harnessing stem cells as drug delivery …vc043cd5324/deveza... · harnessing stem cells as drug...

HARNESSING STEM CELLS AS DRUG DELIVERY VEHICLES FOR

THERAPEUTIC ANGIOGENESIS:

A BIOMATERIALS-MEDIATED APPROACH

BY

Lorenzo R. Deveza

A DISSERTATION SUBMITTED TO THE DEPARTMENT OF BIOENGINEERING AND THE

COMMITTEE ON GRADUATE STUDIES OF STANFORD UNIVERSITY

IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY IN BIOENGINEERING

AT

STANFORD UNIVERSITY

August 2014

http://creativecommons.org/licenses/by-nc/3.0/us/

This dissertation is online at: http://purl.stanford.edu/vc043cd5324

© 2014 by Lorenzo De Rama Deveza. All Rights Reserved.

Re-distributed by Stanford University under license with the author.

This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.

ii



http://purl.stanford.edu/vc043cd5324

I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.

Fan Yang, Primary Adviser


Jennifer Cochran


Ngan Huang

Approved for the Stanford University Committee on Graduate Studies.

Patricia J. Gumport, Vice Provost for Graduate Education

This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.

iii

Harnessing Stem Cells as Drug Delivery Vehicles for Therapeutic Angiogenesis:

A Biomaterials-Mediated Approach

By

Lorenzo R. Deveza

Abstract

Cardiovascular disease (CVD) represents a global medical and economic problem with

high morbidity and mortality rates. CVD is often associated with partial occlusion of the blood

vessels and tissue ischemia, and restoring blood supply to ischemic tissues is critical to prevent

irreversible tissue damage. Therapeutic angiogenesis aims to stimulate the growth of new blood

vessels from pre-existing vessels, which offers a valuable tool for treating CVD. Several

strategies have been developed to promote angiogenesis, including growth factor delivery and

gene therapy. Direct delivery of angiogenic growth factors has the potential to stimulate new

blood vessel growth. However, it is limited by short half-lives in vivo, and uncontrolled

diffusion of angiogenic factors may also cause undesirable side effects. Gene therapy offers an

alternative approach by delivering genes encoding angiogenic factors, but previous approaches

often require the use of viral vectors for efficient gene delivery, and are limited by safety

concerns such as immunogenicity.

The goal of this thesis research is to develop novel strategies for stimulating therapeutic

angiogenesis by harnessing stem cells as drug delivery vehicles and to validate the efficacy of

non-viral engineered stem cells in vivo using mouse models of hindlimb ischemia. Our strategy

takes advantage of the natural homing capacity of stem cells towards ischemic tissues in vivo,

and their ability to secrete paracrine signals to stimulate blood vessel growth. Specifically we

developed two strategies including: (1) isolating and transfecting stem cells ex vivo using

biodegradable polymeric nanoparticles to overexpress therapeutic genes, followed by

transplanting non-viral engineered stem cells back to ischemic tissues; and (2) recruiting and

programming endogenous stem cells in situ using biomaterials-mediated delivery of biologics.

In the first strategy, we have chosen adipose-derived stem cells (ADSCs), an abundantly

available autologous cell source that can be easily obtained in a minimally invasive manner. To

enhance the paracrine signaling of ADSCs for therapeutic angiogenesis, we transfected ADSCs

using in-house developed biodegradable polymeric nanoparticles, which eliminate the

dependence on viruses for efficient gene delivery. Using the optimized polymeric vectors, we

examined the efficacy of ADSCs overexpressing various angiogenic factors or homing factors

iv

on therapeutic angiogenesis in vitro and in vivo. Transplantation of non-viral engineered ADSCs

led to significantly enhanced tissue salvage in a murine model of hindlimb ischemia with faster

restoration of blood reperfusion and muscle regeneration. Our results suggest that stem cells

programmed with biodegradable polymeric nanoparticles can serve as delivery vehicles to

express therapeutic factors in situ to promote therapeutic angiogenesis.

