method, where the ligand attached to the surface is measured after washing awaythe unconjugated material. The second is the indirect method, where the wash contain-ing the unconjugated ligand is collected and analyzed. By comparing this with theinitial amount of ligand used, the conjugated ligand can be determined. There aremany different analytical tools available to quantify ligand conjugation. One tool iscolorimetric assay kits such as the BCA assay for proteins, Picogreen for DNA, orRibogreen for RNA. X-ray photoelectron spectroscopy (XPS) can be used to analyze thesurface chemistry of a nanoparticle and detect the presence of a ligand based on itschemical signature. Ultraviolet (UV) absorbance is a simple technique that is used toquantify small molecules. Ligands can also be labeled with a fluorophore, assuming itdoes not interfere with the conjugation, and quantified using flow cytometry or afluorescence plate reader. When quantifying ligand conjugation, it is necessary toaccount for interference in the assay by the polymer and ligand on the surface due tononcovalent interactions. For polymer interference, nanoparticles without ligand canbe prepared in the exact same conditions as the nanoparticles with ligand to correct forthe nanoparticle signal in the assay. For noncovalent interactions, nanoparticles can be


219

0

0.25

0.5

0.75

1

0 24 48 72 96Time (hours)

PLGA0.17PEG3400

PLGA0.17PEG12000

PLGA0.19PEG3400

PLGA0.19PEG12000

PLGA0.45PEG3400

PLGA0.67PEG12000

PLGA0.67PEG3400

Mas

sfr

actio

nof

doce

taxe

l rel

ease

d

Figure 12.10 Drug release profiles for docetaxel encapsulated in PLGA-PEG nanoparticles withvarying PLGA and PEG molecular weights. For each of the formulations, a biphasic release profile isobserved where an initial burst release occurs followed by a slower release rate. (Reproduced withpermission from [36]. Copyright 2008 National Academy of Sciences, U.S.A.)

incubated with ligands but without the conjugation chemistry reagents. For example,if using thiol-maleimide chemistry to conjugate a thiol-modified protein to the nano-particle, the three samples to analyze would be nanoparticles without protein incuba-tion, nanoparticles incubated with unmodified protein, and nanoparticles incubatedwith thiol-modified protein. Ligand conjugation is calculated using conjugation effi-ciency and ligand weight fraction on the nanoparticle surface. Both are calculated using(12.1) and (12.2) but with the ligand masses used instead of drug masses.

12.4.3 In Vitro Experiments

The effectiveness of the nanoparticle delivery system needs to be evaluated in a cellmodel once the system is fully characterized. The critical parameters are binding selec-tivity and uptake by targeted cancer cells as well as cytotoxicity in both targeted andnontargeted cells.

Binding and uptake experiments are designed to demonstrate that targeted nano-particles selectively enhance interactions with targeted cells (receptor-positivecells) compared with nontargeted nanoparticles. When conducting these experiments,nanoparticles binding to the cell surface and nanoparticles taken up by targeted cellsthrough endocytosis need to be distinguished. For these experiments, the incubationtime and amount of nanoparticles should be optimized to emphasize the differencesbetween targeted and nontargeted nanoparticles. The incubation time should beapproximately 1 to 2 hours, with shorter incubation times preferred for several reasons.First, it reduces background signal by minimizing dye leakage from the nanoparticles.Second, it minimizes nonspecific uptake of cells through fluid-phase endocytosis.Surface-bound and endocytosed nanoparticles can be distinguished through confocalfluorescence microscopy. Using 3-D reconstruction of imaged cells, nanoparticles on thesurface can be distinguished from those inside the cell (Figure 12.11). Nanoparticle posi-tion can be further elucidated using colocalization analysis, in which the position of thenanoparticles is compared with a dye that localizes to specific cellular compartments,such as endosomes or lysosomes [36]. Specificity of targeted nanoparticles can beevaluated using colocalization analysis as well. Using a fluorescent ligand such as afluorescently labeled antibody specific to the targeted receptor, imaging can be used toshow the association of the targeted nanoparticles with the receptor on the cell surfaceor in cellular compartments if endocytosed. Specificity can also be demonstrated using acompetitive binding study, where targeted nanoparticles and free ligand are incubated

with receptor-positive cells. The free ligand is usually in 10–100× molar excess, allowingit to bind to the receptor and block binding of the nanoparticles, demonstrating thenanoparticles’ specificity for that receptor.

Fluorescent nanoparticles allow quantification of binding and uptake using severaldifferent analytical tools. Using flow cytometry, the uptake of nanoparticles can bequantified, but surface-bound nanoparticles are not included because the cells aretrysinized [37]. The trypsin cleaves surface proteins, which should detach surface-boundnanoparticles. However, this tool allows comparison of uptake under differentexperimental conditions as well as analysis of the uptake kinetics. Another analyticaltechnique is to grow the cells on 96-well plates and use a fluorescence plate readerfor quantification. With this technique, both surface-bound and endocytosednanoparticles can be quantified but not distinguished. An alternative to fluorescence is


220

the use of radioactive polymers, radioactive molecules conjugated to polymers, orradioactive molecules encapsulated in nanoparticles to quantify surface-bound andendocytosed nanoparticles using a scintillation counter [36]. Regardless of the methodused, targeted nanoparticles should be compared with nontargeted nanoparticles inboth receptor-positive and receptor-negative cells to fully evaluate the specificity and


221

NP-Apt

NP

NP-Apt

NP

2 hrs 16 hrs

PC3

LNC

aP

(a)

LNCaPA B D E

NP-Apt

C

F

LNCaPA B D EC

F

(b)

Figure 12.11 (a) Binding study of aptamer-targeted PLA-PEG nanoparticles incubated with LNCaP(receptor-positive prostate cancer cells) and PC3 cells (receptor-negative prostate cancer cells) withincubation times of 2 and 16 hours. A rhodamine-dextran dye (red) is encapsulated within thenanoparticles, the cell nuclei is stained with 4’,6’-diamidino-2-phenylindole (blue), and the actincytoskeleton is stained with Alexa-Fluor Phalloidin (green). The samples were imaged using fluores-cence microscopy. (b) 3-D reconstruction of the cell using confocal microscopy rotated along the z-axisthrough the mid z-axis point of the cell to demonstrate that nanoparticles are present inside the cell.(Reproduced with permission from [35]. Copyright 2004 American Association for Cancer Research.)

enhancement in interactions between nanoparticles and cells due to the targetingligand.

Cytotoxicity studies are the other key in vitro experiment to demonstrate the effec-tiveness of the delivery system. The key parameters are the drug concentration and incu-bation time. The incubation time used for the binding and uptake experiments can alsobe used for these experiments since binding or uptake has been demonstrated underthose conditions. Incubation times should be kept as short as possible since longer timesallow more of the drug to be released in the solution before nanoparticles are taken up bythe cells. The free drug in the solution can then contribute to the toxicity. Shorter timeslimit this effect and make the toxicity data more representative. The drug concentrationshould then be varied by preparing drug-loaded nanoparticles and incubating varyingamounts of the nanoparticles with the cells. For cytotoxicity studies, targeted andnontargeted nanoparticles loaded with the drug should be compared in receptor-positive and receptor-negative cells to demonstrate the enhanced toxicity of targetednanoparticles in receptor-positive cells. Further controls include targeted and non-targeted nanoparticles without the drug in both cell types to evaluate whether thenanoparticles themselves are toxic to the cells. Results from cytotoxicty experiments canbe presented in two ways (Figure 12.12). The first way is to show the toxicity at specificconditions which may be most representative of in vivo conditions, such as a specificincubation time and drug concentration [23, 34]. The other way is to present the entiredose-response curve, showing the toxicity as a function of the drug concentration [38,39]. As part of the dose-response curve, the IC-50, which represents the drug concentra-tion where 50% of the cells are killed, can be calculated and compared across differentconditions.


This section will focus on the optimization of the nanoparticle delivery system. Impor-tant physicochemical properties to consider when designing a delivery system includethe particle size, particle shape, surface chemistry, drug loading, drug release, and target-ing. The parameters available for manipulation of the system include the nanoparticleformulation parameters as well as the components of the system, which can be changedindependently due to the modular design of the nanoparticles. Both allow significantcontrol over the physicochemical properties of the nanoparticles and provide flexibilityin the design of the system. By understanding how the components and formulationparameters influence nanoparticle properties, the delivery system can be tailored to thedesign criteria for the application of interest.

12.5.1 Particle Size

Nanoparticle size is a key property of the delivery system that influences biodistributionand blood circulation half-life. An important advantage of using nanoparticles for thetreatment of cancer is that the small size (<150 nm) allows enhanced extravasation intotumor tissues through a phenomenon known as enhanced permeation and retention(EPR). The EPR mechanism allows nanoparticles to passively target tumors due to theformation of leaky vasculature and poor lymphatic drainage in tumor tissue [40].


222

Smaller nanoparticles have also been shown to have reduced protein surface absorption,leading to a reduction in hepatic uptake and longer blood circulation half-life [41, 42].However, a practical lower limit exists for nanoparticle size as particles on the order of 10nm or less exhibit increased renal clearance, significantly reducing the blood circulationhalf-life [43]. The optimal particle size for an application would have to be determinedthrough in vivo experiments.

There are several parameters that can be used to control nanoparticle size. One is themolecular weight of the corona component, which is PEG in this chapter. By decreasingthe molecular weight, the particle size can be reduced [36]. The other parameters are


223

30 minutes 2 hours(viability assessed at 72 hours)

NP-AptNP Dtxl-NP

1

Dtxl-NP-Apt

PN-NP-AptPN-NP

0

0

20

20

40

40

60

60

80

80

100

100

120

120

Cel

lvia

bilit

y(%

)Su

rviv

al(%

)

0.0 0.2 0.4 0.6 0.8 1.0

[Pt] ( M)μ

(b)

(a)

Figure 12.12 (a) Cytotoxicity study using aptamer-targeted PLGA-PEG nanoparticles encapsulatingdocetaxel presented at a fixed drug concentration with two different incubation times fornanoparticles with the cells. Nanoparticles without drug are compared with targeted and nontargetednanoparticles containing drug. *, significance by ANOVA at 95% confidence interval. (Reproducedwith permission from [23]. Copyright 2008 National Academy of Sciences, U.S.A.) (b) Dose responsecurve for receptor-positive prostate cancer cells treated with cisplatin-encapsulating, aptamer-targetedPLGA-PEG nanoparticles. Targeted nanoparticles are compared with nontargeted nanoparticles and thefree drug. Calculated IC50 values were 0.03 μM for the targeted nanoparticles and 0.13 μM for thenontargeted nanoparticles. (Reproduced with permission from [38]. Copyright 2008 National Academyof Sciences, U.S.A.)

associated with the nanoparticle formulation. For the nanoprecipitation method, theparameters include polymer concentration, solvent:water ratio, the solvent used to dis-solve the polymer, and mixing rate [2]. The nanoparticle size is dependent on the rate atwhich organic solvent diffuses into the aqueous phase, with faster diffusion resulting insmaller particles. By decreasing the polymer concentration or the solvent:water ratio,nanoparticle sizes are reduced. Dissolving the polymer in a solvent with higher watermiscibility also reduces nanoparticle sizes. Increasing the mixing rate can result insmaller nanoparticle size as well [44]. For the emulsion methods, addition parameters fornanoparticle size control include the concentration of surfactant, type of surfactant, andsonication intensity [45, 46]. Increasing the surfactant concentration or sonicationintensity generally results in smaller nanoparticles. Surfactants able to reduce surfacetension more efficiently also result in smaller nanoparticles at the same surfactant con-centration. This could be an important consideration as residual PVA on nanoparticlesprepared using emulsion methods has been shown to affect cell uptake [47].

12.5.2 Particle Shape

Particle shape is a physicochemical property of nanoparticles that should be consideredwhen designing a delivery system. Some interesting initial studies have shown that par-ticle shape can have a significant effect on the blood circulation half-life of particles.When comparing spherical and cylindrical particles, cylindrical particles demonstrateincreased circulation half-life which can be controlled to some upper limit by the lengthof the cylinder [48]. This is due to the difficulty that phagocytic cells in the liver andspleen have engulfing particles with cylindrical shapes. The shape of the particle can bemanipulated by changing the volume fraction of the hydrophilic (PEG) and hydropho-bic (PLA or PLGA) polymer blocks. In general, when the volume fraction of the hydro-phobic block is greater than 50%, the particles tend to be spherical, while hydrophobicvolume fractions between 25–50% tend to result in cylindrical particles [49, 50].

12.5.3 Surface Chemistry

Nanoparticle surface chemistry is a critical property that affects nanoparticle uptake bycells of the mononuclear phagocyte system (MPS), significantly influencing blood circu-lation half-life and biodistribution [51]. Surface chemistry includes the composition ofthe corona and the nanoparticle surface charge. The corona composition has a directeffect on protein opsonization on the nanoparticle surface. Proteins absorbed on thenanoparticle surface result in receptor-mediated phagocytosis through interactions withthe MPS in tissues such as the liver or spleen [52–54]. Polymers such as PEG are used toreduce these nonspecific protein interactions through steric repulsion and the hydro-philic nature of the polymer [55, 56]. When designing the corona, parameters to con-sider include the molecular weight of the PEG [41], the surface density of the PEG [57],and the architecture of the polymer (linear versus branched) [58]. The corona alsoimproves particle stability, preventing aggregation through steric repulsion.

The other aspect of surface chemistry to consider is the surface charge and functionalend groups. Positively charged nanoparticles generally demonstrate higher rates ofphagocytosis compared with neutral or negatively charged nanoparticles, resulting in ashorter circulation half-life and biodistribution that favors tissues such as the liver andspleen [51]. Therefore, neutral or negatively charged end groups (sulfate, hydroxyl,


224

carboxylate, maleimide) should be favored over positively charged end groups (primaryamines). The other consideration in terms of surface charge is the effect on conjugation.Certain functional groups are required based on the ligand conjugation strategy chosen.In addition, complexes can form between the ligand and nanoparticle surface based onelectrostatic interactions that may not be as robust as covalent linkages in physiologicalenvironments. The interactions may also orient the ligand in a position that prevents itfrom interaction with the targeted receptor.

12.5.4 Drug Loading

Drug encapsulation depends on the physicochemical properties of the drug and poly-mers used as well as the nanoparticle formation conditions used. The optimal condi-tions for the delivery system of interest will have to be determined empirically; however,some general guidelines are provided next to aid in the optimization. As different condi-tions are evaluated, it is important to consider both the encapsulation efficiency anddrug load of the system and decide which parameter to maximize. Increasing valuesof both are optimal, but in some cases only one can be increased at the expense ofthe other.

Encapsulation of hydrophobic drugs can be improved by varying the nanoparticleformation parameters. Work with paclitaxel, a model hydrophobic drug, has elucidatedhow parameters such as the drug:polymer ratio, organic:aqueous phase ratio, and poly-mer and drug concentrations affect encapsulation [59, 60]. Increasing the drug:polymerratio can result in increased encapsulation, although the encapsulation efficiency maydecline at higher ratios. The organic:aqueous phase ratio generally has a minimal effecton the encapsulation. Higher polymer and drug concentrations improve encapsulationsince lower concentrations may reduce interactions between the polymer and drug.Changing the core polymer can also increase encapsulation by improving the interac-tion between the core polymer and drug. Interactions between the polymer and drugcan be maximized by matching the Flory-Huggins interaction parameters if available[61, 62]. Besides PLA and PLGA, other polymers used in controlled release biodegradablenanoparticle applications include poly(orthoesters) [63], poly(caprolactone) [64],poly(butyl cyanoacrylate) [65], polyanhydrides [66], and poly-N-isopropylacrylamide[67]. The properties of PLGA can also be manipulated by varying the G:L ratio, withglycolic acid having a more hydrophilic character that could improve encapsulation forsome drugs.

The encapsulation of hydrophilic drugs is a greater challenge because hydrophilicdrugs rapidly partition into the aqueous phase during nanoparticle formation. However,there are some parameters that can be used to enhance encapsulation, such as changingthe nanoparticle formation method or manipulating the properties of the drug. Doubleemulsion is typically used for the encapsulation of hydrophilic drugs. However, studieshave been conducted to improve the method. One improvement is to use an organicphase that is partially water-miscible [68]. By changing the solvent from dichloro-methane to ethyl acetate, the encapsulation of alendronate (a hydrophilic, low molecu-lar weight biphosphonate) was significantly improved. The hydrophobicity of thedrug can also be manipulated through several different methods. When encapsulatingprocaine hydrochloride (a water soluble drug) using the nanoprecipitation method,decreasing the ionization state of the drug by increasing the pH of the aqueous phase


225

resulted in improved encapsulation [69]. In addition, switching from the salt form(procaine hydrochloride) to the base form (procaine dehydrate) increased encapsula-tion. The drug itself can also be chemically modified to enhance encapsulation. Thechemotherapeutic drug cisplatin was modified with linear hexyl chains, resulting inimproved encapsulation [38].

For drugs with poor encapsulation, another method for increasing drug load is toconjugate the drug directly to the core polymer. This approach has the added advantageof allowing more precise control over the drug load compared with physical entrap-ment. There are several ways to conjugate a drug to a polymer. In one study, the rela-tively hydrophilic drug doxorubicin was conjugated to PLGA-COOH through a primaryamine group on doxorubicin [70] and formulated into nanoparticles using a singleemulsion method. In another study, paclitaxel-PLA conjugates were formed throughring opening polymerization from a hydroxyl group on the paclitaxel similar to the pro-cedure for the ring opening polymerization of PLA-PEG described earlier [71]. Althoughpaclitaxel is a hydrophobic drug that has reasonable encapsulation efficiency inPLA-PEG nanoparticles, the conjugation approach provides more control over the drugloading and offers greater batch-to-batch consistency. One potential drawback to thisapproach is that conjugating polymers to drugs can inactivate the drug depending onthe drug conjugation site.

12.5.5 Drug Release

Drug release can be controlled through different release mechanisms to obtain a thera-peutically desirable release profile. These mechanisms include the diffusion of theencapsulated drug through the polymer core matrix, bulk or surface degradation of thepolymer, and swelling of the polymer core followed by diffusion of the drug [72, 73].Release rates in nanoparticles can be controlled through the diffusion of the drug, thedegradation rate of the polymer, and the partition coefficient of the drug between thepolymer core and the aqueous environment.

Several parameters can be used to manipulate the release rate of a drug physicallyentrapped in the nanoparticle core. Increasing the molecular weight of the core polymerhas been shown to reduce the release rate due to slower drug diffusion through a denserpolymer core [36]. Increasing the size of the nanoparticles can also slow the release ratebecause there is less total surface area and the release rate is proportional to the particlesurface area. Different biodegradable polymers may be used as well to influence therelease rate. The rate will depend on the affinity of the drug for the core polymer and thedegradation rate of the polymer under physiological conditions. Finally, the rate ofpolymer degradation can be used tune the release rate [74]. Glycolic acid is more hydro-philic than lactic acid, causing it to degrade faster in aqueous environments [75]. Byincreasing the proportion of glycolic acid in PLGA, the release rate can be increased. Theproperties of the drug can also affect the release rate. In general, smaller and more hydro-philic drugs will diffuse faster and result in increased release rates.

For drugs chemically conjugated to the polymer as discussed in Section 12.5.4, therelease rate is dependent on the degradation of the drug-polymer linker and diffusion ofthe drug through the core. The release rate can be tuned by changing the molecularweight of the polymer in this case [71]. This approach provides more precise controlover the release rate in addition to the advantages in drug loading. However, because the


226

drug load and release rate are both dependent on the molecular weight of the polymer,these two properties are not completely independent using the chemical conjugationapproach.

When designing a delivery system, the ideal nanoparticle would retain the drug withminimal leakage while in the bloodstream, then release the drug in a controlled mannerat an effective dose level in the tumor tissue. This ensures that a toxic dose of drug isreleased at the disease site only. This type of release profile is difficult to achieve withphysical entrapment of the drug or conjugation using a hydrolysable linker. Anotherapproach that comes closer to realizing this ideal release profile is the use of an environ-mental stimulus that changes from physiological to pathological conditions to triggerthe release of the drug at the tumor site [76]. Two different mechanisms can be used totrigger drug release. The first is a change in the polymer that destabilizes the core andallows a fast release of the drug. The second is to chemically conjugate the drug to thepolymer using a linker that degrades based on a certain trigger.

There are several different stimuli that can be used to destabilize the polymer core,two of which are temperature and pH. For the temperature stimulus, thermo-responsivepolymers need to be combined with a local hyperthermia treatment at the tumor site.Thermo-responsive polymers such as poly(N-isopropylacrylamide) (PIPAAm) exhibitreversible hydration-dehydration changes in response to small temperature changeswith the transition temperature called the lower critical solution temperature(LCST) [77]. PIPAAm can be combined with a hydrophobic block such as PLA to formP(IPAAm-b-D,L-lactide) with the LCST of the resulting nanoparticle controlled by thePIPAAm block. At low temperatures, the PIPAAm acts as a hydrophilic segment formingthe corona of the nanoparticle. When the temperature increases above the LCST, itbecomes hydrophobic and collapses, distorting the core and resulting in a faster releaserate of the drug. In one study, the polymer was engineered to retain drug at 37°C, butrelease it at an increased rate at 42.5°C, resulting in increased cellular toxicity at the ele-vated temperature [78]. An additional advantage of the local hyperthermia treatment atthe tumor site is that it enhances vascular permeability in the tumor tissue relative tohealthy tissue, resulting in an increase in passive targeting of nanoparticles to the tumorsite [77, 78].

Polymers responsive to pH are also attractive for tumor treatment because the localtumor tissue environment is hypoxic, resulting in a mildly acidic pH of ~6.8 [79, 80].Endosomal and lysosomal cell compartments in the cell are also acidic with a pH of ~5–6[62]. This allows the triggered release of drug in the tumor tissue and after nanoparticleuptake by cells. For these applications, the core polymer should be basic and thenanoparticles should be formulated above the pKa of the protonable group to neutralizethe polymer [62]. When the solution pH decreases, changes in ionization state will causethe polymer to be more hydrophilic and experience electrostatic repulsion, resulting in adestabilization of the core and a triggered release of the drug [81]. One example of thisis poly(L-histidine)-b-PEG (pHis-b-PEG), which has poly(L-histidine) at its core anddestabilizes at pH 7.4. By blending pHis-b-PEG with PLA-b-PEG, the pH of destabilizationcan be decreased to a pH of 6.0–7.2. Several studies with pH-sensitive polymers haveshown pH-responsive triggered release of drugs such as paclitaxel at faster rates than atphysiological conditions, providing a greater initial dose of drug upon entering tumortissue [82, 83]. Temperature and pH can also be used together to improve triggeredrelease in cells or tumors. In one study, a polymer was designed so that the LCST of the


227

nanoparticles dropped below 37°C in acidic environments, triggering the release ofpaclitaxel without the need for hyperthermia treatment [84].

The other approach used to take advantage of differences between physiological andpathological tissue is to conjugate the drug to the polymer though a stimuli-responsivelinker. Several delivery systems have been developed using the acid-labile hydrazonelinkage between the polymer and drug, leading to an accelerated release rates in lowerpH environments [85, 86]. Another example is the use of peptide linkers that aredegradable by specific enzymes present in the tumor environment or in certain cellularcompartments. One example is the use of a short peptide linker degradable by MMP-9, amatrix metalloproteinase (MMP) overexpressed in metastatic tumors [87]. The releaserate was shown to be proportional to the concentration of the enzyme [88].

Another consideration when designing a delivery system is the codelivery of multi-ple drugs using the same nanoparticle. This approach offers several advantages, includ-ing synergistic effects, the ability to suppress drug resistance, and the ability to tune therelative dosage and release rates of various drugs at the level of a single nanoparticle [34].A major obstacle of cancer therapy is the development of multidrug resistance (MDR)in cancer cells [89]. One approach used by researchers to overcome MDR was toencapsulate paclitaxel and ceramide together. Ceramide has been shown to increase thecytotoxic response of cells to antitumor chemotherapeutics [90, 91]. When the twodrugs were codelivered to MDR ovarian cancer cells, the chemo-sensitivity of the cellswas increased. In another study, an anti-angiogenesis drug was codelivered alongwith doxorubicin [70]. The release profiles of the drugs were engineered so thatthe anti-angiogenesis drug was released first to cause vascular collapse and trap thenanoparticle in the tumor tissue. Doxorubicin was then released to kill the tumors. Micereceiving the nanoparticle treatment showed increased survival.

12.5.6 Active Targeting and Ligand Conjugation

There are multiple targeting strategies that can be used to selectively concentratenanoparticles at tumor sites by exploiting differences between normal and malignanttissue. One approach is passive targeting, which is discussed in Section 12.5.1. Anotherapproach is active targeting, where targeting molecules are conjugated to the surface ofnanoparticles to take advantage of molecular recognition events such as ligand-receptoror antibody-antigen interactions. Passive targeting allows nanoparticles to enter tumortissue. Active targeting enhances passive targeting by using ligand interactions with cellsto increase uptake through binding to targeted receptors and endocytosis. Becausenanoparticles have multiple ligands conjugated to the surface, multivalent interactionscan increase the avidity of nanoparticles for surface receptors and enhances the target-ing effectiveness of the ligand [92, 93]. Different ligands can also be conjugated to thesurface to improve targeting since tumor cells typically overexpress multiple types ofsurface receptors, improving selectivity over single-ligand targeting [94].

The challenge with active targeting is to find highly specific and nonimmunogenictargeting molecules. There are many different classes of targeting agents that have beenused for active targeting, including antibodies [95], antibody fragments [32], proteins[96], peptides [97], small molecules [98], carbohydrates [99, 100], and nucleic acids [23].While antibodies and other proteins have been used to successfully target nanoparticles,it also results in increased particle size, complexity and risk of adverse biological


228

reactions. Therefore, it is usually desirable to use small molecule or peptide targeting ifpossible. Peptides and small molecules are also easier to manufacture and, in the case ofpeptides, are less likely to be immunogenic than an antibody or protein.

Covalent conjugation of ligands to the nanoparticle surface is usually preferred tononcovalent interactions because the conjugation is more robust and results inenhanced stability under physiological and pathological conditions. This is achievedthrough succinimidyl ester-amine chemistry or through maleimide-thiol chemistry,both of which were described in Section 12.3.3. Noncovalent strategies include affinityinteractions (streptavidin-biotin) and metal coordination [15].

One issue that can arise with ligand conjugation is the orientation of the moleculeon the nanoparticle surface. For some ligands, such as peptides or aptamers, the conju-gation site can be fixed to ensure the ligand is oriented correctly for interaction with thetarget receptor. However, for antibodies and other proteins, there may be multipleconjugation sites, resulting in multiple orientations, some of which may preventligand-receptor interactions. In this case, a strategy may be required to create a specificconjugation site on the ligand. The heterogeneity of orientations can have an effect onin vivo efficacy, tolerability, and pharmacokinetics [101]. In one study, cysteine residueswere engineered at specific sites in an antibody for drug conjugation. The resulting anti-body-drug complex retained binding and specificity while exhibiting an improvedtherapeutic index.

Another approach for ligand conjugation is to conjugate the ligand to the polymerprior to nanoparticle formation to form a triblock polymer such as PLA-PEG-ligand. Thisallows a one-step targeted nanoparticle synthesis, as opposed to the two-step methodwhere nanoparticles are formed and ligand is subsequently conjugated. For the one-stepmethod, the triblock polymer is mixed with drug or polymer-drug and nanoparticles areformed using nanoprecipitation, yielding a targeted nanoparticle loaded with drugwithout any further modifications. Triblock polymers have been generated with severalligands, including aptamers [36] and small molecules such as folic acid [102]. Thetriblock approach also offers greater control over the ligand surface density andimproved consistency, which is important because ligand density in addition to theligand itself has been shown to have a significant influence on nanoparticlebiodistribution. In a study with aptamer-targeted nanoparticles, it was shown that moreaptamer on the surface increases targeting but also shields the PEG surface. This causesgreater uptake by the MPS and makes the ligand a liability at high densities [36].

One other concept in targeting is the idea of using an adaptable surface to allownanoparticles to overcome multiple barriers for optimal drug delivery. An example ofthis is intracellular targeting, where the drug needs to be released in the cell cytoplasm.There are several examples of cell-penetrating ligands such as peptides and proteins thatallow intracellular delivery of nanoparticles [103]. Intracellular targeting is also possiblethrough surface structure ordering. In one study, it was shown that alternating hydro-philic and hydrophobic domains ordered on gold nanoparticle surfaces resulted in pen-etration of cell membranes, while random organizations of the domains in the sameproportions did not penetrate the cell membranes [104]. However, neither approach isable to target specific cells. One solution to this challenge is the use of an adaptablenanoparticle surface, where one surface is used to overcome a barrier followed by achange in the surface properties based on a stimulus to overcome the next barrier. In onestudy, a targeting ligand was conjugated to a long PEG chain while a cell penetrating


229

peptide was conjugated to a shorter PEG chain. The long PEG chain was attached to thenanoparticle through an acid-labile linker so that the longer PEG chains and targetingligand would be shed after uptake into an endosome, exposing the second ligand forintracellular delivery [105].

12.6 Troubleshooting Tips

This section provides troubleshooting tips for frequently encountered problems. Thetips are summarized in the troubleshooting table.

Troubleshooting Tips

Problem Potential Solution

Polymer conjugation efficiency is low using thecarbodiimide chemistry.

-hydroxysuccinimide (NHS) can be added along withEDC in 10× molar excess. NHS should be prepared freshwith EDC.

Ligand conjugation efficiency is low using themaleimide-thiol conjugation.

Check for free thiol groups on the ligand using Ellman’sReagent. The ratio of Traut’s Reagent can be increased tointroduce more thiol groups on the ligand. If poor conju-gation persists, other cross-linking agents such as SATA orSAT-PEO (Pierce) with spacers can be used to increaseefficiency.

Small aggregate population of nanoparticles areobserved.

Aggregate populations can be removed using a 0.1 or 0.2μm syringe filter. This will result in the loss of some mate-rial as well.

Cells detach from surfaces when washing with PBS. Use cell growth media or Opti-MEM media to wash cells.


There are currently several different polymeric nanoparticle delivery systems in clinicaltrials, although most are nontargeted [51]. The first polymeric nanoparticle deliverysystem to reach Phase II clinical trials in the United States is Genexol-PM [methoxy-PEG-poly(D,L-lactide)Taxol]. The delivery system consists of PLA-PEG with a methoxyend group which encapsulates paclitaxel.

Paclitaxel (Taxol) is a commonly used chemotherapy agent used in the treatment ofseveral different types of cancer, including lung, ovarian, breast cancer, Karposi’s sar-coma, and head and neck malignancies [106]. Because of its water instability, paclitaxelis formulated with the lipid-based solvent Cremophor EL (CrEL). However, CrEL hasbeen shown to cause hypersensitivity reactions and neuropathy. CrEL also significantlyalters the pharmacokinetics of paclitaxel.

Genexol-PM is a polymeric micellar formulation of paclitaxel that alleviates the needfor Cremophor EL. The polymeric formulation improves the water solubility and in vivostability of paclitaxel. The formulation also aids in targeting the drug to tumor tissuethrough passive targeting due to the nanoscale structure (particle size: 20–50 nm) of thepolymeric micelles. Genexol-PM, in preclinical studies with nude mice, was shown tohave a threefold higher maximum tolerated dose (MTD) and the biodistribution showedtwo- to threefold higher levels in a variety of tissues including tumor tissue. In vivoantitumor efficacy was also shown to be greater than that of Taxol [107]. In phase I clini-


230

cal trials, Genexol-PM permitted higher paclitaxel doses than the Taxol formulation andresulted in higher concentrations of paclitaxel in the tumor tissue. The phase I trial iden-tified an MTD of 390 mg/m2 and a recommended dose of 300 mg/m2, which is approxi-mately two times higher than the MTD of CrEL-based paclitaxel [108]. In phase IIstudies, Genexol-PM has shown promising efficacy against metastatic breast cancer,with 75% of patients showing 2 years of overall survival.

12.8 Summary Points

1. The modular design of polymeric nanoparticles allows a significant amount offlexibility in the engineering of a delivery system for a specific application.

2. Encapsulation of hydrophobic and hydrophilic chemotherapeutic drugs innanoparticles is possible either through physical entrapment or chemicalconjugation.

3. Drug release can be controlled through diffusion, polymer degradation, or externalstimuli such as pH and temperature.

4. Nanoparticle targeting strategies consist of both passive (dependent on size andsurface chemistry) and active (dependent on molecular interactions) approaches.

5. Ligands for active targeting of nanoparticles include antibodies, peptides, smallmolecules, carbohydrates, nucleic acids, and surface morphology.

6. Evaluation of the targeting ability and cytotoxicity of the nanoparticle deliverysystem is necessary in an in vitro cell model that expresses the targeted receptor.

Acknowledgements

This work was supported by National Institute of Health Grants CA119349 andEB003647 and a Koch-Prostate Cancer Foundation Award in Nanotherapeutics. EMP issupported by a National Defense Science and Engineering Graduate Fellowship(NDSEG).

References

[1] National Cancer Institute Cancer Nanotechnology Plan, http://nano.cancer.gov/about_alli-ance/cancer_nanotechnology_plan.asp.

[2] Cheng, J., et al., “Formulation of functionalized PLGA-PEG nanoparticles for in vivo targeted drugdelivery,” Biomaterials, Vol. 28, No. 5, 2007, pp. 869–876.

[3] Greish, K., “Enhanced permeability and retention of macromolecular drugs in solid tumors: a royalgate for targeted anticancer nanomedicines,” J. Drug Target, Vol. 15, No. 7-8, 2007, pp. 457–464.

[4] Jain, R.K., “Delivery of molecular and cellular medicine to solid tumors,” Adv. Drug Deliv. Rev., Vol.46, No. 1-3, 2001, pp. 149–168.

[5] Brigger, I., et al., “Nanoparticles in cancer therapy and diagnosis,” Adv. Drug Deliv. Rev., Vol. 54,No. 5, 2002, pp. 631–651.

[6] Gref, R., et al., “Biodegradable long-circulating polymeric nanospheres,” Science, Vol. 263, No.5153, 1994, pp. 1600–1603.

[7] Moghimi, S.M., et al., “Long-circulating and target-specific nanoparticles: theory to practice,”Pharmocol. Rev., Vol. 53, 2001, pp. 283–318.

[8] Langer, R., “Drug delivery and targeting,” Nature, Vol. 392, No. 6679, Suppl., 1998, pp. 5–10.

12.8 Summary Points

231

[9] Moghimi, S.M., “Recent developments in polymeric nanoparticle engineering and their applica-tions in experimental and clinical oncology,” Anti-Cancer Agents Med. Chem., Vol. 6, 2006, pp.553–561.

[10] Pridgen, E.M., et al., “Biodegradable, polymeric nanoparticle delivery systems for cancer therapy,”Nanomed., Vol. 2, No. 5, 2007, pp. 669–680.

[11] LaVan, D.A., et al., “Small-scale systems for in vivo drug delivery,” Nat. Biotechnol., Vol. 21, No. 10,2003, pp. 1184–1191.

[12] Kim, K., et al., “Cell-permeable and biocompatible polymeric nanoparticles for apoptosis imag-ing,” J. Am. Chem. Soc., Vol. 128, No. 11, 2006, pp. 3490–3491.

[13] Montet, X., et al., “Imaging pancreatic cancer with a peptide-nanoparticle conjugate targeted tonormal pancreas,” Bioconjug. Chem., Vol. 17, No. 4, 2006, pp. 905–911.

[14] Nasongkla, N., et al., “Multifunctional polymeric micelles as cancer-targeted, MRI-ultrasensitivedrug delivery systems,” Nano Lett., Vol. 6, No. 11, 2006, pp. 2427–2430.

[15] Farokhzad, O.C., et al., “Aptamers and cancer nanotechnology,” in Nanotechnology for Cancer Ther-apy, M.M. Amiji, (ed.), Boca Raton, FL: CRC Press, 2007.

[16] Shive, M.S., and J.M. Anderson, “Biodegradation and biocompatibility of PLA and PLGAmicrospheres,” Adv. Drug Deliv. Rev., Vol. 28, No. 1, 1997, pp. 5–24.

[17] Saunders, R.A., and E.M. Helveston, “Coated Vicryl (polyglactin 910) suture in extraocular musclesurgery,” Ophthalmic Surg., Vol. 10, No. 7, 1979, pp. 13–18.

[18] Guerin, C., et al., “Recent advances in brain tumor therapy: local intracerebral drug delivery bypolymers,” Invest. New Drugs, Vol. 22, No. 1, 2004, pp. 27–37.

[19] Bazile, D., et al., “Stealth Me.PEG-PLA nanoparticles avoid uptake by the mononuclear phagocytessystem,” J. Pharm. Sci., Vol. 84, No. 4, 1995, pp. 493–498.

[20] Ben-Shabat, S., et al., “PEG-PLA block copolymer as potential drug carrier: preparation and charac-terization,” Macromol. Biosci., Vol. 6, No. 12, 2006, pp. 1019–1025.

[21] Scott, M.D., and K.L. Murad, “Cellular camouflage: fooling the immune system with polymers,”Curr. Pharm. Des., Vol. 4, No. 6, 1998, pp. 423–438.

[22] Hermanson, G.T., Bioconjugate Techniques, 2nd ed., London, U.K.: Academic Press, 2008.[23] Farokhzad, O.C., et al., “Targeted nanoparticle-aptamer bioconjugates for cancer chemotherapy in

vivo,” Proc. Natl. Acad. Sci. USA, Vol. 103, No. 16, 2006, pp. 6315–6320.[24] Lemoine, D., and V. Preat, “Polymeric nanoparticles as delivery system for influenza virus

glycoproteins,” J. Control Release, Vol. 54, No. 1, 1998, pp. 15–27.[25] Mu, L., and S.S. Feng, “A novel controlled release formulation for the anticancer drug paclitaxel

(Taxol): PLGA nanoparticles containing vitamin E TPGS,” J. Control Release, Vol. 86, No. 1, 2003,pp. 33–48.

[26] Fessi, H., et al., “Nanocapsule formation by interfacial polymer deposition following solvent dis-placement,” Int. J. Pharm., Vol. 55, 1989, pp. R1–R4.

[27] Cohen-Sela, E., et al., “Single and double emulsion manufacturing techniques of an amphiphilicdrug in PLGA nanoparticles: formulations of mithramycin and bioactivity,” J. Pharm. Sci., 2008.

[28] Yamamoto, Y., et al., “Long-circulating poly(ethylene glycol)-poly(D,L-lactide) block copolymermicelles with modulated surface charge,” J. Control Release, Vol. 77, No. 1-2, 2001, pp. 27–38.

[29] Simberg, D., et al., “Biomimetic amplification of nanoparticle homing to tumors,” Proc. Natl. Acad.Sci. USA, Vol. 104, No. 3, 2007, pp. 932–936.

[30] Alexis, F., et al., “HER-2-targeted nanoparticle-affibody bioconjugates for cancer therapy,” Chem.Med. Chem., Vol. 3, No. 12, 2008, pp. 1839–1843.

[31] Gindy, M.E., et al., “Preparation of poly(ethylene glycol) protected nanoparticles with variablebioconjugate ligand density,” Biomacromolecules, Vol. 9, No. 10, 2008, pp. 2705–2711.

[32] Sugano, M., et al., “Antibody targeting of doxorubicin-loaded liposomes suppresses the growth andmetastatic spread of established human lung tumor xenografts in severe combinedimmunodeficient mice,” Cancer Res., Vol. 60, No. 24, 2000, pp. 6942–6949.

[33] Abdelwahed, W., et al., “Freeze-drying of nanoparticles: formulation, process and storage consider-ations,” Adv. Drug Deliv. Rev., Vol. 58, No. 15, 2006, pp. 1688–1713.

[34] Zhang, L., et al., “Co-delivery of hydrophobic and hydrophilic drugs from nanoparticle-aptamerbioconjugates,” Chem. Med. Chem., Vol. 2, No. 9, 2007, pp. 1268–1271.

[35] Farokhzad, O.C., et al., “Nanoparticle-aptamer bioconjugates: a new approach for targeting pros-tate cancer cells,” Cancer Res., Vol. 64, No. 21, 2004, pp. 7668–7672.

[36] Gu, F., et al., “Precise engineering of targeted nanoparticles by using self-assembled biointegratedblock copolymers,” Proc. Natl. Acad. Sci. USA, Vol. 105, No. 7, 2008, pp. 2586–2591.

[37] Green, J.J., et al., “Nanoparticles for gene transfer to human embryonic stem cell colonies,” NanoLett., Vol. 8, No. 10, 2008, pp. 3126–3130.


232

[38] Dhar, S., et al., “Targeted delivery of cisplatin to prostate cancer cells by aptamer functionalizedPt(IV) prodrug-PLGA-PEG nanoparticles,” Proc. Natl. Acad. Sci. USA, Vol. 105, No. 45, 2008,pp. 17356–17361.

[39] van Vlerken, L.E., et al., “Modulation of intracellular ceramide using polymeric nanoparticles toovercome multidrug resistance in cancer,” Cancer Res., Vol. 67, No. 10, 2007, pp. 4843–4850.

[40] Maeda, H., “The enhanced permeability and retention (EPR) effect in tumor vasculature: the keyrole of tumor-selective macromolecular drug targeting,” Adv. Enzyme Regul., Vol. 41, 2001,pp. 89–207.

[41] Fang, C., et al., “In vivo tumor targeting of tumor necrosis factor-alpha-loaded stealthnanoparticles: effect of MePEG molecular weight and particle size,” Eur. J. Pharm. Sci., Vol. 27,No. 1, 2006, pp. 27–36.

[42] Nagayama, S., et al., “Time-dependent changes in opsonin amount associated on nanoparticlesalter their hepatic uptake characteristics,” Int. J. Pharm., Vol. 342, No. 1-2, 2007, pp. 215–221.

[43] Venturoli, D., and B. Rippe, “Ficoll and dextran vs. globular proteins as probes for testing glomeru-lar permselectivity: effects of molecular size, shape, charge, and deformability,” Am. J. Physiol.Renal Physiol., Vol. 288, No. 4, 2005, pp. F605–F613.

[44] Karnik, R., et al., “Microfluidic platform for controlled synthesis of polymeric nanoparticles,” NanoLett., Vol. 8, No. 9, 2008, pp. 2906-12.

[45] Astete, C.E., and C.M. Sabliov, “Synthesis and characterization of PLGA nanoparticles,” J. Biomater.Sci. Polymer Edn., Vol. 17, No. 3, 2006, pp. 247–289.

[46] Duan, Y., et al., “Optimization of preparation of DHAQ-loaded PEG-PLGA-PEG nonaparticlesusing central composite design,” J. Mater. Sci.: Mater. Med., Vol. 17, 2006, pp. 559–563.

[47] Sahoo, S.K., et al., “Residual polyvinyl alcohol associated with poly (D,L-lactide-co-glycolide)nanoparticles affects their physical properties and cellular uptake,” J. Control Release, Vol. 82, No. 1,2002, pp. 105–114.

[48] Geng, Y., et al., “Shape effects of filaments versus spherical particles in flow and drug delivery,”Nat. Nanotechnol., Vol. 2, No. 4, 2007, pp. 249–255.

[49] Discher, B.M., et al., “Polymersomes: tough vesicles made from diblock copolymers,” Science,Vol. 284, No. 5417, 1999, pp. 1143–1146.

[50] Discher, D.E., and A. Eisenberg, “Polymer vesicles,” Science, Vol. 297, No. 5583, 2002, pp. 967–973.[51] Alexis, F., et al., “Factors affecting the clearance and biodistribution of polymeric nanoparticles,”

Mol. Pharm., Vol. 5, No. 4, 2008, pp. 505–515.[52] Owens, D.E., 3rd, and N.A. Peppas, “Opsonization, biodistribution, and pharmacokinetics of poly-

meric nanoparticles,” Int. J. Pharm., Vol. 307, No. 1, 2006, pp. 93–102.[53] Romberg, B., et al., “Sheddable coatings for long-circulating nanoparticles,” Pharm. Res., Vol. 25,

No. 1, 2008, pp. 55–71.[54] Vittaz, M., et al., “Effect of PEO surface density on long-circulating PLA-PEO nanoparticles which

are very low complement activators,” Biomaterials, Vol. 17, No. 16, 1996, pp. 1575–1581.[55] Beletsi, A., et al., “Biodistribution properties of nanoparticles based on mixtures of PLGA with

PLGA-PEG diblock copolymers,” Int. J. Pharm., Vol. 298, No. 1, 2005, pp. 233–241.[56] Panagi, Z., et al., “Effect of dose on the biodistribution and pharmacokinetics of PLGA and

PLGA-mPEG nanoparticles,” Int. J. Pharm., Vol. 221, No. 1-2, 2001, pp. 143–152.[57] Gref, R., et al., “‘Stealth’ corona-core nanoparticles surface modified by polyethylene glycol (PEG):

influences of the corona (PEG chain length and surface density) and of the core composition onphagocytic uptake and plasma protein adsoprtion,” Colloids and Surfaces B: Biointerfaces, Vol. 18,2000, pp. 301–313.

[58] Mosqueira, V.C., et al., “Biodistribution of long-circulating PEG-grafted nanocapsules in mice:effects of PEG chain length and density,” Pharm. Res., Vol. 18, No. 10, 2001, pp. 1411–1419.

[59] Dong, Y., and S.S. Feng, “Methoxy poly(ethylene glycol)-poly(lactide) (MPEG-PLA) nanoparticlesfor controlled delivery of anticancer drugs,” Biomaterials, Vol. 25, No. 14, 2004, pp. 2843–2849.

[60] Fonseca, C., et al., “Paclitaxel-loaded PLGA nanoparticles: preparation, physicochemical character-ization and in vitro anti-tumoral activity,” J. Control Release, Vol. 83, No. 2, 2002, pp. 273–286.

[61] Allen, C., et al., “Nano-engineering block copolymer aggregates for drug delivery,” Colloids and Sur-faces B: Biointerfaces, Vol. 16, 1999, pp. 3–27.

[62] Rijcken, C.J., et al., “Triggered destabilisation of polymeric micelles and vesicles by changing poly-mers polarity: an attractive tool for drug delivery,” J. Control Release, Vol. 120, No. 3, 2007,pp. 131–148.

[63] Deng, J.S., et al., “In vitro characterization of polyorthoester microparticles containingbupivacaine,” Pharm. Dev. Technol., Vol. 8, No. 1, 2003, pp. 31–38.

[64] Molpeceres, J., et al., “A polycaprolactone nanoparticle formulation of cyclosporin-A improves theprediction of area under the curve using a limited sampling strategy,” Int. J. Pharm., Vol. 187, No. 1,1999, pp. 101–113.

References

233

[65] Sommerfeld, P., et al., “Long-term stability of PBCA nanoparticle suspensions,” J. Microencapsul.,Vol. 17, No. 1, 2000, pp. 69–79.

[66] Gao, J., et al., “Surface modification of polyanhydride microspheres,” J. Pharm. Sci., Vol. 87, No. 2,1998, pp. 246–248.

[67] Huang, G., et al., “Controlled drug release from hydrogel nanoparticle networks,” J. Control Release,Vol. 94, No. 2-3, 2004, pp. 303–311.

[68] Cohen-Sela, E., et al., “A new double emulsion solvent diffusion technique for encapsulatinghydrophilic molecules in PLGA nanoparticles,” J. Control Release, Vol. 133, No. 2, 2009, pp. 90–95.

[69] Govender, T., et al., “PLGA nanoparticles prepared by nanoprecipitation: drug loading and releasestudies of a water soluble drug,” J. Control Release, Vol. 57, No. 2, 1999, pp. 171–185.

[70] Sengupta, S., et al., “Temporal targeting of tumour cells and neovasculature with a nanoscale deliv-ery system,” Nature, Vol. 436, No. 7050, 2005, pp. 568–572.

[71] Tong, R., and J. Cheng, “Paclitaxel-initiated, controlled polymerization of lactide for the formula-tion of polymeric nanoparticulate delivery vehicles,” Angew Chem. Int. Ed. Engl., Vol. 47, No. 26,2008, pp. 4830–4834.

[72] Langer, R., “Polymeric delivery systems for controlled drug release,” Chem. Eng. Commun., Vol. 6,1980, pp. 15–48.

[73] Siepmann, J., and A. Gopferich, “Mathematical modeling of bioerodible, polymeric drug deliverysystems,” Adv. Drug Deliv. Rev., Vol. 48, No. 2-3, 2001, pp. 229–247.

[74] Avgoustakis, K., et al., “PLGA-mPEG nanoparticles of cisplatin: in vitro nanoparticle degradation,in vitro drug release and in vivo drug residence in blood properties,” J. Control Release, Vol. 79, No.1-3, 2002, pp. 123–135.

[75] Matsusue, Y., et al., “Tissue reaction of bioabsorbable ultra high strength poly (L-lactide) rod. Along-term study in rabbits,” Clin. Orthop. Relat. Res., No. 317, 1995, pp. 246–253.

[76] Schmaljohann, D., “Thermo- and pH-responsive polymers in drug delivery,” Adv. Drug Deliv. Rev.,Vol. 58, No. 15, 2006, pp. 1655–1670.

[77] Chung, J.E., et al., “Inner core segment design for drug delivery control of thermo-responsive poly-meric micelles,” J. Control Release, Vol. 65, No. 1-2, 2000, pp. 93–103.

[78] Kohori, F., et al., “Control of adriamycin cytotoxic activity using thermally responsive polymericmicelles composed of poly(N-isopropylacrylamide-co-N,N-dimethylacrylamide)-b-(poly(D,L-lactide),” Colloids and Surfaces B: Biointerfaces, Vol. 16, 1999, pp. 195–205.

[79] Brown, J.M., and W.R. Wilson, “Exploiting tumour hypoxia in cancer treatment,” Nat. Rev. Cancer,Vol. 4, No. 6, 2004, pp. 437–447.

[80] Engin, K., et al., “Extracellular pH distribution in human tumours,” Int. J. Hypertherm., Vol. 11,1995, pp. 211–216.

[81] Shenoy, D., et al., “Poly(ethylene oxide)-modified poly(beta-amino ester) nanoparticles as apH-sensitive system for tumor-targeted delivery of hydrophobic drugs. 1. In vitro evaluations,”Mol. Pharm., Vol. 2, No. 5, 2005, pp. 357–366.

[82] Lee, E.S., et al., “Polymeric micelle for tumor pH and folate-mediated targeting,” J. Control Release,Vol. 91, No. 1-2, 2003, pp. 103–113.

[83] Lee, E.S., et al., “Poly(L-histidine)-PEG block copolymer micelles and pH-induced destabilization,”J. Control Release, Vol. 90, No. 3, 2003, pp. 363–374.

[84] Seow, W.Y., et al., “Targeted and intracellular delivery of paclitaxel using multi-functional poly-meric micelles,” Biomaterials, Vol. 28, No. 9, 2007, pp. 1730–1740.

[85] Bae, Y., et al., “Preparation and biological characterization of polymeric micelle drug carriers withintracellular pH-triggered drug release property: tumor permeability, controlled subcellular drugdistribution, and enhanced in vivo antitumor efficacy,” Bioconjug. Chem., Vol. 16, No. 1, 2005,pp. 122–130.

[86] Hruby, M., et al., “Polymeric micellar pH-sensitive drug delivery system for doxorubicin,” J. ControlRelease, Vol. 103, 2005, pp. 137–148.

[87] Sarkar, N., et al., “Matrix metalloproteinase-assisted triggered release of liposomal contents,”Bioconjug. Chem., Vol. 19, No. 1, 2008, pp. 57–64.

[88] Elegbede, A.I., et al., “Mechanistic studies of the triggered release of liposomal contents by matrixmetalloproteinase-9,” J. Am. Chem. Soc., Vol. 130, No. 32, 2008, pp. 10633–10642.

[89] Bradley, G., et al., “Mechanism of multidrug resistance,” Biochim. Biophys. Acta, Vol. 948, No. 1,1988, pp. 87–128.

[90] Kolesnick, R., “The therapeutic potential of modulating the ceramide/sphingomyelin pathway,” J.Clin. Invest., Vol. 110, No. 1, 2002, pp. 3–8.

[91] Senchenkov, A., et al., “Targeting ceramide metabolism—a strategy for overcoming drug resis-tance,” J. Natl. Cancer Inst., Vol. 93, No. 5, 2001, pp. 347–357.

[92] Hong, S., et al., “The binding avidity of a nanoparticle-based multivalent targeted drug deliveryplatform,” Chem. Biol., Vol. 14, No. 1, 2007, pp. 107–115.


234

[93] Montet, X., et al., “Multivalent effects of RGD peptides obtained by nanoparticle display,” J. Med.Chem., Vol. 49, No. 20, 2006, pp. 6087–6093.

[94] Saul, J.M., et al., “A dual-ligand approach for enhancing targeting selectivity of therapeuticnanocarriers,” J. Control Release, Vol. 114, No. 3, 2006, pp. 277–287.

[95] Steinhauser, I., et al., “Trastuzumab-modified nanoparticles: optimisation of preparation anduptake in cancer cells,” Biomaterials, Vol. 27, No. 28, 2006, pp. 4975–4983.

[96] Xu, Z., et al., “In vitro and in vivo evaluation of actively targetable nanoparticles for paclitaxeldelivery,” Int. J. Pharm., Vol. 288, No. 2, 2005, pp. 361–368.

[97] Murphy, E.A., et al., “Nanoparticle-mediated drug delivery to tumor vasculature suppresses metas-tasis,” Proc. Natl. Acad. Sci. USA, Vol. 105, No. 27, 2008, pp. 9343–9348.

[98] Yoo, H.S., and T.G. Park, “Folate receptor targeted biodegradable polymeric doxorubicin micelles,”J. Control Release, Vol. 96, No. 2, 2004, pp. 273–283.

[99] Irache, J.M., et al., “Mannose-targeted systems for the delivery of therapeutics,” Expert Opin. DrugDeliv., Vol. 5, No. 6, 2008, pp. 703–724.

[100] Kim, I.S., and S.H. Kim, “Development of polymeric nanoparticulate drug delivery systems: evalua-tion of nanoparticles based on biotinylated poly(ethylene glycol) with sugar moiety,” Int. J. Pharm.,Vol. 257, No. 1-2, 2003, pp. 195–203.

[101] Junutula, J.R., et al., “Site-specific conjugation of a cytotoxic drug to an antibody improves thetherapeutic index,” Nat. Biotechnol., Vol. 26, No. 8, 2008, pp. 925–932.

[102] Kim, S.H., et al., “Target-specific cellular uptake of PLGA nanoparticles coated withpoly(L-lysine)-poly(ethylene glycol)-folate conjugate,” Langmuir, Vol. 21, No. 19, 2005,pp. 8852–8857.

[103] Torchilin, V.P., “Cell penetrating peptide-modified pharmaceutical nanocarriers for intracellulardrug and gene delivery,” Biopolymers, Vol. 90, No. 5, 2008, pp. 604–610.

[104] Verma, A., et al., “Surface-structure-regulated cell-membrane penetration by monolayer-protectednanoparticles,” Nat. Mater., Vol. 7, No. 7, 2008, pp. 588–595.

[105] Sawant, R.M., et al., “‘SMART’ drug delivery systems: double-targeted pH-responsive pharmaceuti-cal nanocarriers,” Bioconjug. Chem., Vol. 17, No. 4, 2006, pp. 943–949.

[106] Kim, S.C., et al., “In vivo evaluation of polymeric micellar paclitaxel formulation: toxicity and effi-cacy,” J. Control Release, Vol. 72, No. 1-3, 2001, pp. 191–202.

[107] Kim, T.Y., et al., “Phase I and pharmacokinetic study of Genexol-PM, a cremophor-free, polymericmicelle-formulated paclitaxel, in patients with advanced malignancies,” Clin. Cancer Res., Vol. 10,No. 11 2004, pp. 3708–3716.

[108] Lee, K.S., et al., “Multicenter phase II trial of Genexol-PM, a Cremophor-free, polymeric micelle for-mulation of paclitaxel, in patients with metastatic breast cancer,” Breast Cancer Res. Treat., Vol.108, No. 2, 2008, pp. 241–250.

References

235

C H A P T E R

1 3Porous Silicon Particles for Multistage Delivery

Ennio Tasciotti,1* Jonathan Martinez,1 Ciro Chiappini,2 Rohan Bhavane,1 and MauroFerrari1,2,3,4

1The Division of Nanomedicine, Department of Biomedical Engineering, The University of Texas HealthScience Center at Houston, Houston, TX 77030

2Department of Biomedical Engineering, The University of Texas, Austin, TX 770303Department of Experimental Therapeutics, The University of Texas MD Anderson Cancer Center,Houston, TX 77030

4Department of Bioengineering, Rice University, Houston, TX 77005

*Corresponding author: Ennio Tasciotti, Ph.D., Assistant Professor, Department of Nanomedicine andBiomedical Engineering, Division of Nanomedicine, Institute for Molecular Medicine, 1825 Pressler Street,Suite 537B, Houston, TX 77030; e-mail: [email protected], Phone: (713) 500-2468, Fax: (713)500-2462

237

Abstract

In this chapter, we present a novel multistage delivery system (MDS) to addressthe inherent complexity involved in drug delivery. The proposed system hasthe potential to revolutionize the delivery of therapeutics at target lesions bydistributing the tasks of biobarrier avoidance, targeting, and therapeutic effectamong different vector stages. The first-stage vector of this MDS system is amicrofabricated nanoporous silicon particle with tailored chemo-physical andgeometrical properties. The subsequent stages can be selected among a widerange of nano-sized carriers or therapeutics. Employing the MDS, investigatorscan concentrate on synthesizing novel and innovative therapies disregardingthe issues of targeting and biobarrier avoidance that will be addressed by thefirst stage of the MDS.

Key terms multistageporous silicondrug deliveryimagingtunable porosity/pore sizenanovectorsbiodegradablebiobarrier avoidance

13.1 Introduction

The detection of trace markers in clinical samples and the localization of carriers to dis-eased body sites are the ultimate goals for effective disease diagnosis/prevention, andtreatment [1, 2]. The ability to obtain sensitive data in a noninvasive manner and to con-centrate therapeutic compounds at the target sites are among the most crucial, break-through applications currently needed in the clinic. Over the last three to five decades,cancer treatment has relied upon surgical removal of the primary tumor, followed by theuse of radiation, and then repeated cycles of the maximum-tolerated doses of a combi-nation of cytotoxic chemotherapeutic agents. Unfortunately, the vast majority ofmalignancies have proven to be resistant to this type of chemotherapeutic intervention,partially due to the requisite dose limitations for preventing adverse effects on nor-mal tissues. Conventional cancer chemotherapeutics gain access to the blood streamthrough intravenous administration and are required to penetrate the extravascularspace in order to present the drug at an adequate concentration such to inflict lethal tox-icity to the tumor lesion. Even the best injectable drug to date retains its specificity ofaction only through its molecular affinity for the ultimate therapeutic substrate whileremaining completely indifferent to its own distribution within the body. Effectivecancer therapy continues to present the drug delivery conundrum of right treatment,right cell, and right dose, with minimal collateral damage. Additionally, 40% of newanticancer compounds fail to enter clinical trials due to solubility or systemic toxicityissues. Despite advances in drug discovery, the transition to the clinical setting remainschallenged by the inability to efficiently deliver the right compound to the best in vivotarget. To address this issue, a plethora of different vectors have been proposed as theideal candidates to the time-honored problem of optimizing the therapeutic index fortreatment (i.e., to maximize efficacy, while reducing health-adverse side effects).

To provide effective drug delivery, the carriers must be capable of reaching and rec-ognizing their target site. Nanomaterial characteristic size, close to that of cell compo-nents, allowed the development of tools capable of interfacing directly with the pillarconstituents of life: nucleic acids, proteins, and biological molecules. Thanks to theseunique features, nanotechnology and the nanotechnology toolset hold great promisesin the field of drug delivery and have the potential to revolutionize this research area,enabling a paradigmatic shift from molecularly targeted therapeutics to cell or sitedirected therapeutics [3]. As a result, the drug diffuses without differentiation amongall the body tissues, becoming activated at nonspecific sites and generating adverseside effects, thus lowering the therapeutic index [4]. Among the various classes ofnanoparticles (NP) developed for drug delivery (dendrimers, liposomes, nanotubes, andso forth), very few are amenable for the optimizations (surface modification, targeting,surface stealthing, and particle size/shape) required to obtain an individualized deliverystrategy and to improve their efficacy. In order to maintain their therapeutic level, thecarriers must be able to efficiently negotiate the biobarriers from the point of entry to thetarget. Hemorheology [5], Reticulo-Endothelial System (RES) cells [6], thrombocytes anderythrocytes [7], attack by lytic enzymes [8], crossing of the endothelial wall [9] or bloodbrain barrier (BBB) [10], diffusion in the perivascular tissue against the interstitial andosmotic pressure [11, 12], and entry into the cell cytoplasm through the cell membrane[13] constitute only part of the many sequential biobarriers that stand between the car-rier and its target site. These mechanisms, intended to oppose harmful entities, do notdiscriminate between potentially beneficial delivery vectors and harmful foreign bodies.

Porous Silicon Particles for Multistage Delivery

238

As a consequence, these barriers pose as insurmountable obstacles for any prototypicaldrug or nano-therapeutic to overcome [14]. Of the vast and diverse array of NPs devel-oped in laboratories, only a handful have made their way to the clinic [15]. Thisshortcoming can be traced to the inability to develop a NP capable of sequentiallynegotiating all the biological barriers in an effective manner. As an example, surfacefunctionalization with poly-ethyleneglycol (PEG) prevents particles from being rapidlyscavenged by the RES [16], but also limits their ability to be recognized and internalizedby the target cells. Recently, the concept of NP engineering has revolutionized the battleagainst the biological barriers. For example, appropriate engineering of NP size andshape allows them to reach tumor sites exploiting the enhanced permeability and reten-tion (EPR) effect (passive targeting) [17]; localization at a lesion site can also be activelysought conjugating targeting molecules chosen from a vast array of antibodies, ligands,peptides, aptamers, or phages [18–24]. Fusion with the cell membrane can be facili-tated by conjugation of the NPs with cell penetrating peptides [25], and release fromlysosomes can be triggered by chemical sensors on the NPs [26]. A successful NP mustthen be endowed with multiple, and often conflicting, functions. In most, if not allcases, it is practically impossible to provide a single NP with all the necessary tools toachieve these goals [27, 28].

The current pharmaceutical and biotechnological paradigm for a successful thera-peutic is to embody in it three critical functions: cytotoxic action, biological recogni-tion, and avoidance of biological barriers. An elegant solution to this challenge is a newclass of drug delivery vectors in which these essential tasks have been distributed amonga larger number of components within a coordinated system. The components can beassembled ex vivo with the objective of addressing the biological barriers in a multi-plexed, sequential, independent yet synergistic way. Our laboratory recently proposed anew class of nanotechnology delivery vectors and a new approach called the MultistageDelivery Systems (MDS) [29] (Figure 13.1). This new vision decouples the tasks requiredfrom therapeutics, assigning to the first-stage delivery vectors the role of biobarrieravoidance, first-order localization, and the conventional biorecognition modalities. Thesecond-stage NPs, nested within the first stage, are endowed with the ability of penetrat-ing into the lesion exploiting EPR or with the help of permeation enhancers andthus selectively directing their cytotoxic payload against target cells and tissues. Thefirst-stage particle’s engineered geometry assists in protecting the second-stage NPswhile the system navigates the vasculature. Once the first-stage particles reach the finaldocking place on the vascular endothelium, adjacent to the target site, the second-stageNPs are released and diffuse in the perivascular tissue, where they can accomplish theirfinal tasks. The nanoporous silicon particles (PSP) constitute the primary vector of theMDS. These first-stage PSPs are biodegradable and biocompatible [30], and their size,shape, porosity, and pore size can be finely tuned during the manufacturing processes[31–33]. The chemo-physical properties of these PSPs are tailored according to rationaldesign guided by a strong mathematical toolset, in order to obtain the desired function-ality. Mathematical modeling of the effects of the geometry on the vascular navigationbehavior of micro and NPs provides optimal solutions for biodistribution, adhesion, andendocytosis [34]. The vast repertoire of chemical functionalizations achievable on sili-con surfaces enables the PSPs to address and overcome some of the aforementioned bio-logical barriers [35–37]. Finally, the release kinetics of the second-stage NPs can be linked

13.1 Introduction

239

to the degradation rates of the first stage PSPs through the adjustment of porosity, poresize, and pore distribution on the silicon carrier [38].

Different approaches can be employed to capture, enclose, and carry the intendedfunctional cargoes, and recently, alternative multifunctional systems for the delivery ofbiological agents have been proposed. These systems take advantage of bacterial strains[39, 40], T lymphocytes [41], and phages [42] to incorporate or target agents andnanoparticles to induce a cytotoxic effect on selected cell types.

Ferrite oxide particles have been enveloped in the cytoplasmic bacterial wall andmodified with polyethylenimine (PEI), a proven effective gene carrier, to improve the


240

(c)

(b)

(a)

Figure 13.1 Schematic depicting the process of the multistage delivery system. (a) The MDS is firstassembled by loading second-stage NPs into the pores of the PSPs. (b) After i.v. injection, these ratio-nally designed PSPs travel through the blood stream and due to their size, shape, and surface modifica-tions avoid RES uptake and finally migrate to the vessel wall where they can adhere to the target’sendothelium (c) Once docked, the PSPs release their payload (second-stage NPs), which will penetratethrough the natural fenestrations of the target’s vasculature and eventually diffuse into the tissue,where they will be taken up and accomplish their final task. (Reproduced with permission from [29]courtesy of Nature Publishing Group.)

transfer of genetic material into target cells. The authors refer to this multifunctionalsystem as bacterial magnetic particles-PEI (BMP-PEI) complexes [40].

The “microbots” (Figure 13.2) are a multistage delivery system that exploits live bac-teria to mediate the delivery of bioactive agents into cells [39]. The choice of bacterialstrains with particular physiological properties allowed for the development of distinctdelivery applications. In this system the cargo is conjugated on the surface of the bacte-ria, rather than being loaded inside the cytoplasm. This type of conjugation avoids bac-terial disruption in order to take advantage of the natural tropism of the bacteria to thehost tissues. However, both these methods are limited in the type of nanoparticles thatcan be effectively delivered. The BMP-PEI system allows for the delivery of DNA and islimited by its lack of flexibility towards the delivery of other nano-agents. Microbotscould not deliver more than an average of 22 200-nm particles per cell, while our sili-con-based MDS method can deliver an average of 20 microparticles per cell, each loadedwith thousands of NPs. Moreover, the microbots, when not immediately internalized bycells, expose the conjugated nanoparticles to the hostile environment within thevasculature, while the MDS protects the payload inside the nanopores. Lastly, thefunctionalization leading to the conjugation of nanoparticles prevents further decora-tion of the bacteria with targeting moieties, rendering it unfit for intravascular deliveryand susceptible to nonspecific interaction with the biological barriers.

13.1 Introduction

241

Bacteria

Figure 13.2 Schematic of a microbot. The nanoparticles, labeled with imaging moieties, are conju-gated on the surface of the bacteria. (Adapted from [39].)

In another MDS, the transport of therapeutic agents is mediated by T lymphocytes[41]. In this system, lymphocytes would be isolated from the patient and then incubatedor electroporated with desired nanoparticles. The cargo is efficiently protected from thebiobarriers once reintroduced into the patient (Figure 13.3). This type of system allowsfor localized drug targeting and detection of metastases and could possibly be combinedwith existing immunotherapies. The method, however, is greatly limited by the incon-sistent and scarcely controllable loading and release kinetics of the nanoparticles. Inaddition, the nanoparticles might have a detrimental effect on the targeting ability ofthe lymphocytes.

The last class of MDS is a network of bacteriophage and gold nanoparticles(Figure 13.4) [42]. The phages are engineered in such a way that each phage displays apeptide. This peptide can be selected among a huge combinatorial library in order to tar-get a specific receptor expressed, for example, on the surface of an endothelial cells.These Au-phage complexes can be designed to specifically target cells for imaging orthermal ablation purposes. Nevertheless, this network of nanoparticles may still bevulnerable to biobarriers due to their overall size, and its best use would probably be asa payload or as a targeting moiety embedded or attached into a larger multifunctionalsystem.

The success of any MDS relies on the ability to accurately engineer its components,decorate its surface, and govern the loading and release of the various stages. In this


242

(a) (b)

(c) (d)

Nanoparticle

Surface Ligand

T lymphocite

Cancer cell

Figure 13.3 (a–d) Schematic of T-lymphocytic delivery of loaded nanoparticles to cancer cells.

sense, the physical (geometry) and chemical (surface functionalizations) features of thefirst-stage vectors are critical to improve the efficiency of any in vivo applications. In thefollowing sections, we will outline a minimal array of methods to address the aforemen-tioned tasks and methodically integrate the MDS system for any application that may beenvisioned. Combining state-of-the-art microelectronics technology with finely con-trolled electrochemical etch, we developed protocols for the high-throughput manufac-turing of highly reproducible PSPs: the first-stage vector (Figure 13.5). The objective ofthis chapter is to succinctly outline and describe validated techniques and proto-cols enabling the use of the MDS and to allow the successful reproduction of theresults obtained. First, it is necessary to describe the typical protocols employed tomicrofabricate the first stage vectors (Figure 13.6). Briefly, in order to produce PSPs of awell-determined shape, desired pore size, and porosity, a silicon wafer is initially pat-terned with the desired two-dimensional shape, through standard lithographic tech-niques. The wafer then undergoes anodic etch in an aqueous solution of hydrofluoricacid. Controlling the parameters of the anodic etch determines the pore size and poros-ity of the material. Finally, the PSPs are released from the bulk silicon wafer by means ofsonication in isopropanol solution. Several techniques can be employed to modify thesurface of the PSPs. We describe how to oxidize the PSPs and then modify them withAPTES (3-aminopropyltriethoxysilane) and with fluorescent dyes. Control of the propersurface modifications, quantification of the number of PSPs, and characterization of theoverall size distribution are critical to properly reproduce experiments and ensurehomogeneity within the PSPs. The techniques described take advantage of a:

1. ZetaPals Zeta Potential Analyzer to evaluate surface charge;

2. Beckman Coulter Counter to count and provide the size distribution of the PSPs [43];

3. Inductively Coupled Plasma-Atomic Emission Spectroscope (ICP-AES) [44] todetermine the degradability of the PSPs by quantifying the amount of silicon in thesolution;

4. Becton Dickinson FACSCalibur to determine the size, shape, and fluorescenceintensity emitted from the PSPs themselves or from any second-stage NP that mighthave been embedded within.

13.1 Introduction

243

Phage Gold

Figure 13.4 Au-phage networks. These multifunctional networks of gold combined with bacterio-phages have the potential to target to the tumor vasculature and then be thermally ablated to treat thetumor. (Adapted from [42].)

We also describe the protocols used to determine the loading and release kinetics ofthe PSPs and to quantify the amount of second-stage NPs loaded in or released from thepores of the PSP. We believe that, by the end of this chapter, the reader should be able to


244

(a) (b)

(c)

Figure 13.5 SEM micrographs of collections of porous silicon particles. (a) Overall view of a largecluster of large pores porous silicon particles after release showing substantial size and shape unifor-mity. (b) Close-up view of a small cluster of small pore porous silicon particles after release showingsize and shape uniformity. (c) 45° tilt view of large pores porous silicon particles before release, show-ing substantial size and shape uniformity. The silicon nitride sacrificial layer is present on the sub-

Figure 13.6 Schematic representation of the protocol steps necessary to fabricate pSi particles.

fabricate, modify, and characterize the first-stage vectors and to control the loading andrelease of any second-stage NPs of choice.

13.2 Fabrication of PSPs

13.2.1 Materials

The entire process is performed in a cleanroom facility with the minimum requirementsof:

• Furnace for the deposition of SiO2.

• Low pressure chemical vapor deposition (LPCVD) furnace for the deposition ofSi3N4.

• Photolitography tools (HMDS oven, spin coater, mask aligner, and so forth).

• White light ellipsometer, or any other tool to measure thin film thickness.

• Reactive ion etch (RIE) tool for the dry etch of Si3N4, SiO2, and Si. CF4, SF6, and HBrgases have been employed in the protocol.

• Aluminum sputtering tool.

• Acid hood.

• Solvent hood with sonic bath.

• Wafer rinsing and drying tools.

The protocol uses two quartz/Cr dark field photolithographic masks with 2-μmcircles patterned with 2-μm pitch, custom ordered from Photosciences, California.

The 100-mm heavily doped p-type silicon wafers with resistivity lower than 0.005Ω-cm have been used. This material allows for the formation of pores in the range of afew nanometers to few hundred nanometers, depending on the details of the anodicetch. The use of P-type Si of different resistivity or of distinct n-type Si wafers grantsaccess to other ranges of pore size and porosities, as described in Table 13.1.

The anodic etch solution is composed of 49% hydrofluoric acid (HF) and absoluteanhydrous ethanol (EtOH). The ratio of HF to EtOH and the applied current density arecrucial in determining the pore size and porosity of the PSPs.

The anodic etch is performed in a custom-made HF resistant tank, schematized inFigure 13.7. The most important features of this chamber are:


245

Table 13.1 Range of Accessible Pore Size Depending on Si Doping Type and Concentration

Wafer Type (Dopant Concentration) Pore Range Illumination*

p–-type (<1015) > 1 mm N/A

p-type (1015–1018) 1–10 nm N/A

p+-type (>1018) 10–100 nm N/A

n-type (<1018) 10 nm–10 mm No

n-type (<1018) 50 nm–10 mm Yes

n+-type (>1018) 10–100 nm No

n+-type (>1018) 50 nm–10 mm Yes* Illumination refers to the ability to irradiate the front side of the wafer with light during the anodic etch process.

• The ability to provide adequate backside electrical contact for the Si wafer. Thebackside contact must not be exposed to the etch solution. The backside contact istypically provided through thin aluminum foil, shaped as the wafer. Thus, the tankmust provide enough mechanical stability to guarantee a uniform contact betweenthe aluminum foil and the backside of the Si wafer.

• The ability to immerse a mesh electrode, facing the wafer and parallel to the wafer,at a fixed, replicable distance. The mesh electrode is usually constituted of Pt, anHF-resistant metal with sufficiently good electrical properties.

• The ability to expose a majority of the front side surface of the Si wafer to the etchsolution, in order to maximize the yield of each etch process.

• The ability to resist acid attack by HF; Teflon and aluminum oxide are the materialsof choice for the realization of the tank.

• The ability to allow for the escape of gaseous species formed during the etchprocess. The tank must have an opening from which the gas can escape, and if thePt mesh is positioned horizontally, the gas bubbles must be able to escape betweenthe grid.

A constant current power supply capable of currents up to 8A is required.

Chemicals:

• AZ-5209 photoresist, or equivalent positive, thin photoresist is required forphotolitography.

• Isopropanol is required for the conservation of the PSPs following their release.

• 49% HF is required for the etch solution.

• Anhydrous 200-proof EtOH is required for the etch solution.

• Acetone, methanol, and isopropanol are required to clean the substrates.


246

Tank ring

Al electrode

Al electrode

Seal O-ring

Seal O-ring

Electrode spacer

Tank ring

Pt mesh electrode

Patterned substrate

Tank base

ScrewScrew

Tank base

Electrode spacer

Pt mesh electrode

(a) (b)

Figure 13.7 View of the anodic etch tank. (a) Disassembled view of the tank components. (b) Schematics ofthe assembly of the components.

Characterization:

• Scanning electron microscope with 0.5-nm resolution;

• Nitrogen absorption analysis tool (Quantasorb 3 from Quantachrome).

13.2.2 Methods

13.2.2.1 Thin Film Deposition

The thin film deposition provides a masking layer to the Si wafers, necessary for the pat-terned anodic etch.

1. A 100-nm Si wafer (substrate) is stripped of eventual organic contaminants in 2:1H2SO4:H2O2 piranha solution in an acid hood.

2. The wafer is rinsed for 5 minutes under flowing deionized water, spin dried.

3. The wafer is transferred to a carrier boat for oxide growth. The boat is placed at thecenter of an open furnace tube. Dry air is flowed into the furnace; the temperature israised to 1,000°C and left there for 40 minutes, growing 50 nm of gate oxide. Theexact thickness of the oxide layer is measured and recorded in a white lightellipsometer.

4. The substrate is transferred to the LPCVD furnace; the wafer is placed in the center ofthe loading boat and two dummy wafers are disposed on each side to guaranteeuniformity of the resulting thin film. Si3N4 is deposited to reach the thickness of 80nm, usually requiring 25 minutes of deposition. The exact thickness of the nitridelayer is measured and recorded in a white light ellipsometer. Knowledge of thenitride thickness is necessary to properly time the dry etch step.

GuidelinesThe uniformity of the thin film layer is the most important aspect of this step. Layers ofuniform thickness (within a 5% maximum variation) are necessary for the success of theprotocol, although 1% uniformity is generally preferred. To ensure the best possible uni-formity, the substrate must always be carefully placed on the boat, in the center of thefurnace, where the temperature is most uniform. The substrate must face away from thegas source and be surrounded by as many dummy wafers as possible.

13.2.2.2 Photolitography

The photolitography transfers the desired 2-μm holes pattern on the photoresist layeron top of the substrate. The patterned photoresist acts as masking layer for the dry etch.

1. The substrate is coated with HMDS to improve photoresist adhesion in an HMDSoven for 5 minutes.

2. AZ-5209 positive photoresist is spun on the substrate using: 500 RPM speed, 1,000RPMS acceleration for 5 seconds, followed by 5,000 RPM/4,000 RPMS/30 seconds,resulting in a resist thickness of approximately 700 nm.

3. The photoresist is soft baked for 8 minutes in an oven at 90°C.

4. The 2-μm pattern is transferred from the photomask to the photoresist using a KarlSuss MA6 Mask Aligner, 70J exposure (approximately 3 seconds) using soft vacuumcontact.


247

5. The transferred pattern is developed in an MIF 726 developer for 20 seconds, and

then inspected for uniformity under a 100× optical microscope.

6. If the pattern is sufficiently uniform, the substrate is hard baked for 8 minutes in anoven at 120°C, to completely crosslink the photoresist.

7. If the pattern is not sufficiently uniform, the photoresist can be removed withacetone under sonication. The acetone residues can then be cleaned by subsequentrinses in methanol and isopropanol. The protocol can then be resumed from point 2.

GuidelinesThe uniformity of the pattern is the most important aspect of this step. Even if a smallportion of the substrate is not properly exposed, the substrate should be reprocessed. Toobtain the best possible uniformity, it is critical to know the UV light source power forthe aligner, which would allow for one to calculate the exposure time necessary toobtain the correct exposure. Since the source power cannot be measured, the best prac-tice is to initially calibrate the exposure/development times on several dummy Si wafers,spun with photoresist, and use the best obtained parameters to pattern the substrate.

13.2.2.3 Dry Etch

The dry etch transfers the desired micrometric pattern from the photoresist to thesilicon. This allows the patterned anodic etch to take place.

1. The substrate is transferred in a plasma etch tool, with the patterned side exposed tothe plasma, where it undergoes the dry etch processes necessary to form a 200-nmtrench into the Si by means of a 4-minute CF4 etch (25 sccm, 200 mTorr, 250W in aPlasmatherm RIE).

2. The substrate is flipped to expose the backside (unpatterned) to the plasma. A4-minute CF4 etch (25 sccm, 200 mTorr, 250W in a Plasmatherm RIE) is employedto expose the bare silicon on the backside and ensure electrical contact for thesuccessive anodic etch.

GuidelinesThe timing and chemistry of the dry etch to obtain the desired trench depth and profileis the crucial aspect of this step. Each tool and etch chemistry will have their specific etchrate for Si3N4 and Si. Using the previously calculated thickness of the Si3N4 layer, it is pos-sible to estimate the correct etch time (in seconds) necessary to form the 200-nm trenchinto the silicon, simply employing:

t R h RSi N Si N Si= ⋅ + ⋅3 4 3 4

200 (13.1)

where RSi N3 4is the etch rate for Si3N4 in nanometers per second, hSi N3 4

is the thickness ofthe Si3N4 sacrificial layer in nanometers as measured by ellipsometry, and RSi is the etchrate for Si in nanometers per second.

13.2.2 Anodic Etch

The anodic etch selectively porosifies the substrate where the silicon is directly exposedto the HF solution, forming PSPs.


248

1. The patterned substrate is stripped of photoresist and organic contaminants in 2:1H2SO4:H2O2 piranha solution for 8 minutes.

2. A 200-nm thin film of sputtered aluminum is deposited on the nonpatternedbackside of the substrate to improve electrical contact. The sputtering is performedfor 12 minutes in a 16-wafer holding Varian sputter.

3. The etch tank is assembled as follows (Figure 13.8):

i. The tank ring is placed upside down, and the wafer is placed on top of the tankring, sitting on the seal o-ring, with the patterned side facing inside the ring,where the solution will be poured.

ii. The aluminum-covered backside of the substrate is placed in conformalcontact with an aluminum foil shaped like a table tennis racket.

ii. The base of the tank is screwed to the tank ring, ensuring sealing of the tank andproviding the pressure necessary to guarantee the electrical contact between thesubstrate and the aluminum foil. The handle of the aluminum foil racket is nowoutside the tank and provides the contact spot to connect to the power supply.

iv. The tank is flipped back in the upright position and the platinum mesh isinserted at a distance of approximately 2.5 cm. An annular Teflon spacerpositioned between the wafer and the platinum mesh determines the distance.

v. The etch solution, specified in Table 13.2, is poured in the etch tank.

vi. The anode (positive lead) of the power supply is connected to the aluminumelectrode.

vii. The cathode (negative lead) of the power supply is connected to the platinummesh electrode.


249

1 2 3 4

5 6 7 8

Figure 13.8 Assembly of the etch tank. (1) Upside-down view of the etch ring with seal o-ringmounted. (2) The substrate is placed on top of the o-ring seal with the nonpatterned backside facingoutside the ring. (3) The aluminum electrode is placed on top of the substrate backside to provide elec-tric contact. (4) The etch tank bottom, placed on top of the aluminum electrode, is screwed togetherwith the tank ring to ensure electric contact and seal the tank. (5) The tank is flipped upright and themesh electrode spacer is inserted in the etch ring. (6) The Pt mesh electrode is inserted in the etch ring.(7) The HF:EtOH solution is poured in the etch tank. (8) The anode is connected to the aluminum elec-trode and the cathode to the platinum mesh electrode.

4. The current is started with the porosification current density and time specified inTable 13.2, to produce PSPs with size, pore size, and porosity specified therein. Thecurrent density is then raised to the release current density value and time specifiedin Table 13.2, forming the release layer.

GuidelinesThis is the most important step of all, where the PSP and the release layer are formed.The choice of the correct current density will produce PSPs with the desired pore size andporosity. Additionally, another critical factor is the current density of the release layer. Ifset too high, the elements will release in the etch solution and be lost; if set too low, theelements will not release from the substrate and be unusable.

13.2.2.5 Release of pSi Elements to Obtain PSPs of Desired Shape/Size/Pores

1. The etch tank is emptied of the etch solution.

2. The tank is rinsed three times with deionized water to reduce HF concentration.

3. The tank is disassembled and the substrate removed.

4. The substrate is rinsed for 5 minutes under running deionized water to completelyremove any HF residues.

5. The substrate is spin dried.

6. The substrate is then inspected visually under a 100× optical microscope. A golden-yellow color of the substrate indicates the successful formation of the porouselements. Observing yellow/purple circles of the appropriate diameter (2 μm)surrounded by a yellow colored corona under the optical microscope is also anindication of the successful formation of the porous elements.

7. The substrate is soaked for 30 minutes in HF to strip the SiO2 and Si3N4 layers.Incomplete removal of these layers will prevent the release of the elements from thesubstrate and/or cause shattering of the elements.

8. The substrate is rinsed for 5 minutes under running deionized water to completelyremove HF residues.

9. The substrate is spin dried.

10. The substrate is inspected visually. A dull yellow-grayish tint is a positive predictorfor the element release from the substrate. A yellow-golden tint as in the previous

inspection is a negative predictor for the element release. Under a 100× opticalmicroscope a grey/purple tint is a positive predictor for the element release, while ayellow/purple tint is a negative predictor for the element release.


250

Table 13.2 Anodic Etch Parameters Used to Obtain Desired Pore Size

Target PoreSize

Etch Solution(HF:EtOH)

Etch CurrentDensity (A/cm2)

Etch CurrentTime (Seconds)

Release Solution(HF:EtOH)

Release CurrentDensity (A/cm2)

Release CurrentTime (Seconds)

6 nm 1:1 0.0129 110 2:5 0.779 6

15 nm 1:3 0.0390 90 1:3 0.620 6

26 nm 1:3 0.0900 45 1:3 0.620 6

Current densities are measured on the effective area of Si exposed to the etch solution.

11. The substrate is transferred to a crystallization dish filled with 40 ml ofisopropanol.

12. The crystallization dish is placed in a sonication bath until the release of theelements, typically 1 minute. The occurred release can be visually determined by asubtle change in tint of the substrate, from grayish/green to shiny gray.

13. After release, the substrate is again inspected under a 100× optical microscope todetermine release efficiency. The presence of dull gray disks of approximately twicethe diameter of the original lithographic pattern indicates a released element. Thepresence of gray/purple or yellow/purple disks indicates nonreleased PSPs.

14. The isopropanol suspension rich in PSPs is then transferred in a 50-ml centrifugetube and stored at 4°C.

13.2.2.6 Scanning Electron Microscopy (SEM) Characterization

A small aliquot of the PSP-rich suspension is spotted on a 17-mm SEM stage. Theisopropanol is allowed to dry and the sample is analyzed in a scanning electronmicroscope. Cross-sectional views of the PSPs can be obtained, cleaving the sub-strate before releasing the PSPs and mounting the substrate piece on a 45° or 90° SEMstage.

13.2.2.7 Nitrogen Absorption/Desorption Characterization

A suspension containing 10 mg of PSPs (corresponding approximately to the product of10 substrates) is centrifuged until the PSPs form a pellet at the bottom of the centrifugetube, and all but 10 ml of the supernatant is removed; the PSPs are resuspended. The sus-pension is transferred to a nitrogen absorption analysis cuvette and dried completely.The cuvette containing the PSP powder is mounted on a nitrogen absorption analyzerand the absorption/desorption curves are collected. Using the provided software, theaverage pore size, the pore distribution, and the porosity for the analyzed PSPs areobtained by means of the Barret-Joyner-Halenda (BJH) model.

13.2.3 Characterization

The PSPs resulting from the described protocol are shown in Figure 13.9. The PSPs areanalyzed by SEM to inspect their overall features and when using the standard. The 2-μmphotolithographic pattern will result in quasi-hemispherical PSPs of 3.2-μm diameterand 1-μm height. The top side of the PSP, from where the porosification began, is charac-terized by a circular nucleation site, surrounded by an external corona. Pores run per-pendicular to the nucleation site surface and parallel to the external corona surface(Figure 13.9). The nucleation layer, which extends 10–20 nm below the nucleation site,is constituted of pores with 2–3 nm in diameter. Right below the nucleation layer, thepore size rapidly increases to the one determined by the anodic etch parameters. Thebottom side of the PSP is bowl-shaped and the pores are normal to the surface and havethe characteristic size imparted by the anodic etch.


251

13.3 Oxidation and Surface Modification with APTESof PSPs

13.3.1 Reagents

• Deionized (DI) water;

• IPA;

• Hydrogen peroxide (H2O2);

• Concentrated (95–98%) sulfuric acid (H2SO4);

• 3-aminopropyltriethoxysilane (APTES).

13.3.2 Methods

13.3.2 Wet Oxidation of PSPs

1. PSPs in isopropyl alcohol (IPA) (or any other organic media in which the PSPs aresuspended) are dried in a glass beaker, on a hot plate (80°C–90°C) in a fume hood.The smallest amount of liquid is desirable for this step, as this reduces the dryingtime for the process.

2. A piranha solution consisting of 1 volume of H2O2 and 2 volumes of H2SO4 is used forthe wet oxidation of the PSPs. H2O2 is added to the dried PSPs and sonicated. Owingto the hydrophobicity of the silicon, the PSPs normally tend to float. Concentrated(95%–98%) H2SO4 is then added slowly to this solution.

3. The PSP suspension is then heated to 100°C–110°C for 2 hours with intermittentsonication in a bath sonicator to disperse the PSPs. Utmost precautions should betaken during these steps, and the process should be carried out in a fume hood.Sonication helps not only in dispersing the PSPs but also in dislodging any airpockets within the pores of the PSPs.

4. The particulate suspension is then transferred to centrifuge tubes, and the PSPs arespun down at ~3,000g. The supernatant is discarded and the PSPs are resuspended indeionized (DI) water and transferred to microcentrifuge tubes and spun down again.


252

Backside Front side

(b)

(a)

Cross section Pores (back side) Pores (cross section)

Figure 13.9 SEM images of PSPs. (a) Fabricated according to the parameters in the last row of Table13.2. (b) Fabricated according to the parameters in the first row of Table 13.2. The front side shows thecircular nucleation site surrounded by the external corona. The cross section shows the poresdirectionality from the nucleation site to the particle back side. (Reproduced with permission from [29]courtesy of Nature Publishing Group.)

This process is referred to as washing the PSPs and is critical for the proper removal ofany unreacted substrates. In this way the PSPs are washed five to six times in DI wateruntil the pH of the suspension is approximately around 5 to 6. PSPs may then betransferred to an appropriate buffer (if used immediately) or sorted in IPA or DI waterand refrigerated at 4°C until further use.

13.3.2.2 Surface Modification of PSPs with APTES

1. PSPs that are oxidized by the piranha method are washed thoroughly in water andthen washed in IPA three to four times. After the washings, PSPs are resuspendedin IPA.

2. PSPs are then transferred to a solution of IPA containing 0.5% (v/v) of APTES for 45minutes to 2 hours, at room temperature. The PSPs are sonicated intermittently in abath sonicator and placed on a tabletop shaker for the duration of modification.

3. The chemical modification is usually performed in a microcentrifuge tube. Thereaction volumes used are below 0.8 ml. The lower volumes are ideal for themodification of micron-sized PSPs, as this consumes lower reagents during themodification and subsequent washing steps.

4. The PSPs are washed with IPA four to six times as described earlier and stored at 4°C.Alternatively, aliquots can be taken, dried, and stored under vacuum and desiccantuntil further use.

Figure 13.10 shows the schematic of the surface modification by the APTES

Useful Tip: It is difficult to spin down PSPs in aqueous media completely; most PSPstend to stick to the walls of the tube or remain in suspension. This leads to hugelosses of PSPs, especially if they undergo several cleaning steps to remove suspendingmedia or reactants. In order to recover the maximum amount of PSPs during thecentrifugation step, adding a small amount of detergent (like TritonX-100) assists inthe formation of a nice PSP pellet. Typically 1–2 μl of 1% TritonX-100 in 300–600 μlof aqueous media should do the trick.

Make sure that the Triton is removed before proceeding to any further work with thePSPs. This is normally done by removing the supernatant after the PSPs have been spundown and then adding media to the pellet slowly, attempting not to disturb the PSPs.The PSPs are spun down again, the supernatant is discarded, and fresh media is addedagain. This can be done two to three times, depending on the discretion of theresearcher.