In the second strategy, we seek to circumvent the need of isolating and manipulating

stem cells ex vivo by directly recruiting and transfecting endogenous progenitor cells in situ at

the site of ischemia. To achieve this, we developed a biomaterials-mediated delivery platform

for sequential release of stem cell homing factors and DNA encoding therapeutic genes. Our

results show that biomaterials-mediated release of homing factors followed by delayed DNA

delivery enhanced recruitment of endogenous progenitor cells and improved limb salvage in a

mouse model of hindlimb ischemia. Finally, we demonstrate the potential of using microfluidic-

synthesized microspheres to aid therapeutic angiogenesis using a biomaterials-mediated drug

delivery depot.

Thesis Advisor: Fan Yang, PhD

Title: Asst. Professor in Bioengineering and Orthopaedic Surgery

v

vi

Acknowledgments

I wish to dedicate this thesis to my wife and son, Shenna and Eli. We are in this together and

will continue to grow as we traverse life’s terrain. I am grateful to Fan Yang for her mentorship

and support in helping me to develop as a scientist and a professional; guiding me from Medical

Scholars research to Ph.D. research. I am thankful for all those that have worked with me on

this thesis, especially Jeffrey Choi who deserves wholeheartedly to be recognized for his

contributions. Also, I wish to acknowledge Jothikritka Ashoken and Gloria Castaneda, who

spent the last year working with us on the angiogenesis team! Of course, I also wish to

acknowledge all the members of the Fan Yang lab for all your help and great times.

Table of Contents

1. Introduction……………………………………………………………………………………..1

2. Background…………………………………………………………………………..................4

3. Overexpressing Angiogenic Growth Factors in Adipose-Derived Stem Cells using

Biodegradable Polymeric Nanoparticles………………………………………………………31

4. Paracrine Release from Non-viral Engineered Adipose-Derived Stem Cells Promotes

Endothelial Cell Survival and Migration in vitro……………………………………................49

5. CXCR4-overexpressing Stem Cells Enhance Tissue Regeneration in a Mouse Model of

Hindlimb Ischemia………………………………………….…………………………………70

6. Adipose-Derived Stem Cells Co-expressing Hepatocyte Growth Factor and CXCR4 Promotes

Ischemic Tissue Repair………………………………………………………………………..93

7. Recruiting and Programming Endogenous Progenitor Cells in situ for Therapeutic

Angiogenesis…………………………………………………………………………………115

8. Microfluidic Synthesis of Biodegradable Polyethylene-Glycol Microspheres for Controlled

Delivery of Growth Factors and DNA Nanoparticles…..…………………………………….135

9. Summary, Limitations and Future Work……………………………………………………..154

vii

List of Figures

Figure 2.1 Methods for incorporating biological signals into scaffolds for controlled release..24

Figure 2.2 Polymeric scaffolds for dual delivery of VEGF and PDGF……………...………....25

Figure 2.3 Effects of VEGF temporal presentation on angiogenesis…………………………..26

Figure 2.4 Environmental-responsive controlled release system………………………………27

Figure 2.5 Cell-triggered VEGF delivery using biomimetic hydrogels………………………..28

Figure 2.6 Combined stem cell and gene therapy approach for therapeutic angiogenesis……...29

Figure 2.7 Exploiting cell homing for treating acute myocardial infarction…………………...30

Figure 3.1. Synthesis of end-modified poly(β-amino esters) (PBAE)…………………………42

Figure 3.2. Effects of polymer on transfection and viability of human adipose-derived stem cells,

mouse adipose-derived stem cells, and human embryonic kidney cells….……………………43

Figure 3.3. Enhancement of hADSC proliferation and transfection in the presence of basic

fibroblast growth factor…………………………………………………………..……………44

Figure 3.4. Effects of culture conditions post-transfection…………………………………….45

Figure 3.5. Transfection of cells with different plasmids encoding for multiple genes………...46

Figure 3.6. Effects of transfected cell paracrine release on human umbilical vein endothelial

cell (HUVEC) and human aortic smooth muscle cell (HASMC) viability……………………..47

Figure 4.1. Efficiency of gene delivery to human adipose-derived stem cells (hADSC) utilizing

poly(β-amino esters) (PBAE) nanoparticles…………………………………………………...63

Figure 4.2 Effects of nanoparticle dose on transfection efficiency and cell viability…………..64