13.3 Oxidation and Surface Modification with APTES of PSPs

253

OH O

O

O

OH CH C

OH

OH

Hydroxylatedsilicon surface APTES APTES modified surface

2 3

OH CH C2 3

OH CH C

Si(CH ) NH Si(CH ) NH2

2 2

3

3 32 2+

Figure 13.10 Schematic showing the modification of silicon surface with 3-aminopropyl triethoxysilane (APTES).

13.4 Fluorescent Dye Conjugation of PSPs

PSPs modified with APTES can be conjugated with any commercially available fluores-cent dyes that have a hydroxy-succinimidyl ester (NHS) conjugated to them. The NHSester readily reacts with primary amines. NHS conjugated dyes are commonly used to tagproteins and antibodies and can be purchased from Invitrogen and Pierce.

13.4.1 Reagents

• 10 mM Phosphate buffer (PB);

• pH~ 7.3 (for the conjugation);

• 1% Triton X-100 (for washing unconjugated dye).

13.4.2 Methodology

1. The APTES-modified PSPs are washed and suspended in the conjugation buffer.

2. The fluorescent dye to be conjugated is dissolved in the buffer and mixed with thePSP suspension.

3. The mix is sonicated and reacted for up to 1 hour. (The conjugation protocolprovided by the supplier can also be followed.)

4. After reaction the PSPs are washed three times in 1% TritonX-100 followed by five tosix washes in PB.

13.5 Zeta Potential Measurement

13.5.1 Equipment

• ZetaPals Zeta Potential Analyzer (Brookhaven Instruments Corp., Southborough,Massachusetts)

13.5.2 Reagents

• 10 mM Phosphate buffer (PB);

• pH~ 7.3 (for suspending the PSPs for performing the Zeta potential measurements).

13.5.3 Methodology

1. The application window for the zeta potential measurement (ZetaPals) is opened inorder to power on the laser. After 15 minutes (for laser warm-up), the zeta potentialmeasurements can be done.

2. The cuvette for holding the particulate suspension is rinsed with filtered (0.2 μmfilter) buffer. The cuvette is filled with 1.5–2 ml of buffer. A small amount of PSPs issuspended in the buffer and well mixed either using a pipette or a brief sonication.

3. Make sure the electrodes are cleaned and rinsed with the buffer in which themeasurements are performed.

4. After the electrodes are placed in the cuvette, the measurement for the zeta potentialis started.


254

5. Typically, three runs of 25 cycles per run are performed, but for more consistentresults during each run, the number of cycles can be increased based on thediscretion of the user.

6. The counts per second (cps) during data acquisition should be above 20 Kilo countsper second (Kcps), and below 700 Kcps. The instrument will automatically register ifthe quality of the sample for measurement is good or bad.

7. A detailed explanation of operating the equipment can be found in the manual ofthe instrument or from a training session with a Brookhaven Instruments scientist.

13.5.4 Results

After oxidation the PSPs charge is negative. The negative charges depend on the numberof hydroxyl groups that are formed on the surface of the PSPs. After modification withAPTES, the PSPs become less negative due to the surface coverage by the silane. A com-plete multilayer APTES coverage leads to PSS with higher positive Z-potential.

Table 13.3 shows the typical results of zeta potential measurement on PSPs.

13.6 Count and Size Analysis of PSPs13.6.1 Materials

13.6.1.1 Reagents

1. ISOTON II Diluent (Beckman Coulter);

2. Accuvettes (Beckman Coulter);

3. Standard cuvette (VWR);

4. 10-mL syringe (BD);

5. Single-use 0.20 mm syringe filter (Sartorius Stedim Biotech).


1. Z2 COULTER COUNTER Cell and Particle Counter (Beckman Coulter);

2. PC Computer with AccuComp Software (Beckman Coulter);

3. 50 μm Ampoule Aperture Tube (Beckman Coulter);

4. Sonicator (Branson).

13.6.2 Methods

1. 20 mL of ISOTON diluent into a CLEAN Accuvette.

2. Aliquot filter the ISOTON.

3. Clean a cuvette using 1 mL of the filtered ISOTON to remove dust/debris inside ofthe cuvette.

4. Aliquot 2 mL of filtered ISOTON into the clean cuvette.

13.6 Count and Size Analysis of PSPs

255

Table 13.3 Zeta Potential of PSPs as Measured in 10 mM PB

PSP Sample Zeta Potential (mV)

Oxidized PSPs From –29 to –34APTES PSPs From +5 to +11

5. Using a concentrated sample of PSPs, aliquot a small volume (between 0.5 to

4 μL) into the cuvette (concentrated samples are usually in the range of 2 × 108

particles/mL).

6. Sonicate cuvette to ensure homogeneity within sample.

7. Touch SETUP on the control panel of the Z2.

8. Place sample into machine for measurement. Do not allow the probe to go all theway to the bottom of the cuvette. The cuvette is placed onto an Accuvette cap so thatthe probe can reach the sample.

9. Adjust and examine the “Aperture Viewer” so that during the experiment one canobserve any possible blocking of the aperture.

10. Input the upper and lower size limits:

i. For 3.2 mm PSPs (seen as 2 μm): 1.1–2.8 μm.

11. Touch SETUP again, scroll down to Optimize Settings, and move cursor to say YES.

12. Touch START/STOP and review settings.

13. Touch START/STOP again, and the Z2 should begin the measurement.

14. Observe the Concentration on the control panel of Z2; if it is too high, considerdiluting the sample.

15. When the measurement is finished, import the run into PC using AccuCompsoftware.

16. Inspect the graph for one central peak; then using the software, calculate the numberof PSPs measured/counted.

17. Remove the sample, sonicate briefly, and repeat steps 7–16 four more times.

18. Average the counts and find the standard deviation.

19. To get to your overall count, multiply the number/mL of the measured sample by thedilution factors used when the sample was prepared.

Note: Each measurement can be further analyzed to give size distribution, overlayingruns, and averaging multiple runs into one file/graph. Furthermore, Beckman Coulterhas recently released new counters called “Multisizers” that have aperture sizes thatrange from 20 to 2,000 mm and thus can count particles as small as 400 nm.

13.6.3 Data Acquisition, Anticipated Results, and Interpretation

When the sample is measured at the Z2 Analyzer, the resulting signal is calculated intothe volume of diluent displaced per event. This gives an idea of the change in morphol-ogy of the particles over time and of the total number of readable particles present ateach time point. Figure 13.11 is a sample Z2-generated graph using the AccuComp soft-ware. In this figure, we have the cell/particle diameter versus the number of particlescounted per milliliter. One can move the cursor to whatever location to display thenumber of particles per milliliter at that particular volume; furthermore, one can select awhole area between two volumes to find the number of particles counted in that section.The six menu options at the top left-hand corner of the figure allows for the manipula-tion of items within the figure. Under the “Run File” menu, users can save graphs gener-ated, add overlays, and export critical data to Excel. In “Graph,” users can customize theoptions displayed on the graphs from a pull-down menu with a list of possible x and yvalues. The “Analyze” menu allows the user several options to interpret the data


256

and determine the total number of particles counted and measured in that particularfile.

13.7 Using Inductively Coupled Plasma–Atomic EmissionSpectroscopy (ICP-AES) to Determine the Amount ofDegraded Silicon in Solution

13.7.1 Materials

13.7.1.1 Reagents

1. 0.45 μm Nylon Filter Tubes (VWR);

2. 15 and 50 mL Polypropylene Conical Tubes (BD FALCON);

3. ISOTON II Diluent (Beckman Coulter);

4. Distilled H2O;

5. Yttrium;

13.7 Using Inductively Coupled Plasma–Atomic Emission Spectroscopy (ICP-AES)

257

(b)

500

10001500

20002500

300035004000

4500

5000

5500600065007000

75008000

85009000

9500

10000

00 0.2 0.4 0.6 0.9 1 1.2 2

Cell diameter ( m)μ

2.2 2.4 2.6 2.8 31.4 1.6 1.9

(a)

100

200

300

400

500

600

700

800

00 0.2 0.4 0.6 0.9 1 1.2 2

Cell diameter ( m)μ

2.2 2.4 2.6 2.8 31.4 1.6 1.9

900

1000

1100

1200

1300Differential number

Differential number

Figure 13.11 Graphs produced by AccuComp displaying the size distribution of a PSP. (a) Typical profile forvectors with extra large pores (60–80 nm). (b) Profile for vectors that have been broken; notice the typical peakat ~2 mm.

6. Silicon.


1. ICP-AES/OES (Varian);

2. Autosampler (Varian);

3. Argon Saturator Accessory (Varian, suggested).

13.7.2 Methods

1. Collect sample (i.e., 100 μL). Highly recommend collecting sample in triplicate.

2. Place sample into nylon filter tube.

3. Centrifuge sample at 4,200 rpm for 10 minutes.

4. Remove filter and collect the solution that flowed through.

5. Prepare a diluted sample to be analyzed by ICP.

i. For each individual sample, aliquot 5 mL of a solution that contains distilledwater and 1 ppm of Yttrium into a 15-mL conical tube.

ii. Aliquot a known amount of sample from the flow through into the conical tube.Keep this amount consistent (i.e., 50 μL from the ~100 μL).

6. Prepare known concentrations of silicon “standards” with 1 ppm of Yttrium. Suggestpreparing 0, 25, 50, 100, 250, 500, and 1,000 ppm solutions of silicon.

7. Briefly shake sample and silicon standards.

8. Load samples and standards in autosampler*.

9. Set up template for acquisition. ICP starts each run by running the knownconcentrations and finishes by running a calibration off one of the standards.Suggest using 50 ppm of silicon as the control calibration and rerun the knownconcentrations after 15–20 samples have been analyzed.

10. When all the samples have been measured, examine the data of each sample and, ifnecessary, mask any run that may have extremely high standard deviations.

11. Export data to Excel or any other spreadsheet application.

12. Analyze data to determine silicon concentration of samples.

*Note: It is highly advisable that the operation and measurement of samples usingthe ICP-AES/OES machine be done by an operator that is highly proficient in runningthe machine.


13.7.3.1 Data Acquisition

The data received by the user will be in spreadsheet format, with the first row showingthe data labels, as shown in Table 13.4.

Table 13.4 Data Labels for ICP Data

Tube Sample Labels Si 250.690 Si 251.432 Si 251.611 Si 288.158 Y 360.074 Y 371.029

The first column, “Tube,” designates in what rack and position (Rack:Position)within the autosampler the machine is measuring from. In this particular setup there are


258

two racks each with 60 positions available to hold the samples (racks 1 and 2) and a thirdrack that holds the known concentrations. The next column is the “Sample Label” thatis used at the start of the experiment to designate what sample is being measured. Thenext four columns that start with “Si” correspond to the four wavelengths used to mea-sure the concentration of Silicon within the sample. The last two columns are used toshow the measurement of Yttrium within the sample. Yttrium measurement is impor-tant to assess the stability of measurements in time. The tool normalizes the Yttriumreading to 1.00 for the first sample in both wavelengths and then uses this value to cali-brate for the decay in concentration found in the subsequent samples. Calibration to theYttrium standard is crucial, since, depending on the number of samples, measurementscan take up to several hours and may need to run overnight (60 samples take about 3–4hours), and thus measurements would need to be adjusted for any decay in the readings.

Each sample will have its own row, including the known concentrations given inμg/L (Table 13.5). The known concentrations are set to the actual value and are used tobuild a standard curve/line. The sample’s values are then extrapolated from the curveobtaining a numerical value. At the end of a cycle, the machine runs the calibration con-centration, labeled as “Cont. Calib. Verif.,” to verify proper calibration. In this particular

13.7 Using Inductively Coupled Plasma–Atomic Emission Spectroscopy (ICP-AES)

259

Table 13.5 Actual Values from ICP-AES Analysis (in μg/L)


3:1 Blank 0 0 0 0 0.8 0.93:2 Si Standard A 25 25.0000e 25 25 0.8 0.93:3 Si Standard B 100 100 100 100 0.8 0.93:4 Si Standard C 250 250 250 250 0.8 0.93:5 Si Standard D 500 500 500 500 0.8 0.93:6 Si Standard E 1,000 1,000 1,000 1,000 0.8 0.92:1 61 56.8 70.9 40.0 45.6 0.7 0.82:2 62 28.9 68.5 36.1 32.0 0.7 0.82:3 63 28.2 60.1 uv 36.3 31.2 0.7 0.82:4 64 65.2 64.6 69.7 62.3 0.7 0.82:5 65 55.5 77.6 49.4 53.2 0.7 0.82:6 66 55.1 50.1 50.7 45.1 0.7 0.82:7 67 42.2 42.6 55.9 50.5 0.7 0.82:8 68 51.9 60.0 37.7 44.9 0.7 0.82:9 69 30.4 63.4 43.6 33.8 0.7 0.82:10 70 73.3 84.4 72.3 68.2 0.7 0.72:11 71 55.2 75.5 57.8 53.4 0.7 0.72:12 72 24.4 28.1 uv 7.9 9.1 0.7 0.82:13 73 4.0 47.9 9.7 11.9 0.7 0.82:14 74 2.7 39.6 uv 15.6 12.6 0.7 0.82:15 75 7.0 67.8 29.6 29.3 0.7 0.72:16 76 22.0 27.3 uv 18.3 17.2 0.7 0.82:17 77 11.7 49.3 24.0 17.9 0.7 0.82:18 78 38.2 50.9 34.6 36.9 0.7 0.82:19 79 30.6 40.4 36.6 37.1 0.7 0.82:20 80 37.3 56.0 31.3 26.2 0.7 0.83:7 Cont. Calib. Verif. 49.5 72.1 Q 52.1 45.8 0.8 0.8

Values from acceptable wavelengths are in red.

example the reference sample is set to 50 ppm or 50 μg/L. Upon inspection of the sampledata, some values are observed to have additional letters or even negative sign, such asQ, uv, and e. Q and uv are used to designate values that are found to be under detectionlimits. The operator designates the letter e to a particular value that had to be edited. Thisis preformed only when a value has an extremely high internal standard deviation, in anattempt to mitigate the effect of any stray data within the run of that particular sample.

13.7.3.2 Analyzing data

Only data from wavelengths for which the reference sample was measured within 10%of the actual value should be used for analysis. Inspecting the output file, it can beobserved that a “Q” is placed next to any value in the control calibration row that doesnot fall within 10% of the known value of 50 μg/L. Thus, only wavelengths with areference sample value between 45 and 55 μg/L are used. The wavelength’s control con-centration values are then rescaled such that the reference sample concentration mea-surement is set equal to the expected concentration of 50 μg/L. For example, if thecontrol calibration sample value for a given wavelength were to be 48, then every valuein the column would be multiplied by 50/48. Then using the resulting values that havebeen rescaled or normalized to a value of 50 from each sample and averaging with theother samples (since each sample was run in triplicate) will result in the final concentra-tion of silicon of that particular sample in μg/L (Table 13.6).

However, one still needs to account for the dilution factors involved in the prepara-tion of the sample. Thus, the true concentration of silicon in the sample can be foundmultiplying by the dilution factor, as seen in (13.2):

VV

sol

sample

×2 (13.2)

where

Vsol = volume of solution;

Vsample = volume of the sample.

Multiply by two, since we only used half of the sample to be measured.This simple calculation provides the amount of silicon in your sample in μg/L.Depending on the preference of the user, there are two alternative methods to inter-

pret the results. The most beneficial, for most applications, is displaying the amount ofsilicon released into solution as a percentage of the total amount of silicon that can bereleased per PSP (Figure 13.12(a)). This type of interpretation facilitates the display ofminor degradation rate changes within the different PSPs. Using this interpretation,PSPs that degrade quicker would show a steeper slant during their “linear” degradation.The other interpretation displays the total amount of silicon that is in solution(Figure 13.12(b)). This analysis would be useful in showing the different amounts of sili-con contained in the different PSP types. Ideally, it would show that once degraded, PSPswith larger pores would release a lower amount of silicon in solution.


260

13.7 Using Inductively Coupled Plasma–Atomic Emission Spectroscopy (ICP-AES) to Determine the Amount ofDegraded Silicon in Solution

261

Table 13.6 Normalized Values from ICP-AES Analysis


3:1 Blank 0 0 0 0 0.8 0.93:2 Si Standard A 25 25.0000e 25 25 0.8 0.93:3 Si Standard B 100 100 100 100 0.8 0.93:4 Si Standard C 250 250 250 250 0.8 0.93:5 Si Standard D 500 500 500 500 0.8 0.93:6 Si Standard E 1,000 1,000 1,000 1,000 0.8 0.92:1 61 57.3 70.9 38.4 49.8 0.7 0.82:2 62 29.1 68.5 34.6 35.0 0.7 0.82:3 63 28.5 60.1 uv 34.8 34.1 0.7 0.82:4 64 65.8 64.6 66.9 68.0 0.7 0.82:5 65 56.1 77.6 47.4 58.1 0.7 0.82:6 66 55.6 50.1 48.6 49.3 0.7 0.82:7 67 42.7 42.6 53.6 55.1 0.7 0.82:8 68 52.4 60.0 36.2 49.1 0.7 0.82:9 69 30.7 63.4 41.9 36.9 0.7 0.82:10 70 74.0 84.4 69.4 74.5 0.7 0.72:11 71 55.7 75.5 55.5 58.4 0.7 0.72:12 72 24.6 28.1 uv 7.6 10.0 0.7 0.82:13 73 4.1 47.9 9.3 13.0 0.7 0.82:14 74 2.8 39.6 uv 15.0 13.7 0.7 0.82:15 75 7.1 67.8 28.4 32.0 0.7 0.72:16 76 22.2 27.3 uv 17.5 18.8 0.7 0.82:17 77 11.8 49.3 23.0 19.5 0.7 0.82:18 78 38.6 50.9 33.2 40.3 0.7 0.82:19 79 30.9 40.4 35.1 40.5 0.7 0.82:20 80 37.7 56.0 30.1 28.6 0.7 0.83:7 Cont. Calib. Verif. 50.0 72.1 Q 50.0 50.0 0.8 0.8

Values that have been normalized are in blue.

0 00 0

12 1224 2436 3648 4860 6072 72

20 50

40 100

60 150

80 200

100 250120 300

140 350

Suin

saln

(%of

tota

l)

Si(

g)μ

MP1 MP1MP2 MP2

Time (hrs) Time (hrs)

(a) (b)

Figure 13.12 ICP graphs produced in Excel showing multiple ways to display the amount of silicondissolved into the solution. MP1 and MP2 refer to a PSP of medium-sized pores with 10 and 15 nm,respectively. (a) Displaying the amount of silicon in solution by using the percentage of total possiblesilicon in solution. (b) Displaying the amount of silicon by showing the amount, by mass, of silicon.

13.8 Flow Cytometry to Characterize PSP Shape, Size,and Fluorescence Intensity13.8.1 Materials13.8.1.1 Reagents

1. Sodium chloride;

2. Distilled water;

3. Phosphate buffered saline, PBS pH 7.2 (Gibco);

4. 5-mL polystyrene round bottom tubes (BD Falcon).


1. FACSCalibur (Becton Dickinson, BD);2. Computer running CellQuest software (Becton Dickinson, BD).

13.8.2 Methods

1. Start system. Power on Calibur, then computer.

i. Allow 15 minutes for machine to warm up in “STNDBY.”

2. Mix 9g of NaCl into 100 mL of distilled water, thus 9% NaCl.

3. Make a 1:10 dilution of 9% NaCl in water (sheath fluid).

4. Load sheath fluid into proper compartment and empty out the waste, if necessary.

5. Set up acquisition parameters:

i. Parameters include detectors/amps, instrument settings, file names, location ofsaved file, compensation, and threshold.

ii. For particles only, see Table 13.7 for reference settings.

6. Prepare sample. Aliquot ~5 × 105 PSPs into 500 mL of PBS into a polystyrene tube. Wesuggest to run the samples in triplicate, by either reading same sample three times or

running three different samples each with at least 1 × 105 PSPs.

7. Briefly vortex sample.

8. Load polystyrene tube, press RUN, and select a flow rate:

i. LOW: 12 μL/min (information obtained from the BD Web site on January 21,2009);

ii. MID: 35 μL/min (suggested to start here if using same concentration as above);

iii. HI: 60 μL/min.

9. After each sample, briefly run some distilled water through the machine until thereare no events recorded in the acquisition plots.

10. Repeat steps 7–9 for each sample.

11. Analyze samples using CellQuest software.


262

Table 13.7 Recommended Instrument Settings for Particle Measurement

Detector Voltage (V) Mode

FSC E-1 LOGSSC 475 LOGFL1 800 LOGFL2 750 LOGThreshold Primary: FSC 30 (for other applications,

leave at 52)Secondary: SSC 52


The CellQuest software allows the user to customize the type of data that can be col-lected and displayed. Certain graphs can only be used for acquisition or analysis or for

acquiring and analyzing (Acquisition → Analysis) the data simultaneously. The softwareis able to display data in five basic graphs: histogram, density, dot, contour, and 3-D

plot. The first three can be set to Acquisition, Analysis or Acquisition → Analysis, whilethe last two, contour and 3-D plot, can only be used for the Analysis display and datashould be acquired through other graphs (Figure 13.13). Histograms may also be over-laid, thus allowing users to compare several curves on the same plot.

Quantitative results from CellQuest can be obtained for histograms and regions(available in square, polygon, or circle) or gates of interest (both are manually drawn byuser). Statistics are selected by choosing the appropriate type from the “Stats” pull-downmenu located in the toolbar at the top of the screen. This results in an embedded boxthat can be resized or moved with the user-selected statistics inside. This box contains

13.8 Flow Cytometry to Characterize PSP Shape, Size, and Fluorescence Intensity

263

(c) (d)

(a) (b)

Figure 13.13 Using flow cytometry to study the PSPs’ size and fluorescence. (a) Depicts the relativesize (FSC) and shape (SSC) of the PSPs thorough a contour plot, Region R1 represents the gating region.(b) 3-D plot showing the distribution of PSPs gated in (a), where the z-axis represents the total numberof counts/particles. (c) Histogram showing the background fluorescence of unloaded PSPs and is usedto set up M1 so that at least 99% of events are captured here and M2 captures the rest, and (d) theincrease in fluorescence distribution after loading the particles, keeping M1 and M2 regions the sameas in part (c). (Reproduced with permission from [29] courtesy of Nature Publishing Group.)

several parameters including (but not limited to) mean, median, C.V., standard devia-tion, peak, total event, and events gated.

For the analysis of PSP shape and size, a bivariate plot (dot, contour, or density can beused) graphing forward scatter (FSC) versus side scatter (SSC) is used. This type of analy-sis can also be used to exclude events by defining a polygonal region of interest aroundthe population of interest and analyzing the statistics within that region to obtain valuesfor the geometric mean in both the X and Y parameters.

In addition, the FACSCalibur has the capabilities for fluorescence analysis. In rela-tion to this procedure, only two colors will be described: FL1 (green) and FL2 (red). Thegreen fluorescence (FL1) can detect FITC and QDot 525, as an example, using a 530/30bandpass filter. The red fluorescence (FL2) can detect QDot 565 using a 575/26 bandpassfilter. If single color detection is needed, color compensation can be set to zero. How-ever, when detecting dual green-red color, FL1 compensation is set to 25% FL2, and FL2compensation is set to 35% FL1 using the Compensation palette under the Cytometerpull-down menu in the CellQuest window.

This type of fluorescent setup allows users to characterize and quantify the amountof second-stage NPs loaded into the PSPs. To accomplish this, first the region of interestis located within a dot plot of FSC versus SSC. Then a histogram displaying the detector(FL1 or FL2) is created. This plot selectively displays the events within the defined regionthat correspond to the second-stage NP used.

13.9 Loading and Release of Second-Stage NPs from PSPs

13.9.1 Loading of NP into PSPs

13.9.1.1 Materials

Reagents

1. Nanoparticles (i.e., QDots, SWNT);

2. 1.5-mL low-binding polypropylene centrifuge tubes (VWR International);

3. DI water;

4. Tris(hydroxymethyl) aminomethane (Tris-HCl).

Facilities/Equipment

1. Thermo Scientific Barnstead LabQuake Tube Rotators (Thermo Scientific).

13.9.1.2 Methods

1. Put 3.0 × 105 PSPs in low-binding polypropylene tubes in 3 mL of DI water.

2. Adjust Tris-HCl to a pH of 7.3.

3. Add NPs and adjust the final solution to 20 mL using Tris-HCl (i.e., 2 mM QDots: 5μL Qdots + 3 μL H2O + 12 μL Tris-HCl, or 20 ng/μL PEG-FITC-SWNTs: 9 μL SWNTs + 3μL H2O + 8 μL Tris-HCl).

4. Incubate samples on tube rotator (~20 r.p.m.) for 15 minutes at room temperature.

5. Dilute samples with Tris-HCl to final volume of 150 μL and measure fluorescenceintensity using flow cytometry.


264

13.9.2 Release of NPs from PSPs

13.9.2.1 Materials

Reagents

1. Nanoparticles (i.e., QDots, SWNT);

2. 1.5 mL low-binding polypropylene centrifuge tubes (VWR International);

3. DI water;

4. Tris(hydroxymethyl) aminomethane (Tris-HCl);

5. Sodium chloride.

Facilities/Equipment

1. Thermo Scientific Barnstead LabQuake Tube Rotators (Thermo Scientific).

13.9.2.2 Methods

1. Combine 2.1 × 106 PSPs at pH 7.3 with a final solution of 140 mL:

i. 2 μM QDots in 200 mM Tris-HCl;

ii. 20 ng μL–1 PEG-FITC-SWNT in 20 mM Tris-HCl;

iii. 1 μM QDots + 10 ng μL–1 PEG-FITC-SWNT in 50 mM Tris-HCl.

2. Incubate samples on tube rotator (~20 r.p.m.) for 15 minutes at room temperature.

3. Wash samples in 1.4 mL of DI water.

4. Centrifuge for 5 minutes at 4,200 r.p.m.

5. Remove supernatant and resuspend in 70 μL DI water. Use 10 μL from each sample toassess fluorescence intensity using flow cytometry. Record intensity at time 0 andthen over several time points (i.e., 30, 60, 90, 180, 360, 1,200 minutes).

6. Dilute residual 60 μL to 3 mL using 20 mM Tris-HCl 0.9% NaCl (release buffer).

7. Incubate at 37°C on tube rotator (~20 r.p.m.) for your defined amount of time.

8. After each time period has expired, centrifuge the sample for 5 minutes at 4,200r.p.m. and measure fluorescence using flow cytometry.


Determining the amount of agent that is loaded or released is critical for any deliverysystem. Proper characterization of the second-stage NPs is necessary for the optimalloading into the first stage vector. The knowledge of the second stage’s surface charge,size, and concentration will greatly impact the choice of the pore size and surface chargeof the first stage to be used to optimize the loading and release of these second-stage NPs(Figure 13.14). To characterize the first stage PSPs after they have been loaded and todetermine the kinetics of second-stage NPs release both flow cytometry and confocalmicroscopy can be used. These are extremely useful tools when the loaded NPs arefluorescently tagged. Flow cytometry can characterize the amount of loaded NPs basedon fluorescence intensity (Figure 13.15). To properly evaluate the amount of loaded NPs,it is necessary to compare two samples: PSPs with pores whose size will not allow theloading of the NPs (pore size too small), and PSPs with an adequate pore size to properlyload the NPs of choice. This type of analysis allows for the discrimination of the amount

13.9 Loading and Release of Second-Stage NPs from PSPs

265

of fluorescence that can be attributed to NPs adhering to the surface of the PSPs, and thefluorescence due to the NPs loaded inside the pores.

Confocal microscopy is useful in determining the distribution of the second-stageNPs within the first-stage PSPs (Figure 13.16) and in quantifying the amount of fluores-cence attributed to the embedded NPs. However, a large sample population would beneeded to get a statistically significant average intensity and thus flow cytometry wouldbe more appropriate. The distribution of the NPs within the PSPs can be detected by sim-ply zooming in on the PSP and then defining a ROI around that PSP. The next step is todraw an intensity profile/line covering the diameter of the PSPs, a graph that displaysthe intensity versus the length of the line is produced and showing the fluorescenceintensity of the NPs distributed in the PSP. For example, when PSPs are loaded simulta-neously with two types of NPs, it was concluded that the larger NPs were exclusivelyfound in the central area of the PSP (associated with the larger pores), while smaller NPswere found throughout the entire vector but with a primary accumulation on the borderof the PSP (associated with the smaller pores) (Figure 13.16).

The release of the NPs from within the pores of the PSPs can also be characterizedusing flow cytometry (Figure 13.15). This is achieved by indirectly measuring the resid-ual fluorescence of PSPs after they have released the second stage NPs. Carefully choos-ing time points and displaying the data as percentage released of the optimal loadingcan give crucial data regarding the release kinetics from within the pores and thus assistin on the best choice of PSP characteristics needed for optimal delivery.


266

(e)

(c)

(a)

(f)

(d)

(b)

Figure 13.14 Models representing the three major mechanisms responsible for the optimal loadingand release of second-stage NPs from PSPs. Size, dose, and charge are critical factors that govern theamount of NPs that can be loaded within the PSPs. Size dependency and the size of the pores deter-mine the types of NPs that can be preferentially loaded in PSPs. (a) NPs that are too big remain outside.(b) NPs that are smaller than the size of the pores are loaded into the PSPs. Dose dependency: (c) alower concentration of NPs in the loading solution results in reduced loading into the pores while (d)an increased concentration will result in increased number of NPs loaded within the pores. Chargedependency: (e) NPs with a surface charge opposite to PSPs are strongly attracted into the pores, while(f) NPs with a similar charge to that of the PSPs will result in NPs being partially or completely repelledfrom loading into the pores. (Reproduced with permission from [29] courtesy of Nature PublishingGroup.)


This chapter describes a novel multistage delivery system (MDS) based on PSPs capableof sequentially negotiating biobarriers and improve targeted delivery of imaging and


267

Carboxl Q-dots

00 15 30

LP oxidized

45 60

400

800

1200

1600

2000

Mea

nflu

ores

cenc

e

Time (minutes)

Amino Q-dots PEG-FITC-SWNTs(a)

Carboxl Q-dots

00 15 30

LP APTERS

45 60

400

800

1200

1600

2000

Mea

nflu

ores

cenc

e

Time (minutes)

Amino Q-dots PEG-FITC-SWNTs

Carboxl Q-dots

00 15 30

SP oxidized

45 60

40

80

120

160

200

Mea

nflu

ores

cenc

e

Time (minutes)

Amino Q-dots PEG-FITC-SWNTs(b)

Carboxl Q-dots

00 15 30

SP APTES

45 60

40

80

120

160

200

Mea

nflu

ores

cenc

eTime (minutes)

Amino Q-dots PEG-FITC-SWNTs

(c) (d)

Carboxl Q-dots

00.5 1 1.5

LP oxidized

3 6

20

40

60

80

100

Rele

ased

pay

load

(%)

Time (minutes)

PEG-FITC-SWNTs

20

(e)Carboxl Q-dots

00.5 1 1.5

LP oxidized

3 6

20

40

60

80

100

Rele

ased

pay

load

(%)

Time (minutes)

PEG-FITC-SWNTs

20

(f)

Carboxl Q-dots

00.5 1 1.5

LP APTES

3 6

20

40

60

80

100

Rele

ased

pay

load

(%)

Time (minutes)

PEG-FITC-SWNTs

20

(g)Carboxl Q-dots

00.5 1 1.5

SP APTES

3 6

20

40

60

80

100

Rele

ased

pay

load

(%)

Time (minutes)

PEG-FITC-SWNTs

20

(h)

Figure 13.15 Loading and release of second-stage NPs from PSPs. (a–d) Four different types of PSPswere loaded with different second-stage NPs and their mean fluorescence measured by flow cytometryover time were measured: (a) LP oxidized, (b) SP oxidized, (c) LP APTES, and (d) SP APTES. (e) Releaseof Q-dots and PEG-FITC-SWNTs from LP oxidized, (f) SP oxidized, (g) LP APTES, and (h) SP APTES wasmeasured over time and expressed as a percentage of the total amount of second-stage NP payloadreleased from the PSPs for every time interval, after optimal loading. (Reproduced with permissionfrom [29] courtesy of Nature Publishing Group.)

therapeutics. The versatility and ease of modification of the MDS are one of its majoradvantages over competing multistage delivery technologies. The methods previouslyoutlined constitute the core of the MDS technology, but few crucial guidelines mustalways be kept in mind when attempting to implement this system:

• Each step of the PSP fabrication process must be thoroughly controlled and vali-dated to obtain satisfactory results and replicability. As with any silicon manufac-turing process, good manufacturing practice is the key to a high-throughput,high-yield process producing functional devices according to specifications.

• The surface modification process of the PSPs can lead to a significant loss of PSPsduring the several steps required. To minimize this loss, it is highly recommendedthat a small amount of detergent (i.e., Triton X-100) is added to pellet down thePSPs. In a volume of 300–600 μL, 1–2 μL of Triton can help recovering millions of


268

(h) (i) (j)

(d)

(a)

(e)

(b)

(f)

(c)

(g)

Figure 13.16 Simultaneous loading and release of Q-dots and PEG-FITC-SWNT. (a) FACS histo-gram-overlay of unloaded PSPs; PSPs loaded with PEG-FITC-SWNTs (+SWNTs), with Q-dots (+Q-dots)and with both Q-dots and SWNTs (+Q-dots +SWNTs). Flow cytometry analysis of (b) simultaneousloading and (c) release of second-stage NPs. (d–g) Confocal microscopy images show the localization ofPEG-FITC-SWNTs (green) and Q-dots (red) in a single PSP: (d) bright-field, (e) green and red (f) fluores-cence, and (g) overlay are shown. (h, i) Fluorescence intensity profiles of each channel along theorange dashed lines in (e) (PEG-FITC-SWNTs) and (f) (Q-dots) are shown, respectively. (j) The greenand red arrows incorporated into the SEM image confirm the spatial distribution of fluorescence in thePSP. White scale bars in (d–g) are 3 mm. (Reproduced with permission from [29] courtesy of NaturePublishing Group.)

PSPs that may have otherwise been discarded with the supernatant. However, atthe user’s earliest convenience, this detergent should be removed from the surfaceof the PSP since it may inhibit the further modifications required.

• The loading and release kinetics of second-stage NPs can be controlled by tailoringthe first-stage PSP’s features. Confocal microscopy can be used to confirm optimalloading conditions, determine the distribution of multiple second-stage NPs, andensure homogeneity within each first-stage PSP analyzed.

• Procedures described include using flow cytometry to determine shape, size, andintensity, Z2 Coulter Counter to analyze the concentration and size distribution,and ICP-AES to quantify the amount of silicon in solution. These methods providedetails regarding the status of the PSPs and therefore need to be calibrated usingcontrol samples of known and defined nature prior to each analysis.

The quantification of trace amounts of silicon must be performed meticulously toobtain an accurate quantification. When preparing samples, it is essential that notool/material comes into contact with glass. In the construction of the standard curve,the selection of the correct concentrations (standards have to adequately represent therange of expected values) greatly increases the accuracy of the measurements.

After successfully replicating the methods outlined earlier, the reader can modifyand expand them in order to better suit its specific application. The versatility of theMDS technology allows users to easily build upon the core methods and to adapt themto a variety of different drug delivery scenarios. In particular:

• The design of the PSP size and shape can be optimized to enhance the PSP functionusing proprietary mathematical algorithms developed in our laboratory [45, 46].While other NPs follow the laminar flow through the center of the capillary, thePSP tumbles along the wall of the capillary and eventually binds to markers on thetumor associated endothelial capillary wall.

• The PSPs can be surface modified with peptide sequences used to target tumor cellsincorporated into the tumor vasculature. The PSP can use humanized monoclonalantibodies or peptide sequences and specific aptamers in order to avoid antibodytargeting limitations and increase system stability.

• The possibility to maximize drug/second-stage NP loading and release through themodulation of PSP external and internal surface charges.

• Controlling the details of the pore structure, the PSP can be engineered to deliverdrug or secondary NPs only in the direction of the endothelium. This minimizesthe amount of NPs swept away in the bloodstream immediately after their release.It is also possible to obtain a PSP where only the external corona is functionalizedwith targeting moieties, the nucleation layer has been removed, and a small porelayer has been formed on the opposite side of the PSP. Such a PSP will attach to theendothelium with the nucleation layer facing the endothelial cells, and the releaseof NPs will occur only in one direction: from the nucleation layer towards thevessel wall.

• PSPs can be engineered to deliver drug and/or second-stage NPs upon endothelialbinding or at a tuned delivery rate. Enzymatically degradable cross-linking peptidesor pH-responsive polymers could be dispersed within the porous matrix of the PSPalongside NPs for environmentally triggered release.


269

• Functionalization with permeation enhancers will enable the PSPs to open tightjunctions of the endothelial lining, through which NPs can pass to augment and/orcreate appropriate EPR conditions.

• The MDS is capable of codelivering drug cocktails. Many chemotherapy protocolsinvolve a combination of drugs given together or in sequence. The PSP payloadvolume is large enough to carry a cocktail of free drugs and/or drugs containingNPs, together with thermal ablation agents and imaging NPs.

• The MDS enables In Silico Delivery Design to create a personalized therapy for eachdrug/disease combination. As the pharmaceutical industry has utilized large com-binatorial compound libraries to identify new drug candidates, similarly, the MDScan be assembled in a combinatorial way optimizing shape, size, chemistry, surfacetargeting modalities, and charge modifications of the PSP, for the wide choice ofavailable NPs.



Si3N4 film is not uniform. Nonuniform gas distribution during LPCVD.Nonuniform temperature during LPCVD.

Add more dummy wafers. Move the rela-tive position of the substrate to the gassource. Flip the substrate facing the direc-tion with respect to the gas source.Change the position of the substratewithin LPCVD tube. Wait longer for tem-perature stabilization before gas insertion.Improve temperature uniformity in thetube tuning the Si3N4 deposition recipe.

Litographic pattern is:1) too small or absent.2) too large or photoresistis absent.3) nonuniform.

The pattern is:1) underexposed or under-developed.2) overexposed or overdeveloped.3) improperly exposed or developed.ORThe mask or substrate is contaminated with dust.

1) Increase exposure or development time.2) Decrease exposure or developmenttime.3) Vary exposure or development time.ORClean the mask/substrate: acetone-metha-nol-isopropanol or piranha.

PSP is:1) too flat.2) too rounded.3) too thin.4) too thick.5) cracked.6) is not released.7) released ahead of time.

1) Dry etch is too shallow.2) Dry etch is too deep.3) Porosification time is too short.4) Porosification time is too long.5) Porosification or release current density is too high.6) Release current density is too low.7) Release current density is too high.

1) Increase dry etch time.2) Decrease dry etch time.3) Increase porosification time.4) Decrease porosification time.5) Reduce release current density orporosification current density.6) Increase release current density.7) Reduce release current density.

Pore size is:1) too big.2) too small.

Porosification current density is:1) too high.2) too low.

1) Reduce porosification current density.2) Increase porosification current density.

Z2’s aperture is blocked. Dirty cuvette or ISOTON. Hit “UNBLOCK” on control panel;ORRemove sample, wash aperture, loadAccuvette with filtered ISOTON, hit“FUNCTION” → “FLUSH APERTURE”


270


Z2’s software shows morethan one central peak[Figure 13.11(b)].

Small second peak: Large number of PSPs stickingtogether;ORLarge second (or more) peak(s): PSPs breaking upwithin sample being measured (seen with PSPs withhigh porosity).

Briefly sonicate the sample longer.Check sonicator water level; the waterneeds to be set at the marked operatinglevel.

“RUN” button onFACSCalibur is not greenafter pushed or the Statusis Standby.

The machine cannot build up enough pressure to createthe proper flow rate to introduce PSPs into the system.

If tube does not fit properly and you hearpressure/gas leaving the top of tube, try anew tube or replace the o-ring.

FACSCalibur’s softwareshows a high noise orbackground acquisition.

The sensitivity of the machine is set such that it candetect extremely small particles/events, which inher-ently results in the detection of any dust or dirt thatmay be present in the system due to previous samplesor poor cleaning.

Allow the machine to aspirate distilledwater, and measure the tube with waterand observe the number of events; iflarge, repeat.If after several cycles of aspirating waterdoes not work, check sheath fluid leveland replenish if necessary.

Acknowledgments

The authors would like to recognize M. Landry for excellent graphical support, Dr. D.L.Haviland for his superior expertise and experience with flow cytometry, Dr. Glen Snyderfor his technical support at ICP-AES, Dr. Kaushal Rege for his continual support anduseful commentary when compiling this chapter, and all present and past members ofThe Division of NanoMedicine for useful discussion and assistance.

References

[1] Pope-Harman, A., et al., “Biomedical nanotechnology for cancer,” The Medical Clinics of NorthAmerica, Vol. 91, No. 5, September 2007, pp. 899–927.

[2] Moghimi, S.M., A.C. Hunter, and J.C. Murray, “Nanomedicine: current status and future pros-pects,” FASEB Journal, Vol. 19, 2005, pp. 311–330.

[3] Debbage, P., “Targeted drugs and nanomedicine: present and future,” Curr. Pharm. Des., Vol. 15,No. 2, 2009, pp. 153–172.

[4] Northfelt, D.W., et al., “Doxorubicin encapsulated in liposomes containing surface-bound poly-ethylene glycol: pharmacokinetics, tumor localization, and safety in patients with AIDS-relatedKaposi’s sarcoma,” J. Clin. Pharmacol., Vol. 36, No. 1, 1996, pp. 55–63.

[5] Decuzzi, P., et al., “The effective dispersion of nanovectors within the tumor microvasculature,”Ann. Biomed. Eng., Vol. 34, No. 4, 2006, pp. 633–641.

[6] Müller, R.H., et al., “Phagocytic uptake and cytotoxicity of solid lipid nanoparticles (SLN) stericallystabilized with poloxamine 908 and poloxamer 407,” J. Drug Target., Vol. 4, No. 3, 1996,pp. 161–170.

[7] Ten Tije, A.J., et al., “Pharmacological effects of formulation vehicles : implications for cancer che-motherapy,” Clin. Pharmacokinet., Vol. 42, No. 7, 2003, pp. 665–685.

[8] Katragadda, S., et al., “Role of efflux pumps and metabolising enzymes in drug delivery,” ExpertOpin. Drug Deliv., Vol. 2, No. 4, 2005, pp. 683–705.

[9] Bassingthwaighte, J.B., C.Y. Wang, and I.S. Chan, “Blood-tissue exchange via transport and trans-formation by capillary endothelial cells,” Circ. Res., Vol. 65, 1989, pp. 997–1020.

[10] Silva, G.A., “ Nanotechnology approaches to crossing the blood-brain barrier and drug delivery tothe CNS,” BMC. Neurosci., Vol. 9, 2008, pp. Suppl 3:S4.

[11] Jang, S.H., et al., “Drug delivery and transport to solid tumors,” Pharm. Res., Vol. 20, No. 9, 2003,pp. 1337–1350.

[12] Jain, R.K., “Transport of molecules, particles, and cells in solid tumors,” Annu. Rev. Biomed. Eng.,Vol. 1, 1999, pp. 241–263.

Acknowledgments

271

[13] Nies, A.T., “The role of membrane transporters in drug delivery to brain tumors,” Cancer Letters,Vol. 254, 2007, pp. 11–29.

[14] Sakamoto, J., et al., “Antibiological barrier nanovector technology for cancer applications,” ExpertOpin. Drug Deliv., Vol. 4, 2007, pp. 359–369.

[15] Peer, D., et al., “Nanocarriers as an emerging platform for cancer therapy,” Nat. Nanotechnol., Vol.2, No. 12, 2007, pp. 751–760.

[16] Gabizon, A.A., “Stealth liposomes and tumor targeting: one step further in the quest for the magicbullet,” Clin. Cancer Res., Vol. 7, No. 2, 2001, pp. 223–225.

[17] Decuzzi, P., et al., “Intravascular delivery of particulate systems: does geometry really matter?”Pharm. Res., Vol. 26, No. 1, 2009, pp. 235–243.

[18] Eckelman, W.C., and C.A. Mathis, “Targeting proteins in vivo: in vitro guidelines,” Nucl. Med. Biol.,Vol. 33, No. 2, 2006, pp. 161–164.

[19] Brannon-Peppas, L., and J.O. Blanchette, “Nanoparticle and targeted systems for cancer therapy,”Adv. Drug Deliv. Rev., Vol. 56, 2004, pp. 1649–1659.

[20] Yezhelyev, M.V., et al., “Emerging use of nanoparticles in diagnosis and treatment of breast can-cer,” Lancet Oncol., Vol. 7, 2006, pp. 657–667

[21] Lin, M.Z., M.A. Teitell, and G.J. Schiller, “The evolution of antibodies into versatile tumor-target-ing agents,” Clin. Cancer Res., Vol. 11, 2005, pp. 129–138.

[22] Farokhzad, O.C., J.M. Karp, and R. Langer, “Nanoparticle-aptamer bioconjugates for cancer target-ing,” Expert Opin. Drug Deliv., Vol. 3, No. 3, 2006, pp. 311–324

[23] Simberg, D., et al., “Biomimetic amplification of nanoparticle homing to tumors,” Proc. Natl. Acad.Sci. USA, Vol. 104, No. 3, 2007, pp. 932–936.

[24] Yang, X., et al., “Selection of thioaptamers for diagnostics and therapeutics,” Ann. N. Y. Acad. Sci. ,Vol. 1082, 2006, pp. 116–119.

[25] Allen, T.M., “ Ligand-targeted therapeutics in anticancer therapy,” Nat. Rev. Cancer, Vol. 2, No. 10,2002, pp. 750–763.

[26] Duncan, R., “Designing polymer conjugates as lysosomotropic nanomedicines,” Biochem. Soc.Trans., Vol. 35, Pt. 1, 2007, pp. 56–60.

[27] Ferrari, M., “Nanovector therapeutics,” Curr. Opin. Chem. Biol., Vol. 9, No. 4, 2005, pp. 343–346.[28] Ferrari, M., “Cancer nanotechnology: opportunities and challenges,” Nature Rev. Cancer, Vol. 5,

No. 3, 2005, pp. 161–171.[29] Tasciotti, E., et al., “Mesoporous silicon particles as a multistage delivery system for imaging and

therapeutic applications,” Nat. Nanotechnol., Vol. 3, No. 3, 2008, pp. 151–157.[30] Canham, L.T., et al., “Derivatized mesoporous silicon with dramatically improved stability in sim-

ulated human blood plasma,” Adv. Mater., Vol. 11, No. 18, 1999, pp. 1505–1507.[31] Cohen, M.H., et al., “Microfabrication of silicon-based nanoporous particulates for medical appli-

cations,” Biomedical Microdevices, Vol. 5, No. 3, 2003, pp. 253–259.[32] Serda, R.E., et al., “Porous silicon particles for imaging and therapy of cancer,” Nanomaterials for the

Life Sciences, 2009.[33] Canham, L.T., Properties of Porous Silicon: Crystal Research and Technology, Vol. 34, New York:

Wiley-VCH, 1999.[34] Decuzzi, P., et al., “Adhesion of microfabricated particles on vascular endothelium: a parametric

analysis,” Ann. Biomed. Eng., Vol. 32, 2004, pp. 793–802[35] Zhang, M., T. Desai, and M. Ferrari, “Proteins and cells on PEG immobilized silicon surfaces,”

Biomaterials, Vol. 19, No. 10, 1998, pp. 953–960.[36] Nashat, A.H., M. Moronne, and M. Ferrari, “Detection of functional groups and antibodies on

microfabricated surfaces by confocal microscopy,” Biotechnol. Bioeng., Vol. 60, No. 2, 1998,pp. 137–146.

[37] Nijdam, A.J., et al., “Physicochemically modified silicon as a substrate for protein microarrays,”Biomaterials, Vol. 28, No. 3, 2007, pp. 550–8.

[38] Anglin, E.J., et al., “Porous silicon in drug delivery devices and materials,” Adv. Drug Deliv. Rev., Vol.60, 2008, pp. 1266–1277.

[39] Akin, D., et al., “Bacteria-mediated delivery of nanoparticles and cargo into cells,” NatureNanotechnology, Vol. 2, No. 7, 2007, pp. 441–449.

[40] Xiang, L., et al., “Bacterial magnetic particles (BMPs)+PEI as a novel and efficient non-viral genedelivery system,” The Journal of Gene Medicine, Vol. 9, 2007, pp. 679–690.

[41] Steinfeld, U., et al., “T lymphocytes as potential therapeutic drug carrier for cancer treatment,” Intl.Jour. of Pharm., Vol. 311, 2006, pp. 229–236.

[42] Souza, G.R., et al., “Networks of gold nanoparticles and bacteriophage as biological sensors andcell-targeting agents,” Proc. Natl. Acad. Sci. USA, Vol. 103, No. 5, 2006, pp. 1215–1220.

[43] “Z2 COULTER COUNTER Cell and Particle Counter,” Beckman Coulter, http://www.beckman.com/products/instrument/partChar/pc_z2.asp, last accessed on February 2, 2009.


272

[44] Rahil-Khazen, R., et al., “Validation of inductively coupled plasma atomic emission spectrometrytechnique (ICP-AES) for multi-element analysis of trace elements in human serum,” Scand. J. Clin.Lab. Invest., Vol. 60, No. 8, 2000, pp. 677–686.

[45] Gentile, F., et al., “The effect of shape on the margination dynamics of non-neutrally buoyant par-ticles in two-dimensional shear flows,” Jour. of Biomech., Vol. 41, 2008, pp. 2312–2318.

[46] Decuzzi, P., et al., “A theoretical model for the margination of particles within blood vessels,” Ann.Biomed. Eng., Vol. 33, No. 2, 2005, pp. 179–190.

Acknowledgments

273

C H A P T E R

1 4Mathematical Modeling of NanoparticleTargeting

Elena V. Rosca1, 2 and Michael R. Caplan1, 2*

1Harrington Department of Bioengineering, Arizona State University, 2Center for InterventionalBiomaterials, Arizona State University

*Corresponding Author: Michael R. Caplan, Harrington Department of Bioengineering, Arizona State Uni-versity, P.O. Box 879709, Tempe, AZ 85287-9709, Phone: 480-965-5144, Fax: 480-727-7624, e-mail:[email protected].

275

Abstract

Mathematical models based on the principle of conservation of mass cangreatly enhance understanding of the behavior of and lead to design principlesfor nanoparticles used for drug or image contrast agent targeting. Implement-ing such models can be performed at the molecular scale, tissue scale, andorganism scale, or at combinations of these scales. Molecular scale modeling isfocused on changes in concentrations of bound and unbound nanoparticleswith respect to time using chemical kinetics. Tissue scale modeling adds con-vection and diffusion within tissues along with reaction terms as in molecularscale modeling. Organism scale modeling uses compartmental models withrates of mass exchange between compartments. Once the model is capable ofgenerating accurate predictions of the system’s behavior under conditions notyet studied, the equations on which the model is based most likely incorporatethe physical phenomena important to the behavior of the nanoparticles.

Key terms mathematical modelingnanoparticlesdrug deliverymass transportprotein bindingligandscell surface receptors

14.1 Introduction

All branches of science and engineering rely on some type of modeling to analyze, inter-pret, or explain data. Therefore, models serve diverse functions from aiding scientists inorganizing data to deciding what data mean and developing an understanding of com-plex phenomena [1]. For example, understanding a complex event from empiricalexperimentation might prove to be a difficult and daunting task involving multiple tri-als to uncover the complex interplay of the principles involved. A theoretical model can-not only be helpful but sometimes critical to understanding the complex interplay ofimportant factors affecting a system’s behavior. Mathematical models are a class of mod-els that involve the use of mathematics to describe a set of physical phenomena quanti-tatively. Such models allow a researcher to simulate one possible set of relationshipsamong the components that he or she deems important. Comparison of the simulationresults to experimental data can indicate that the factors that the scientist deemedimportant are indeed working the way modeled if the model and data produce similarresults. If a large disparity between predicted and experimental data is observed, themodel is perhaps too simplistic (omitting major underlying phenomena) or the interac-tions may be modeled incorrectly. If the discrepancy is relatively small, perhaps someparameters are estimated inaccurately.

This can be thought of as using a model as a hypothesis generator. The model is infact a statement of the hypothesis: that the physical components of the system relate asdescribed in the mathematics. The model is then used to simulate what would happenunder various sets of conditions to find a set of theoretical results that, if found to existin reality, would lend credence to the relationships being as they are described in themodel. The experiments are then performed, and the experimental results are comparedto the theoretical predictions as described above. If there is a good fit between predictionand data, it is possible that the phenomena are accurately described in the model. How-ever, the normal caveats about experimental validation of hypotheses apply, namely,that one test of a hypothesis does not prove the hypothesis to be true. Additionally, aswe will discuss later in this chapter, there is the added caveat that a large number of fittedparameters can make a model fit many sets of data even if the model is not an accuratedescription of the physical phenomena.

Targeting with nanoparticles is a complex problem that encompasses multiple phe-nomena: interaction of the particle with the target cells, delivery throughout the tissueof interest, stability of targeting moieties (ligands), clearance by various organs, andothers. Modeling these factors can assist in the rational development of more effectivetargeting particles. In particular, modeling can help researchers deal with tradeoffsinherent to the design process such as those between dose of particles and specificity [2].Here we describe methods for modeling at three different length scales: (1) interactionsof the particles with the target at the molecular/cellular scale, (2) delivery and diffu-sion/convection through tissue at the tissue scale, and (3) systemic delivery, clearance,and biodistribution at the organism scale. At each of these scales, this chapter discussesthe available modeling techniques applicable to that length scale, provides in-depth dis-cussion of how to apply those techniques, and indicates how these techniques can be orhave been applied to advance targeting of nanoparticles.

Molecular/cellular scale modeling is mainly concerned with interaction between thenanoparticles and cell surface receptors. Typically nanoparticles are carriers of specificmolecules (ligands) able to interact with cell surface receptors effectively creating

Mathematical Modeling of Nanoparticle Targeting

276

multivalent constructs [3–8]. Modeling at this level has been focused on understandingthe effects of multivalent interactions. These studies suggest that multivalent interac-tions exhibit increased avidity (overall increased binding of the constructs of highervalency), which is predicted to result in greater targeting specificity [2]. Chemicalkinetics are used to describe the interactions between such multivalent particles andcells. Thermodynamics can be used to better estimate parameters for these biophysicalmodels [8, 9].

Tissue scale modeling adds diffusion and/or convection of the particles through thetissue in which the target cells reside. One method of delivery to tumor can be via pas-sive transport from the blood due to high permeability and multiple fenestrations intumor vasculature [10–12]. Mass transport in these cases is a function of diffusion, inter-stitial pressure, and tumor pressure. Tissue heterogeneity and anisotropy are also factorsthat affect fluid distribution. A different approach to delivery consists of local deliveryfollowed by diffusion and/or perfusion [13, 14]. Models in this case are concerned withbulk fluid flow velocities, tissue permeability, filtration of nanoparticles, and otherparameters that can influence nanoparticle distribution within the tissue.

Last, modeling at the organism scale involves a much broader view of the issue athand. This scale typically seeks to address biostability, biodistribution, and clearancerates of the nanoparticles. Some parameters important in organism scale modeling arethe size of the particles, injection volume and location, dose frequency, and concentra-tion [15]. Such issues are often studied using compartmental models in which the organsor tissues encountered by the particle are modeled as compartments that are intercon-nected through rates of transfer from one compartment to another.

14.2 Molecular/Cellular Scale

14.2.1 Methods

In molecular/cellular scale modeling of nanoparticles, the model describes binding ofthe particle with the target cell via the cells’ surface receptors or other surface-boundmarkers. The most widely used method to study biophysics at this scale is chemicalkinetics. Also known as receptor-ligand modeling, this approach was first adapted tostudy binding of molecules to cell surface receptors by Perleson [16] and has since beenextensively reviewed by Lauffenburger and Linderman [17]. The foundation for thistype of model is a single binding event between a cell surface receptor and a solubleligand (such as a growth factor) forming a bound complex. This event can be describedand simulated mathematically using the principle of conservation of mass with thefollowing set of equations:

dLdt

k LR k Cf r= − + (14.1)

dCdt

k RL k Cf r= − (14.2)

R R C= −0 (14.3)


277

where L is the concentration of the ligand, R is the concentration of the unbound recep-tor, C is the concentration of the receptor-ligand complex, kf is the association rate, kr isthe dissociation rate, and R0 is the total density of receptors. Equations such as these canbe written using the following procedure:

1. Determine the number of terms in each equation. Equations (14.1) and (14.2) havetwo terms each because each species (L for (14.1) and C for (14.2)) participates in two

reactions (association L + R → C and dissociation C → L + R).

2. Determine the species variables (concentrations) that must be in each term. Thereactants always determine kinetic order (note: these must be mechanistic reactions,not overall stoicheometry). The first term, describing association, is written withsecond-order kinetics since two freely moving molecules must collide for associationto occur. The second term, describing dissociation, is written with first-order kineticsbecause only the presence of receptor-ligand complexes (no collision) is necessaryfor these events to occur. The appropriate rate constant is then added to each term.

3. Determine the sign of each term. The signs of each term are written to describe

whether association or dissociation adds (+) or removes (−) ligand or complexes fromthe system. It can be seen by adding (14.1) and (14.2) that the overall change in massof the system with time is zero; thus, mass is conserved. Mass must be conserved forthe overall system.

These equations are inserted into a program which can solve ordinary differentialequations such as MATLAB (Mathworks) as follows:

1. Enter each parameter value by naming them p.name with the syntax “p.kf = 1e6;” for

the example of setting the association rate to 1 × 106. Also enter initial conditions andthe time at which the simulation will end (p.tf) using the same syntax. Note that theuser must make sure units (e.g., meters, seconds, and so forth) are consistent.

2. Enter “[t y] = ode15s(@equationfile, [0 p.tf], y0, options, p);” where “ode15s” is theordinary differential equation solver chosen, “equationfile” is the name of thefunction where the equations are defined, “y0” is a row vector containing the initialconditions, options are defined as in MATLAB help, and “p” calls the parametervalues defined above.

3. Output variables can be calculated. For example the number of unbound receptors

could be calculated by “R = p.R0 − y(:,2);” where “y(:,2)” denotes y values in thesecond column of the [t y] matrix as a function of time. Figures can be plotted basedon these calculated values or on the raw data as desired.

4. The equations are defined in a file beginning with “function yp = equationfile (t, y,p);” where “equationfile” must match the name supplied in step 2 exactly. Thevariables are defined as “L=y(1);” and “C=y(2)”. Immediately after this, any variablescalculated with algebraic equations should be calculated, in this case R(t) is definedas “R = p.R0 – C;”.

5. Finally the ordinary differential equations are defined as “yp(1) = -p.kf * L * R + p.k2 *C;” and “yp(2) = p.kf * L * R – p.kr * C;”.

Perleson and DiLisi extended this model by applying it to receptor clustering andbinding of multivalent ligands (such as antibodies) to oligomeric receptors of B cells[16]. Since nanoparticles have many ligands bound to their surfaces, they likely behavesimilarly to these multivalent molecules. Converting Perleson and DiLisi’s model into


278

the notation used in this chapter, L0 represents the total concentration of divalent mole-cules, C1 represents the concentration of divalent molecules bound by one ligand to thecell, and C2 represents the concentration of divalent molecules bound to the cell by twoligands, which also corresponds to the concentration of cross-links.

( ) ( ) ( )dLdt

k R t L t k C tf r= − +2 1 (14.4)

( ) ( ) ( ) ( ) ( ) ( )dCdt

k R t L t k C t k C t R t k C tf r x x1

1 1 22 2= − − + − (14.5)

( ) ( ) ( )dCdt

k C t R t k C tx x2

1 22= − − (14.6)

( ) ( ) ( )R R t C t C t0 1 2= + + (14.7)

where kx and k-x are the association and dissociation rate constants of the second ligandto bind (thus forming the crosslink). These equations are generated in the same manneras described for the single ligand, but the coefficient 2 is necessary in (14.4) and (14.5) toadjust the probability of collision since there are two ligands on the divalent moleculeand in (14.5) and (14.6) because a C2 species occupies two receptors either of which candissociate.

The rate constants, kx and k-x, differ from kf and kr by a factor accounting for theincreased effective concentration of the ligand when it is tethered to the cell surface bythe first receptor-ligand bond. Shewmake et al. [18] defined a factor, VR, which accountsfor the increased effective concentration. This binding enhancement factor corrects theassociation rate constant of secondary binding events in relation to the first bindingevent. Shewmake’s work is based on work by Krishnamurthy, Whitesides, and coworkers[8], who modeled an inhibitor tethered to the enzyme which it inhibits. Their model cal-culates Ceff, which is similar to VR*C1 in (14.11), as a function of the root-mean-squared

distance between the ends of the polymeric linker, Rg = ⟨r2⟩1/2, and the distance betweentether site and binding site, a. Shewmake et al. applied this to multivalent targeting forseveral cases, including a random-coil model for linkers between ligands resulting in:

V IhR = ϕ (14.8)

( )( )

IR

aR

g g

= −⎛

⎝⎜⎜

⎞

⎠⎟⎟

3

2

32

1 2

1 2

2

2πexp (14.9)

where ϕ is a scalar accounting for excluded volume and h is the ratio of interstitial fluidand the cell surface area.

Caplan and Rosca applied this model to multivalent targeting by allowing for two ormore different cell types that differ only in the number of receptors expressed (cell typeswith different R0 values). Using this model they investigated the binding of targetingmolecules with various valence (monovalent, divalent, trivalent, and tetravalent), ofwhich various concentrations were applied to cells, and for constructs targeting onereceptor type (homovalent) or two receptor types (heterovalent). For the homo,bivalent


279

model, (14.4) is modified to allow binding of the unbound construct to two different celltypes, M and N:

( ) ( )dLdt

k C C k R Rr M N f M N= + − +1 1 2 (14.10)

Two sets of (14.5) to (14.7) are created, one set for constructs bound to cell type Mand another set for constructs bound to cell type N. An additional difference from (14.4)to (14.7) is the introduction of the parameter VR, the binding enhancement factor. Theintroduction of this parameter allows the replacement of kx and k-x with kf and kr becauseVR encompasses the effects of secondary binding events of the multivalent constructs so,for instance, the equation for C1,M (C1 binding to cell type M) becomes:

dC

dtk LR k C k V C R k CM

f M r M f R M M r M1

1 1 22,, , ,= − − + (14.11)

in which the association rate between C1,M and an additional receptor is multiplied by VR.This model can be used to test the dominant premise of targeting, that more drug or

imaging molecules will be bound to the target cell, by taking the ratio of constructsbound to the target cell (C1,M + C2,M) versus the number of constructs bound to nontargetcells (C1,N + C2,N). This ratio, defined as specificity, provides a quantitative description ofhow effective the targeting would be under such conditions.


The equations in and of themselves are the mathematical representation of the physicalphenomena, but most often they are a means to an end rather than the ultimate goal. Inthis case, Caplan and Rosca sought to use the mathematical model to elucidate princi-ples for rational design of such multivalent constructs. By considering the possible waysin which the constructs could be designed or employed, several points of controlbecame apparent by which designers can modify constructs. In this system these includethe affinity of the receptor-ligand bond, the number of receptors on the target cell, theratio of the receptors between target and nontarget cells, the concentration of the con-struct (dose), number of ligands on the construct (valence), and the properties of thelinker between ligands. These correspond to parameters in the equations or initial con-ditions KD (kr/kf ), R0,M, R0,M/R0,N, L0, n (as in Cn), and VR, respectively.

Caplan and Rosca varied the receptor number on the target cell (R0,M), construct con-centration (L0), and valence (n) while keeping the other parameters constant. Theydeveloped sets of equations for homo,trivalent, homo,tetravalent, hetero,divalent (twoligands of each type), and hetero,trivalent (three ligands of each type) constructs similarto those shown above. Results from these models, shown in Figure 14.1, depict the simu-lated binding specificities of multivalent constructs when the initial construct concen-tration (Figure 14.1(a)) and the number of receptors on the target cells (Figure 14.1(b))are varied.

14.2.3 Discussion and Commentary

Illustrating the purpose of modeling at this scale, it is instructive to note several things.First, Caplan and Rosca were able to narrow the scope of their experimental study from


280

all possible variations in the design of these constructs to those aspects of the design thathad direct correlations to parameters in the mathematical equations. Likewise, the needto quantify a measure of output, in this case specificity, highlighted the need to studybinding to two different cell types. In vitro characterization of targeted constructs pre-dating this work studied binding of constructs to the target cell type and control experi-ments were typically constructs with a nonfunctional ligand. Modeling showed thatmultivalency could achieve increased avidity for the target cell type without necessarilyincreasing specificity for the target cell type. Thus, merely writing the equations andchoosing how to quantify the output of the model provided an advance to the field inthe form of clarifying this metric.

Second, the model yielded insights that would not be available by intuition alone or,if intuition could have achieved them, were not intuited prior to the application of thismodel. Modeling provides a formalism for breaking very complex problems downinto manageable pieces which can then be assembled into the mathematical modelsdescribed. The molecular scale models shown here are broken down into equations foreach species of interest (e.g., L, C1, C2, and so forth), and each of these equations is fur-ther broken down into a summation of terms which each represent an association or dis-sociation event. When these pieces were reassembled and parameters varied, the resultsproduced provided insights that were not initially obvious.

For example, when the concentration of construct was varied (Figure 14.1(a)), thespecificities of multivalent constructs at high concentration were no different thanthose for monovalent. At low concentration, however, the expected trend for whichspecificity increases as valence increases was predicted. Since the model keeps track ofthe various individual species (C1, C2, and so forth), Caplan and Rosca were able to deter-mine that this was due to the prevalence of C1 species at high concentration, which ineffect made all binding monovalent due to saturation of the available receptors evenwhen the constructs were multivalent. In a similar manner, when receptor density onthe target cell was varied, a biphasic trend was observed with specificity increasing atlower receptor number and decreasing at higher receptor number. Again, availability ofinformation on the individual species revealed that specificity is mostly a function ofthe percentage of constructs bound by most or all of the ligands (C2 for divalent, C3 fortrivalent, and so forth). At lower receptor numbers, the percentage of these species


281

Log of construct concentration (M)

Bind

ing

spec

ifici

ty

−11 −91

1.5

2

2.5

3

3.5

−7 −5 −3

(a) (b)

Bind

ing

spec

ifici

ty

1

1.5

2

2.5

3

10 100Receptors on target cell (#/cell)

1000 10000 100000

Figure 14.1 Binding specificity of different constructs (monovalent is depicted by a thick dashedline, divalent is depicted by a thick solid line, trivalent is depicted by a thin dashed line, andtetravalent is depicted by a thin solid line) as (a) construct concentration and (b) receptor numbers arevaried. (From: [2]. © 2005 Reproduced with permission from Elsevier.)

increases more rapidly on target cells than on nontarget cells; however, at higherreceptor numbers, the percentage on target cells approaches 100%, so the percentageincreases more rapidly on nontarget cells. Thus, although the avidity monotonicallyincreases with increasing receptor numbers, specificity is predicted to be biphasic.

These insights provide general design principles that can be used to increase thelikelihood of successful application of nanoparticle targeting. The results shown inFigure 14.1 indicate that the receptor-ligand binding affinity should be two to threeorders of magnitude weaker than the required dose if multivalency is to achieve specific-ity in excess of the ratio of receptors. Additionally, a receptor target which expresses amid-range number of receptors must be chosen even though a receptor at very low orvery high copy number might have a greater ratio of expression between target andnontarget. Experiments must then be performed to validate such design principles, butmodeling can provide the initial impetus to perform such experiments and indicate howone should carry out the experiment to see the predicted result.

14.3 Tissue Scale

14.3.1 Methods

Tissue scale modeling can be used to address spatial variations in tissues. An example ofthis level of modeling is diffusion/convection modeling of nanoparticles delivereddirectly to tissue containing a tumor. Models at the tissue scale can also account for spa-tial variations in tissue or construct that arise either due to tissue architecture, such asthe growth of a tumor in the tissue, or through delivery of the construct in a particularway (e.g., systemically through the blood or injected directly into the tissue). The appli-cation of the principle of conservation of mass to such convection/diffusion problemshas a very long history, but recently these principles have been applied directly tonanoparticle targeting.

Morrison et al. [19] developed a model describing the injection of macromoleculesinto brain tissue in which the macromolecules can convect with fluid flow, diffuse, bedriven across a capillary wall into the blood stream, or be inactivated by metabolism.

( ) ( ) ( )( ) ( )[ ]RCt

D C vC L s p p e C k Cd e p e iPemv

irr

∂

∂φ φ σ= ∇⋅ ∇ − ∇⋅ − − − − −1 1 (14.12)

where Rd accounts for the distribution of the macromolecule between the intracellularand extracellular space, De is the effective diffusion coefficient, φ is the volume fractionnot filled by cells or extracellular matrix, Lp is the vascular hydraulic conductivity, s is thecapillary surface area per volume of tissue, pi and pe are the interstitial and Starlingpressures, Pemv is the microvascular Peclet number, and kirr is the rate constant fordegradation of the macromolecule. This equation accounts for accumulation of thebiomacromolecule with time (left side), diffusion (first term, right), convection (secondterm, right), loss to the blood stream (third term, right), or deactivation (fourth term,right). Solving (14.12) requires one initial condition (in this case C = 0 at t = 0) and twoboundary conditions. One boundary condition at the injection site (r = 0) is set so thatthe concentration of the macromolecule in the injection is held constant (C = C0). A

typical second boundary condition used as r → ∞ is that the concentration remains


282

unchanged (C = 0 at r = ∞). Morrison et al. use a simplified version of this equation, inwhich the third term is omitted, to model convection-enhanced delivery to the brain ofa therapeutic molecule which cannot cross the blood-brain barrier. Similar models canbe applied to delivery of nanoparticles to tissue if parameter values are known for De andthe retardation coefficient (σ) of the particles in the tissue. Rosca et al. [20] have best fit

these values for targeted polymers and quantum dots and found that De values of 6 × 10−6

and 1 × 10−6 and filtration coefficients (1 − σ) of 1 and 0.25, respectively, describe the dif-fusion/convection of these particles in an agarose mock of brain tissue.

Stukel et al. [21] have incorporated the molecular scale binding interactions dis-cussed above into Morrison et al.’s model of convection-enhanced delivery. In thestudy, brain tissue was modeled using a nodal network with a region of healthy cells anda subdomain of tumor cells. The method presented here is the finite difference schemeused by Stukel et al.; however, it is possible to perform similar modeling using COMSOLMultiphysics which is a finite element simulation. The method described for molecularscale modeling is modified as follows:

1. Equations are derived as in molecular scale modeling; however, there are additional

terms for diffusion, D∇2L, and convection, –v∇L, which can be modeled in Cartesiancoordinates with Taylor series expansions:

D L DL

xL

yD

L L Lh

L Li i i j j∇ = +⎡

⎣⎢

⎤

⎦⎥ = + − +

++ − +22

2

2

21 1

212∂

∂

∂

∂− −⎡

⎣⎢⎤⎦⎥

12

2L

kj (14.13)

( ) ( ) ( )1 1 1 1− ∇ = − +⎡

⎣⎢

⎤

⎦⎥ = − −⎛

⎝⎜⎞⎠⎟

+−σ σ∂

∂

∂

∂σv L v

Lx

vLy

vL L

hx y xi i v

L L

kyj j−⎛

⎝⎜⎞⎠⎟

⎡

⎣⎢

⎤

⎦⎥

−1 (14.14)

where h and k are the distance between nodes in the x and y coordinates respectively,Li is the construct concentration at x-position i, and Lj is the construct concentrationat y-position j.

2. These equations are now nested in a loop structure which varies i and j from 1 to nand 1 to m, respectively, where nh and mk are the dimensions of the tissue. Thetumor is defined as several i,j pairs and distinguished by a greater p.R0 value.

3. Boundary conditions are set at i = 0, i = n + 1, j = 0, and j = m + 1. Concentrationboundary conditions are set as L0,j = 1e-9, for the example of a constant con-centration boundary condition at i = 0. No flux boundary conditions can be set bydeclaring L0,j = L1,j since there will be no flux at this boundary because there can beno concentration gradient.

Stukel et al.’s model is intended to represent a catheter placed within brain tissuethrough which a solution of drug-targeting construct is injected and the fluid velocity isoriented radially outward from the source. Equations (14.15) and (14.16) describe thetransport of drug-targeting constructs including convection. Equation (14.15) describesthe equation in Cartesian coordinates, while (14.16) shows the equation in sphericalcoordinates for which the Cartesian equation is a 2-D simplification.

∂

∂

∂

∂

∂

∂α

∂

∂

�

�

�

�

�

�

� � �

�

�

Lt

Lx

Ly

RL C vLxx= +

⎡

⎣⎢

⎤

⎦⎥ − + −

2

2

2

2 13 +⎡

⎣⎢

⎤

⎦⎥v

Lyy

∂

∂

�

�

(14.15)

14.3 Tissue Scale

283

∂

∂

∂

∂

∂

∂α

β ∂�

��

�

�

�

� � �

�

�Lt r r

rLr

RL Cr

=⎛

⎝⎜⎜

⎞

⎠⎟⎟ − + −1

322

1 2

Lr∂ �

(14.16)

where �L is the dimensionless concentration of unbound concentration (concentration

scale is R0),�R is the dimensionless unbound receptor density, �C1 is the dimensionless den-

sity of complexes with one ligand bound, �vx and �vy are the dimensionless Cartesian com-

ponents of vr = Q/4πr2 for which the flow rate, Q, is held constant at 3 μL min-1 (β =( )/( )/Q k Dr 4 3 2π = 6,906.59), and �x, �y, and �r are dimensionless coordinates ( D k/ is the

length scale). α is a dimensionless parameter (α = R0/KD) describing the relationshipbetween receptor density and receptor-ligand affinity; β is a dimensionless parameterdependent on the radial velocity, from which the x and y velocity components, vx and vy,are calculated for each time and matrix location. The equations for C1, C2, C3, and R arecalculated for each node at each time point, and these equations remain the same as inthe model discussed in the molecular scale section. Boundary conditions are set at the

catheter edge (r = 0.64 cm) to be � �L Linjectate= for 0 < t < tc, and �L = 0 for tc = t < tf where tc is

43,200 seconds. This simulates the injection of nanoparticles for some duration tc andthen injecting an artificial cerebrospinal fluid afterwards. The external boundary (edgesfar from catheter tip) is set to no-flux for all times. The source was placed in the center of

the matrix. Initial condition for the tissue is �L = 0 at t = 0.


Results from this diffusion/convection model of nanoparticle targeting are tracked astotal constructs (L + C1 + C2 + C3) at each node because imaging and/or therapy woulddepend on the total amount of construct—not just the amount bound to the cells. Fig-ure 14.2 illustrates that enhancement of contrast occurs only when unbound constructis washed away from the tissue. Diffusion alone can accomplish this, but the timerequired is impractically long. This model predicts that convection-enhanced deliverycan dramatically decrease the time required to achieve desirable levels of contrastbetween target and nontarget tissue. Figure 14.2 shows these results as well as demon-strates the volume of tissue that can be effectively probed using this approach.


The results from this diffusion/convection model reveal several points about nano-particle targeting of cancer, particularly in the brain. First, even when tumor location isunknown, the model predicts that it is possible to achieve contrast in excess of 10:1 fortumor tissue versus surrounding tissue. Second, the time required for constructs to beconvected to the tumor and then for unbound construct to be convected away from thetumor is large relative to typical imaging procedures but is reasonable for a clinical pro-cedure. Combined with the third prediction, that concentration must be less than thereceptor-ligand affinity to achieve high contrast, this severely limits the choice of con-trast agents that can be used. Typical magnetic resonance imaging (MRI) contrast agentsare long-lived but require high concentration. Conversely typical positron emissiontomography (PET) contrast agents can be used at low concentration but are veryshort-lived (minutes). The convection-enhanced delivery model shown here quantifies


284

the problems of applying multivalent targeting to cancer imaging, but it also provides ameans to study possible solutions to these issues.

14.4 Organism Scale

14.4.1 Methods

Models at the organism scale also make use of the principle of conservation of mass;however, using the approach discussed in the tissue scale modeling section would beimpractical. This is due to several reasons including that the number of nodes requiredto accurately reflect whether tissue/organ architecture would be very large, architecturewould require having regions in which diffusion dominates and regions in which con-vection dominates, and several other problems. Instead, when one needs to modelnanoparticle targeting on the scale of the whole organism, compartmental models aretypically used in which each compartment is modeled using one of the techniques dis-cussed above and the connections between the compartments are typically modeledusing mass transfer rates between compartments.

One recent example is the work of Davis et al., who investigated the efficacy of tar-geting and delivering siRNA to tumors using transferrin-targeted nanoparticles [22]. Themodel is comprised of three interconnected compartments: plasma, tumor interstitialvolume, and tumor intracellular volume. Concentrations (mol/L) of siRNA in each ofthese spaces are defined as C1 (plasma), C2 (interstitial tumor), and C3 (intracellulartumor). The equations governing the concentrations of the nanoparticles in thesecompartments are:

14.4 Organism Scale

285

Figure 14.2 Concentration of targeting constructs achieved via convection-enhanced delivery at dif-ferent time points and locations. Panels represent total construct concentration (z axis, molecules/cell)at each position (x and y axes, cm) at: (a) 12,000 seconds, (b) 43,000 seconds, (c) 86,000 seconds, (d)172,000 seconds, (e) 432,000 seconds, and (f) 864,000 seconds. Contrast is visible at (d) t = 172,000seconds and reaches maximum at (f) 864,000 seconds. (From: [21]. © 2008 Reprinted with permissionfrom Elsevier.)

( )dCdt

k CVV

k k Ce1

21 22

112 1= − + lim (14.17)

( )dCdt

k CVV

k CVV

k k C212 1

1

232 3

3

221 23 2= + − + (14.18)

dCdt

k CVV

k C323 2

2

332 3= − (14.19)

These equations are derived similarly to the method described for molecular scalemodeling with the difference that, instead of mechanistic reactions, the terms in eachequation describe rates of transfer from one compartment to another and are typicallywritten as first-order events. As can be seen in the schematic depiction of this model(Figure 14.3(a)), this is a relatively simple model in which the terms multiplied by k12

represent transfer from the blood to the tumor interstitial space, k21 the opposite, k23

represents uptake into tumor cells, k32 the opposite. The only additional term is the elim-ination of particles from the blood (kelim). V1, V2, and V3 are the volumes corresponding toeach compartment, and the ratio of these volumes must be accounted for because thetransfer between compartments is in mass per time; however, the variables beingcalculated are in concentration units.


This model was validated against in vivo data by fitting the extravasion rate (k12) and set-ting the rate of return to blood (k21) and tumor uptake (k23) to zero. As can be seen inFigure 14.3(b), the model result for nanoparticles in the tumor fits the experimental datavery well if a dilution rate of 25 min−1 is included in the definition of total particles in the


286

(a) (b)

Figure 14.3 Compartmental modeling of tumor-specific targeting. (a) The three-compartment modelthat was used to derive the equations describing tumor targeting. (b) A comparison of model predic-tions to experimental data collected. (From: [22]. © 2007 Reprinted with permission from PNAS.)

tumor. It should be noted that the dilution effect would have been more accurately han-dled by adding convection terms in (14.17) accounting for particles being injected in

(none in the dilutant, so this term is zero) and washed out (−Q C1/Vtv), where Q is thevolumetric flow rate of the dilutant. If Q is set to a nonzero value for t < 25 minutes and

to zero for t > 25 minutes, this would more accurately reflect the mass transport of theexperiment performed by Davis et al.


The strength of organism scale modeling is that it gives a description of the behavior ofthe targeting construct within the overall study system, the organism. However, thelimitation that is usually found in this type of model is the lack of mechanistic descrip-tion of the physical meaning of the parameters. In this case, the model yields generalinformation about the relative importance of the various transfer terms. For example,since the data can be fit by setting k21 and k23 to zero, we can reasonably conclude thatuptake into the cells does not affect the concentration of particles in the tumor and that,once the particles enter the tumor, their rate of transfer back into the blood is negligible.We also see that the data can be fit reasonably well with a first-order rate of transfer fromthe blood to the tumor tissue. However, this sort of compartmental model does notprovide any information as to why, mechanistically, the transfer from blood to tumor isfirst-order.

This limitation can potentially be overcome by combining a compartmental modelwith the molecular or tissue scale models discussed in the previous sections. For exam-ple, a model in which three compartments represent: (1) blood, (2) nontarget cells, and(3) tumor cells could be used. The rate of transfer from the blood to compartments 2 or 3could be modeled using a term similar to the third term of the right side of (14.18) orusing the terms representing binding of unbound constructs to cells as in (14.4). Once ineither compartment 2 or 3, the equations describing the biophysics of multivalent inter-actions could track unbound (L) and the various bound constructs (C1, C2, and so forth).The only constructs which could be exchanged with the blood would be unbound con-structs. Such a multiple-scale model could perhaps provide both the overall descriptionof nanoparticle performance while also providing mechanistic detail that is necessary touse modeling as a design tool.

14.5 Model Validation and Application

14.5.1 Statistical Guidelines

Mathematical models of physiological systems or processes are approximations andestimations of the real system. The process of creating the model can generate error dueto either under-parameterization or over-parameterization. An under-parameterizedmodel, a too simplistic representation of the system, will give inaccurate predictions dueto having made simplifying assumptions that are not quite true; thus, the predictionswill be inaccurate if important phenomena were omitted due to such simplifications. Anover-parameterized model, complex relative to the prior knowledge that the modelerhas about the system, contains many parameters for which there is little to no priorinformation upon which to estimate those parameters. These parameters need to be fit

14.5 Model Validation and Application

287

to data, and in many cases it is possible to fit experimental data even if the underlyingequations are not accurate descriptions of the physical behavior of the system. This isprobably the most common mistake because it is usually hidden under the impressionthat the equations provide a very good fit of the system’s behavior [23]. To avoid the riskof over-parameterization, two general rules to follow are: (1) the number of data pointsshould considerably exceed the number of parameters to be fitted, and (2) the technicalbehavior of the optimization process will improve as the ratio of data to parametersincreases [24].

Fitting a model to data entails the adjustment of model parameters to achieve a con-cordance between the model prediction and the actual data. However, parameter esti-mation can be accomplished independently of fitting from previously existing data,and, if this estimated value is not adjusted in the fitting process, model validation ismore meaningful [25]. Model fitting is often used to indicate the predictive value of themodel; however, there is a clear distinction between the two. Model fitting takes a modelthat is missing several key parameter values and then trains the model by finding thoseparameter values that allow the model to best describe the data. As discussed earlier,if the model includes the phenomena important to the function of the system, itshould be able to match the data closely. It is possible, particularly if the model isover-parameterized (fitting too many parameters), to match the data closely despite thefact that the model does not accurately describe the underlying phenomena. In such acase, if the model were to be used to predict what would happen if the conditions werechanged and the experiment run again, it would predict poorly.

The procedure for best fitting parameters is as follows:

1. Create either a spreadsheet or a matrix with experimental data in one column andthe model value for conditions identical to each experimental point in anothercolumn.

2. Subtract the model result from the data or vice versa.

3. Square the difference. This is the square of the error

4. Sum the “squares of the error.”

5. Vary parameter values either manually or through an automated method (somesoftware will have a feature that does this, but to do this in MATLAB requires writinga simple code to vary the parameters). Find the parameter set that minimizes the sumof squares of the error. These are the best-fit parameter values.

The true test of whether the model accurately reflects the phenomena important tothe function of the system is to use the model to make a prediction under conditions notused to fit the parameters in the model. This process of predictive validation is closelyrelated to hypothesis testing of an experimental hypothesis.

1. Use the model to make a prediction of what data will result under certain, previouslyunmeasured, conditions.

2. Perform experiments under those conditions to measure data. Perform sufficientreplicates so that 95% confidence intervals are of reasonable size (this will dependon the level of accuracy desired in the model and variance in the experimentalsystem).


288

3. For each condition predicted/measured, compute the t-statistic between the average

of the data (x) and the model predicted value (x) using the equation tx x

n= −

( / )σwhere

σ is the standard deviation of the data and n is the number of replicates.

4. Compare the value of the t-statistic with the established t-value corresponding to thedesired level of significance and degrees of freedom. If no statistical differences arefound, the hypothesis that the model prediction was different from the data was notfound to be valid, which is one indication that the model may be valid.

Determination of statistical significance by the method in step 3 is mathematicallyidentical to plotting the experimental data with their confidence intervals (i.e., 95%,99% confidence intervals) and the model prediction on the same plot and then visuallyinspecting to determine if the model predictions do or do not lie within the confidenceintervals (this will only work with confidence intervals—not standard deviations orstandard error of the mean). It is important to note that this approach will never rejectthe alternative hypothesis, and “not rejection” of the null hypothesis does not necessarymean that the null hypothesis is true—only that there is not sufficient evidence againstit. Also rejecting the null hypothesis does not mean that the alternative hypothesis istrue—only that it is more accurate given the data. Similarly, with this approach one cannever prove that the model is true—only that the conditions used to test the model didnot demonstrate a flaw in the model.


Problem Potential Solution

Code will not run. Check syntax (i.e., parenthesis, operator, variable names).Function name/call do not match (also dashes or numbers in the name may cause this error).

Suspension on time steps. Make equations dimensionless so that variables are on the same order of magnitude (~1).Try a different ordinary differential equation solver.Adjust tolerances.Check equations.

Concentrations are negative. Check loop structure (for finite differences).Check for sign error in equations.Check order of reaction.

Results do not seem correct. Check the predictions of the model against a case for which an analytical solution is known.Check the values of the parameters.

14.6 Summary Points

1. Mathematical models based on the principle of conservation of mass can greatlyenhance understanding of the behavior of and lead to principles for design ofnanoparticles used for targeting.

2. Implementing such models can be performed at the molecular scale, tissue scale, andorganism scale, or at combinations of these scales.

3. Molecular scale modeling is focused on changes in concentrations of species withrespect to time using chemical kinetics.

4. Tissue scale modeling adds convection and diffusion within tissues along withreaction terms as in molecular scale modeling.

14.6 Summary Points

289

5. Organism scale modeling uses compartmental modeling with rates of massexchange between compartments.

6. Once the model is capable of generating accurate predictions of the system’sbehavior under conditions not yet studied, the equations on which the model isbased most likely incorporate the physical phenomena important to the behavior ofthe nanoparticles.

Acknowledgments

The authors thank our funding sources: NIH (R21 NS051310, K22 DE014386) and Ari-zona Biomedical Research Commission Grant (#0707).

References

[1] Lubicher, D., and M. B. G. Manfred, Modeling Biology Structures, Behaviors, Evolutions. Cambridge,MA: The MIT Press, 2007, p. 396.

[2] Caplan, M. R., and E. V. Rosca, “Targeting drugs to combinations of receptors: a modeling analysisof potential specificity,” Ann. Biomed. Eng., Vol. 33, No. 8, 2005, pp. 1113–1124.

[3] West, J. L., and N. J. Halas, “Engineered nanomaterials for biophotonics applications: improvingsensing, imaging, and therapeutics,” Ann. Rev. Biomed. Eng., Vol. 5, 2003, pp. 285–292.

[4] Lowery, A., A. M. Gobin, D. S. Emily, J. N. Halas, and J. West, “Immunonanoshells for targetedphotothermal ablation of tumor cells,” International Journal of Nanomedicine, Vol. 1, No. 2, 2006.

[5] Gao, X., L. Yang, J. A. Petros, F. F. Marshall, J. W. Simons, and S. Nie, “In vivo molecular and cellu-lar imaging with quantum dots,” Curr. Opin. Biotechnol., Vol. 16, No. 1, 2005, pp. 63–72.

[6] Handl, H. L., J. Vagner, H. I. Yamamura, V. J. Hruby, and R. J. Gillies, “Lanthanide-basedtime-resolved fluorescence of in cyto ligand-receptor interactions,” Anal. Biochem., Vol. 330, No. 2,2004, pp. 242–250.

[7] Balthasar, S., K. Michaelis, N. Dinauer, H. von Briesen, J. Kreuter, and K. Langer, “Preparation andcharacterisation of antibody modified gelatin nanoparticles as drug carrier system for uptake inlymphocytes,” Biomaterials, Vol. 26, No. 15, 2005, pp. 2723–2732.

[8] Krishnamurthy, V. M., V. Semetey, P. J. Bracher, N. Shen, and G. M. Whitesides, “Dependence ofeffective molarity on linker length for an intramolecular protein-ligand system,” J. Am. Chem. Soc.,Vol. 129, No. 5, 2007, pp. 1312–1320.

[9] Mammen, M., S. Choi, and G. M. Whitesides, “Polyvalent interactions in biological systems: impli-cations for design and use of multivalent ligands and inhibitors,” Angew. Chem. Int. Ed., Vol. 37,1998, pp. 2754–2794.

[10] Folkman, J., “Tumor angiogenesis: therapeutic implications,” N. Engl. J. Med., Vol. 285, No. 21,1971, pp. 1182–1186.

[11] Dvorak, H. F., J. A. Nagy, and A. M. Dvorak, “Structure of solid tumors and their vasculature: impli-cations for therapy with monoclonal antibodies,” Cancer Cells, Vol. 3, No. 3, 1991, pp. 77–85.

[12] Dreher, M. R., W. Liu, C. R. Michelich, M. W. Dewhirst, F. Yuan, and A. Chilkoti, “Tumor vascularpermeability, accumulation, and penetration of macromolecular drug carriers,” J. Natl. Cancer Inst.,Vol. 98, No. 5, 2006, pp. 335–344.

[13] Morrison, P. F., M. Y. Chen, R. S. Chadwick, R. R. Lonser, and E. H. Oldfield, “Focal delivery duringdirect infusion to brain: role of flow rate, catheter diameter, and tissue mechanics,” Am. J. Physiol.,Vol. 277, No. 4, Pt. 2, 1999, pp. R1218–R1229.

[14] Jain, R. K., K. D. Janda, and W. M. Saltzman, “Drug discovery and delivery,” Mol. Med. Today, Vol. 1,No. 1, 1995, p. 4.

[15] Wang, Y., and F. Yuan, “Delivery of viral vectors to tumor cells: extracellular transport, systemicdistribution, and strategies for improvement,” Ann. Biomed. Eng., Vol. 34, No. 1, 2006,pp. 114–127.

[16] Perelson, A. S., and C. DeLisi, “Receptor clustering on a cell surface. I. Theory of receptor cross-link-ing by ligands bearing two chemically identical functional groups,” Mathematical Biosciences, Vol.48, 1980, pp. 71–110.

[17] Lauffenburger, D. A., and J. J. Linderman, Receptors Models for Binding, Trafficking and Signaling, NewYork, Oxford University Press, 1993, p. 365.


290

[18] Shewmake, T., F. Solis, and M. R. Caplan, “Effects of linker properties on multivalent targeting,”Biomacromolecules, Vol. 9, No. 11, 2008, pp. 3057–3064.

[19] Morrison, P. F., D. W. Laske, H. Bobo, E. H. Oldfield, and R. L. Dedrick, “High-flow microinfusion:tissue penetration and pharmacodynamics,” Am. J. Physiol., Vol. 266, No. 1, Pt. 2, 1994, pp.R292–R305.

[20] Rosca, E. V., J. M. Stukel, R. J. Gillies, J. Vagner, and M. R. Caplan, “Specificity and mobility ofbiomacromolecular, multivalent constructs for cellular targeting,” Biomacromolecules, Vol. 8,No. 12, 2007, pp. 3830–3835.

[21] Stukel, J. M., J. J. Heys, and M. R. Caplan, “Optimizing delivery of multivalent targeting constructsfor detection of secondary tumors,” Ann. Biomed. Eng., Vol. 36, No. 7, 2008, pp. 1291–1304.

[22] Bartlett, D. W., H. Su, I. J. Hildebrandt, W. A. Weber, and M. E. Davis, “Impact of tumor-specific tar-geting on the biodistribution and efficacy of siRNA nanoparticles measured by multimodality invivo imaging,” Proc. Natl. Acad. Sci. USA, Vol. 104, No. 39, 2007, pp. 15549–15554.

[23] Lemmon, A. R., and E. C. Moriarty, “The importance of proper model assumption in bayesianphylogenetics,” Syst. Biol., Vol. 53, No. 2, 2004, pp. 265–277.

[24] Garfinkel, D., and K.A. Fegley, “Fitting physiological models to data,” Am. J. Physiol., Vol. 246,No. 5, Pt. 2, 1984, pp. R641–R650.

[25] Landaw, E. M., and J. J. DiStefano, 3rd, “Multiexponential, multicompartmental, andnoncompartmental modeling. II. Data analysis and statistical considerations,” Am. J. Physiol.,Vol. 246, No. 5, Pt. 2, 1984, pp. R665–R677.

Acknowledgments

291

C H A P T E R

1 5Techniques for the Characterization ofNanoparticle-Bioconjugates

Benita J. Dair,1 Katherine Tyner,2 and Kim E. Sapsford3*1Division of Chemistry and Materials Science, Office of Science and Engineering, Center for Devices andRadiological Health, U.S. Food and Drug Administration. 2Division of Applied Pharmacology Research,Office of Testing and Research, Office of Pharmaceutical Science, Center for Drug Evaluation andResearch, U.S. Food and Drug Administration. 3Division of Biology, Office of Science and Engineering,Center for Devices and Radiological Health, U.S. Food and Drug Administration. 10903 New HampshireAvenue, Silver Spring, MD 20993, U.S.A. *Contact Author: [email protected]

293

Abstract

There are a variety of well-developed analytical tools that have been success-fully applied to unmodified/native nanoparticle (NP) characterization. Thequestion addressed here is whether these same technologies can be used for theanalysis of NP-bioconjugates, given the added complexity of their compositestructure, and if they can provide the additional information sought by theuser. The short answer is, of course, yes, but as found with unmodified NP anal-ysis, it is fair to say that no one technique can provide a complete characteriza-tion of engineered NP-bioconjugates. Rather, a combination of techniquesmust be used to characterize the many metrics associated with the NP scaffolditself and also the overall NP-bioconjugate assembly. The aim of this chapter isto provide the reader with an overview of the general principles and potentialinformation available from each technology, along with some pertinent exam-ples which highlight both the potential advantages and/or drawbacks of eachparticular technique.

Key terms reviewnanoparticlebiomoleculebioconjugationseparationmicroscopyspectroscopicmass spectroscopythermal

15.1 Introduction

Nanotechnology is a rapidly expanding, multidisciplinary field of research with thepotential to revolutionize many fundamental and applied aspects of science. In particu-lar, nanoparticles (NPs) modified with biological molecules are emerging in areas as var-ied as biomedical therapeutic and diagnostic research [1–4], the study of fundamentalbiological processes/interactions [1], in vivo and in vitro biosensors for clinical, food,and biodefense applications [1, 2, 5–7], bioelectronics [5, 6], nanodelivery systemsused in the food industry [8], and novel functional bioassembled architectures/macro-structures [9, 10]. Essential to reliably predicting the function of these novel hybridnanomaterials is intimate knowledge, and hence extensive characterization, of both theNP and the biomolecular layer [12–15]. A schematic highlighting some of the compo-nents that make up a typical NP-bioconjugate is shown in Figure 15.1, and descriptionsare provided in Table 15.1.

The exact nature of the NP-bioconjugate is highly dependent on the particular sys-tem under investigation. For example, the biomolecule can be inside the particle, ratherthan outside, and in some instances the biomolecule is larger than the NP. Generally,the NP-bioconjugate will be comprised of: (1) the nanoparticle scaffold, with or withoutan additional shell layer, which may have either an active or passive role in the desiredapplication, (2) various ligands added to make the nanoparticle soluble in an aqueousenvironment, biocompatible (especially polyethylene glycol -PEG species), and/or reac-tive to aid in bioconjugation (such as -NH2, -COOH, or -SH), and (3) the biological mole-cule, such as antibodies, peptides, DNA, and carbohydrates, used to sense/target/treatcan either bind directly to the NP surface, via an intermediate linker, or be sequestered

Techniques for the Characterization of Nanoparticle-Bioconjugates

294

(a) (b)

(e) (f)(g) (h)

(c)

(d)

Figure 15.1 Schematic of the various potential NP-bioconjugate components and configurations(not to scale). (a) Biomolecule interacting with NP core. (b) Biomolecule interacting with NP core viaintermediate ligands. (c) Biomolecule interacting with NP shell layer that surrounds the NP core. (d)Biomolecule interacting with NP shell layer—NP core via intermediate ligands. (e) Porous NP core con-taining entrapped biomolecules. (f) Porous or hollow NP core containing entrapped biomolecules sur-rounded by a NP shell layer. (g) NP core (or NP core/NP shell structures) particles smaller in size thanthe much larger biomolecule. (h) NP core (or NP core/NP shell structures) particles smaller in size thanthe much larger biomolecule attached via intermediate ligands.

inside the core of the NP. The specifics of NP modification and bioconjugation havebeen the subject of a number of excellent reviews and book chapters [7, 16–18].

There are a variety of NPs and NP-bioconjugate physicochemical metrics that areimportant to address. These include NP size and size distribution, shape, topology,molecular weight, aggregation state, purity, chemical composition, surface characteris-tics, functionality, Zeta potential (overall charge), stability, and solubility [11, 19, 20].Bioconjugation of NPs typically occurs via stochastic synthesis, resulting in a distribu-tion of NPs functionalized with different populations of biomolecules. This can be ofparticular concern when single biomolecule labeling of the NP is desired, as is the casefor many bioassembly-based applications [1]. Bioconjugation to the NP surfaces there-fore pose additional questions and metrics that need to be addressed, including: (1) con-firmation of biomolecule attachment, (2) average ratio of NP-to-biomolecule and ratiodistribution, (3) hydrodynamic radius, (4) structure and orientation of the biomoleculeupon attachment, and (5) stability of bioconjugation to NP environment for theintended application. Structure, orientation, and stability of the biomolecule are of par-ticular interest as these govern how well the NP-bioconjugate functions in its intendedapplication. Correct orientational control of the biomolecule, such as antibodies, forexample, will prevent blockage of the active site and prevent mixed avidity that canoccur if random orientations are present [21].

There are a variety of well-developed tools that have successfully been applied tocharacterization of the NP themselves and NP-bioconjugates with the exact choicesomewhat dependent on the physical properties of the species under investigation [8,19, 22, 23]. The Nanotechnology Characterization Laboratory (NCL) [24], in particular,has developed a variety of standardized analytical tests, termed the assay cascade, used tocharacterize not only the physicochemical characteristics, but also the in vitro and invivo properties of NP materials used in cancer research [25].

While the ultimate test of successful NP-bioconjugation is, of course, functionalityin the desired application, where activity infers the presence and activity of thebiomolecule on the NP surface, this may not provide specific details of theNP-bioconjugate architecture. The aim of this chapter is to provide the reader an

15.1 Introduction

295

Table 15.1 Nanoparticle Components

NP Core NP Shell Surface Ligands Targeting Biomolecules

Can be a solid, porous orhollow environment.

A shell layer surrounding asolid or hollow core.

Typically bifunctional, interactingwith both the NP core/shell surfaceand its surrounding environment.The terminal moiety can be stabi-lizing, provide aqueous solubilityand/or reactive, allowing subse-quent bioconjugation.

Interacts either directly withthe NP core/shell or reactswith surface ligands. Isresponsible for the uniquespecificity of theNP-bioconjugate.

Examples Examples Functional Groups ExamplesSolid: metallic NPs, semicon-ductor NPs, QDsPorous: polymer, dendrimersHollow: Carbon NPs, goldnanoshell NPs, viral NPs*,liposomes

Gold nanoshell NPs, Silicananoshell NPs (hollow orsolid—magnetic or Au corescommon), Semiconductor QDcore/shell NPs, Carbon NPs,viral NPs, mixed metalliccore/shell NPs (Ag/Au struc-tures), liposomes

Carboxylic acids (-COOH), amines(-NH2), thiols (SH), -PEG, hydroxyls(-OH)Reactive Chemistries: “click chem-istry,” affinity-based (biotin-avidin,nickel NTA-poly-histidene,succinimidyl esters, maleimides

Antibodies, peptides,proteins, carbohydrates,aptamers, nucleic acids(DNA/RNA), enzymes, simplemolecules (biotin, small tox-ins, drugs), biomimics, recep-tors, cofactors, substrates

*Viral NP typically refers to the coat protein cage that surrounds and protects the viral genes of a number of different viruses.

informed review of the characterization methods available, expressly focused onNP-bioconjugates, along with some pertinent examples which highlight both thepotential advantages and/or drawbacks of each particular technique. The techniquesdiscussed have been grouped under six main categories based upon the intrinsic typeof analysis performed: separation-based, scattering, microscopy, spectroscopic, massspectroscopy, and thermal.

15.2 Methods

15.2.1 Separation-Based Techniques

Separation-based techniques such as chromatography, electrophoresis, and centrifu-gation are routinely used to purify NP-bioconjugates. However, in many cases they canalso provide approximate hydrodynamic radius, purity of product, NP-to- biomoleculeratio, and stability (e.g., postproduction degradation).

Chromatography is a separation technique that relies upon differing affinities of themultiple sample components for the chosen chromatographic mobile and station-ary/solid phase. There are many types of chromatographic techniques and likewisenumerous detectors available for measuring the eluting fractions including: UV-visibleabsorbance, light scattering, fluorescence, refractive index measurements, and massspectroscopy [26]. Column-based liquid chromatography techniques, in particularhigh performance liquid chromatography (HPLC), have been used extensively for NP-bioconjugate separations [26, 27]. HPLC is often preferred over classical (gravity or lowpressure) chromatography due to improved peak resolving power [26, 27]. The ability ofsize exclusion-based HPLC to explore the size and shape polydispersity of various quan-tum dot (QD) materials was recently demonstrated [28]. Of the many varieties of chro-matography columns available, reverse-phase [29, 30], ion-exchange [31], and sizeexclusion (SEC) [13, 26, 27, 32, 33] are the most common for NP-bioconjugate studies.In most cases chromatography techniques are capable of purifying NP-bioconjugatesboth from unmodified NP and free biomolecules, as demonstrated for amine-modifiedgold NP-cytochrome c conjugates [29] and polymer-coated QD-antibody complexes[27]. Care should be taken to limit nonspecific interactions with the solid phase matrixwhich can be problematic. In some instances optimized HPLC has demonstrated theexquisite ability to resolve NP-bioconjugates with different NP-to-biomolecule ratios,providing both the distribution and overall average ratio of NP-to-biomolecule per sam-ple [30, 31]. Reverse-phase HPLC has been used to determine the distribution of ligandsper dendrimer for (3-(4-(prop-2-ynyloxy) phenyl) propanoic acid) conjugated to the pri-mary surface amines of dendrimer NPs [30]. Anion exchange HPLC was used to investi-gate DNA-gold NP conjugates and demonstrated the superior resolving power of HPLCover gel electrophoresis for separating 5-nm gold NPs labeled with 1, 2, or 3 PolyT DNA(see Figure 15.2(a)) [31]. In contrast, agarose gel electrophoresis achieved higher resolv-ing capabilities compared to SEC-HPLC for PEG functionalized QDs [13], highlightingthe need to tailor techniques for each particular NP-bioconjugate system under investi-gation. Chromatography is likewise a powerful tool for investigating NP-bioconjugatestability postproduction, as demonstrated for nanohydrogel materials used for drugdelivery [34].


296

Hydrodynamic chromatography (HDC) uses a nonporous stationary phase and apressure-driven mobile phase to fractionate mixtures in a channel [35]. Larger particlesreside in the faster-moving central region of the parabolic flow profile, while smallerspecies readily diffuse to the slower-moving regions near the channel walls, result-ing in efficient separation. This technique has demonstrated characterization of lipidnanocapsules [35] and fluorescently labeled polystyrene NPs [36], but has yet to beapplied to NP-bioconjugates, where its resolving capabilities maybe limited to separat-ing unconjugated biomolecules from NP-bioconjugate.

15.2 Methods

297

0 .1 .2 .5 1 2 4

Valence# of MBP/QD

+

->2210

5

+

->2210

(a)

(b) (c)

cadded(DNA) /c(Au) »

010005002501306332167.93.92.00.990.490.250.12

- +satwell 7 6 5 4 3 2 1 0

cadded»

10005002501306332167.93.92.00.990.490.250.12

- +satwell 7 6 5 4 3 2 1 0 N-maltose binding protein His -COOH

(iii)(i)

(ii)

0

0

1

2

3

A520

5 10 15 20Time (min)

Figure 15.2 Separation techniques. (a) Comparison of (i) agarose gel electrophoresis and (iii) anionexchange high-performance liquid chromatography (AE-HPLC) purification of polyT DNA conjugatedto 5-nm gold NPs. (i) Agarose gel electrophoretic separation of gold NP-DNA bioconjugatesfunctionalized with 0, 1, 2, and 3 DHA strands. (ii) Optical density analysis of the agarose gel electro-phoresis bands in (i) demonstrating band overlap and limited resolution. (iii) AE-HPLC purification ofthe same gold-NP bioconjugates illustrating the superior resolving power, especially at the peak base,of the technique. Images kindly provided by Dr. Claridge (Berkeley). Reprinted with permission from[31], Copyright 2008 American Chemical Society. (b) Agarose gel separation of different DNA-conju-gated gold NPs at various modification ratios. Note, at the lower DNA-to-gold NP ratios multiple dis-tinct narrow bands are observed in the gel representing modification ratios of 1, 2, 3, and so forth.However, at higher ratios of DNA-to-gold NP, broader bands, which move increasingly slower in thegel, were observed, reflecting the increase in DNA loading and concurrently larger overall hydrody-namic size of the DNA-gold NP. Images kindly supplied by Dr. Parak (Lugwig-Maximilians-Universität).Reprinted with permission from [49], Copyright 2003 American Chemical Society. (c) Agarose gelcharacterization of maltose binding protein (MBP)-QD bioconjugates. The gel image clearly shows theseparation of QD conjugates with different numbers of MBP protein-per-QD. Due to the Poisson distri-bution, smaller ratios demonstrate several mobility bands which merge into a single band as the ratioincreases, suggesting a more homogeneous product. Images kindly provided by Dr. Mattoussi (U.S.Naval Research Laboratory). Reprinted with permission from [51], Copyright 2006 American ChemicalSociety.

Field flow fractionation (FFF) encompasses a continually evolving family of analyticalseparation techniques. The sample is introduced into a pressure-driven mobile phasecontained within an open channel (no stationary phase), comprising a parabolic flowprofile, and a field is applied perpendicular to the direction of flow [37, 38]. Through acombination of complex effects, the sample components separate into different laminaeregions of the parabolic flow above an accumulation wall and hence separate due todiffering transport velocities. The main applied fields are crossflow (flow, includingasymmetric flow), centrifugal (sedimentation), electrical fields (electrical) and ther-mal/temperature gradients (thermal), although magnetic and dielectrophoric fieldshave also been used. Detection of eluted fractions is typically achieved by couplingthe FFF channel outlet to a UV-visible detector or multiangle light scattering (MALS)measurements.

Both sedimentary and flow FFF, which separate based on effective mass and diffu-sion coefficients, respectively, have predictable retention times that depend on variousphysical parameters of the particle constituents, including effective mass, hydrody-namic diameter, density, and/or volume [37]. The resulting fractograms can provideboth size (peak height) and size distribution (peak width) information. Precalibrationwith known “size” NP standards is often desired; however, in the case of flow, FFF is notalways necessary if all geometric dimensions of the fractionation channel are accuratelyknown [39]. Thermal FFF has also demonstrated the ability to separate according to bothsize and surface potential, as demonstrated using silica NPs [40].

To date, the FFF family has found limited use for NP-bioconjugate analysis, withapplications mainly concerning polymer NPs modified with targeting peptides [41], bio-degradable polymer NPs for drug delivery [39, 42–44], and QD-DNA conjugates [45].Magnetic FFF has been used to characterize dextran-coated magnetic NPs [46]. As withany separation technique, nonspecific binding can occur at the FFF accumulation walland hence optimization is required to obtain the desired separation [39]. Optimizationcan include varying the buffer type and ionic strength as well as the choice of membraneused as the accumulation wall, including the membrane molecular weight cutoff(MWCO) and material [47]. FFF has not yet demonstrated the ability to resolve NPsfunctionalized with varying numbers of biomolecules, but this may in part be due to itscurrent limited application in this field, as opposed to a fundamental lack of ability.

Slab gel- and capillary-electrophoresis are the two main types of electrophoretic tech-niques successfully applied to the characterization of NPs. Slab gel electrophoresis mea-sures the electrophoretic mobility of charged species in a gel matrix when an electricfield is applied. For NPs both the overall size and charge density will influence the direc-tion and distance moved in the gel. In many, but not all, cases bioconjugation has lim-ited influence on the overall surface charge and therefore the electrophoretic mobility isdominated by the hydrodynamic size. Gel electrophoresis, when combined with appro-priate controls, is a powerful tool for demonstrating biomolecule attachment to the NPscaffold through sensitive changes in mobility [13, 48, 49–55]. As the NP-bioconjugatesbecome larger in size they tend to migrate at slower rates in the gel matrix, as illustratedin Figure 15.2 [31, 49, 51]. Gel electrophoresis has demonstrated exquisite resolutionunder optimal conditions and is able to separate NPs labeled with 1, 2, 3, and so forthbiomolecules (see Figure 15.2) [31, 49, 51, 53, 54].

On the small scale the technique is routinely used to separate and purify NP-bioconjugates, and extracted particles can be further characterized using techniques


298

such as AFM and mass spectrometry [13]. Gel electrophoresis has not only demonstratedthe ability to separate NPs based on size and shape [56], but also revealed unanticipatedNP-biomolecule nonspecific binding, which blocked the biomolecule active site [57].Agarose and polyacrylamide gel electrophoresis (PAGE: SDS and native) represent themain gel matrices used for NP-bioconjugates’ characterization. Colored NPs (gold, car-bon nanotubes, silver) can be visualized by eye, while fluorescent NPs (quantum dots orfluorescently labeled) are detected using an appropriate excitation source and detector(e.g., CCD camera). Proteins and DNA present on the NP surface can be detected viastaining, using Coomassie Blue for proteins and ethidium bromide (or SYBR dyes) forDNA, which is typically performed after NP measurement to demonstrate comobilitywith the NP.

The use of gel electrophoresis for determining absolute hydrodynamic diameters wasinvestigated in a comprehensive study by Parak and coworkers characterizing gold-DNAconjugates [49, 54]. While extremely sensitive to the extent of NP-bioconjugation, gelelectrophoresis suffered several limitations with respect to absolute effective diametersderived using either a calibration curve or Ferguson plots [49–51, 54, 58]. Both methodsuse gold NPs of increasing “known” size to calibrate the mobility-diameter relationshipresulting in calibrations based on rigid size increases and not the flexible “soft” increasesmore likely to occur from DNA attachment to a NP surface [49, 54]. Although not sim-ple to design or prepare, more appropriate calibration materials may eliminate someinherent limitations.

Capillary electrophoresis (CE) measures the electrophoretic mobility of chargedspecies in an open capillary (no solid matrix) filled with a liquid electrolyte, when anelectric field is applied. Through a combination of electrophoresis of the sample compo-nents and the electro-osmotic flow (EOF) of the electrolyte buffer, the sample compo-nents are transported from the positive anode to the cathode with separation based onthe species size-to-charge ratio [59]. EOF of the electrolyte buffer is observed when thecapillary wall is charged and NP studies to date use fused-silica capillaries rendered nega-tively charged through ionization via exposure to a basic solution. Under these condi-tions, positively charged, species generally elute the fastest, and in the case of NPs,smaller diameters elute first, as demonstrated for gold and gold/silver NPs [60–62]. If sep-aration based on pure electrophoresis of the sample components is desired, the capillarywalls can be coated with a neutral polymer which suppresses EOF and likewise any inter-action of the sample components with the interior wall [110]. UV-visible absorbanceand, where appropriate, fluorescence (specifically, laser induced fluorescence-LIF) repre-sent common methods for species detection in CE, with mass spectroscopy occasionallyused. Extensive optimization of the electrolyte components, including surfactants andpH, is required for effective CE [62, 64]. Variations on the traditional CE theme exist,including capillary gel electrophoresis, where the capillary is filled with gel matrix,and micellar and microemulsion electrokinetic chromatographies, which aid in theseparation of neutral species [65].

CE has been applied to the study of a number of NP-bioconjugates, includingQD-BSA [64], silicon NP-streptavidin [66], and iron oxide NP-protein/antibody conju-gates [67]. CE has also been used to study the drug loading abilities of poly(lactic acid)NPs [65], plasma protein absorption to PEGylated polymer NPs [68], and IgM interac-tions with QD-anti-IgM bioconjugates [69]. However, to date, CE has mainly been usedin a qualitative sense to demonstrate bioconjugation through the differing mobilities of

15.2 Methods

299

the NP-bioconjugates compared to the free NPs, and its quantitative capabilities remainunproven but of high potential as the technology evolves.

Analytical ultracentrifugation (AUC) is a separation method that is often used to deter-mine sample purity and average molecular weight in liquid-based dispersions withoutthe need for special solvents, such as those found commonly in sucrose/glycerol gradi-ent centrifugation. AUC consists of a high-speed centrifuge rotor with cell compart-ments and an optical system (usually UV) used to measure concentration gradients ofthe sample when centrifugal force is applied [70]. Two main modes are used in AUC—sedimentation velocity and sedimentation equilibrium—and they can be used sequen-tially to provide information about the individual NPs and NP interactions. Size, size dis-tribution, and shape of NPs can be calculated with AUC, with no assumption about thedimensions of the particle needed, as opposed to light scattering techniques (see Section15.2.2). In addition to the basic structure of a NP material, AUC can provide both struc-tural and conformational information about conjugated biomolecules. AUC theory andbasic techniques have been reviewed in the literature [71, 72]. The main formulism forAUC analysis involves determining the sedimentation coefficient, s, which containsinformation about the particle’s physical properties, described by

( )[ ]M Nf u r s1 2− = =νρ ω (15.1)

where M is the molar weight of the solute (in g/mol), N is Avogadro’s number, ν is thepartial specific volume of the particle (in mL), ρ is the density of the solvent (g/mL), ω isthe angular velocity (in radians per second), r is the distance of the particle from the axisof rotation, f is the frictional coefficient, and u is the particle velocity.

AUC can elucidate the binding of small molecules to NPs, NP self-association oraggregation (as demonstrated for human serum albumin NPs [73]), and interactionsbetween heterogeneous NPs as each noninteractive species is separated into aunique boundary [73–76]. In addition, the size of the individual portions of theNP-bioconjugate and the overall size of the complex may be determined. Goodagreement between sizes determined from the AUC sedimentation coefficients andthose observed in transmission electron microscopy (TEM) (see Section 15.2.3) for un-conjugated gold NPs have been found and the average stoichiometry upon proteinligand (lactose repressor) bioconjugation to the gold NP determined without the need tofirst remove the unconjugated protein [74]. Stoichiometry of the NP-to-ligand may alsobe determined with this method as demonstrated for bovine serum albumin (BSA) modi-fied QDs [75]. Researchers have also used AUC to size CdSe/ZnS core/shell QDs as well asstudy their bioconjugation with dihydrolipoic acid (DHLA) and poly(ethyelene glycol)[75] and to study protein interactions with silica NP cores [76]. Benefits to the AUC

method include small sample sizes (20 μL) and a wide range of usable concentrations. Inaddition, AUC is nondestructive and the sample may be recovered for subsequent analy-sis, thus allowing for detailed testing of post-formulated products.

15.2.2 Scattering Techniques

Scattering techniques, as the name implies, measure the scattering of radiation (e.g.,light or particles) through its interaction with a sample, and information about theNP structure, morphology, hydrodynamic size, and aggregation state, as well as thebiomolecule conformation and the NP-bioconjugate stability, can be obtained.


300

Dynamic light scattering (DLS), also known as photon correlation spectroscopy (PCS) orquasi-elastic light scattering (QRLS), is a nondestructive technique used to size particles insolution from the nanometer-to-micron size scale [77–81]. When a sample area is sub-jected to incident light, the total light that reaches the detector located a specific dis-tance and angle away from the sample is the sum of the scattered waves from all of theilluminated particles. Small particles in solution undergo Brownian motion or thermalfluctuations in which they continuously vibrate, move, rotate, and collide with oneanother. This motion causes the distances between the scattering particles to change,resulting in constructive and destructive interference of the scattered light over time andintensity fluctuations in the detected signal. The time dependence of the fluctuations,

and notably the decay rate, Γ, can be fitted to give a diffusion coefficient, D, for the parti-cles, as described by

( ) ( )Γ = =Dq q n20 04 2where π λ θsin (15.2)

where n0 is the index of refraction of the solvent, λ0 = wavelength of incident light, and θ

= angle of measurement. Using the Stokes-Einstein relationship, the hydrodynamicradius of the molecules, Rh, can be calculated from the diffusion coefficient D using

R kT Dh o= 6πη (15.3)

where T is the temperature in Kelvin, k is the Boltzmann constant, and ηo is the viscosityof the solvent. A monodisperse sample gives rise to a single decay rate, which is ratherstraightforward to analyze for particle size. However, polydisperse samples give rise to aseries of exponential decays, which are analyzed for size distributions by fitting toassumed distribution functions, which may or may not represent the actual particle dis-tribution [77].

DLS measurements are very common in characterizing NP solutions and NP-bioconjugates [51, 82–84]. For instance, Figure 15.3(a, b) presents DLS data of luciferase(Luc8)-conjugated quantum dots (QDs), along with corresponding TEMs, to comparethe size of the conjugated particles before and after modification [82]. Pons and cowork-ers used DLS to investigate the hydrodynamic dimension variations of QDs capped withvarious surface ligands, including various PEG-polymer ligands, and maltose bindingprotein (MBP) [51].

One current drawback of DLS is that the hydrodynamic radius reported assumes aspherical particle; therefore, hydrodynamic radii reported for nonspherical shapes maynot reflect the true size of the particles. There are, however, models being developed inthe DLS literature for nonspherical shapes, such as rods [85]. DLS also has problems dis-tinguishing between two species close in hydrodynamic radius, and given that the scat-tering intensity is proportional to the sixth power of the particle radius, care must betaken when interpreting data from samples containing a wide range of size distributions,since the scattering signal will be heavily weighted to small numbers of larger particles[86]. Sample preparation is extremely important and care must be taken to remove largecontaminating particles, such as dust particles, which are highly scattering. Filtrationprior to analysis, or the use of prefiltered solvents, is commonly employed to reduce thisissue. The benefits of DLS are numerous and include rapid sample analysis, taking a fewminutes at most, inexpensive technology compared to other characterization tech-niques, and many benchtop models are commercially available. Sample preparation is

15.2 Methods

301

relatively simple and measurements can be made in any media or solvent of interest.DLS is very sensitive, capable of measuring very dilute solutions (~0.01% w/v), and thetechnology is improving to allow measurement of concentrated samples, therebyreducing the requirement for dilution. This technique is also particularly useful formonitoring stability of particles or formulations postproduction providing valuableinformation regarding shelf life.

Fluorescence correlation spectroscopy (FCS) is similar to DLS in that it measuresfluctuations due to diffusion, aggregation, and interactions, but rather than scattering,FCS measures fluorescence. The technique commonly uses optical microscope instru-mentation, in particular confocal microscopy, to excite (using single- or multiphotonexcitation) and measure fluorescence within a confined optical volume. Data is essen-tially derived from monitoring sample transit times in a known confined excitation vol-ume. Note that confocal microscopy can be used in the backscatter mode to studynonfluorescent particles, as demonstrated for gold and latex particles [87]. While FCS isbest used for rapidly diffusing fluorescence molecules (such as organic dyes), it has been


302

Figure 15.3 Dynamic light scattering (DLS) and Zeta potential. (a, b) The DLS and transmission elec-tron microscope (TEM) characterization (inserts) of QD-luc8 NP-bioconjugates, pre- and post- poly-meric encapsulation, respectively. Polymeric encapsulation cross-linked two to three QD-luc8 NPsresulting in an overall increase in the diameter, as observed in the DLS analysis. (Images kindly pro-vided by Dr. Xing and Dr. Rao (Stanford University). Reprinted from Biochemical and BiophysicalResearch Communications, [82], Copyright 2008, with permission from Elsevier.) (c) Zeta potentialcharacterization. Schematic of a charged particle and its associated potentials, including zeta potential.(Reprinted with permission granted by Malvern Instruments Ltd. UK [97].) (d) Simultaneous measure-ment of both the Zeta potential (squares) and the dynamic light scattering (DLS) determined particlesize (triangles) of streptavidin modified silica NPs during a pH titration in water. The streptavidin mod-ified silica NPs are found to be unstable above pH 7.0 where clearly the mean particle size rapidlyincreases (suggesting agglomerate formation) and the Zeta potential drops below 20 mV. (Imagesreprinted with permission from the Hindawi Publishing Corporation (DOI#10.1155/2008/712514)[98].)

successfully used to accurately size the hydrodynamic radius of slower diffusingNPs such as QDs and fluorescent beads [13, 87–91] and characterize the photophysicalproperties of QDs [92, 93]. FCS has also been used to determine binding kinetics between100-nm unilamellar vesicles and fluorescently labeled peptides [94]. Fluorescencelifetime correlation spectroscopy has been used to monitor the metal enhancedfluorescence (MEF) resulting from Cy5-labeled DNA hybridizing to DNA-modified silverNPs [95].

Electrokinetic potential, or zeta (ζ) potential, characterizes the surface charge of aparticle, which can influence its stability, dispersability, and agglomeration [96]. Acharged particle in solution will have a layer of opposite charges around its surface calledthe double diffuse layer, consisting of an inner core layer of tightly bound charges (theStern layer), and a more diffuse outer layer of charges, within which is a boundaryreferred to as the slipping plane (see Figure 15.3(c)) [97]. Within the slipping plane, theparticle and those associated diffuse ions can be considered to move as a single entity.The potential difference between this point and the bulk solvent is the Zeta potential.

Zeta potential is determined by applying an electric field across a sample and measur-ing the velocity at which charged species move towards the electrode. The velocity,which is proportional to the zeta potential, is measured as a phase or frequency shift inthe incident light using the technique of laser Doppler velocimetry (LDV) [51]. Theresulting electrophoretic mobility, μE, is calculated using (15.4), where v is the velocityand E is the applied field [96, 97]:

μE v E= (15.4)

From the electrophoretic mobility, the Zeta potential (ζ) is determined using theHenry equation [96]:

( )ζ ημ ε= 3 2E f Ka (15.5)

where η is the viscosity, ε is the dielectric constant and f(Ka) is the Henry function, forwhich one of two common approximations are generally assumed: (1) 1.5 for aqueousmedia containing particles larger than 200 nm dispersed in an electrolyte media (theSmoluchowski approximation), or (2) 1.0 for particles in a low dielectric constant media(the Huckel approximation) [96, 97].

Zeta potential can be used as a measure of particle stability, with a value of +/− 25 mVoften selected as an arbitrary delineation of stability. Absolute values larger than 25 mVindicate stability and represent highly charged particles that repel one another, whilevalues <25 mV indicate particle instability with a propensity to coagulate, flocculate,and/or agglomerate. Factors which affect Zeta potential (and hence stability) includepH, ionic strength of the solution, and concentration in solution. Zeta poten-tial measurements have been used to study the stability and particle size ofstreptavidin-functionalized silica NPs as a function of pH (see Figure 15.3(d)) [98], andcharacterize the surface coverage of cytochrome c bioconjugated to gold NPs [99] andvarious QD capping agents [51]. Zeta potential has become fairly easy to measure andhas many of the same advantages as DLS; in fact, oftentimes the two techniques areavailable on the same instrument.

Raman spectroscopy measures the inelastic scattering of monochromatic radiation(UV, visible, or near IR) by a sample, where the incident light becomes either Stokes or

15.2 Methods

303

anti-Stokes shifted in wavelength, and the resulting signals are associated with vibra-tional states within the material. Relative to the elastic Rayleigh scattering, the Ramansignal intensity is much weaker. The resulting sharp fingerprint Raman bands, however,provide unique information about the material, complementary to infrared (IR) spec-troscopy (see Section 15.2.4). Carbon nanotubes, for example, exhibit strong Ramanscattering found to be sensitive to isotope composition (see Figure 15.4(a)) [100] andbiomolecular interactions [101]. Various methods exist to enhance Raman scatteringsignals, including resonance Raman (RR), where the incident monochromatic lightwavelength coincides with an absorption band of the material under investigation[102], surface-enhanced Raman scattering (SERS), and surface-enhanced resonanceRaman scattering (SERRS). SERS typically occurs when materials adsorbed onto struc-tured (roughened) metallic surfaces experience enhanced local electromagnetic fieldsthat result when the incident light matches the surface plasmon band of the metallicsurface. Metallic NPs, in particular silver NPs, offer the unique possibility of localizedSERS and have been used to study hemoproteins (heme-containing proteins such ashemoglobin, myoglobin, and cytochrome c) bioconjugation to gold and/or silver NPs[103, 104]. To date, however, Raman characterization of NP-bioconjugates is limited.Raman NP imaging tags are increasingly being proposed for biomedical imaging applica-tions using Raman microscopy [105]. Carbon nanotubes have demonstrated potential[100] (see Figure 15.4(a)), but more common tags are comprised of small moleculecompounds with unique Raman signatures conjugated or encapsulated to gold or silver


304

BA

i

ii

(i)

(ii)

(iii)

(iv)

(v)

Figure 15.4 Scattering techniques used in NP-bioconjugate characterization. (a) (i) Schematic ofthree single walled carbon nanotubes (SWNTs) Raman tags, comprised of different isotopes of carbon13C, 12C/13C, and 12C and each modified with different biomolecules. (ii) Raman spectra, taken using785-nm laser excitation, of the three SWNTs, comprised of the different isotope compositions, demon-strating the three different G-band peak spectral positions observed. (Reprinted with permission from[100], Copyright 2008 American Chemical Society.) (b) X-ray diffraction (XRD) spectra of biologicalmolecules intercalated between the sheets of a layered NP host. Powder XRD patterns and resultingschematic representation of layered double hydroxides (LDH) separation for (i) as preparedMg2Al-NO3-LDH, (ii) DNA modified-LDH, (iii) adenosine triphosphate (ATP)-LDH, (iv)fluorescein-5-isothiocyanate (FITC)-LDH, and (v) c-myc antisense oligonucleotide (As-myc)-LDH.(Images reprinted from Angewandte Chemie International Edition [118]. 2000. Copyright Wiley-VCHVerlag GmbH & Co. KGaA. Reproduced with permission.)

NPs and imaged through the SERS mechanism [105–107]. Such increased interest in thetechnology may result in more widespread use as a tool for characterization.

X-ray diffraction (XRD) is a nondestructive technique that can provide structuralinformation about a crystalline sample. In wide angle X-ray scattering, X-rays are scat-tered from microcrystals in a sample, providing information about the distance betweenplanes of atoms in the crystal. This data can provide information about the overall mate-rial, including polymorph information and quantification. For samples where individ-ual crystallite diameters fall below ~ 200 nm, additional information about the size andshape of the crystallite can also be estimated. XRD methodology is frequently used inthe analysis of materials containing nano-sized components embedded in an ex-tended matrix, such as those found in tissue scaffolds and bone cements [108, 109] ornanobioconjugate layered materials such as nanobiohybrids [110, 111], where biologi-cal species are intercalated between sheets of a nanomaterial. By taking advantage ofBragg’s law, (15.6), structural and compositional information about the material may beobtained.

n dλ θ= 2 sin (15.6)

where n is the integer of the order of reflection, λ is the wavelength of the incident X-raybeam, d is the distance between atomic layers in a crystal, and θ is the angle of incidence.Crystal forms of both the NP and the biological entity may be determined through com-parison of known polymorph patterns and the resulting NP-bioconjugate may be evalu-ated, as demonstrated by the comparison of sodium montmorillonite and chitosanbefore and after intercalation [112]. In addition, the amount of each crystal phase pres-ent in a formulated product containing NP-bioconjugates may be determined throughanalysis of the diffraction patterns and applying the relevant phase calculations [113].

The Scherrer equation (15.7) uses peak broadening, measured at the full width at halfmaximum (FWHM) of the characteristic crystal reflection peaks to obtain approximatecrystallite diameters [114, 115]. Before sample peak broadening can be analyzed,however, instrumental broadening must first be accounted for, using FWHM2

observed =FWHM2

instrument + FWHM2

size+strain (with FWHM in radians).

D K FWHMsize~ cosλ θ (15.7)

where K represents a unit cell geometry dependent-constant (typically between0.85-0.99), D is the particle diameter (volume weighted), λ is the wavelength of the inci-dent X-ray beam, θ is the angle of incidence, and FWHMsize is the corrected peak width.This equation assumes that no strain is present in the system. Strain in the crystalline lat-tice can cause shifts in the lattice parameter d (see (15.6), Bragg’s law), resulting in varia-tions in the peak position of the crystalline reflections, as well as peak broadening andcalculation errors as high as 20%. If NP and biomolecule peaks are sufficiently separatedthen by analyzing peak shape, height, and relative intensities, the Scherrer equation canprovide a rough estimate of different crystallite/particle sizes of both the NP andbiomolecule [116].

NP-bioconjugate structures or nanobiohybrid architectures often make use of lay-ered materials in order to deliver, protect, or stabilize a biomolecule [117]. By analyzingthe d-spacing changes during the intercalation of a biomolecule into the NP host, thecompletion of the reaction maybe assessed. For example, the d-spacing changes of the

15.2 Methods

305

nitrate form of a magnesium aluminum layered double hydroxide nanomaterial wasmonitored as it was exposed to different biomolecules such as DNA, oligonuclueotides,or tracer molecules (see Figure 15.4(b)) [118]. In addition, by analyzing the d-spacingchanges and the structure of the biomolecule, the orientation of the biomoleculebetween nanomaterial layers was estimated. XRD has also been used to assess thepolymorph stability of solid lipid NPs containing vitamin A, and the aggregation effectsof these NP-bioconjugates were then monitored over time by correlating polymorphtransitions to observed aggregation [119]. XRD may provide an estimate of the size andgrowth of crystallites in the sample, as demonstrated by for TiO2 NPs formed by a tem-plate synthesis [120] and hydroxyapatite crystals formed on fibrin protein-modifiedgold NPs [121].

Small angle X-ray scattering (SAXS) uses X-rays, on the order of 0.1–0.2 nm in wave-length, to characterize macromolecular (polymer and/or biologic) size, shape, and dis-tances [122, 123]. Typically “small angle” scattering techniques refer to those structureswhose scattering angles are less than 10°. X-rays are scattered by electrons; therefore,electron density inhomogeneities on the size scale of 0.5- to 150-nm scatter X-rays atsmall angles [124] and provide information about size, shape, structure, periodicities,and orientations of structures in the sample (while electron density fluctuations on theatomic and crystallographic scale give rise to wide angle scattering). A SAXS instrumentusually consists of a monochromatic X-ray source (typically a CuKα line of wavelength1.54Å), a sample chamber or holder, and a detector. The X-ray source can either be a lab-oratory source (e.g., rotating anode tube) or a synchrotron source, which has higher fluxand is better focused. When the X-ray beam hits the sample, a portion of the X-raysinteract with the sample and are scattered, while the remainder passes through thesample. The resulting two-dimensional (2-D) scattered intensities are typically recordedon a plate detector located behind the sample (see Figure 15.5(a)).

One difficulty of SAXS measurements is distinguishing the weak scattering intensi-ties from the very strong unscattered beam. Analysis becomes increasingly more difficultthe smaller the angle of interest, a fact complicated by X-ray sources that producedivergent beams. Synchrotron X-ray sources address this problem by focusing the beamusing mirrors, while laboratory X-ray sources use either pinhole collimation or linecollimation geometries to confine the beam. SAXS data alone cannot be used to deter-mine the structure or morphology of the samples. Rather, the results are plotted as scat-tered intensity versus scattering vector, and models of the structures, either known apriori or constructed, are applied to the data. Figure 15.5(b) shows the SAXS pattern ofDNA-modified gold NPs [125]. The circularly integrated pattern shows peaks in inten-sity, whose ratios to the first peak are determined and subsequently matched to abody-centered cubic (bcc) structure.

SAXS has been used for analyzing NP-bioconjugate systems to elucidate structureand morphology and measure characteristic spacings. Examples include elucidating thestructure of block copolymers conjugated to oligonucleotides [126] and QD NPs con-jugated with proteins [127], investigating the packing structure of DNA-modified goldNPs [168], and extended DNA-gold NP structures [128], studying the aggregationof DNA-modified gold NPs [129] and observing morphological temperature inducedtransitions in DNA-surfactant complexes [130]. Limitations of using SAXS for NP-bioconjugates characterization include the susceptibility of biological entities to radia-


306

tion damage and sample preparation can be difficult and is limited to sample thicknessesof about 1–2 mm to allow X-ray penetration.

Small angle neutron scattering (SANS) monitors the interaction of neutrons with mate-rials and can be used to probe soft materials (polymers and biologics) providing informa-tion about size, shape, and orientation of structures on the nanometer up to the micronscale [131–134]. Unlike X-rays and other electromagnetic wavelengths that are scatteredby electrons, neutrons are scattered only by the atomic nucleus. The nuclear scatteringcross section varies irregularly with Z, and even some isotopes of the same element havevery different scattering cross sections. The largest difference between isotopes occursfor hydrogen (H or 1H) and deuterium (D or 2H), making SANS a powerful technique forstudying polymers and biological materials [135, 136]. Exchanging D for H has minimaleffect on the structural properties of the polymeric or biologic system being probed, butmakes them differentiable by SANS and contrast-matching techniques can be employedto elucidate the structures of components of interest. Neutrons interact weakly withmatter and are not scattered to the same extent as electromagnetic waves, such as X-rays.Neutrons therefore penetrate deeply into samples and can be used to measure bulk prop-erties, solution properties, and/or samples within sample cells or containers. Addition-ally, neutron energy is much smaller than that of X-rays; for a neutron of comparablewavelength to a Cu-Kα X-ray source, the resulting neutron energy is five orders of mag-

15.2 Methods

307

X-raysource

SampleDetector

T-raybeam Scattered

beam

Beam stop

A

(i) (ii)

B

Figure 15.5 (a) Schematic of a typical small angle X-ray scattering (SAXS) setup. X-rays from a sourceinteract with a sample and produce a SAXS pattern that is measured at the detector. (b) (i) VariousDNA-gold NP complements are hybridized to form larger cubic assemblies; the resulting SAXS patternof one such higher-order structure is shown. (ii) Integrated SAXS data determined from the SAXSpattern in (i) is fitted to a body-centered cubic (bcc) crystal structure model, demonstrating successfulassembly of higher-order cubic structures from the DNA-gold NP building blocks. (Images kindly pro-vided by Dr. Mirkin (Northwestern University). Reprinted by permission from Macmillan PublishersLtd: Nature [125], copyright 2008.)

nitude lower. Neutrons can therefore be used to probe sensitive materials, such as poly-mers and biologics containing delicate C-C bonds that would otherwise be destroyedduring similar X-ray analysis. Several studies have used SANS in combination with DLSand static light scattering (SLS) to elucidate the structure and core morphology ofNP-bioconjugates [137–140]. To date, however, structural information about the coreNP seems to be the biggest application of SANS for these NP-bioconjugate materials. Thebiggest drawback of SANS is that the technique requires a neutron source and accordingto the “World Directory of SANS Instruments,” there are only 22 SANS instrumentsworldwide [141]. These are large, costly facilities that require travel and whose beamtimes are usually oversubscribed.

15.2.3 Microscopy

Microscopy techniques involve visualizing a sample using light, electrons, or a scanningprobe [142–145]. Electron microscopy, including transmission electron microscopy(TEM) and scanning electron microscopy (SEM), and scanning probe microscopy, suchas atomic force microscopy (AFM), readily obtain single particle resolution and are rou-tinely used for unmodified NP characterization, including NP size and shape.

Atomic force microscopy (AFM) is a high-resolution technique that makes use of acantilever scanned across a sample surface to obtain a wide range of information on thenanometer scale. As opposed to the bulk methods of most scattering and spectroscopictechniques, AFM can analyze individual NPs, and unlike many of the electronmicroscopy techniques (TEM and SEM), which are best performed with conductingor semi-conducting samples under vacuum conditions, AFM can be applied tononconductive, wet, and soft samples, allowing for many different types of materials tobe analyzed in physiological environments [144, 146, 147]. In addition, the substrate onwhich the measurement is made can be varied, ranging from mica to glass cover slidesand even biological substrates such as skin [15, 146]. The type of information that AFMcan provide depends mainly on the interaction of the cantilever with the sample surface(mode) and the type of cantilever, which in essence traces the nanoscale three-dimen-sional (3-D) outline of the sample into a 3-D image. The cantilever can be in constantcontact with the surface (contact mode), intermittent contact (tapping or AC mode), orno contact at all (noncontact mode), each providing three-dimensional surface topogra-phy. The size and shape of surface features maybe further analyzed with a range of otherparameters including phase and force measurements. Phase measurements are typicallyderived from the changes in the phase of cantilever oscillation when the cantilever is inintermittent contact with the surface. As the cantilever is scanned and encounters differ-ent regions of the surface (e.g., hard versus soft regions, sticky versus slippery regions,and so forth), its oscillation varies, which highlights differences in the properties of thesample surface. Force measurements are determined through force-distance curves thatarise from the deflection of the cantilever when it encounters the sample surface [148].The forces and resulting parameters that can be obtained include sample adhesion, elas-ticity, Young’s modulus, and molecular stretching parameters (especially useful forcharacterizing DNA and RNA stretching, protein folding-unfolding, or biomoleculeattachment to a NP surface) [15,147,148]. Beyond changing modes and measuringforces, the cantilever tip may also be modified. Specific molecular interactions (anti-gen/antibody interactions), molecular analysis, and chemical bonding information can


308

all be determined and mapped into three-dimensional surface profiles and compared tothe standard topography images [149]. As with most techniques, AFM does have somedisadvantages, including: (1) analysis is usually limited to the exterior of the NP (surfacetechnique) and also requires immobilization of the sample onto a substrate surface, suchthat it will not be moved around by the tip, (2) small areas are typically analyzed due totime, tip, sample, and program constraints, so care must be taken to provide an accuraterepresentation of the sample, (3) image acquisition times may be slow as the sample isscanned using the tip, resulting in lengthy scan times that can prevent analysis of tran-sient events, and (4) images can be difficult to interpret and extensive optimization istypically required to obtain meaningful results and avoid artifacts [147, 148]. The field isconstantly evolving to overcome some of these difficulties, as reviewed by Midgley andDurkan [144] and Liu [150].

The different conformational forms that DNA may take when conjugated to goldNPs have been studied by topography analysis [151] and AFM combined with QD label-ing was used to analyze single-stranded DNA conformation when associated with car-bon nanotubes [152]. Topography images in conjunction with cross-section height andphase analysis have provided NP-bioconjugate heights and morphological characteriza-tion for streptavidin functionalized QDs [153]. TiO2 modified DNA nanocompositesstructures [154] and gold NP-cytochrome c [99] have also been studied and hence suc-cessful bioconjugation confirmed with topographical AFM analysis. In NP-bioconjugateinteractions, the study of adhesion forces has played a large role in the analysis ofNP-bioconjugates. Adhesion forces have been used to map the hydrophilic and hydro-phobic portions of surface functionalized gold NPs [155]. By attaching ligand or receptormolecules to the tip of the AFM cantilever, recognition events and binding forcesbetween the cantilever and the sample surface can be measured using force-distancecurves [156]. Figure 15.6(a) reflects AFM topography and phase measurements thatallowed researchers to distinguish between the core and ligand-corona shell of aProbucol NP and determine the thickness of the ligand shell [157]. The sensitivity andversatility of AFM have allowed researchers to map ligands attached to a surface in 3-D[158]. Due to the wide range of information that can be obtained with AFM under physi-ological conditions, many different aspects of NP-bioconjugates can be monitored.General surface topography, extent of surface coverage, and the strength of thebioconjugation may all be analyzed with this technique.

Transmission electron microscopy (TEM) can be used to image nanoparticle shapes and,to some extent, their sizes [159–161]. TEM uses the wave nature of electrons as anillumination source to “image” a sample [143]. Because electrons and their resultingwavelengths are much smaller than atoms, they can be used to probe and provide infor-mation about a sample’s atomic structure, with the capability to image features down tothe Angstrom scale [162]. Electrons emitted from the source are focused to a thin beamusing electromagnetic lenses before passing through a vacuum column to the sample.Depending on the density of each area, some of the electrons are scattered, while othersare transmitted through the sample, with only electrons transmitted or minimally scat-tered reaching the imaging detector. The image is a function of the accelerating voltage,sample thickness, and material under study. The thicker and/or denser the sample orarea analyzed (e.g., the higher the atomic number), the less likely electrons will be ableto pass through the sample and the more likely they are to be scattered away from themain path, thereby producing darker regions in the image. Reducing scatter and

15.2 Methods

309

increasing electron penetration through the sample can be achieved using a higheraccelerating voltage, giving the electrons more energy.

TEM can be a useful tool for characterizing the size and shape of the NP core prior tobioconjugation [163]. Figure 15.6(b) is a TEM image of gold NPs, both prior to and afterconjugation with hemoproteins, showing the uniformity in shape and size (for thegiven field of view) of the NPs themselves [104]. The direct imaging capability of theTEM is particularly useful for NPs with nonspherical shapes and maybe useful to obtainmeasurements of aspect ratios [104]. In comparison, a particle size analyzer (such as DLS)may report a wide distribution of sizes for nanorods or other nonspherical shapesdepending on the angle at which the analyzer has detected the individual particles, andtherefore not provide a true indication of size and shape.

It is possible, using TEM at low accelerating voltages, to visualize biomoleculesattached to a NP core as demonstrated in Figure 15.6(c), for polylactide NP prior to andafter tethering adenoviruses to the surface [164]. The tethered biomolecule appears as alight halo around the high contrast, dark, NP core. At higher accelerating voltages, theimage of the core will be sharper, giving a more accurate sizing measurement; however,the bioconjugates will be less visible under these conditions. Imaging the bioconjugates


310

BA(A) (B)

0 100 200 300 400 500

(nm)

0 100 200 300 400 500

(nm)

(nm

)

500

300

400

200

100

0

500

300

400

200

100

0

(nm

)

(nm

)

40

20

0

-20

-40

(deg

)

1.5

1.0

0

0.5

-0.5

-1.0

-1.5

(A) (B)(A) (B)

0 100 200 300 400 500

(nm)

0 100 200 300 400 500

(nm)

(nm

)

500

300

400

200

100

0

500

300

400

200

100

0

(nm

)

(nm

)

40

20

0

-20

-40

(deg

)

1.5

1.0

0

0.5

-0.5

-1.0

-1.5

0 100 200 300 400 500

(nm)

0 100 200 300 400 500

(nm)

(nm

)

500

300

400

200

100

0

500

300

400

200

100

0

500

300

400

200

100

0

500

300

400

200

100

0

(nm

)

(nm

)

40

20

0

-20

-40

40

20

0

-20

-40

(deg

)

1.5

1.0

0

0.5

-0.5

-1.0

-1.5

(deg

)

1.5

1.0

0

0.5

-0.5

-1.0

-1.5

i ii

C D

Figure 15.6 Microscopy techniques. (a) Atomic force microscopy (AFM) characterization ofNP-biomolecule conjugates. (i) AFM topography image and (ii) AFM phase image ofProbucol/Polyvinylpyrrolidone (PVP)-K17/sodium dodecyl sulfate (SDS) NPs. Probucol is a choles-terol-lowering agent and used as a model drug here to demonstrate improved aqueous solubility of thedrug when formulated as a NP by co-grinding with PVP and SDS. (Images kindly provided by Dr.Moribe (Chiba University) and reprinted with permission from Pharmaceutical Society of Japan [157].)(b) TEM of Hemoglobin-gold NPs, showing uniform sized particles. The gold core-protein shell struc-ture is shown in the under focused image of NP-bioconjugates (inset). (Images kindly supplied by Dr.Pradeep (Indian Institute of Technology Madras). Reprinted with permission from [104], Copyright2008 American Chemical Society.) (c) TEM of a free PLA-NP (left) compared to a PLA-NP–Adenovirusbioconjugates (right), demonstrating successful complex formation. Negative staining with 2% uranylacetate aided adenovirus imaging. (Reprinted by permission from Macmillan Publishers Ltd: MolecularTherapy [164], copyright 2006.) (d) SEM of perforated aggregates obtained from polystyrene-b-horseradish peroxidase (HRP) bioconjugates (bar scale = 200 nm). (Images kindly supplied by Dr.Sommerdijk (Eindhoven University of Technology). Images reprinted from Angewandte Chemie Inter-national Edition [175]. 2002. Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with per-mission.)

may also be facilitated by staining the biomolecules with a contrast agent as demon-strated for Cowpea Mosaic viral (CPMV) NPs following negative staining [165].

There are several limitations to TEM characterization of NP-bioconjugates. First, theelectron beam is composed of ionizing radiation and can damage the sample, especiallysoft, biologic, or polymeric materials and certain NPs such as those made of hydrogels.In such cases the image changes over time, and care should be exercised for interpreta-tion of images taken while the electron beam has been continuously focused on thesame sample area. Second, one of the drawbacks of the very high magnification used forTEM measurements is the resulting small field of view. To obtain an adequate size mea-surement of NPs, samples of many particles (>100), and therefore many TEM images,must be surveyed followed by averaging. Alternatively, a complementary methodshould be employed in conjunction with TEM to provide an average size measurementfor the sample. Third, because samples are placed into the vacuum column for imaging,they must be dry. Typical TEM samples are made by drop-drying a solution onto aholey/lacey C film Cu-grid and either allowing the solution to dry or actively wickingaway excess solution with absorbent film. However, drying out the NP-bioconjugatesmay cause the biologic or polymeric molecules to collapse on themselves and the parti-cles to aggregate. Thus, the resulting TEM image of the NP-bioconjugates may not be atrue reflection of its native state or size, and care must be taken to interpret the images.Complementary methods are therefore commonly employed to get a true indication ofthe morphology of the NP-bioconjugates in their intended media. Cryogenic-TEM is amore recent advancement in the technology and has been applied to the study of tran-sient nanostructures such as lipid micelles, vesicles, and bilayers [166], and may presenta truer reflection of a NP-bioconjugate native state.

Scanning electron microscopy (SEM) can also be used to image NP shapes and sizes. Thetechnique uses a high energy electron beam, ranging from a few hundred eV up to ~100keV, which is rastered across the surface energy secondary electrons, and/or X-rays aregenerated at each point, and each provides different types of information about the sam-ple. The intensities of the secondary electrons are a function of both the sample compo-sition and the topographic geometry of the sample. Only the low energy secondaryelectrons generated near the surface are able to escape the sample and be measured, andhence SEM is primarily a topographical technique. Magnification of SEM can range from

~25× to ~250 k×, and features down to ~0.5–5 nm can be resolved depending on the spotsize of the beam and its interactions with the sample. In addition to topography details,the high energy backscattered electrons (BEI mode) also give information about thecomposition, with contrast arising from differences in atomic number. The SEM (andTEM) can also be combined with energy dispersive X-ray spectroscopy analysis (EDX,EDS, or EDAX) which characterizes the generated X-rays to provide elemental composi-tion. For a more in-depth description of the SEM technique and instrumentation, thereader is referred to [132, 167, 168].

There are several limitations to the SEM technique. First, there is a need to coatnonconductive samples with a conductive layer (e.g., gold), such that the sample doesnot build up charge from the electron beam, so care must be taken when interpretingSEM images from coated samples. If the gold coating is thick, for example, the details ofthe sample surface would be that of the gold and its coating process rather than the NPsample. Like TEM, traditional SEM requires dry samples that do not outgas and, if inpowder form, adhere to the sample mount. Environmental SEM (ESEM) does allows

15.2 Methods

311

sample imaging under low pressure, fairly high humidity and without the requirementfor a conducting overcoat and has been particularly useful for imaging biological sam-ples [169, 170] but has not been applied to NP-bioconjugates to date. Traditional SEMtechniques have been used to characterize NP shapes and sizes prior to and afterbioconjugation [171–173], but like TEM these are more commonly used to characterizethe NP core as opposed to the conjugated biomolecules. SEM has the added advantage ofa larger imaging field of view than the TEM [174]. Figure 15.6(d) shows an SEM of pro-tein-conjugated polystyrene NPs with a perforated microstructure [175]. The combina-tion of SEM and the elemental analysis of EDX has been used to characterize a fungalprotease-gold NP bioconjugate [172] and the growth of hydroxyapatite crystals ofphysiologically clotted fibrin modified gold NPs [121].

Traditional transmission optical light microscopy involves sample illumination frombelow coupled with detection/observation from above and typically measures the reflec-tion or absorption of the light. However, unlike electron microscopy, traditional lightmicroscopy cannot typically resolve nanoscale features <200 nm due to diffraction limi-tations, and as such has limited use in characterizing individual NP-bioconjugate prop-erties. Recently, however, ultrahigh-resolution optical systems have been designed toresolve features less than 100 nm [145, 176–178].

Fluorescence microscopy, a derivative of the traditional optical microscopy method-ology, in contrast offers the possibility of single-molecule-based measurements [145].Single-molecule microscopy techniques, encompassing both fluorescence and AFMmethodologies, have recently been reviewed [179, 180]. Fluorescence correlation spec-troscopy (see Section 15.2.2), which takes advantage of confocal scanning microscopy,has studied the size and photophysical properties of a number of NP types, with singlemolecule resolution. These single molecule microscopy techniques have been used tostudy the metal-enhanced fluorescence (MEF) resulting from Cy5-labeled DNA hybridiz-ing to DNA-modified silver NPs [181, 182], photobleaching of fluorescently labeledproteins attached to luminescent NPs [183], and various Förster resonance energy trans-fer (FRET) studies [179, 184, 185]. Single-molecule microscopy techniques and opticalmicroscopy, in general, however, have had limited application to date for characteriza-tion of individual NP-bioconjugates, in part due to the relatively recent emergence ofappropriate technologies. They have, however, been extensively used to track andimage a number of NP-bioconjugates in cell and small animal studies which rely onfunctional-based responses.

15.2.4 Spectroscopic

Spectroscopic techniques study the interaction of electromagnetic radiation with asample material, resulting in the wavelength-dependent absorption, and in the caseof fluorescence re-emission, of radiation. Typically, a wavelength dependent spectrumis produced with characteristic absorption/emission peaks inherent to the sample[186]. Spectroscopic techniques can provide a range of information about the NP-bioconjugate, including the confirmation of successful NP-biomolecule conjugation,the conformational state of the biomolecule once attached to the NP, the averageNP-to-biomolecule ratio, and the stability of the resulting NP-bioconjugate.

A number of NPs have intrinsic optical properties such as UV-visible absorbance spec-tra, which can be used to characterize NP properties, such as concentration, size, and


312

aggregation state. QDs have size-sensitive absorption profiles, which, although usefulfor characterizing the quality of the inorganic NP itself, are found to be insensitive toligand addition and refractive index changes in the medium [187–189], but the same isnot always true for the QD emission profiles. Metal NPs, and in particular gold and sil-ver, exhibit strong absorption in the visible region, known as the surface plasmon reso-nance (SPR) band (although often shortened to SP band). The SP band is dependent on anumber of factors and is found to be sensitive to shape, composition (i.e., Ag, Au,nanoshell structures), aggregation state, and also refractive index changes within surfaceproximity [99, 104, 187, 190–192]. A number of researchers have looked at the effect ofhemoproteins binding to gold and silver NPs, finding small (~5 nm) shifts in the UV-vis-ible measured SP band as a result of protein binding and in some cases larger shifts due tosubsequent aggregation [99, 104, 193]. The tunable, sensitive nature of the gold or sil-ver NP SP band has resulted in applications in numerous, diverse fields such assurface enhanced Raman and fluorescence measurements as well as a number ofaggregation-based sensing assays and biomedical imaging [191, 194].

Apart from characterizing the NP itself, UV-visible spectroscopy has also found appli-cation in analyzing NP-bioconjugates using both direct and indirect approaches. Directcharacterization is possible when the biomolecule has a distinct UV-visible profile thatremains discernable upon conjugation to the NP. Protein absorption bands at 280–290nm and the soret bands (410 nm) for hemoproteins, such as cytochrome c, have beenused to directly quantify the average amount of protein immobilized on a NP surface[12, 29, 195]. Proteins coeluting with NPs can also be viewed in gel electrophoresis usingcolorimetric stains, such as Coosamine Blue, which can then be quantified with appro-priate instrumentation (such as a fiber optic spectrophotometer: Ocean Optics) to deter-mine the average NP-to-biomolecule ratio. Alternatively, using a more indirect method,the amount of protein present in solution before and after NP exposure can be quanti-fied using either protein absorbance at 280–290 nm [196] or a number of reactivecolorimetric assays, including the Bradford reagent and bicinchoninin acid (BCA) assays[197, 198], as demonstrated for gold-coated magnetic particles modified with anti-body fragments [199] and single-walled carbon nanotubes (SWNTs) functionalizedwith enzymes [200]. Many of these tests, although not specifically designed fornanomaterials, may be applied to their analysis. As NPs have been demonstrated tointerfere with certain assays and tests, appropriate analysis of controls and the “naked”particles should be undertaken prior to interpretation of the results. While providing anaverage NP-to-biomolecule ratio, such tests do not give a NP-bioconjugate productdistribution profile or information regarding biomolecule conformation once attachedto the NP surface.

Circular dichroism (CD) measures the ability of optically active materials to differen-tially absorb circularly polarized light (usually UV) and has been applied to theconformational analysis of biomolecules, in particular proteins [201–204] and, to asomewhat lesser extent, nucleic acids [205]. Far-UV (< 250 nm) CD provides structuralinformation concerning the protein, including the degree of α-helical, β-sheet, or otherstructure (random coil). Changes in the CD spectra induced by external conditions, suchas temperature, pH, salt, or various denaturants, are representative of conformationalchanges within the protein’s secondary structure. CD has successfully been applied tothe study of protein conformational changes that occur following both initial adsorp-tion and later stable interactions with a variety of NP surfaces, including gold [99, 193,

15.2 Methods

313

206], silica [76, 196], iron oxide [12, 29], carbon nanotubes [207], and QDs [208]. Theobserved conformational changes of proteins interacting specifically or nonspecificallywith NP surfaces have been found to be dependent on a number of factors including pH[206], surface density of the protein [196], additional ligands present on the NP surface,and temperature [29]. Advantages of CD analysis include that it is nondestructive andstudies are typically performed in physiologically relevant aqueous environments. How-ever, to correctly interpret the protein response, extensive CD characterization of theparticular protein in solution (no NP) under natural and denatured conditions is pre-ferred and ideally confirmed using a high-resolution technique such as X-ray crystallog-raphy [201, 202]. CD analysis also requires buffers that are nonabsorbing in the UVrange to allow measurements below 200 nm. Oxygen is a known interferent (as itabsorbs strongly below 170 nm) and since most NPs scatter or absorb in the UV region,the NP concentration must be limited to reduce noise in the CD spectra [12, 203]. Anumber of recent developments in CD technology offer exciting prospects. Synchrotronradiation CD (SRCD) offers more detailed spectral information at wavelengths <190 nmdue to higher light intensities relative to conventional light sources in this region,although the main disadvantage is its limited availability [203]. Vibrational CD (VCD)uses polarized light in the infrared (IR) region, looking predominately at variations inthe amide I and amide II bands of proteins, although it does require replacement of H2Owith D2O in solution measurements, due to the strong absorption of water in thisregion [209].

Fluorescent spectroscopy, both steady-state and time-resolved, is a powerful and sensi-tive technique for determining a number of the parameters associated with the immobi-lization of biomolecules to a NP surface, including fluorophore local environment,biomolecule-NP coupling ratio, conformational state, and in some instances moleculardistances. Fluorescence techniques are, of course, limited to NP-bioconjugate compo-nents that have some form of either intrinsic or extrinsic fluorescence; however, giventhe variety of fluorescent NPs and the vast array of commercially available biomolecularreactive fluorescent dyes this should not be considered a limitation. A number ofresearchers have used the intrinsic fluorescence from tryptophan (Trp) residues, com-monly found in protein sequences, to obtain information about local changes in tertiarystructure upon NP binding [196, 206, 207, 210]. The Trp fluorescence of bovine serumalbumin was significantly quenched and the fluorescence emission maximum blueshifted in wavelength upon binding to 15-nm gold NPs [206]. The fluorescence quench-ing is expected due to efficient energy transfer with the metal NP surface while the slightblue wavelength shift is characteristic of Trp residues transferring to a more hydropho-bic environment. β-Lactoglobulin absorbing to silica NP surfaces was found to increasethe intrinsic Trp fluorescence and cause a red wavelength shift suggesting the protein isunfolding and exposing the Trp residues to a more hydrophilic environment. Moreinterestingly, the extent of protein unfolding was dependent on its surface concentra-tion, with very little unfolding observed at high surface coverage [196].

The biomolecules themselves can be extrinsically labeled with fluorophores to aid inNP-bioconjugation characterization and optimization. Fluorescently labeled DNA wasused to determine the effects of salt concentration, spacer composition, NP size, andsonication on the average DNA coverage per gold NP [211]. Since gold is a knownquencher of fluorescence emission, once purified, the DNA was displaced from thegold NP surface using dithiothreitol (DTT), a multithiolated species, and the solution


314

fluorescence measured once the gold NPs had been removed via centrifugation. Thisdirect displacement assay takes advantage of the thiol exchange mechanism that occurson gold (and other noble metal) surfaces and is therefore unique to metallic NPs. A simi-lar exchange method was used to determine the number of fluorescently labeledpeptides conjugated via disulfide linkages to semiconductor QD surfaces [48]. Unreacteddopamine was quantified following a dopamine-QD coupling reaction, usingo-phthaldialdehyde (OPA) in the presence of β-mercaptothanol, which produces ahighly fluorescent product upon reaction with primary amines [212]. While thesemethods provide information about the average biomolecule coverage on the NP sur-face, distribution information is still lacking. In single-molecule studies the proteinα-bungarotoxin, mono-labeled with an Alexa488 dye, was coupled to a lanthanide-iondoped siloxane-based luminescent NP and stepwise photobleaching of the Alexa488and wide-field fluorescence microscopy used to quantify the NP-protein ratio for eachindividual NP and hence obtain ratio distributions (see Figure 15.7(a)) [183].

Förster resonance energy transfer (FRET) is a unique configuration of fluorescencespectroscopy that can be used to extract very specific information regarding NP-bioconjugates. FRET is a nonradiative process that occurs between an excited state donor(typically fluorescent) and ground state acceptor species (fluorescent or nonfluorescent).A number of extensive reviews concerning FRET can be found in the literature [213,

15.2 Methods

315

FT-IR?

A B

0 1000 2000 3000 4000 5000 6000

0

5

10

15

20

25

01

2

3

4

5

>5

0 1

2

34

5

>5

1

3 Alexa/NP

24

5

2

1

Alexa/proteinAlexa/NP

proteins/NP

Nu

mb

er

of

em

iss

ion

sp

ots

0 1000 2000 3000 4000 5000 6000

0

5

10

15

20

25

01

2

3

4

5

>5

0 1

2

34

5

>5

1

3 Alexa/NP

24

5

2

1

Alexa/proteinAlexa/NP

proteins/NP

Initial count number

Nu

mb

er

of

em

iss

ion

sp

ots

(i)

(ii)

(iii)

(i)

(ii)

(iii)

(iv)

Figure 15.7 Spectroscopic techniques for NP-bioconjugate characterization. (a) (i) Wide-field fluores-cence microscopy images of lanthanide-ion doped oxide-NPs labeled with an Alex488-labeled protein.The left image NP emission and right image shows Alex488 emission. Stepwise photo-bleaching of theAlex488 is observed for individual NP-protein bioconjugates and the number of bleaching stepscounted to precisely measure the number of proteins-per-NP. The data is summarized in (ii), whichshows the distribution of Alex488-proteins-per-NP. (iii) Pie graphs summarizing the distribution ofAlex488-per-NP, Alex488-per-protein and Alex488-proteins-per-NP. (Images kindly supplied by Dr.Casanova (Ecole Polytechnique). Reprinted with permission from [183], Copyright 2008 AmericanChemical Society.) (b) Successive FT-IR spectra of the functionalization of detonated nanodiamondsamples with biotin. (i) FT-IR of pristine nanodiamond (arrow shows the carbonyl band of thenanodiamond material which disappears upon reduction), (ii) hydroxylated nanodiamond, producedthrough reaction of (i) with borane, (iii) silanization of (ii) with (3-aminopropyl) trimethoxysilane, fol-lowed by biotinylation to produce biotinylated nanodiamond (iv). The presence of characteristicbonds confirms successful modification at each stage of the synthesis. (Image kindly provided by Dr.Krueger (Christian-Albrechts-Universität zu Kiel). Reprinted with permission from [224], Copyright2008 American Chemical Society.)

214]. Suffice to say that the FRET phenomenon is highly dependent on a number of fac-tors, most importantly the extent of donor/acceptor spectral overlap and the distancebetween the two. The underlying process has been likened to a molecular ruler with sep-aration sensitivities for donor/acceptor distances proportional to r6 and usually falls inthe 1–10-nm range. Medintz and coworkers have demonstrated the unique abilities ofsemiconductor QDs as donors in a variety of FRET formats [215–218]. In an elegant FRETstudy, they used six mutants of maltose binding protein (MBP), each labeled at a differ-ent unique site with a rhodamine dye, to determine FRET distances for each mutant andestablish the orientation of MBP immobilized to a QD NP [215]. FRET has also been usedto monitor the binding of fluorescently labeled proteins or peptides to the surface ofQDs [55, 219], but is more commonly used as a signal transduction mechanism in func-tional assays. A number of researchers have used gold NP acceptors as quenchers inenergy transfer studies with fluorescent donor species and the resulting surface energytransfer (SET) process has been shown to have a nontraditional r4 distance dependency,essentially extending the reach of the molecular ruler [220–222]. Sen and coworkersrecently used Trp-gold SET to probe conformational changes that occur when BSA bindsgold NPs [222].

Infrared (IR) spectroscopy measures the absorption of IR radiation by a sample result-ing from vibrational stretching and bending modes within the molecule. Technicaladvances in IR spectroscopy, notably Fourier transform-IR (FT-IR), have resulted in itsnow-routine use in the characterization of protein structures [223]. Many researchershave used FT-IR spectroscopy to demonstrate NP bioconjugation through the appear-ance of characteristic spectral bands, including biotin to diamond NPs (see Figure15.7(b)) [224], dextran or albumin to gold NPs [206, 225], hemoproteins on gold and sil-ver NPs [104, 193], streptavidin to silicon [66], and β-lactoglobulin adsorbed on silicaNPs [196]. In the case of globular proteins, careful interpretation of the stretching andbending vibrations in the amide regions can provide secondary structural informationregarding α-helical, β-sheet, turns, and “other” (also referred to as unordered or randomcoil) strands [196, 200, 206, 223]. The amide I band (1,600–1,700 cm–1) region, in partic-

ular, is found to show substantial changes related to α-helical → β-sheet structuralconformational changes. Samples are prepared by either depositing within solid KBr pel-lets [104, 200, 224] or dissolving in an aqueous solvent [197, 206, 225]. Liquid samplesare measured using either special IR optical cuvettes [225] or an attenuated total reflec-tion (ATR) attachment [196, 206]. As with many of the techniques discussed, appropri-ate background spectra must be acquired prior to sample analysis and subsequentlysubtracted from the sample spectra. This is particularly important for aqueous sampleswhere water is found to be strongly absorbing in the IR region [223]. Some researchers

perform H2O → D2O exchange prior to the measurement, although this can result inband frequency shifts and incomplete exchange which can complicate spectral interpre-

tation [223]. The rate of H2O → D2O exchange monitored in the amide II region(1,510–1,580 cm−1), resulting from increased exposure of internal protein peptides to theexternal aqueous environment, was used to infer tertiary conformational changes ofβ-lactoglobulin adsorbed on silica NPs [196].

Nuclear magnetic resonance (NMR) spectroscopy and nuclear magnetic imaging (MRI)measure the intrinsic magnetic moment of certain nuclei in the presence of an appliedmagnetic field [226]. While a number of atomic nuclei comprise odd numbers of neu-trons or protons, and hence intrinsic magnetic moments, hydrogen-1 (1H), carbon-13


316

(13C), and occasionally nitrogen-15 (15N) isotopes represent the most commonly used inNMR studies. Application of an external magnetic field causes splitting of nuclear spinstate energy levels within these magnetic nuclei and, assuming a Boltzmann distributionbetween the states, absorption (typically in the radio frequency range) of electromag-netic radiation can occur and effect transition between these magnetically split energystates. Through these unique chemical shifts, measured in ppm versus an internal stan-dard (commonly tetramethylsilane-TMS), and the peak splitting (Zeeman effect) thatoccur, NMR spectroscopy can provide physical, chemical, and structural/environmentalinformation about the species under study. MRI measures relaxation rates referred to aseither T1 or T2 values, which correspond to different relaxation mechanisms—spin-lat-tice and spin-spin, respectively. Superparamagnetic materials such as iron oxide NPs andNPs doped or labeled with gadolinium (Gd) are commonly used in MRI as contrastagents for biomedical applications [227, 228]. NMR spectroscopy is nondestructive andhas been used to determine the structure and dynamic interactions of many biologicalmolecules, including proteins and nucleic acids [226]. NMR spectroscopy has also beenused to characterize PEG-stabilized lipid NPs [229] and gold NPs [230] as well as QD capexchange reactions [51] and dendrimer NP-surfactant interactions [231]. High-resolu-tion 2-D 1H-15N NMR (see [226] for an overview of 2-D NMR) was used to compare theinteractions of human carbonic anhydrase (HCA) I and II with silica NPs, demonstratingenzyme-dependent conformational changes upon interaction with the NP [232]. Bylooking at the environmentally sensitive NMR peak shifts, splitting, and/or relaxationrates (T1 and T2), researchers are able to elicit information concerning the mechanism ofinteraction, highlighting NMR as a technique capable of providing dynamic and struc-tural NP-bioconjugate conformation information. Sample preparation is often key,as NMR is relatively insensitive and requires relatively high concentrations of purematerials to obtain quality spectra, with deuterated solvents preferred. That said, giventhe different NMR techniques available, such as high-resolution 2-D NMR, NuclearOverhauser Effect (NOE) NMR, and transverse relaxation optimized spectroscopy(TROSY), NMR represents a powerful and underutilized tool available to researchers forNP-bioconjugate characterization [226].

15.2.5 Mass Spectroscopy

Mass spectroscopy (MS) comprises a family of analytical technologies that analyzesamples based on their mass-to-charge ratio. The basic instrument arrangement includesan ionizing source, the mass analyzer, and the detector, with the various types of MSreferring to different ionizing and/or mass analyzer technologies. MS has been used tostudy protein structure by measuring either the intact protein, denatured protein, orenzyme digested protein, and in particular demonstrates its utility in proteomics[233]. When dealing with protein-containing samples, electrospray ionization (ESI) andmatrix-assisted laser desorption/ionization (MALDI) represent the MS techniques ofchoice [233]. MS has been used to characterize NP-bioconjugates and has found particu-lar utility in the analysis of protein based NPs, such as viral NPs, where mass increases inthe viral coat proteins due to the addition of biotin or fluorophore species was success-fully monitored using MALDI–time of flight (TOF)-MS [234–236]. Through the mea-sured mass increase, the stochiometry of the additional species added per virus NP couldreadily be determined. MALDI-TOF MS has also been used to qualitatively demonstrate

15.2 Methods

317

hemoprotein binding to gold and silver NPs [104]. Inductively coupled plasma (ICP)-MSwas used to determine TiO2 NPs binding a dopamine ligand used to complex a gadolin-ium MRI contrast agent [227]. The application of MS techniques is currently fairly lim-ited for NP-bioconjugate characterization; this may in part be due to the relative cost ofthe instrumentation and the required level of expertise needed to run analyses.

15.2.6 Thermal Techniques

Thermal gravimetric analysis (TGA) is a method that utilizes a high-precision balance todetermine changes in the weight of a bulk sample relative to changes in temperature. Bymodifying the temperature and the rate of heating, information can be obtained aboutthe sample. Such information includes determining the relative amounts of inorganicversus organic components or the amount of adsorbed water or other solvents presentwithin the material. In terms of NP-bioconjugates, TGA may aid in determining theamount of conjugate biomolecules as well as their thermal stability as shown inFigure 15.8(a) for magnetic NPs functionalized with PEG-based polymers [237]. Theamount of dendrons attached to gold NPs and subsequent surface reactions have alsobeen monitored with TGA [238], as well as the amount of an active therapeutic within aNP-bioconjugate, as demonstrated for paclitaxel bound to gold NPs [239]. Further calcu-lations can reveal information about the average number of ligands attached per NP andthe extent of surface functionalization as demonstrated for PEG/lactose ligands on goldNPs [240].


318

Figure 15.8 Thermal analysis of NP-bioconjugates. (a) Thermogravimetric analysis (TGA) used todetermine the thermal stability and organic component of inorganic/organic NP-bioconjugates. Ironoxide magnetic NPs were functionalized with poly(ethylene glycol) monomthacrylate (PEGMA) via asilane initiator, [4-(chloromethyl)phenyl]triclorosilane (CTS), by applying a copper-mediated atomtransfer radical polymerization (ATRP). The distinct TGA curves of (i) as prepared magnetic NPs, (ii)CTS- magnetic NPs, and (iii, iv, v) polymerized-(PEGMA)-magnetic NPs after 1-, 2-, and 4-hour poly-merization times, respectively, provide indications of the amount of CTS and polymerized-(PEGMA)present on the magnetic NPs. (Images kindly provided by Dr. Neoh (National University of Singapore).Reprinted with permission from [237], Copyright 2008 American Chemical Society.) (b) Differentialscanning calorimetry (DSC) of different components of bovine serum albumin (BSA)-Zn2+ NPs formedin polyethylene glycol (PEG) solutions. DSC thermograms of (i) zinc acetate control, (ii) PEG control,(iii) BSA alone, and (iv) BSA-Zn2+ NPs are shown. (Images kindly supplied by Dr. Yuan (ShanghaiJiaotong University School of Pharmacy) and reprinted by permission from IOP Publishing Ltd:Nanotechnology [244], copyright 2007.)

As with TGA, differential scanning calorimetry (DSC) and isothermal titration calorimetry(ITC) are thermal methods that can provide bulk information about the NP-bioconjugate. The basic measurement monitors the difference in the amount of heatrequired to increase/decrease the temperature of a sample versus a reference material.DSC is used to study various transitions including melting, crystallization, glasstransition, and decomposition. Subsequent analysis can indicate the state of theNP-bioconjugate such as the stability of the biomolecule, structural information of boththe NP and biomolecule including crystallinity, and how the different components areinteracting with each other. Researchers have used DSC to elucidate the structure andstability of surface coatings of NP-bioconjugates as well as the state of therapeutic pay-loads. For example, DSC has been used to determine the state (order versus disordered,interdigitated, and so forth) of dodecylamine and cetyltrimethylammonium bromide(CTAB) ligands bound to gold NPs [241] and the stability of solid, lipid NP-insulin com-plexes [242], and to investigate the physical state of paclitaxel incapsulated insidepoly(lactic-coglycolic acid) NPs [243]. DSC has also been used to study how individualcomponents of a NP-bioconjugate system (Zn nanoparticles and bovine serum albumin)interact with each other [244] (see Figure 15.8(b)). ITC has the potential to investigatethe stoichiometry, affinity, and enthalpy of the NP-biomolecule interaction, as demon-strated by various polymeric NPs binding proteins, but as an analytical technique stillremains vastly underutilized [14].

Thermophoresis or thermodiffusion involves local heating of a sample and monitoringthe resulting motion of the particles due to the temperature gradient [13, 245, 246]. Sim-ilar to thermal FFF (Section 15.2.1), the direction that particles move in a temperaturegradient is found to be dependent both on the overall size and on the surface potentialof the particle. The effective diameter of the particle can be estimated by conversion ofthe measured diffusion coefficients. Thermophoresis has found limited application forNP-bioconjugate characterization to date, demonstrating the ability to determine thesize of various PEG-functionalized QDs [13] and colloidal suspensions [246]. That said, itis very much an evolving technology and has demonstrated potential.

15.3 Summary Points

As the design of NP-bioconjugates and their subsequent applications become morediverse and complex, it is essential that researchers have at hand techniques capable ofintimately characterizing these specialized hybrid materials. This chapter has endeav-ored to highlight some of the major characterization techniques currently available toNP-bioconjugate researchers. All these techniques have associated advantages and dis-advantages including relative cost, ease of use, resolution capabilities, sample prepara-tion, ease of data interpretation, versatility, and bulk versus single particle analysis.Probably the most important issues from a NP-bioconjugate characterization viewpointare: (1) confirmation of biomolecule attachment to the NP, (2) determination of theaverage ratio of NP-to-biomolecule, which includes the individual NP-to-biomoleculeratio and resulting ratio distribution, (3) the NP-bioconjugate hydrodynamic radius andaggregation state, and finally (4) the activity of the biomolecule upon NP attachment (asrelated to its orientation, structure, and stability). Clearly there are a number of tech-niques available to the researcher that can, at least to some extent, address some of these

15.3 Summary Points

319

questions. For example, chromatography and electrophoresis techniques are relativelycheap and widely available, can readily confirm biomolecule attachment to the NP sur-face, and can provide purification and characterization (hydrodynamic size) of theNP-bioconjugate product, and in some instances even provide the ability to resolve NPswith different ratios of biomolecules attached. DLS and Zeta potential characterizationof NP-bioconjugates are relatively cheap and simple to perform providing hydrody-namic radius, aggregation state, and surface potential information. The electron micro-scope techniques, SEM and TEM, are mainly used for characterization of the NP core(not so much the biomolecule to date) and, while relatively expensive equipment- andmaintenance-wise, characterize the size and shape of the NP on an individual particlebasis. AFM, in contrast, can divulge a range of information about both the NP and thebiomolecule again on a single-particle basis. Many of the spectroscopic techniques pro-vide bulk analysis of the NP-bioconjugate, with NMR and IR spectroscopy demonstrat-ing the ability to characterize biomolecule conformational states. Also desirable aretechniques with the ability to characterize the NP-bioconjugate under physiologicalenvironments as well as their stability postproduction. Characterizing the NP-bioconjugate stability to sterilization will likewise become increasingly important for invivo applications. Many of the techniques described require dried samples or samplessuspended in ultrapure liquids, which may result in nonphysiological states and resultin perturbed properties of the NP-bioconjugate, so interpretation should be erred on theside of caution. It is apparent that while there are many techniques, no one techniquecan address all questions or apply to all types of NP-bioconjugates. Full characterizationwill require a combination of techniques and the exact choice and the extent of tailoringrequired will depend on the particular NP-bioconjugates under characterization. Suchcharacterization may require collaboration between researchers, since a number of thetechniques discussed require specialized facilities and/or training. As mentioned previ-ously, the NCL [24] offers to perform extensive characterization of nanoparticle materi-als to researchers involved in the areas of cancer therapy and diagnostics. Alternatively,there are a limited number of commercial companies that offer nanoparticle character-ization services, such as nanoComposix [247]. Many of the technologies described arecontinually evolving to meet the demands of nanoscale characterization, and whilebulk analysis will continue to play a significant role, additional focus should be placedon techniques capable of purifying and characterizing individual NP-bioconjugatepopulations. Many of the technologies described here will play a pivotal role in thefuture development of these novel and increasingly complex NP-bioconjugates and areindispensable to the future of this field.

Acknowledgments

The authors would like to thank CDRH/OSEL/Division of Biology and Division ofChemistry and Materials Science and CDER/OPS/OTR/Division of Applied Pharmacol-ogy Research. The mention of commercial products, their sources, or their use in con-nection with material reported herein is not to be construed as either an actual orimplied endorsement of such products by the Department of Health and HumanServices.


320

References

[1] Niemeyer, C.M., and C.A. Mirkin, (eds.), Nanobiotechnology: Concepts, Applications, and Perspectives,New York: Wiley-VCH, 2004.

[2] Medintz, I.L., Mattoussi, H., and Clapp, A.R., “Potential clinical applications of quantum dots,”International Journal of Nanomedicine, Vol. 3, 2008, pp. 151–167.

[3] Gonsalves, K.E., Halberstadt, C.R., Laurencin, C.T., and Nair, L.S., (eds.), Biomedical Nanostructures,New York: John Wiley & Sons, 2008.

[4] Cai, W., and Chen, X., “Nanoplatforms for targeted molecular imaging in living subjects,” Small,Vol. 3, 2007, pp. 1840–1854.

[5] Katz, E., and Willner, I., “Integrated nanoparticle-biomolecule hybrid systems: synthesis, proper-ties, and applications,” Angewandte Chemie International Edition, Vol. 43, 2004, pp. 6042–6108.

[6] Willner, I., Baron, R., and Willner, B., “Integrated nanoparticle-biomolecule systems for biosensingand bioelectronics,” Biosensors and Bioelectronics, Vol. 22, 2007, pp. 1841–1852.

[7] Pandey, P., Datta, M., and Malhotra, B.D., “Prospects of nanomaterials in biosensors,” AnalyticalLetters, Vol. 41, 2008, pp. 159–209.

[8] Luykx, D.M.A.M., Peters, R.J.B., Van Ruth, S.M., and Bouwmeester, H., “A review of analyticalmethods for the identification and characterization of nano delivery systems in food,” Journal ofAgricultural and Food Chemistry, Vol. 56, 2008, pp. 8231–8247.

[9] Sotiropoulou, S., Sierra-Sastre, Y., Mark, S.S., and Batt, C.A., “Biotemplated nanostructured materi-als,” Chemical Materials, Vol. 20, 2008, pp. 821–834.

[10] Wang, Y., Angelatos, A.S., and Caruso, F., “Template synthesis of nanostructured materials vialayer-by-layer assembly,” Chemical Materials, Vol. 20, 2008, pp. 848–858.

[11] Gaumet, M., Vargas, A., Gurny, R., and Delie, F., “Nanoparticles for drug delivery: The need for pre-cision in reporting particle size parameters,” European Journal of Pharmaceutics andBiopharmaceutics, Vol. 69, 2008, pp. 1–9.

[12] Aubin-Tam, M.-E., and Hamad-Schifferli, K., “Structure and function of nanoparticles-protein con-jugates,” Biomedical Materials, Vol. 3, 2008, Article# 034001.

[13] Sperling, R.A., Liedl, T., Duhr, S., Kudera, S., Zanella, M., Lin, C.-A.J., Chang, W.H., Braun, D., andParak, W.J., “Size determination of (bio)conjugated water-soluble colloidal nanoparticles: a com-parison of different techniques,” Journal of Physical Chemistry C, Vol. 111, 2007, pp. 11552–11559.

[14] Cedervall, T., Lynch, I., Berggard, T., Thulin, E., Nilsson, H., Dawson, K.A., and Linse, S., “Under-standing the nanoparticle-protein corona using methods to quantify exchange rates and affinitiesof proteins for nanoparticles,” Proceedings of the National Academy of Sciences, Vol. 104, 2007,pp. 2050–2055.

[15] Liu, H., and Wedster, T.J., “Nanomedicine for implants: a review of studies and necessary experi-mental tools,” Biomaterials, Vol. 28, 2007, pp. 354–369.

[16] Howard, M.D., Jay, M., Dziublal, T.D., and Lu, X.L., “PEGylation of nanocarrier drug delivery sys-tems: state of the art,” Journal of Biomedical Nanotechnology, Vol. 4, 2008, pp. 133–148.

[17] Medintz, I.L., Uyeda, H.T., Goldman, E.R., and Mattoussi, H., “Quantum dot bioconjugates forimaging, labeling and sensing,” Nature Materials, Vol. 4, 2005, pp. 435–446.

[18] Hermanson, G.T., (ed.), Bioconjugate Techniques, 2nd ed., San Diego, CA: Academic Press, 2008.[19] Powers, K.W., Brown, S.C., Krishna, V.B., Wasdo, S.C., Moudgil, B.M., and Roberts, S.M., “Research

strategies for safety evaluation of nanomaterials. Part VI. Characterization of nanoscale particlesfor toxicological evaluation,” Toxicological Sciences, Vol. 90, 2006, pp. 296–303.

[20] Warheit, D.B., “How meaningful are the results of nanotoxicity studies in the absence of adequatematerial characterization?” Toxicological Sciences, Vol. 101, 2008, pp. 183–185.

[21] Bonroy, K., Frederix, F., Reekmans, G., Dewolf, E., De Plama, R., Borghs, G., Declerck, P., andGoddeeris, B., “Comparison of random and oriented immobilization of antibody fragments onmixed self-assembled monolayers,” Journal of Immunological Methods, Vol. 312, 2006, pp. 167–181.

[22] Hassellöv, M., Readman, J.W., Ranville, J.F., and Tiede, K., “Nanoparticle analysis and characteriza-tion methodologies in environmental risk assessment of engineered nanoparticles,” Ecotoxicology,Vol. 17, 2008, pp. 344–361.

[23] Tiede, K., Boxall, A.B.A., Tear, S.P., Lewis, J., David, H., and Hassellöv, M., “Detection and charac-terization of engineered nanoparticles in food and the environment,” Food Additives and Contami-nants, Vol. 25, 2008, pp. 795–821.

[24] http://ncl.cancer.gov/[25] http://ncl.cancer.gov/working_technical_reports.asp[26] Otto, D.P., Vosloo, B.C.M., and de Villiers, M.M., “Application of size exclusion chromatography

in the development and characterization of nanoparticulate drug delivery systems,” Journal of Liq-uid Chromatography and Related Technologies, Vol. 30, 2007, pp. 2489–2514.

References

321

[27] Fernández-Argüelles, M.T., Costa-Fernández, J.M., Pereiro, R., and Sanz-Medel, A., “Simplebio-conjugation of polymer-coated quantum dots with antibodies for fluorescence-basedimmunoassays,” Analyst, 133, 2008, pp. 444–447.

[28] Wilcoxon, J.P., and Provencio, P.P., “Chemical and optical properties of CdSe and CdSe/ZnSnanocrystals investigated using high-performance liquid chromatography,” Journal Physical Chem-istry B, Vol. 109, 2005, pp. 13461–13471.

[29] Aubin-Tam, M.-E., Zhou, H., and Hamad-Schifferli, K., “Gold nanoparticle-cytochrome c com-plexes: The effect of nanoparticle ligand charge on protein structure,” Langmuir, Vol. 21, 2005,pp. 12080–12084.

[30] Mullen, D.G., Desai, A.M., Waddell, J.N., Cheng, X., Kelly, C.V., McNerny, D.Q. Majoros, I.J.,Baker, J.R., Sander, L.M., Orr, B.G., and Holl, M.M.B., “The implications of stochastic synthesis forthe conjugation of functional groups of nanoparticles,” Bioconjugate Chemistry, Vol. 19, 2008,pp. 1748–1752.

[31] Claridge, S.A., Liang, H.W., Basu, S.R., Fréchet, J.M.J., and Alivisatos, A.P., “Isolation of discretenanoparticle-DNA conjugates for plasmonic applications,” Nano Letters, Vol. 8, 2008,pp. 1202–1206.

[32] Wang, S., Mamedova, N., Kotov, N.A., Chen, W., and Studer, J., “Antigen/antibodyimmunocomplex from CdTe nanoparticle bioconjugates,” Nano Letters, Vol. 2, 2002, pp. 817–822.

[33] Krueger, K.M., Al-Somali, A.M., Mejia, M., and Colvin, V.L., “The hydrodynamic size of polymerstabilized nanocrystals,” Nanotechnology, Vol. 18, 2007, Article# 475709.

[34] Gao, D., Xu, H., Philbert, M.A., and Kopelman, R., “Bioelimination nanohydrogels for drug deliv-ery,” Nano Letters, Vol. 8, 2008, pp. 3320–3324.

[35] Yegin, B.A., and Lamprecht, A., “Lipid nanocapsule size analysis by hydrodynamic chromatogra-phy and photon correlation spectroscopy,” International Journal of Pharmaceutics, Vol. 320, 2006,pp. 165–170.

[36] Blom, M.T., Chmela, E., Oosterbroek, R.E., Tijssen, R., and van den Berg, A., “On-chip hydrody-namic chromatography separation and detection of nanoparticles and biomolecules,” AnalyticalChemistry, Vol. 75, 2003, pp. 6761–6768.

[37] Giddings, J.C., “Field-flow fractionation: Analysis of macromolecular, colloidal, and particulatematerials,” Science, Vol. 260, 1993, pp. 1456–1465.

[38] Schimpf, M., Caldwell, K.D., and Giddings, J.C., (eds.), Field Flow Fractionation Handbook, NewYork: Wiley Interscience, 2000.

[39] Contado, C., Dalpiaz, A., Leo, E., Zborowski, M., and Williams, P.S., “Complementary use of flowand sedimentation field-flow fractionation techniques for size characterizing biodegradablepoly(lactic acid) nanospheres,” Journal of Chromatography A, Vol. 1157, 2007, pp. 321–335.

[40] Pasti, L., Agnolet, S., and Dondi, F., “Thermal field-flow fractionation of charged submicrometerparticles in aqueous media,” Analytical Chemistry, Vol. 79, 2007, pp. 5284–5296.

[41] Andersson, M., Fromell, K., Gullberg, E., Artursson, P., and Caldwell, K.D., “Characterization ofsurface-modified nanoparticles for in vivo biointeraction. A sedimentation field flow fractionationstudy,” Analytical Chemistry, Vol. 77, 2005, pp. 5488–5493.

[42] Fraunhofer, W., Winter, G., and Coester, C., “Asymmetrical flow field-flow fractionation andmultiangle light scattering for analysis for gelatin nanoparticle drug carrier systems,” AnalyticalChemistry, Vol. 76, 2004, pp. 1909–1920.

[43] Augsten, C., Kiselev, M.A., Gehrke, R., Hause, G., and Mäder, K., “A detailed analysis of biodegrad-able nanospheres by different techniques—a combined approach to detect particle sizes and sizedistributions,” Journal of Pharmaceutical and Biomedical Analysis, Vol. 47, 2008, pp. 95–102.

[44] Kang, D.Y., Kim, M.J., Kim, S.T., Oh, K.S., Yuk, S.H., and Lee, S., “Size characterization ofdrug-loaded polymeric core/shell nanoparticles using asymmetrical flow-field fractionation,” Ana-lytical Bioanalytical Chemistry, Vol. 390, 2008, pp. 2183–2188.

[45] Rameshwar, T., Samal, S., Lee, S., Kim, S., Cho, J., and Kim, I.S., “Determination of the size ofwater-soluble nanoparticles and quantum dots by field-flow fractionation,” Journal of Nanoscienceand Nanotechnology, Vol. 6, 2006, pp. 2461–2467.

[46] Carpino, F., Zborowski, M., and Williams, P.S., “Quadrupole magnetic field-flow fractionation: Anovel technique for the characterization of magnetic nanoparticles,” Journal of Magnetism and Mag-netic Materials, Vol. 311, 2007, pp. 383–387.

[47] Lyvén, B., Hassellöv, M., Haraldsson, C., and Turner, D.R., “Optimization of on-channelpreconcentration in field-flow fractionation for the determination of size distributions of lowmolecular weight colloidal material in natural waters,” Analytica Chimica Acta, Vol. 357, 1997,pp. 187–196.

[48] Derfus, A.M., Chen, A.A., Min, D.-H., Ruoslahti, E., and Bhatia, S.N., “Targeted quantum dot conju-gates for siDNA delivery,” Bioconjugate Chemistry, Vol. 18, 2007, pp. 1391–1396.


322

[49] Parak, W.J., Pellegrino, T., Micheel, C.M., Gerion, D., Williams, S.C., and Alivisatos, A.P., “Confor-mation of oligonucleotides attached to gold nanocrystals probed by gel electrophoresis,” Nano Let-ters, Vol. 3, 2003, pp. 33–36.

[50] Park, S., Brown, K.A., and Hamad-Schifferli, K., “Changes in oligonucleotide conformation onnanoparticles surfaces by modification with mercaptohexanol,” Nano Letters, Vol. 4, 2004,pp. 1925–1929.

[51] Pons, T., Uyeda, H.T., Medintz, I.L., and Mattoussi, H., “Hydrodynamic dimensions, electropho-retic mobility, and stability of hydrophilic quantum dots,” Journal of Physical Chemistry B, 110,2006, pp. 20308–20316.

[52] Bagalkot, V., Zhang, L., Levy-Nissenbaum, E., Jon, S., Kantoff, P.W., Langer, R., and Farokhzad,O.C., “Quantum dot-aptamer conjugates for synchronous cancer imaging, therapy, and sensing ofdrug delivery based on bi-fluorescence resonance energy transfer,” Nano Letters, Vol. 7, 2007, pp.3065–3070.

[53] Kovacs, E.W., Hooker, J.M., Romanini, D.W., Holder, P.G., Berry, K.E. and Francis, M.B., “Dual-sur-face-modified bacteriophage MS2 as an ideal scaffold for viral capsid-based drug delivery system,”Bioconjugate Chemistry, Vol. 18, 2007, pp. 1140–1147.

[54] Pellegrino, T., Sperling, R.A., Alivisatoes, A.P., and Parak, W.J., “Gel electrophoresis of gold-DNAnanoconjugates,” Journal of Biomedicine and Biotechnology, 2007, Article# 26796.

[55] Medintz, I.L., Pons, T., Delehanty, J.B., Susumu, K., Brunel, F.M., Dawson, P.E. and Mattoussi, H.,“Intracellular delivery of quantum dot-protein cargos mediated by cell penetrating peptides,”Bioconjugate Chemistry, Vol. 19, 2008, pp. 1785–1795.

[56] Hanauer, M., Pierrat, S., Zins, I., Lotz, A., and Sönnichsen, C., “Separation of nanoparticles by gelelectrophoresis according to size and shape,” Nano Letters, Vol. 7, 2007, pp. 2881–2885.

[57] Warnement, M.R., Tomlinson, I.D., Chang, J.C., Schreuder, M.A., Luckabaugh, C.M., andRosenthal, S.J., “Controlling the reactivity of ampiphillic quantum dots in biological assaysthrough hydrophobic assembly of custom PEG derivatives,” Bioconjugate Chemistry, Vol. 19, 2008,pp. 1404–1413.

[58] Rodbard, D., and Chrambach, A., “Unified theory for gel electrophoresis and gel filtration,” Pro-ceedings of the National Academy of Sciences, Vol. 65, 1970, pp. 970–977.

[59] Rodriguez, M.A. and Armstrong, D.W., “Separation and analysis of colloidal/nano-particles includ-ing microorganisms by capillary electrophoresis: a fundamental review,” Journal of ChromatographyB, Vol. 800, 2004, pp. 7–25.

[60] Liu, F.-K., Tsai, M.-H., Hsu, Y.-C., and Chu, T.-C., “Analytical separation of Au/Ag core/shellnanoparticles by capillary electrophoresis,” Journal of Chromatography A, Vol. 1133, 2006,pp. 340–346.

[61] Liu, F.-K., “A high-efficiency capillary electrophoresis-based method for characterizing the sizes ofAu nanoparticles,” Journal of Chromatography A, Vol. 1167, 2007, pp. 231–235.

[62] Lin, K.-H., Chu, T.-C., and Liu, F.-K., “On-line enhancement and separation of nanoparticles usingcapillary electrophoresis,” Journal of Chromatography A, Vol. 1161, 2007, pp. 314–321.

[63] Hjertén, S., “High-performance electrophoresis: Elimination of electroendosmosis and soluteadsorption,” Journal of Chromatography A, Vol. 347, 1985, pp. 191–198.

[64] Huang, X., Weng, J., Sang, F., Song, X., Cao, C., and Ren, J., “Characterization of quantum dotbioconjugates by capillary electrophoresis with laser-induced fluorescent detection,” Journal ofChromatography A, Vol. 1113, 2006, pp. 251–254.

[65] Helle, A., Hirsjärvi, S., Peltonen, L., Hirvonen, J., and Wiedmer, S.K., “Quantitative determinationof drug encapsulation in poly(lactic acid) nanoparticles by capillary electrophoresis,” Journal ofChromatography A, Vol. 1178, 2008, pp. 248–255.

[66] Choi, J., Wang, N.S. and Reipa, V., “Conjugation of the photoluminescent silicon nanoparticles tostreptavidin,” Bioconjugate Chemistry, 19, 2008, pp. 680–685.

[67] Wang, F.-H., Yoshitake, T., Kim, D.-K., Muhammed, M., Bjelke, B., and Kehr, J., “Determination ofconjugation efficiency of antibodies and proteins to the superparamagnetic iron oxidenanoparticles by capillary electrophoresis with laser-induced fluorescence detection,” Journal ofNanoparticle Research, Vol. 5, 2003, pp. 137–146.

[68] Kim, H.R., Andrieux, K., Delomenie, C., Chacun, H., Appel, M., Desmaële, D., Taran, F., Georgin,D., Couvreur, P., and Taverna, M., “Analysis of plasma protein adsorption onto PEGylatednanoparticles by complementary methods: 2D-CE, CD and protein lab-on-chip system,” Electro-phoresis, Vol. 28, 2007, pp. 2252–2261.

[69] Feng, H.-T., Law, W.-S., Yu, L.J., and Li, S.F.-Y., “Immunoassay by capillary electrophoresis withquantum dots,” Journal of Chromatography A, Vol. 1156, 2007, pp. 75–70.

[70] Jamison, J.A., Krueger, K.M., Yavuz, C.T., Mayo, J.T., LeCrone, D., Redden, J.J., and Colvin, V.L.,“Size-dependent sedimentation properties of nanocrystals,” ACS Nano, Vol. 2, No. 2, 2008,pp. 311–319.

References

323

[71] Ralston, G., Introduction to Analytical Ultracentrifugation, Fullerton, CA: Beckman Instruments,1993.

[72] Liu, J., Andya, J.D., and Shire, S.J., “A critical review of analytical ultracentrifugation and field flowfractionation methods for measuring protein aggregation,” The American Association of Pharmaceu-tical Scientists’ Journal, Vol. 8, No. 3, 2006, pp. E580–E589.

[73] Langer, K., Anhorn, M.G., Steinhauser, I., Dreis, S., Celebi, D., Schrickel, N., Faust, S., and Vogel, V.,“Human serum albumin (HAS) nanoparticles: reproducibility of preparation process and kineticsof enzymatic degradation,” International Journal of Pharmaceutics, Vol. 347, 2008, pp. 109–117.

[74] Calabretta, M., Jamison, J.A., Falkner, J.C., Liu, Y., Yuhas, B.D., Mathews, K.S., and Colvin, V.L.,“Analytical ultracentrifugation for characterizing nanocrystals and their bioconjugates,” NanoLetters, Vol. 5, 2005, pp. 963–967.

[75] Lees, E.M., Gunzburg, M.J., Nguyen, T.L., Howlett, G.J., Rothacker, J., Nice, E.C., Clayton, A.H.A.,and Mulvaney, P., “Experimental determination of quantum dot size distributions, ligand packingdensities, and bioconjugation using analytical ultracentrifugation,” Nano Letters, Vol. 8, 2008,pp. 2883–2890.

[76] Lundqvist, M., Sethson, I., and Jonsson, B.-H., “Protein adsorption onto silica nanoparticles:conformational changes depend on the particles’ curvature and the protein stability,” Langmuir,Vol. 20, 2004, pp. 10639–10647.

[77] Berne, B. J., and Pecora, R., Dynamic Light Scattering: With Applications to Chemistry, Biology, andPhysics, New York: Courier Dover Publications, 2000.

[78] Schartl, W., Light Scattering from Polymer Solutions and Nanoparticle Dispersions, New York:Springer-Verlag, 2007.

[79] Pecora, R., “Dynamic light scattering measurement of nanometer particles in liquids,” Journal ofNanoparticle Research, Vol. 2, 2000, pp. 123–131.

[80] Xu, R., Particle Characterization: Light Scattering Methods, New York: Springer, 2000.[81] Takahashi, K., Kato, H., Saito, T., Matsuyama, S., and Kinugasa, S., “Precise measurement of the size

of nanoparticles by dynamic light scattering with uncertainty analysis,” Particle and Particle SystemsCharacterization, 25, 2008, pp. 31–38.

[82] Xing, Y., So, M.K., Koh, A.L., Sinclair, R., and Rao, J., “Improved QD-BRET conjugates for detectionand imaging,” Biochemical and Biophysical Research Communications, Vol. 372, 2008, pp. 388–394.

[83] Ipe, B.I., Shukla, A., Lu, H., Zou, B., Rehage, H., and Niemeyer, C.M., “Dynamic light-scatteringanalysis of the electrostatic interaction of hexahistidine-tagged cytochrome P450 enzyme withsemiconductor quantum dots,” Chemical Physical Chemistry, Vol. 7, 2006, pp. 1112–1118.

[84] Khlebtsov, N.G., Bogatyrev, V.A., Khlebtsov, B.N., Dykman, L.A., and Englebienne, P., “Amultilayer model for gold nanoparticle bioconjugates: application to study of gelatin and humanIgG adsorption using extinction and light scattering spectra and the dynamic light scatteringmethod,” Colloid Journal, Vol. 65, 2003, pp. 622–635.

[85] Rodríquez-Fernández, K., Pérez-Juste, J., Liz-Marzán, L.M., and Lang, P.R., “Dynamic light scatter-ing of short Au rods with low aspect ratios,” Journal of Physical Chemistry C, Vol. 111, 2007, pp.5020–5025.

[86] Van de Hulst, H.C., Light Scattering by Small Particles, New York: John Wiley & Sons, 1957, pp 1–80.[87] Kuyper, C.L., Fujimoto, B.S., Zhao, Y., Schiro, P.G., and Chiu, D.T., “Accurate sizing of

nanoparticles using confocal correlation spectroscopy,” Journal of Physical Chemistry B, Vol. 110,2006, pp. 24433–24441.

[88] Zhang, P., Li, L., Dong, C., Qian, H., and Ren, J., “Sizes of water-soluble luminescent quantum dotsmeasured by fluorescence correlation spectroscopy,” Analytica Chimica Acta, Vol. 546, 2005, pp.46–51.

[89] Wu, B., Chen, Y., and Müller, J.D., “Fluorescence correlation spectroscopy of finite-size particles,”Biophysical Journal, Vol. 94, 2008, pp. 2800–2808.

[90] Müller, C.B., Loman, A., Richtering, W., and Enderlein, J., “Dual-focus fluorescence correlationspectroscopy of colloidal solutions: influence of particle size,” Journal of Physical Chemistry B,Vol. 112, 2008, pp. 8236–8240.

[91] Gadd, J.C., Kuyper, C.L., Fujimoto, B.S., Allen, R.W., and Chiu, D.T., “Sizing subcellular organellesand nanoparticles confined within aqueous droplets,” Analytical Chemistry, Vol. 80, 2008,pp. 3450–3457.

[92] Larson, D.R., Zipfel, W.R., Williams, R.M., Clark, S.W., Bruchez, M.P., Wise, F.W., and Webb,W.W., “Water-soluble quantum dots for multiphoton fluorescence imaging in vivo,” Science, Vol.300, 2003, pp.1434–1436.

[93] Doose, S., Tsay, J.M., Pinaud, F., and Weiss, S., “Comparison of photophysical and colloidal proper-ties of biocompatible semiconductor nanocrystals using fluorescence correlation spectroscopy,”Analytical Chemistry, Vol. 77, 2005, pp. 2235–2242.


324

[94] Rusu, L., Gambhir, A., McLaughlin, S., and Rädler, J., “Fluorescence correlation spectroscopy stud-ies of peptide and protein binding to phospholipid vesicles,” Biophysical Journal, Vol. 87, 2004,pp. 1044–1053.

[95] Ray, K., Zhang, J., and Lakowicz, J.R., “Fluorescence lifetime correlation spectroscopic study offluorophore-labeled silver nanoparticles,” Analytical Chemistry, Vol. 80, 2008, pp. 7313–7318.

[96] Hunter, R.J., Zeta Potential in Colloid Science: Principles and Applications, San Diego, CA: AcademicPress, 1981.

[97] http://www.malvern.co.uk[98] Bergman, L., Rosenholm, J., Öst, A.-B., Duchanoy, A., Kankaanpää, P., Heino, J., and Lindén, M.,

“On the complexity of electrostatic suspension stabilization of functionalized silica nanoparticlesfor biotargeting and imaging applications,” Journal of Nanomaterials, 2008, Article# 712514.

[99] Gomes, I., Santos, N.C., Oliveira, M.A., Quintas, A., Eaton, P., Pereira, E., and Franco, R., “Probingsurface properties of cytochrome c at Au bionanoconjugates,” Journal of Physical Chemistry C,Vol. 112, 2008, pp. 16340–16347.

[100] Liu, Z., Li, X., Tabakman, S.M., Jiang, K., Fan, S., and Dai, H., “Multiplexed multicolor Raman imag-ing of live cells with isotopically modified single walled carbon nanotubes,” Journal of the AmericanChemical Society, Vol. 130, 2008, pp. 13540–13541.

[101] Yang, Q., Shuai, L., Zhou, J., Lu, F., and Pan, X., “Functionalization of multiwalled carbonnanotubes by pyrene-labeled hydroxypropyl cellulose,” Journal of Physical Chemistry B, Vol. 112,2008, pp. 12934–12939.

[102] Efremov, E.V., Ariese, F., and Gooijer, C., “Achievements in resonance Raman spectroscopy reviewof a technique with a distinct analytical chemistry potential,” Analytica Chimica Acta, Vol. 606,2008, pp. 119–134.

[103] Keating, C.D., Kovaleski, K.M., and Natan, M.J., “Protein:colloid conjugates for surface enhancedRaman scattering: stability and control of protein orientation,” Journal of Physical Chemistry B, Vol.102, 1998, pp. 9404–9413.

[104] Tom, R.T., Samal, A.K., Sreeprasad, T.S., and Pradeep, T., “Hemoprotein bioconjugates of gold andsilver nanoparticles and gold nanorods: structure-function correlations,” Langmuir, Vol. 23, 2007,pp. 1320–1325.

[105] Qian, X., Peng, X.-H., Ansari, D.O., Yin-Goen, Q., Chen, G.Z., Shin, D.M., Yang, L., Young, A.N.,Wang, M.D., and Nie, S., “In vivo tumor targeting and spectroscopic detection with sur-face-enhanced Raman nanoparticle tags,” Nature Biotechnology, Vol. 26, 2008, pp 83–90.

[106] Lutz, B., Dentinger, C., Sun, L., Nguyen, L., Zhang, J., Chmura, A.J., Allen, A., Chan, S., andKnudsen, B., “Raman nanoparticle probes for antibody-based protein detection in tissues,” Journalof Histochemistry and Cytochemistry, Vol. 56, 2008, pp. 371–379.

[107] Qian, X., Zhou, X., and Nie, S., “Surface-enhanced Raman nanoparticle beacons based onbioconjugated gold nanocrystals and long range plasmonic coupling,” Journal of the AmericanChemical Society, Vol. 130, 2008, pp. 14934–14935.

[108] Fu, A., Zhou, N., Huang, W., Wang, D., Zhang, L., and Li, H., “Effects of nano HAP on biologicaland structural properties of glass bone cement,” Journal of Biomedical Materials Research Part A, 74A,2, 2005, pp. 156–163.

[109] Brunner, T.J., Bohner, M., Dora, C., Gerber, C., and Stark, W.J., “Comparison of amorphous TCPnanoparticles to micron-sized α-TCP as starting materials for calcium phosphate cements,” Journalof Biomedical Materials Research Part B-Applied Biomaterials, Vol. 84B, 2008, pp. 350–362.

[110] Tyner, K.M, Roberson, M.S., Berghorn, K.A., Li, L., Gilmour Jr., R.F., Batt, C.A., and Giannelis, E.P.,“Intercalation, delivery, and expression of the gene encoding green fluorescence protein utilizingnanobiohybrids,” Journal of Controlled Release, Vol. 100, 2004, pp. 399–409.

[111] Khan, A.I., Lei, L., Norquist, A.J., and O’Hare, D., “Intercalation and controlled release of pharma-ceutically active compounds from a layered double hydroxide,” Chemical Communication, Vol. 22,2001, pp. 2342–2343.

[112] Wang, X., Du, Y., and Luo, J., “Biopolymer/montmorillonite nanocomposite: preparation,drug-controlled release property and cytotoxicity,” Nanotechnology, Vol. 19 2008, Article# 065707.

[113] Rastogi, S.K., and Suryanarayanan, R., “Characterization of delivery systems, X-ray powderdiffractometry,” in Encyclopedia of Controlled Drug Delivery Volumes 1 & 2, E. Mathiowitz, (ed.), NewYork: John Wiley & Sons, 1999, pp. 275–285.

[114] Meier, M., “Crystallite size measurement using X-ray diffraction.” 2004,http://www.matsci.ucdavis.edu/MatSciLT/EMS-162L/Files/XRD-CSize1.pdf.

[115] Jenkins, R., and Snyder, R.L., Introduction to X-Ray Powder Diffractometry, New York: John Wiley andSons, 1996.

[116] Sherwood, D., and Emmanuel, B., “Computing shapes of nanocrystals from X-ray diffractiondata,” Crystal Growth and Design, Vol. 6, 2006, pp. 1415–1419.

[117] Tyner, K.M., Schiffman, S.R., and Giannelis, E.P., “Nanobiohybrids and delivery vehicles forcamptothecin,” Journal of Controlled Release, Vol. 95, 2004, pp. 501–514.

References

325

[118] Choy, J.H., Kwak, S.Y., Jeong, Y.J., and Park, J.S., “Inorganic layered double hydroxides as nonviralvectors,” Angewandte Chemie International Edition, Vol. 39, No. 22, 2000, pp. 4041–4045.

[119] Jenning, V., Schafer-Korting, M., and Gohla, S., “Vitamin A-loaded solid lipid nanoparticles fortopical use: drug release properties,” Journal of Controlled Release, Vol. 66, 2000, pp. 115–126.

[120] Sewell, S.L., and Wright, D.E., “Biomimetic synthesis of titanium dioxide utilizing the R5 peptidederived from cylindrotheca fusiformis,” Chemistry of Materials, Vol. 18, 2006, pp. 3108–3113.

[121] Sastry, T.P., Sundaraseelan, J., Swarnalatha, K., Liji Sobhana, S.S., Uma Makheswari, M., Sekar, S.,and Mandal, A.B., “Growth of hydroxyapatite on physiologically clotted fibrin capped goldnanoparticles,” Nanotechnology, Vol. 19, 2008, Article# 245604.

[122] Levine, J.R., Cohen, J.B., Chung, Y.W., and Georgopoulos, P., “Grazing-incidence small-angleX-ray scattering: new tool for studying thin film growth,” Journal of Applied Crystallography, Vol. 22,1989, pp. 528–532.

[123] Aubin-Tam, M.-E., Zhou, H., and Hamad-Schifferli, K., “Structure of cytochrome c at the interfacewith magnetic CoFe2O4 nanoparticles,” Soft Matter, Vol. 4, 2008, pp. 554–559.

[124] Glatter, O., and Kratky, O., (eds.), Small Angle X-Ray Scattering, San Diego, CA: Academic Press,1982.

[125] Park, S.Y., Lytton-Jean, A.K.R., Lee, B., Weigand, S., Schatz, G.C., and Mirkin, C.A., “DNA-program-mable nanoparticle crystallization,” Nature, 451, 2008, pp. 553-556.

[126] Noro, A., Nagata, Y., Tsukamoto, M., Hayakawa, Y., Takano, A., and Matsushita, Y., “Novel synthe-sis and characterization of bioconjugate block copolymers having oligonucleotides,”Biomacromolecules, Vol. 6, 2005, pp. 2328–2333.

[127] Meziani, M.J., Pathak, P., Harruff, B.A., Hurezeanu, R., and Sun, Y.-P., “Direct Conjugation of Semi-conductor Nanoparticles with Proteins,” Langmuir, Vol. 21, 2005, pp. 2008–2011.

[128] Park, S.-J., Lazarides, A.A., Storhoff, J.J., Pesce, L., and Mirkin, C.A., “The structural characterizationof oligonucleotide-modified gold nanoparticles networks formed by DNA hybridization,” Journalof Physical Chemistry, Vol. 108, 2004, pp. 12375–12380.

[129] Yamakoshi, S., Sakai, Y., Shinohara, Y., Amemiya, Y., Kanayama, N., Takrada, T., Maeda, M., andIto, K., “SAXS measurement of aggregate of DNA modified gold nanoparticles,” Nucleic Acids Sym-posium Series, Vol. 51, 2007, pp. 335–336.

[130] Hsu, W.-L., Li, Y.-C., Chen, H.-L., Liou, W., Jeng, U.-S., Lin, H.-K., Liu, W.-L., and Hsu, C.-S., “Ther-mally-induced order-order transition of DNA-cationic surfactant complexes,” Langmuir, Vol. 22,2006, pp. 7521–7527.

[131] Kahovec, J., (ed.), Macromolecular Symposia: Scattering Methods for the Investigation of Polymers,Weinheim, Germany: Wiley-VCH, 2002.

[132] Wachtman, J.B., (ed.), Characterization of Materials, Boston, MA: Butterworth-Heinemann, 1993.[133] Pynn, R., “Neutron scattering: a primer,” Los Alamos Science, Vol. 19, 1990, pp. 1–31.[134] Wignall, G.D., and Melnichenko, Y.B., “Recent applications of small-angle neutron scattering in

strongly interacting soft condensed matter,” Reports on Progress in Physics, Vol. 68, 2005,pp. 1761–1810.

[135] Teixeira, S.C.M., et al. “New sources and instrumentation for neutrons in biology,” Chemical Phys-ics, 345, 2008, pp. 133–151.

[136] Jacrot, B., “The study of biological structures by neutron scattering from solution,” Reports on Prog-ress in Physics, Vol. 39, 1976, pp. 911–953.

[137] Serefoglou, E., Oberdisse, J., and Staikos, G., “Characterization of the soluble nanoparticles formedthrough coulombic interaction of bovine serum albumin with anionic graft copolymers at lowpH,” Biomacromolecules, 2007, Vol. 8, pp. 1195–1199.

[138] Paul, A., Vicent, M.J., and Duncan, R., “Using small-angle neutron scattering to study the solutionconformation of N-(2-Hydroxypropyl)methacrylamide copolymer-doxorubicin conjugate,”Biomacromolecules, Vol. 8, 2007, pp. 1573–1579.

[139] Castelletto, V., Krysmann, M.J., Clifton, L.A., Lambourne, J., and Noirez, L., “Structural study ofBSA/poly(ethylene glycol) lipid conjugate complexes,” Journal of Physical Chemistry B, Vol. 111,2007, pp. 11330–11336.

[140] Hamley, I.W., Krysmann, M.J., Newby, G.E., Castelletto, V., and Noirez, L., “Orientational order-ing in the nematic phase of a polyethylene glycol-peptide conjugate in aqueous solution,” PhysicalReviews E, Vol. 77, 2008, Article# 062901.

[141] http://www.ill.eu/lss/more/world-directory-of-sans-instruments.[142] Bonnell, D., Scanning Probe Microscopy and Spectroscopy: Theory, Techniques, and Applications, 2nd

ed., New York: Wiley-VCH, 2000.[143] Egerton, R.F., Physical Principles of Electron Microscopy: An Introduction to TEM, SEM, and AEM, New

York: Springer, 2008.[144] Midgley, P.A., and Durkan, C., “The frontiers of microscopy,” Materials Today, Microscopy Special

Issue, 2008, pp. 8–11.


326

[145] Dedecker, P., Hofkens, J., and Hotta, J., “Diffraction-unlimited optical microscopy,” MaterialsToday, Microscopy Special Issue, 2008, pp. 12–21.

[146] Gaczynska, M., and Osmulski, P. A., “AFM of biological complexes: what can we learn?” CurrentOpinion in Colloid & Interface Science, Vol. 13, 2008, pp. 351–367.

[147] Cohen, S.R., and Bitler, A., “Use of AFM in bio-related systems,” Current Opinion in Colloid & Inter-face Science, Vol. 13, 2008, pp. 316–325.

[148] Carpick, R.W., and Salmeron, M., “Scratching the surface: Fundamental investigations of tribologywith atomic force microscopy,” Chemical Reviews, Vol. 97, 1997, pp. 1163–1194.

[149] Turner, Y.T.A., Roberts, C.J., and Davies. M.C., “Scanning probe microscopy in the field of drugdelivery,” Advanced Drug Delivery Reviews, Vol. 57, 2007, pp. 1453–1473.

[150] Liu, C., “Parallel scanning probe arrays: their applications,” Materials Today, Microscopy SpecialIssue, 2008, pp. 22–29.

[151] Li, P., Zhang, L., Ai, K., Li, D., Liu, X., and Wang, E., “Coating didodecyldimehtyl-ammonium bro-mide onto Au nanoparticles increase the stability of its complex with DNA,” Journal of ControlledRelease, Vol. 129, 2008, pp. 128–134.

[152] Campbell, J.F., Tessmer, I., Thorp, H.H. and Erie, D.A., “Atomic force microscopy studies ofDNA-wrapped carbon nanotube structure and binding to quantum dots,” Journal of the AmericanChemical Society, 130, 2008, pp. 10648–10655.

[153] Nehilla, B.J., Vu, T.Q., and Desai, T.A., “Stoichiometry-dependent formation of quantum dot-anti-body bioconjugates: a complementary atomic force microscopy agarose gel electrophoresis study,”Journal of Physical Chemistry B, Vol. 109, 2005, pp. 20724–20730.

[154] Paunesku, T., Rajh, T., Wiederrecht, G., Maser, J., Vogt, S., Stojicevic, N., Protic, M., Lai, B., Oryhon,J., Thurnauer, M., and Woloschak, G., “Biology of TiO2-oligonucleotide nanocomposites,” NatureMaterials, 2, 2003, pp. 343–346.

[155] Xu, L.P, Pradhan, S., and Chen, S., “Adhesion force studies of janus nanoparticles,” Langmuir,Vol. 23, 2007, pp. 8544–8548.

[156] Barattin, R., and Voyer, N., “Chemical modifications of AFM tips for the study of molecular recog-nition events,” Chemical Communications, 2008, pp. 1513–1532.

[157] Moribe, K., Wanawongthai, C., Shudo, J., Higashi, K., and Yamamoto, K., “Morphology and sur-face states of colloidal probucol nanoparticles evaluated by atomic force microscopy,” Chemicaland Pharmaceutical Bulletin, Vol. 56, 2008, pp. 878–885.

[158] Ebner, A., Kienberger, F., Kada, G., Stroh, C.M., Geretschlager, M., Kamruzzahan, A.S.M., Wildling,L., Johnson, W.T., Ashcroft, B., Nelson, J., Lindsay, S.M., Gruber, H.J., and Hinterdorfer, P., “Local-ization of single avidin-biotin interactions using simultaneous topography and molecular recogni-tion imaging,” Chemical Physical Chemistry, Vol. 6, 2005, pp. 897–900.

[159] Williams, D.B., and Carter, C.B., Transmission Electron Microscopy: A Textbook for Materials Science,New York: Plenum Press, 1996.

[160] Dykstra, M.J., and Reuss, L.E., Biological Electron Microscopy: Theory, Techniques, and Troubleshooting,New York: Springer, 2003.

[161] Pyrz, W.D., and Buttrey, D.J., “Particle size determination using TEM: A discussion of image acqui-sition and analysis for the novice microscopist,” Langmuir, Vol. 24, 2008, pp. 11350–11360.

[162] Smith, D.J., “Ultimate resolution in the electron microscope,” Materials Today, Microscopy SpecialIssue, 2008, pp. 30–38.

[163] Zhao, X., Hilliard, L.R., Mechery, S.J., Wang, Y., Bagwe, R.P., Jin, S., and Tan, W., “A rapid bioassayfor single bacterial cell quantitation using bioconjugated nanoparticles,” Proceedings of the NationalAcademy of Sciences, Vol. 101, 2004, pp. 15027–15032.

[164] Chorny, M., Fishbein, I., Alferiev, I.S., Nyanguile, O., Gaster, R., and Levy, R.J., “Adenoviral genevector tethering to nanoparticle surfaces results in receptor-independent cell entry and increasedtransgene expression,” Molecular Therapy, Vol. 14, 2006, pp. 382–391.

[165] Zhan, W., Barnhill, H.N., Sivakumar, K., Tian, H., and Wang, Q., “Synthesis of hemicyanine dyesfor ‘click’ bioconjugation,” Tetrahedron Letters, Vol. 46, 2005, pp. 1691–1695.

[166] Lee, J., Jha, A.K., Bose, A., and Tripathi, A., “Imaging new transient nanostructures using amicrofluidic chip integrated with a controlled environment vitrification system for cryogenictransmission electron microscopy,” Langmuir, Vol. 24, 2008, pp. 12738–12741.

[167] Goldstein, J., Newbury, D.E., Echlin, P., Lyman, C.E., Joy, D.C., Lifshin, E., Sawyer, L.C., andMichael, J.R., Scanning Electron Microscopy and X-Ray Microanalysis, 2nd ed., New York: Springer,2003.

[168] Gabriel, B.L., SEM: A User’s Manual for Materials Science, Metals Park, OH: American Society for Met-als, 1985.

[169] Thiberge, S., Zik, O., and Moses, E., “An apparatus for imaging liquids, cells and other wet samplesin the scanning electron microscopy,” Review of Scientific Instruments, Vol. 75, 2004,pp. 2280–2289.

References

327

[170] Gatti, A.M., Kirkpatrick, J., Gambarelli, A., Capitani, F., Hansen, T., Eloy, R., and Clermont, G.,“ESEM evaluations of muscle/nanoparticles interface in a rat model,” Journal of Materials Scienceand Materials Medicine, Vol. 19, 2008, pp. 1515–1522.

[171] Smith, J.E., Wang, L., and Tan, W., “Bioconjugated silica-coated nanoparticles for bioseparationand bioanalysis,” Trends in Analytical Chemistry, Vol. 25, 2006, pp. 848–855.

[172] Gole, A., Dash, C., Ramakrishnan, V., Sainkar, S.R., Mandale, A.B., Rao, M., and Sastry, M. “On thepreparation, characterization, and enzymatic activity of fungal protease-gold colloidbioconjugates,” Bioconjugate Chemistry, Vol. 12, 2001, pp. 684–690.

[173] Skrabalak, S.E., Chen, J., Au, L., Lu, X., Li, X., and Xia, Y., “Gold nanocages for biomedical applica-tions,” Advanced Materials, Vol. 19, 2007, pp. 3177–3184.

[174] Rayavarapu, R.G., Petersen, W., Ungureanu, C., Post, J.N., van Leeuwen, T.G., and Manohar, S.,“Synthesis and bioconjugation of gold nanoparticles as potential molecular probes for light-basedimaging techniques,” International Journal of Biomedical Imaging, 2007, Article# 29817.

[175] Boerakker, M.J., Hannink, J.M., Bomans, P.H.H., Frederik, P.M., Nolte, R.J.M., Meijer, E.M., andSommerdijk, N.A.J.M., “Giant amphiphiles by cofactor reconstitution,” Angewandte Chemie Inter-national Edition, Vol. 41, 2002, pp. 4239–4241.

[176] Vainrub, A., Pustovyy, O., and Vodyanoy, V., “Resolution of 90 nm (λ/5) in an optical transmissionmicroscope with an annular condenser,” Optics Letters, Vol. 31, 2006, pp. 2855–2857.

[177] Klar, T.A., Engle, E., and Hell, S.W., “Breaking Abbe’s diffraction resolution limit in fluorescencemicroscopy with stimulated emission depletion beams of various shapes,” Physical Review E, Vol.64, 2001, pp. 1–9.

[178] Skebo, J.E., Grabinski, C.M., Schrand, A.M., Schlager, J.J., and Hussain, S.M., “Assessment of metalnanoparticle agglomeration, uptake and interaction using high-illuminating system,” InternationalJournal of Toxicology, Vol. 26, 2007, pp. 135–141.

[179] Tinnefeld, P., and Sauer, M., “Branching out of single-molecule fluorescence spectroscopy: Chal-lenges for chemistry and influence on biology,” Angewandte Chemie International Edition, Vol. 44,2005, pp. 2642–2671.

[180] Walter, N.G., Huang, C.-Y., Manzo, A.J., and Sobhy, M.A., “Do-it-yourself guide: how to use themodern single-molecule toolkit,” Nature Methods, Vol. 5, 2008, pp. 475–489.

[181] Zhang, J., Fu, Y., Chowdhury, M.H., and Lakowicz, J.R., “Metal-enhanced single molecule fluores-cence on silver particle monomer and dimer: coupling effect between metal particles,” Nano Let-ters, Vol. 7, 2007, pp. 2101–2107.

[182] Zhang, J., Fu, Y., Chowdhury, M.H., and Lakowicz, J.R., “Single-molecule studies on fluorescentlylabeled silver particles: effects of particle size,” Journal of Physical Chemistry C, Vol. 112, 2008, pp.18–26.

[183] Casanova, D., Gaiume, D., Moreau, M., Martin, J.-L., Gacoin, T., Boilot, J.-P., and Alexandrou, A.,“Counting the number of proteins coupled to single nanoparticles,” Journal of the American Chemi-cal Society, Vol. 129, 2007, pp. 12592–12593.

[184] Okamoto, K., and Terazima, M., “Distribution analysis for single molecule FRET measurement,”Journal of Physical Chemistry B, Vol. 112, 2008, pp.7308–7314.

[185] Pons, T., Medintz, I.L., Wang, X., English, D.S., and Mattoussi, H., “Solution-phase single quantumdot fluorescence resonance energy transfer,” Journal of the American Chemical Society, Vol. 128,2006, pp. 15324–15331.

[186] Pavia, D.L., Lampman, G.M., Kriz, G.S., and Vyvyan, J.A., Introduction to Spectroscopy, 4th ed.,Brookes Cole, 2008.

[187] Biju, V., Itoh, T., Anas, A., Sujith, A., and Ishikawa, M., “Semiconductor quantum dots and metalnanoparticles: syntheses, optical properties, and biological applications,” Analytical andBioanalytical Chemistry, Vol. 391, 2008, pp. 2469–2495.

[188] Murray, C.B., Kagan, C.R., and Bawendi, M.G., “Synthesis and characterization of monodispersenanocrystals and close-packed nanocrystal assemblies,” Annual Review of Materials Science, Vol. 30,2000, pp. 545–610.

[189] Leatherdale, C.A., Woo, W.K., Mikulec, F.V., and Bawendi, M.G., “On the absorption cross sectionof CdSe nanocrystal quantum dots,” Journal of Physical Chemistry B, Vol. 106, 2002, pp. 7619–7622.

[190] Uechi, I., and Yamada, S., “Photochemical and analytical applications of gold nanoparticles andnanorods utilizing surface plasmon resonance,” Analytical Bioanalytical Chemistry, Vol. 391, 2008,pp. 2411–2421.

[191] Schwartzberg, A.M., and Zhang, J.Z., “Novel optical properties and emerging applications of metalnanostructures,” Journal of Physical Chemistry C, Vol. 112, 2008, pp. 10323–10337.

[192] Schwartzberg, A.M., Olson, T.Y., Talley, C.E., and Zhang, J.Z., “Synthesis, characterization, andtunable optical properties of hallow gold nanospheres,” Journal of Physical Chemistry B, Vol. 110,2008, pp. 19935–19944.


328

[193] Jiang, X., Jiang, J., Jin, Y., Wang, E., and Dong, S., “Effect of colloidal gold size on theconformational changes of adsorbed cytochrome c: Probing by circular dichroism, UV-visible, andinfrared spectroscopy,” Biomacromolecules, Vol. 6, 2005, pp. 46–53.

[194] Qian, X.M., and Nie, S.M., “Single-molecule and single-nanoparticle SERS: from fundamentalmechanisms to biomedical applications,” Chemical Society Reviews, Vol. 37, 2008, pp. 912–920.

[195] Scodeller, P., Flexer, V., Szamocki, R., Calvo, E.J., Tognalli, N., Troiani, H., and Fainstein, A.,“Wired-enzyme core-shell Au nanoparticle biosensor,” Journal of the American Chemical Society,Vol. 130, 2008, pp. 12690–12697.

[196] Wu, X., and Narsimhan, G., “Characterization of Secondary and Tertiary Conformational Changesof β-lactoglobulin Adsorbed on silica nanoparticles Surfaces,” Langmuir, Vol. 24, 2008, pp.4989–4998.

[197] Olson, B.J.S.C., and Markwell, J., “Assays for determination of protein concentration,” Current Pro-tocols in Protein Science, Vol. 3.4, 2007, pp. 3.4.1–3.4.29.

[198] Hermanson, G.T., Mallia, A. K., and Smith, P., Immobilized Affinity Ligand Techniques, San Diego,CA: Academic Press, 1992.

[199] Zhang, H., and Meyerhoff, M.E., “Gold-coated magnetic particles for solid-phase immunoassays:Enhancing immobilized antibody binding efficiency analytical performance,” Analytical Chemis-try, Vol. 78, 2006, pp. 609–616.

[200] Karajanagi, S.S., Vertegel, A.A., Kane, R.S., and Dordick, J.S., “Structure and function of enzymesadsorbed onto single-walled carbon nanotubes,” Langmuir, Vol. 20, 2004, pp. 11594–11599.

[201] Sreerama, N., and Woody, R.W., “Estimation of protein secondary structure from circulardichroism spectra: comparison of CONTIN, SELCON, and CDSSTR methods with an expanded ref-erence set,” Analytical Biochemistry, Vol. 287, 2000, pp. 252–260.

[202] Whitmore, L., and Wallace, B.A., “Protein secondary structure analysis from circular dichroismspectroscopy: methods and reference databases,” Biopolymers, Vol. 89, 2007, pp. 392–400.

[203] Miles, A.J., and Wallace, B.A., “Synchrotron radiation circular dichroism spectroscopy of proteinsand applications in structural and functional genomics,” Chemical Society Reviews, Vol. 35, 2006,pp. 39–51.

[204] Cai, X.M., and Dass, C., “Conformational analysis of proteins and peptides,” Current Organic Chem-istry, Vol. 7, 2003, pp. 1841–1854.

[205] Richards, A.D., and Rodger, A., “Synthetic metallomolecules as agents for the control of DNA struc-ture,” Chemical Society Reviews, Vol. 36, 2007, pp. 471–483.

[206] Shang, L., Wang, Y., Jiang, J., and Dong, S., “pH-dependent protein conformational changes inalbumin: gold nanoparticle bioconjugates: a spectroscopic study,” Langmuir, Vol. 23, 2007,pp. 2714–2721.

[207] Asuri, P., Bale, S.S., Pangule, R.C., Shah, D.A., Kane, R.S., and Dordick, J.S., “Structure, function,and stability of enzymes covalently attached to single-walled carbon nanotubes,” Langmuir,Vol. 23, 2007, pp. 12318–12321.

[208] Mamedova, N.N., Kotov, N.A., Rogach, A.L., and Studer, J., “Albumin-CdTe nanoparticlebioconjugates: preparation, structure, and interunit energy transfer with antenna effect,” Nano Let-ters, Vol. 1, 2001, pp. 281–286.

[209] Shanmugam, G., Polavarapu, P.L., Kendall, A., and Stubbs, G., “Structures of plant viruses fromvibrational circular dichroism,” Journal of General Virology, Vol. 86, 2005, pp. 2371–2377.

[210] Gole, A., Dash, C., Ramakrishnan, V., Sainkar, S.R., Mandale, A.B., Rao, M., and Sastry, M., “Pep-sin-gold colloid conjugates: preparation, characterization, and enzymatic activity,” Langmuir,Vol. 17, 2001, pp. 1674–1679.

[211] Hurst, S.J., Lytton-Jean, A.K.R., and Mirkin, C.A., “Maximizing DNA loading on a range of goldnanoparticles sizes,” Analytical Chemistry, Vol. 78, 2006, pp. 8313–8318.

[212] Clarke, S.J., Hollmann, C.A., Aldaye, F.A., and Nadeau, J.L., “Effect of ligand density on the spec-tral, physical, and biological characteristics of CdSe/ZnS quantum dots,” Bioconjugate Chemistry,Vol. 19, 2008, pp. 562–568.

[213] Sapsford, K.E., Berti, L., and Medintz, I.L., “Materials for fluorescence resonance energy transferanalysis: beyond traditional donor-acceptor combinations,” Angewandte Chemie International Edi-tion, Vol. 45, 2006, pp. 4562–4588.

[214] Lakowicz, J.R., Principles of Fluorescence Spectroscopy, 2nd ed., New York: Springer, 2006.[215] Medintz, I.L., Konnert, J.H., Clapp, A.R., Stanish, I., Twigg, M.E., Mattoussi, H., Mauro, J.M., and

Deschamps, J.R., “A fluorescence resonance energy transfer-derived structure of a quantumdot-protein bioconjugates nanoassembly,” Proceedings of the National Academy of Sciences, Vol. 101,2004, pp. 9612–9617.

[216] Clapp, A.R., Medintz, I.L., Mauro, J.M., Fisher, B.R., Bawendi, M.G., and Mattoussi, H., “Fluores-cence resonance energy transfer between quantum dot donors and dye-labeled protein acceptors,”Journal of the American Chemical Society, Vol. 126, 2004, pp. 301–310.

References

329

[217] Clapp, A.R., Medintz, I.L., Uyeda, H.T., Fisher, B.R., Goldman, E.R., Bawendi, M.G., and Mattoussi,H., “Quantum dot-based multiplexed fluorescence resonance energy transfer,” Journal of the Ameri-can Chemical Society, Vol. 127, 2005, pp. 18212–18221.

[218] Clapp, A.R., Medintz, I.L., and Mattoussi, H., “Forster resonance energy transfer investigationsusing quantum-dot fluorophores,” Chemical and Physical Chemistry, Vol. 7, 2006, pp. 47–57.

[219] Algar, W.R., and Krull, U.J., “Quantum dots as donors in fluorescence resonance energy transfer forthe bioanalysis of nucleic acids, proteins, and other biological molecules,” Analytical andBioanalytical Chemistry, Vol. 391, 2008, pp. 1609–1618.

[220] Yun, C.S., Javier, A., Jennings, T., Fisher, M., Hira, S., Peterson, S., Hopkins, B., Reich, N.O., andStrouse, G.F., “Nanometal surface energy transfer in optical rulers, breaking the FRET barrier,” Jour-nal of the American Chemical Society, Vol. 127, 2005, pp. 3115–3119.

[221] Pons, T., Medintz, I.L., Sapsford, K.E., Higashiya, S., Grimes, A.F., English, D.S., and Mattoussi, H.,“On the quenching of semiconductor quantum dot photoluminescence by proximal goldnanoparticles,” Nano Letters, Vol. 7, 2007, pp. 3157–3164.

[222] Sen, T., Haldar, K.K. and Patra, A., “Au nanoparticles-based surface energy transfer probe forconformational changes of BSA protein,” Journal of Physical Chemistry C, Vol. 112, 2008,pp. 17945–17951.

[223] Kong, J., and Yu, S., “Fourier transform infrared spectroscopic analysis of protein secondary struc-tures,” Acta Biochimica et Biophysica Sinica, Vol. 38, 2007, pp. 549–559.

[224] Krueger, A., Stegk, J., Liang, Y., Lu, L., and Jarre, G., “Biotinylated nanodiamonds: simple and effi-cient functionalization of detonated diamond,” Langmuir, Vol. 24, 2008, pp. 4200–4204.

[225] Nath, S., Kaittanis, C., Tinkham, A., and Perez, J.M., “Dextran-coated gold nanoparticles for theassessment of antimicrobial susceptibility,” Analytical Chemistry, Vol. 80, 2008, pp. 1033–1038.

[226] Cavanagh, J., Fairbrother, W.J., Palmer III, A.G., and Skelton, N.J., (eds.), Protein NMR Spectroscopy,San Diego, CA: Academic Press, 1996.

[227] Endres, P.J., Paunesku, T., Vogt, S., Meade, T.J., and Woloschak, G.E., “DNA-TiO2 nanoconjugateslabeled with magnetic resonance contrast agents,” Journal of the American Chemical Society, Vol.129, 2007, pp. 15760–15761.

[228] Paunesku, T., Ke, T., Mascheri, N., Wu, A., Lai, B., Vogt, S., Maser, J., Thurn, K., Szolc-Kowalska, B.,Larson, A., Bergan, R.C., Omary, R., Li, D., Lu, Z.-R., and Woloschak, G.E., “Gadolinium-conju-gated TiO2-DNA oligonucleotide nanoconjuagtes show prolonged intracellular retention periodand T1-weighted contrast enhancement in magnetic resonance images,” Nanomedicine:Nanotechnology, Biology, and Medicine, Vol. 4, 2008, pp. 201–207.

[229] Garcia-Fuentes, M., Torres, D., Martin-Pastor, M., and Alonso, M.J., “Application of NMR spectros-copy to the characterization of PEG-stabilized lipid nanoparticles,” Langmuir, Vol. 20, 2004, pp.8839–8845.

[230] Zheng, M., Li, Z., and Huang, X., “Ethylene glycol monolayer protected nanoparticles: synthesis,characterization, and interactions with biological molecules,” Langmuir, Vol. 20, 2004,pp. 4226–4235.

[231] Cheng, Y., Li, Y., Wu, Q., and Xu, T., “New insights into the interactions between dendrimers andsurfactants by two dimensional NOE NMR spectroscopy,” Journal of Physical Chemistry B, Vol. 112,2008, pp. 12674–12680.

[232] Lundqvist, M., Sethson, I., and Jonsson, B.-H., “High-resolution 2D 1H-15N NMR characterization ofpersistent structural alternations of proteins induced by interactions with silica nanoparticles,”Langmuir, Vol. 21, 2005, pp. 5974–5979.

[233] Chapman, J.R., (ed.), Mass Spectrometry of Proteins and Peptides, Mahweh, NJ: Humana Press, 2000.[234] Barnhill, H.N., Reuther, R., Ferguson, P.L., Dreher, T., and Wang, Q., “Turnip yellow mosaic virus

as a chemoaddressable bionanoparticle,” Bioconjugate Chemistry, Vol. 18, 2007, pp. 852–859.[235] Barnhill, H.N., Claudel-Gillet, S., Ziessel, R., Charbonniere, L.J., and Wang, Q., “Prototype protein

assembly as scaffold for time-resolved fluoroimmuno assays,” Journal of the American Chemical Soci-ety, Vol. 129, 2007, pp. 7799–7806.

[236] Bruckman, M.A., Kaur, G., Lee, L.A., Xie, F., Sepulveda, J., Breitenkamp, R., Zhang, X., Joralemon,M., Russell, T.P., Emrick, T., and Wang, Q., “Surface modification of tobacco mosaic virus withclick chemistry,” ChemBioChem, Vol. 9, 2008, pp. 519–523.

[237] Hu, F., Neoh, K. G., Cen, L., and Kang, E.T., “Cellular response to magnetic nanoparticlesPEGylated via surface-initiated atom transfer radical polymerization,” Biomacromolecules, Vol. 7,2006, pp. 809–816.

[238] Shon, Y.S., Choi, D., Dare, J., and Dinh, T., “Synthesis of nanoparticle-core dendrimers by conver-gent dendritic functionalization of monolayer-protected nanoparticles,” Langmuir, Vol. 24, 2008,pp. 6924–6931.

[239] Gibson, J.D., Khanal, B.P., and Zubarev, E.R., “Paclitaxel-functionalized gold nanoparticles,” Jour-nal of the American Chemical Society, Vol. 129, 2007, pp. 11653–11661.


330

[240] Takae, S., Akiyama, Y., Otsuka, H., Nakamura, T., Nagasaki, Y., and Kataoka, K., “Ligand densityeffect on biorecognition by PEGylated gold nanoparticles: regulated interaction of RCA120 lectinwith lactose installed to the distal end of tethered PEG strands on gold surface,” Biomacromolecules,Vol. 6, 2005, pp. 818–824.

[241] Swami, S., Kumar, A., and Sastry, M., “Formation of water-dispersible gold nanoparticles using atechnique based on surface-bound interdigitated bilayers,” Langmuir, Vol. 19, 2003,pp. 1168–1172.

[242] Liu, J., Gong, T., Wang, C., Zhong, Z., and Zhang, Z., “Solid lipid nanoparticles loaded with insulinby sodium cholate-phosphatidylcholine-based mixed micelles: preparation and characterization,”International Journal of Pharmaceutics, Vol. 340, 2007, pp. 153–162.

[243] Feng, S.S., Mu, L., Win, K.Y., and Huang, G., “Nanoparticles of biodegradable polymers for clinicaladministration of paclitaxel,” Current Medicinal Chemistry, Vol. 11, 2004, pp. 413–424.

[244] Yuan, W., Wu, F., Geng, Y., Xu, S., and Jin, T., “An effective approach to prepare uniform pro-tein-Zn2+ nanoparticles under mild conditions,” Nanotechnology, Vol. 18, 2007, Article# 145601.

[245] Duhr, S., and Braun, D., “Why molecules move along a temperature gradient,” Proceedings of theNational Academy of Science, Vol. 103, 2006, pp. 19678–19682.

[246] Piazza, R., and Parola, A, “Thermophoresis in colloidal suspensions,” Journal of Physics CondensedMatter, Vol. 20, 2008, Article# 153102.

[247] http://www.nanocomposix.com/services.htm.

References

331

About the Editors

Kaushal Rege, Ph.D. is an assistant professor of chemical engineering at ArizonaState University in Tempe, AZ. He received his Ph.D. in chemical engineering fromRensselaer Polytechnic Institute in Troy, NY and did his postdoctoral research at theCenter for Engineering in Medicine at Massachusetts General Hospital and HarvardMedical School in Boston, MA. Dr. Rege works in the areas of cancer nanotechnology,synergistic cancer therapeutics, and molecular engineering.

Igor L. Medintz, Ph.D. is a research biologist in the Center for Bio/Molecular Scienceand Engineering at the U.S. Naval Research Laboratory in Washington D.C. He receivedhis Ph.D. in molecular, cellular, and developmental biology from the Graduate Schooland University Center of the City University of New York in 1999. Dr. Medintz’sresearch interests lie in the development of methods to bridge the inorganic/organicmolecular interface in the pursuit of nanosensors and other active nanomaterials.

333

List of Contributors

334

List of ContributorsFrank AlexisMIT-Harvard Center for Cancer Nanotechnology Excellenceand Harvard-MIT Division of Health Sciences and Technol-ogy Cambridge, MA 02139Labortatory of Nanomedicine and Biomaterials, Depart-ments of Anesthesiology, Brigham and Women’s Hospitaland Harvard Medical School, Boston, MA 02115

Ardalan ArdeshiriDepartment of Biomedical EngineeringOregon Health and Science UniversityPortland, OR 97239

Prashanth AsuriDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180

Shyam Sundhar BaleDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180

Akhilesh BanerjeeDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180

Rohan BhavaneThe Division of NanomedicineDepartment of Biomedical EngineeringThe University of Texas Health Science Center at HoustonHouston, TX 77030

Michael R. CaplanHarrington Department of BioengineeringCenter for Interventional BiomaterialsArizona State UniversityPO Box 879709Tempe, AZ 85287-9709E-mail: [email protected]

Jeffrey J. ChalmersDepartment of Chemical and Biomolecular EngineeringDirector, University Cell Analysis and Sorting CoreThe Ohio State University125 Koffolt Laboratories140 West 19th AvenueColumbus, OH 43210Telephone: (216) 292-2727Fax: (216) 292-3769E-mail: [email protected]

Ciro ChiappiniDepartment of Biomedical EngineeringThe University of TexasAustin, TX77030

Aaron R. ClappDepartment of Chemical and Biological EngineeringIowa State UniversityAmes, IA 50014E-mail: [email protected]

Benita J. DairDivision of Chemistry and Materials ScienceOffice of Science and EngineeringCenter for Devices and Radiological HealthU.S. Food and Drug Administration10903 New Hampshire AvenueSilver Spring, MD 20993

Jonathan S. DordickDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180E-mail: [email protected]

Omid C. FarokhzadAssistant Professor of AnesthesiologyHarvard Medical SchoolDepartment of AnesthesiologyBrigham and Women’s Hospital75 Francis StreetBoston, MA 02115E-mail: [email protected]

Mauro FerrariThe Division of Nanomedicine, Department ofBiomedical Engineering, The University of Texas HealthScience Center at Houston, Houston, TX 77030Department of Biomedical Engineering, The Universityof Texas, Austin, TX 77030Department of Experimental Therapeutics, The Univer-sity of Texas MD Anderson Cancer Center, Houston, TX77030Department of Bioengineering, Rice University,Houston, TX 77005E-mail: [email protected]

Katye, M. FichterDepartment of Biomedical EngineeringOregon Health and Science UniversityPortland, OR 97239

André M GobinAssistant Professor - BioengineeringUniversity of LouisvilleLouisville, KY 40292E-mail: [email protected]

Tapan K. JainDepartment of Biomedical EngineeringLerner Research InstituteCleveland Clinic, Cleveland, OH 44195

Farouc A. JafferCardiovascular Research Center, Cardiology DivisionHarvard Medical School and Massachusetts GeneralHospital149 13th St., 4th FloorCharlestown, MA 02129

Amit JoshiDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th Street,Troy, NY 12180

List of Contributors

335

Ravi S. KaneDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180E-mail: [email protected]

Amit A. KaleCenter for Pharmaceutical Biotechnology andNanomedicineNortheastern University312 Mugar Hall360 Huntington AvenueBoston, MA 02125

Vinod LabhasetwarDepartment of Biomedical EngineeringLerner Research Institute9500 Euclid AvenueCleveland Clinic, Cleveland, OH 44195E-mail: [email protected]

Robert S. LangerDepartment of Chemical Engineering, MassachusettsInstitute of Technology, Cambridge, MA 02139MIT-Harvard Center for Cancer Nanotechnology ExcellenceCambridge, MA 02139Harvard-MIT Division of Health Sciences and TechnologyCambridge, MA, 02139

Jonathan MartinezThe Division of NanomedicineDepartment of Biomedical EngineeringThe University of Texas Health Science Center at HoustonHouston, TX 77030

Hedi MattoussiU.S. Naval Research LaboratoryCenter for Bio/Molecular Science and EngineeringCode 69004555 Overlook Avenue, SWWashington, D.C. 20375

Jason R. McCarthyCenter for Molecular Imaging ResearchHarvard Medical School and MassachusettsGeneral Hospital149 13th St., Rm 5406Charlestown, MA 02129, USAE-mail: [email protected] L. MedintzU.S. Naval Research LaboratoryOptical Sciences Division, Code 56114555 Overlook Avenue, SWWashington, D.C. 20375

Prabhas V. MogheDepartment of Chemical and Biochemical EngineeringDepartment of Biomedical EngineeringRutgers UniversityPiscataway, NJ 08854E-mail: [email protected]

Rajesh R. NaikNanostructured and Biological Materials BranchMaterials and Manufacturing DirectorateAir Force Research LaboratoryWright-Patterson AFB, OH 45433-7750E-mail: [email protected]

Dominik J. NaczynskiDepartment of Chemical and Biochemical EngineeringRutgers UniversityPiscataway, NJ 08854

Ravindra C. PanguleDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180

Emily PawelskDepartment of Biomedical EngineeringRutgers UniversityPiscataway, NJ 08854

Eric M. PridgenDepartment of Chemical EngineeringMIT-Harvard Center for Cancer NanotechnologyExcellenceMassachusetts Institute of TechnologyCambridge, MA 02139

Elena V. RoscaHarrington Department of BioengineeringCenter for Interventional BiomaterialsArizona State UniversityPO Box 879709Tempe, AZ 85287-9709

María Pía RossiNew Jersey Center for BiomaterialsDepartment of Chemical and Biochemical EngineeringRutgers UniversityPiscataway, NJ 08854

Kim E. SapsfordDivision of BiologyOffice of Science and EngineeringCenter for Devices and Radiological HealthU.S. Food and Drug Administration10903 New Hampshire AvenueSilver Spring, MD 20993, U.S.A.E-mail: [email protected]

Dhiral A. ShahDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th StreetTroy, NY 12180

Mei ShaoDepartment of Chemical and BiomolecularEngineeringThe Ohio State University125 Koffolt Laboratories140 West 19th AvenueColumbus, OH 43210

Ram I. SharmaDepartment of Chemical and Biochemical EngineeringRutgers UniversityPiscataway, NJ 08854

Joseph M. SlocikNanostructured and Biological Materials BranchMaterials and Manufacturing DirectorateAir Force Research LaboratoryWright-Patterson AFB, OH 45433-7750

About the Editors

336

Andrew StineDepartment of Biomedical EngineeringLerner Research InstituteCleveland Clinic, Cleveland, OH 44195

Ennio TasciottiThe Division of NanomedicineDepartment of Biomedical EngineeringThe University of Texas Health Science Center at HoustonHouston, TX 77030

Xiaodong TongBiotechnology Institute and Department of Bioproductsand Biosystems EngineeringUniversity of MinnesotaSt Paul, MN 55108

Vladimir P. TorchilinCenter for Pharmaceutical Biotechnologyand NanomedicineNortheastern University312 Mugar Hall360 Huntington AvenueBoston, MA 02125E-mail: [email protected]

Katherine TynerDivision of Applied Pharmacology ResearchOffice of Testing and ResearchOffice of Pharmaceutical ScienceCenter for Drug Evaluation and ResearchU.S. Food and Drug Administration10903 New Hampshire AvenueSilver Spring, MD 20993

David VanceDepartment of Chemical and Biological EngineeringRensselaer Polytechnic Institute110 8th Street,Troy, NY 12180

Tania, Q. VuDepartment of Biomedical EngineeringOregon Health and Science UniversityPortland, OR 97239E-mail: [email protected]

Ralph WeisslederCenter for Molecular Imaging ResearchHarvard Medical School and Massachusetts GeneralHospital149 13th St., Rm 5406Charlestown, MA 02129

Susan WesterfieldDepartment of Biomedical EngineeringLerner Research InstituteCleveland Clinic, Cleveland, OH 44195

Ping WangBiotechnology Institute and Department of Bioproductsand Biosystems EngineeringUniversity of MinnesotaSt Paul, MN 55108E-mail: [email protected]

Songtao WuBiotechnology Institute and Department of Bioproductsand Biosystems EngineeringUniversity of MinnesotaSt Paul, MN 55108

Ying XiongDepartment of Chemical and BiomolecularEngineeringThe Ohio State University125 Koffolt Laboratories140 West 19th AvenueColumbus, OH 43210

Maciej ZborowskiDepartment of Biomedical EngineeringCleveland Clinic9500 Euclid AvenueCleveland, OH 44195

Index

A

Acetal linkage, 172Active targeting, 228–30

ability evaluation, 231challenge, 228–29concepts, 229ligands for, 228–30, 231

Albumin nanoparticles (ANPs), 85–103benefits of, 88cell attachment assay, 95cell culture, 94creation, 93displaying ligand on, 96enhanced cell migration, 95–97enhanced ECM assembly, 97–99experimental design, 88fabrication, 89–91fibroblast extracellular matrix assembly,

94–95functionalization, 89, 91–93introduction to, 86–88keratinocyte morphology and migration,

94materials, 88–89methods, 89–95microscale plasma initiated patterning, 89pattern creation, 93pitfalls, 100–102preparations, 90results, 95–99summary points, 102–3three-dimensional presentation, 101–2unfunctionalized, 97

Analytical ultracentrifugation (AUC), 300Anodic etch, 248–50

defined, 248–50guidelines, 250nitrogen absorption/desorption

characterization, 251parameters, 250SEM characterization, 251tank, 246tank assembly, 249

See also Porous silicon particles (PSPs)Antibodies

binding, 112concentration evaluation, 163epitopes, 33–34quantification on nanoshells, 160–61

Atherosclerosis, 138Atomic force microscopy (AFM), 35, 308–9

cantilever, 309defined, 308information provided by, 308with QD labeling, 309sensitivity/versatility, 309See also Microscopy techniques

Au-phage networks, 243Avidin-biotin affinity chromatography,

184, 188

B

Bacteria, 116Bacterial magnetic particles -PEI (BMP-PEI),

241Batch magnetic separators, 110Bicinchroninic acid (BCA), 8Binding enhancement factor, 279Biocatalysts, 48Biomolecule conjugation, 61–65

dye-labeled, 64protocol, 64–65

Boltzmann constant, 48Buffers, 28Butyl vinyl ether (BVE), 173

C

Capillary electrophoresis (CE), 298, 299–300application, 299defined, 299species detection, 299

Carbon nanotubes (CNTs), 2acid oxidation of, 19adsorption of proteins onto, 2biofunctionalization of, 18

337

Carbon nanotubes (CNTs) (continued)covalent attachment of proteins onto,

5–7covalently attached protein conjugates,

characterization, 13–18functionalized, 2multiwalled (MWNTs), 3physical adsorption proteins, 3–4physical adsorption proteins,

characterization, 7–11protein assisted solubilization, 4–5protein-assisted solubilization,

characterization, 11–13single-walled (SWNTs), 2solubilizing, 18uniform dispersion of, 18See also CNT-protein conjugates

Carotid atheroma, 147Cell attachment assay, 95Cell-penetrating peptide (CPP), 169Cells

antibody-conjugate binding, 112binding and uptake studies, 214–15double diffusion, 126patterning, with human fibroblasts, 100variability, 114

CellSearch system, 108Chromatography, 296–97

high performance liquid (HPLC), 296hydrodynamic, 297

Circular dichroism (CD), 1, 313–14defined, 313protein structure determination with,

14–15spectra changes, 313use of, 35vibrational (VCD), 314

Cis-aconityl linkage, 171CNT-protein conjugates, 1–21

anticipated results, 7–18application notes, 19–20data acquisition, 7–18discussion and commentary, 18–19interpretation of data, 7–18introduction to, 2–3materials, 3methods, 3–7summary points, 21troubleshooting table, 19

Coil-coil peptide-NP assembly, 28–31disassembly, 30–31gold NP, 28–29gold-QD heterostructures, 29–30

Colocation analysis, 220

COMSOL, 283Confocal microscopy, 266Core-shell QDs, 57Covalently attached CNT-protein conjugates,

5–7characterization, 13–18characterization with tryptophan

fluorescence, 15–16determination with CD spectroscopy,

14–15Hammett analysis, 13–14operational and storage stability, 17–18thermostabilization, 17

Critical micelle concentration (CMC), 133Crosslinked iron oxide nanoparticles (CLIO),

137concentration, 148purification, 148synthesis, 141See also Theranostic nanoparticles

Cytospin, 119Cytotoxicity studies, 215–16, 222

D

Differential scanning calorimetry (DSC), 319Dihydrolipoic acid (DHLA), 59, 60

capped QDs, 61preparation, 60thiols and, 62

Double diffusion cells, 126Double emulsion method, 207, 209Doxorubicin (DOX), 124Drug encapsulation, 211–12

efficiency, 211of hydrophilic drugs, 225of hydrophobic drugs, 225physiochemical properties, 225

Drug-loaded MNPs, 124anticipated results, 132–33antiproliferative activity, 131–32application notes, 134characterization, 129data acquisition, 132–33discussion and commentary, 133–34DOX*HCI conversion, 129experimental design, 124–26facilities and equipment, 127interpretation, 132–33kinetics of DOX release, 130materials, 126–27methods, 128–32multiple drugs, 124reagents, 126

Index

338

summary points, 134synthesis, 128–32synthesis schematic, 125troubleshooting table, 134See also Magnetic nanoparticles (MNPs)

Drug loading, 129–30, 212determining, 130protocol for quantification, 212

Drug release, 226–28control, 226, 231parameters, 226stimuli, 227

Drug release studies, 212–13drug mass and, 213with high-solubility drug, 213with low-solubility drug, 212–13See also Polymeric nanoparticle delivery

systemsDry etch, 248Dynamic light scattering (DLS), 35, 91, 301–2

defined, 301drawbacks, 301–2illustrated, 302measurements, 301

E

Electronic cell counting system, 114Electrophoresis, 298–300

capillary (CE), 299–300slab gel, 298–99types of, 298

Electrophoretic mobility, 303Electrospray ionization (ESI), 317Electrospun scaffolds, 102Enhanced fluorescent proteins (EFPs), 76Enhanced permeability and retention (EPR),

169Enrichment process, 116–17

flow, 118illustrated, 117performance, 118

Enzyme-attached polystyrene nanoparticles,41

Enzymesattachment, porous silica coating for, 42in biocatalyst preparation, 49entrapment of, 41–42immobilized, efficiency, 40loading and activity assay, 42–44as proteins, 40

Ethylene glycol vinyl ether (EGVE), 173Extracellular matrix (ECM), 86

fibroblast assembly, 94–95ligands, 86, 87proteins, 86

F

Ferrite oxide particles, 240Fibroblasts

cell patterning with, 100ECM assembly, 94–95use of, 94

Field flow fractionation (FFF), 298Flow cytometry, 215

anticipated results, 263–64data acquisition, 263–64illustrated, 263interpretation, 263–64materials, 262methods, 262See also Porous silicon particles (PSPs)

Fluorescein, 76Fluorescence correlation spectroscopy (FCS),

302–3Fluorescence energy transfer. See Förster

resonance energy transfer (FRET)Fluorescence measurements, 65–66

continuous excitation, 65spectra acquisition protocol, 65–66time-resolved mode, 65

Fluorescence microscopy, 214–15, 312Fluorescent dye conjugation, 254Fluorescent spectroscopy, 314–15Focal macrophage ablation, 150Förster resonance energy transfer (FRET),

54, 312, 315–16defined, 315efficiency, 67, 68, 71measurements, 55as nonradiative process, 315phenomenon, 316See also QD-based FRET

Forward scatter (FSC), 264Fourier transform infrared (FT-IR), 1, 10–11

G

Gel electrophoresis, 298–99Gel permeation chromatography (GPC), 204,

216Gold nanoparticle assembly, 28–29Gold-quantum dot heterostructures, 29–30,

34Gold/silica core nanoshells, 157–58

histological analysis, 161–62OCT image analysis, 161results, 161–63survival following imaging/therapy, 162

H

Hammett analysis, 13–14

Index

339

High performance liquid chromatography(HPLC), 296

Hydrazide-activated phospholipids, 176Hydrazone-based mPEG-HZ-PE conjugates

synthesis, 176–84aliphatic aldehyde-derived, 176–77application, 185–86aromatic aldehyde-derived, 177–80aromatic ketone-derived, 180–83half-lives, 187pH sensitivity, 188success, 185time required, 187

Hydrazone linkages, 173hydrolytic kinetics, 192hydrolytic stability, 192See also PH-sensitive linkages

Hydrodynamic chromatography (HDC), 297Hydrogels, 46

I

Immunomagnetic labeling, 117Inductively coupled plasma-atomic emission

spectroscopy (ICP-AES), 257–60data acquisition, 258–60data analysis, 260graphs, 261materials, 257–58methods, 258normalized values from analysis, 261See also Porous silicon particles (PSPs)

Infrared (IR) spectroscopy, 316Integrins, 86Intravital fluorescence microscopy (IVFM),

139, 143–44, 148, 149In vitro cell culture study, 184, 188Isothermal titration calorimetry (ITC), 319

K

Keratinocytesincubation of, 96migration, 94morphology, 94

L

Ligand conjugation, 95, 228–30active targeting and, 228–30approaches, 229covalent, 229

Ligandsfor active targeting, 228–30, 231densities, 98displaying on ANPs, 96

ECM, 87nanoscale presentation effect on, 98presentation on cytoskeletal

organization, 99Light-based therapy, 144–45Light microscopy, 312Light scattering techniques, 208Loading NPs into PSPs, 264, 267Lower critical solution temperature (LCST),

227Low pressure chemical vapor deposition

(LPCVD), 245Lyophilization, 213

M

Macrophages, 138Magnetic cell separation, 107–20

in bacteria, 116batch, 110CellSearch system, 108data acquisition, 117–19delivery to breast cancer cell-line, 125discussion and commentary, 120earliest reports, 108enrichment process, 116–17examples, 115–16immunomagnetic labeling, 117interpretation, 117–19introduction to, 108–16materials and methods, 116–17partial flow-through, 110principle, 110–15quantification of performance, 114–15in rare cancer cell detection, 115–16red cell lysis step, 117results, 117–19in stem cell isolation, 115step, 117summary points, 120in T cell depletion, 115

Magnetic field (MF), 123, 131–32Magnetic forces, 110–11

on cells without labeling, 111on labeled cells, 111

Magnetic nanoparticles (MNPs), 42, 123–34characterization, 124, 129conjugated antibodies, 109defined, 123diffusion, 123dispersion, 133dividing, 128DOX*HCI conversion, 129drug-loaded, 123–34for enzyme attachment, 42, 46–47enzyme immobilization, 46–47

Index

340

interaction between, 111–14kinetics of DOX release, 130outer layer, 124results, 46–47static charge, 133synthesis, 124, 128, 133targeting of, 124, 133TEM image, 47yield, 132See also Nanoparticle-enzyme hybrids;

NanoparticlesMagnetic resonance imaging (MRI), 124, 284Magnetotgactic bacteria, 116Maltose binding protein (MBP), 63Mass spectroscopy (MS), 317–18

defined, 317inductively-coupled plasma (ICP), 318

Mathematical models, 275–90best fitting parameters, 288as hypothesis generators, 276introduction to, 276–77molecular/cellular scale, 276–82organism scale, 275, 285–87statistical guidelines, 287–89summary points, 289–90tissue scale, 277, 282–85troubleshooting table, 289validation and application, 287–89

MATLAB, 278Matrix-assisted laser desorption/ionization

(MALDI), 317Maximum tolerated dose (MTD), 230, 231Membrane molecular weight cutoff

(MWCO), 298Mercaptoundecanoic acid (MUA), 59Metal ion-peptide recognition, 32–33Metal nanoshells. See NanoshellsMicrobots, 241Microscale plasma initiated patterning,

89, 93schematic, 101spatial guidance, 100–101

Microscopy techniques, 308–12atomic force microscopy (AFM), 35, 308–9fluorescence microscopy, 312illustrated, 310light microscopy, 312scanning electron microscopy (SEM), 49,

91, 218, 251, 311–12transmission electron microscopy (TEM),

29, 35, 49, 218, 300, 308, 309–11See also NP-bioconjugates

Molecular/cellular scale modeling, 276,277–82

anticipated results, 280

data acquisition, 280discussion and commentary, 280–82interpretation, 280methods, 277–80model description, 277summary points, 289See also Mathematical models

Monocrystalline iron oxide nanoparticles(MION), 138

MPEG-HZ-PE conjugates, 176–77aromatic aldehyde-derived

hydrazone-based, 177–80aromatic ketone-derived

hydrazone-based, 180–83in vitro pH-dependent degradation of,

184Multidrug resistance (MDR), 228Multifunctional peptides, 31–32Multiphysics, 283Multistage delivery system (MDS), 237

classes, 241–42defined, 239gold/bacteriophage nanoparticle

network, 242PSPs in, 237–71schematic, 240silicon-based, 241success, 242–43transport of therapeutic agents, 242versatility and ease of modification, 268

Multiwalled carbon nanotubes (MWNTs), 3acid-treated, 20long oxidized, 17operational and storage stability, 18sonication, 6

Murine monoclonal antibodies (MoAb), 171MWNT-DNAzyme conjugates, 20

N

Nanocrystals, 72Nanomaterials, 2Nanoparticle-enzyme hybrids, 39–49

application notes, 49discussion and commentary, 47–49enzyme-attached polystyrene

nanoparticles, 41, 44–45enzyme loading, 42–44fluorescent quantum dot, 77introduction to, 40ligand-conjugated, 87magnetic nanoparticles, 42, 46–47materials, 40methods, 41–44polyacrylamide hydrogel nanoparticles,

41–42, 45–46

Index

341

Nanoparticle-enzyme hybrids (continued)results, 44–47summary points, 49troubleshooting, 49

Nanoparticle formation, 201, 207–9double emulsion protocol, 209nanoprecipitation protocol, 208single emulsion protocol, 208–9See also Polymeric nanoparticle delivery

systemsNanoparticles, 27

albumin, 85–103biocatalysts, 48biochemical cues, 102components, 295enzyme-attached polystyrene, 41extravasation into malignant tissue, 198fluorescent, 220–22gold assembly, 28–29magnetic, 42, 123–34palladium decorated gold, 32physiochemical metrics, 295physiochemical properties, 197polyacrylamide hydrogel, 41–42polymeric, 197–231production, 156shape, 224size, 222–24surface chemistry, 224–25synthesis precursors, 27theranostic, 137–50toxicity, 102unmodified/native characterization, 293See also NP-bioconjugates

Nanoprecipitation, 207protocol for, 208requirement, 207

Nanoshellsaccumulation in tissue, 159biomedical applications of, 154–55cell culture, 157for combined imaging and therapy

in vivo, 158–59for combined optical contrast/therapeutic

application, 155conjugation of biomolecules to, 160defined, 153experimental design, 156gold/silica core, 157–58, 161–63introduction to, 154–55materials, 156–57mediated cancer therapy, 155metal, 153–66methods, 157–61nanoparticle production, 156

OCT scanning, 158passivation with PEG, 159pitfalls, 163–65protein conjugation to surface, 156–57quantification of antibodies on, 160–61results, 161–63statistical analysis, 165–66therapeutic laser irradiation, 159troubleshooting table, 166in vitro assays, 157in vivo model, 158

Nanotechnology, 294Nanotechnology Characterization

Laboratory (NCL), 295Near infrared light-activated therapeutic

(NILAT) agents, 139, 142Near infrared (NIR) light

gold nanoparticles and, 163heating and, 155in imaging large body sections, 165scattering, 161

Near infrared (NIR) resonant compositenanoparticles. See Nanoshells

NP-bioconjugates, 293–320architecture, 295characterization of, 293introduction to, 294–96mass spectroscopy (MS), 317–18methods, 296–319microscopy techniques, 308–12physiochemical metrics, 295potential, schematic, 294scattering techniques, 300–308separating unconjugated biomolecules

from, 297separation-based techniques, 296–300spectroscopic techniques, 312–17summary points, 319–20thermal techniques, 318–19

Nuclear magnetic resonance (NMR), 49, 204,216, 316–17

defined, 316environmentally sensitive peak shifts, 317as nondestructive, 317

O

OCT images, 158analysis, 161intensity quantification, 166quantification of, 163representative, 162scanning, 158statistical analysis, 165–66

Oleic acid (OA) coating, 123Optical density (OD), 145

Index

342

Organic fluorescence dyes, 76Organism scale modeling, 276, 285–87

anticipated results, 286–87data acquisition, 286–87discussion and commentary, 287interpretation, 286–87methods, 285–86summary points, 290tumor-specific targeting, 286See also Mathematical models

P

Paclitaxel (Taxol), 230Partial flow-through separators, 110Particle size analysis, 35PEG, 239

bifunctional polymer, 160chains, 230linker, 154passivation of nanoshells with, 159

PE-PEG1000-TATp conjugate synthesis, 183–84,186

Peptide-nanoparticle assemblies, 25–36antibody epitopes, 33–34anticipated results, 34–35application notes, 36coil-coil mediated assembly, 28–31data acquisition, 34–35discussion and commentary, 35interpretation, 34–35introduction to, 26–27materials, 27–28mediated by metal ion-peptide

recognition, 32–33methods, 28–32summary points, 36synthesis of hybrid structures, 31–32troubleshooting table, 36

Peptides, 27as antibody epitopes for nanoparticle

assembly, 33–34coil-coil NP assembly, 28–31enzymatically degradable cross-linking

peptides, 269multifunctional, 31–32RGD, 86

Photobleaching, 76Photolitography, 247–48Photon correlation spectroscopy (PCS).

See Dynamic light scattering (DLS)PH-sensitive linkages, 169–92

acetal, 172approaches, 170avidin-biotin affinity chromatography,

184, 188

chemicals, 174–75cis-aconityl, 171conclusion, 191discussion and commentary, 185–91in drug release, 171hydrazone, 173hydrazone-based mPEG-HZ-PE conjugates

synthesis, 176–84, 183–86introduction to, 170–74materials, 174–75methods, 176–85in new function appearance, 171PEG-TATp-liposome-pGFP complexes, 190PE-PEG1000-TATp conjugate synthesis,

183–84, 186pGFP complexed liposomal formulations,

175polyketal, 172poly(ortho-esters), 173in protective “coat” removal, 171rhodamine-labeled liposomal

formulations, 175summary points, 192syntheses, 175TATp-bearing, 175thiopropionates, 173–74trityl, 172troubleshooting table, 192vinyl ether, 172–73in vitro cell culture study, 184, 188in vitro pH-dependent degradation of

PEG-HZ-PE conjugates, 184, 186–88in vivo study, 185, 188–89in vivo transfection with pGFP, 185,

189–91Physical adsorption of proteins, 3–4

characterization of, 7–11determination, 9–10determination with FT-IR, 10–11harsh conditions, 10loading by BCA assay, 8retention of activity, 8–10

PLA-PEG copolymers, 203conjugation efficiency, 217synthesis of, 204–7

PLGA-PEG copolymers, 203conjugation efficiency, 217synthesis of, 204–7

Pluronic coating, 133Polyacrylamide hydrogel nanoparticles,

41–42entrapped enzymes, 45–46hydrogels, 46results, 45–46

Polyethylene oxide (PEO), 86

Index

343

Polyketal linkage, 172Polymer characterization, 216–17Polymeric nanoparticle delivery systems,

197–231active targeting, 228–30application notes, 230–31cell binding and update experiments,

202–3cell binding and uptake studies, 214–15components, 199cytotoxicity studies, 203, 215–16data acquisition, 216–22design criteria, 199discussion and commentary, 222–30drug encapsulation, 211–12drug loading, 225–26drug release, 226–28drug release studies, 212–13illustrated components, 199interpretation, 216–22ligand conjugation, 201, 228–30materials, 200–203methods, 203–16nanoparticle characterization, 217–20nanoparticle formation, 201, 207–9overall procedure, 203particle shape, 224particle size, 222–23PLA-PEG and PLGA-PEG synthesis, 204–7polymer characterization, 216–17polymer synthesis, 200–201post-formulation treatment, 202, 213–14quantification of drug encapsulation,

201–2release experiments, 202results, 216–22summary points, 231surface chemistry, 224–25targeting ligand conjugation, 209–11troubleshooting table, 230in vitro experiments, 220–22

Poly(ortho-esters), 173Polystyrene-enzyme hybrid nanoparticles, 41

results, 44–45SEM image, 45synthetic route, 44See also nanoparticle-enzyme hybrids

Porous silicon particles (PSPs), 237–71anodic etch, 248–50characterization, 251chemo-physical properties, 239count and size analysis, 255–57defined, 239discussion and commentary, 267–71

dry etch, 248fabrication, 245–51fabrication steps, 268first-stage, 239flow cytometry for, 260–64fluorescent dye conjugation, 254homogeneity within, 243inductively coupled plasma-atomic

emission spectroscopy (ICP-AES),257–60

introduction to, 238–45in isopropyl alcohol (IPA), 252loading kinetics, 244loading of NPs into, 264, 267materials, 245–47methods, 247–51for multistage delivery, 237–71oxidation and surface modification with,

252–53photolitography, 247–48release kinetics, 244release of NPs from, 265SEM micrographs of, 244, 252surface modification, 243, 268–69surface modification with peptide

sequences, 269thin film deposition, 247troubleshooting table, 270–71zeta potential measurement, 254–55

Positron emission tomography (PET) contrastagents, 284

Post-formulation treatment, 202, 213–14Prostate specific membrane antigen

(PSMA), 154Proteins

covalent attachment of, 5–7, 13–18ECM, 86enzymes as, 40physical adsorption on carbon nanotubes,

3–4, 7–11solubilization of carbon nanotubes, 4–5,

11–13thermostabilization of, 17

Protein-solubilized CNTs, 4–5characterization of, 11–13Raman spectroscopy for, 12–13with UV-Vis spectroscopy, 11–12

Prototypical laser scanning fluorescencemicroscope, 141

Q

QD-based FRET, 53–72biomolecule conjugation, 61–65conclusions, 72

Index

344

data analysis and interpretation, 66–71donor-acceptor distances, 68–70fluorescence measurements, 65–66interaction with dye pairs and, 56materials, 56methods, 56–66organic dyes and, 63quantum dot synthesis, 56–58reaction rates of surface-bound

substrates, 70–71summary points, 72surface ligand exchange, 58–61

Quantum dot bioconjugates, 70core-shell, 55forming, 79treating cells with, 79

Quantum dotsautomated tracking program, 80best fit, 67in biological applications, 77BSA-modified, 300composite dye signal, 67conjugated, 301control spectrum, 66core-only, 59core-shell, 57, 59DHLA-capped, 61disperse dried, 60as fluorescent tags, 54for FRET-based applications, 53–72functionalization of, 29gold heterostructures, 29–30, 34materials, 296molecular dynamics, 81nanoparticles, 77photophysical properties, 303real-time dynamics, 80for stability, 72synthesis, 56–58trajectory, 80use limitations, 78

Quasi-elastic light scattering (QRLS).See Dynamic light scattering (DLS)

R

Raman spectroscopy, 12–13, 303–5Rare cancer cell depletion, 115–16Receptor-ligand modeling, 277Red cell lysis step, 117Releasing NPs from PSPs, 265, 267Resonance Raman (RR), 304RGD-functionalized gold nanodots, 86Rhodamine, 76Ring opening polymerization (ROP), 204

S

Scanning electron microscopy (SEM), 49, 91,218, 311–12

characterization, 251defined, 311limitations, 311–12See also microscopy techniques

Scattering techniques, 300–308defined, 300dynamic light scattering (DLS), 301–2fluorescence correlation spectroscopy

(FCS), 302–3illustrated, 304Raman spectroscopy, 303–5small angle X-ray scattering, 306–7X-ray diffraction (XRD), 305–6See also NP-bioconjugates

Scherrer equation, 305Separation-based techniques, 296–300

analytical ultracentrifugation (AUC), 300chromatography, 296–97electrophoresis, 298–300field flow fractionation (FFF), 298illustrated, 297types of, 296See also NP-bioconjugates

Side scatter (SSC), 264Signal intensities (SI), 146Single emulsion method, 208–9Single-walled carbon nanotubes (SWNTs),

2, 313aqueous dispersion of, 4disperse purified, 5functionalization of, 2protein adsorbed, 5purified HIPCO, 3SBP adsorbed onto, 10UV-Vis spectrum for, 11

Slab gel electrophoresis, 298–99Slipping plane, 303Small angle X-ray scattering, 306–7

defined, 306measurements, 306schematic, 307uses, 306–7

Spectral deconvolution algorithms, 72Spectroscopic techniques, 312–17

circular dichroism (CD), 313–14defined, 312fluorescent spectroscopy, 314–15Förster resonance energy transfer (FRET),

315–16illustrated, 315infrared (IR) spectroscopy, 316

Index

345

Spectroscopic techniques (continued)nuclear magnetic resonance (NMR),

316–17UV-visible spectroscopy, 312–13See also NP-bioconjugates

Squamous cell carcinoma of the head andneck (SCCHN), 116

Static light scattering (SLS), 308Stem cell isolation, 115Sucrose lyoprotection, 214Surface-enhanced Raman scattering (SERS),

304Surface-enhanced resonance Raman

scattering (SERRS), 304Surface modification (PSPs), 252–53Surface plasmon resonance (SPR), 313

T

Targeting ligand conjugation, 209–11chemistry selection, 209protocol via carbodimide chemistry,

210–11via maleimide-thiol chemistry, 211

Target-to-background ratio (TBR), 146, 149T cell depletion, 115Tetrameric antibody complex (TAC), 117Theranostic nanoparticles, 137–50

alternative reagents and equipment, 141animal experimentation, 146animal model, 141anticipated results, 148characterization of, 145–46data acquisition, 145–47experimental design, 139–40facilities/equipment, 140–41functionalization, 150intravital fluorescence microscopy,

143–44, 146–47introduction to, 138–39light-based therapy, 144–45materials, 140–41optimization for application, 150reagents, 140statistical analyses, 147summary points, 149–50synthesis of, 141–43, 148troubleshooting table, 149See also Nanoparticles

Therapeutic laser irradiation, 159Therapeutic (NILAT) agents for therapy, 137Thermal gravimetric analysis (TGA), 318Thermal techniques, 318–19

DSC, 319illustrated, 318ITC, 319

TGA, 318thermophoresis, 319See also NP-bioconjugates

Thermodiffusion, 319Thermophoresis, 319Thermostabilization, of proteins, 17Thin film deposition, 247Thiopropionates, 173–74Tissue scale modeling, 277, 282–85

anticipated results, 284data acquisition, 284discussion and commentary, 284–85interpretation, 284methods, 282–84summary points, 289uses, 282See also Mathematical models

Tracking single biomolecules, 75–82discussion and commentary, 81introduction to, 76–78materials, 78–79methods, 79troubleshooting table, 82

Transesterification activity, 43, 44Transmission electron microscopy (TEM), 29,

35, 49, 218, 300, 308, 309–11defined, 309limitations, 311at low accelerating voltages, 310uses, 310See also Microscopy techniques

Trityl linkage, 172Troubleshooting tables

CNT-protein conjugates, 19drug-loaded MNPs, 134mathematical models, 289nanoshells, 166peptide-nanoparticle assemblies, 36pH-sensitive linkages, 192polymeric nanoparticle delivery systems,

230porous silicon particles (PSPs), 270–71theranostic nanoparticles, 149tracking single biomolecules, 82

Tryptophan fluorescence, 15–16Two-step labeling, 111, 112

U

UV-visible spectroscopy, 35, 312–13

V

Vibrational CD (VCD), 314Vinyl ether linkage, 172–73Vitro assays, 157

Index

346

X

X-ray diffraction (XRD), 305–6

Z

Zeta potential, 295defined, 303determining, 303measurement, 254–55, 303uses, 303

Index

347

methods in bioengineering - muhammad1988adeel · 2012-02-15 · 11.2.3 preparation of the...

Documents