Figure 4.3. VEGF release from human adipose-derived stem cells (hADSCs) after transfection

using polymeric nanoparticles…………………………………………….…………………...65

Figure 4.4 Effects of hypoxia on VEGF production and cell viability of non-viral engineered

ADSCs………………………………………………………………………………………...66

Figure 4.5. Effects of paracrine signals from non-viral engineered ADSCs on endothelial cell

viability and apoptosis…………………………………………………………….…………...67

Figure 4.6. Paracrine release from PBAE/VEGF engineered human adipose-derived stem cells

(hADSCs) enhanced endothelial cell migration……………………………………………….68

Figure 4.7. Paracrine release from PBAE/VEGF engineered human adipose-derived stem cells

(hADSCs) enhanced endothelial cell tube formation………………………………………….69

Figure 5.1. Genetic upregulation of CXCR4 and/or VEGF in ADSCs by PBAE nanoparticles.86

Figure 5.2. Improvement of cell fate and therapeutic efficacy due to CXCR4-overexpressing

ADSCs………………………………………………………………………………………...87

viii

Figure 5.3. CXCR4-overexpressing ADSCs enhanced muscle regeneration and mediated host

angiogenic and inflammatory response………………………………………………………..88

Figure 5.4. CXCR4-overexpression in ADSCs promotes a pro-angiogenic and anti-

inflammatory phenotype under hypoxia………………………………………………..……...89

Figure S5.1. Improvement of transfection efficiency and low cytotoxicity by PBAE

nanoparticles…………………………………………………………………………………..90

Figure S5.2. Cell engraftment of transfected ADSCs………………………………………….90

Figure S5.3. Reperfusion of the ischemic hindlimb at one week post-procedure………………91

Figure 6.1. Non-viral gene delivery to ADSCs via DNA polyplexes composed of a

biodegradable poly(β-amino ester) and pCK-HGF-IRES-CXCR4 plasmid………………….108

Figure 6.2. Production of functional therapeutic protein from genetically modified ADSCs...109

Figure 6.3. Survival of genetically modified ADSCs post-transplantation in vivo…………...110

Figure 6.4. Angiogenic response in ischemic hindlimbs post-transplantation in vivo………111

Figure 6.5. Analysis of limb status and muscle health………………………………………..112

Figure S6.1 – Plasmid Design of bi-directional and bicistronic vectors………………………113

Figure S6.2 – Optimization of D32-122 transfection by polymer to DNA weight ratio………114

Figure 7.1 SDF-1α Enhances Cell Recruitment Post-ischemic Induction……………………129

Figure 7.2 SDF-1α Promotes In Situ Transfection of recruited endogenous cells……………130

Figure 7.3 Timing of Plasmid DNA Delivery Affects the Level and Duration of Transfection131

Figure 7.4 Enhanced HGF Transfection In Situ by SDF-1α Delivery………………………...132

Figure 7.5 Angiogenic Improvement in the Ischemic Hindlimb by HGF overexpression…...133

Figure 7.6 Promotion of limb salvage………………………………………………………...134

Figure 8.1. Schematics of microfluidic chip design for microsphere synthesis………………148

Figure 8.2. Control of microsphere size by adjusting microchip in put channel dimensions…149

Figure 8.3. Effect of varying PEG concentration on microsphere property including swelling

ratio and microgel mesh size…………………………………………………………………150

Figure 8.4. Protein encapsulation in biodegradable PEG microspheres of differing size and

polymer composition……………………………………………………………...…………151

Figure 8.5. Release kinetics and bioactivity of bFGF from biodegradable PEG microspheres of

varying size and PEG concentration………………………………………………………….152


varying size and PEG concentration………………………………………………………….153

ix

List of Tables

Table 3.1: List of plasmids used for transfection………………………………………………48

Table 3.2: List of primers used for quantitative real time-polymerase chain reaction………….48

Table S5.1: List of primers used for quantitative real time-polymerase chain reaction…….…..92

x

1. Introduction

Peripheral arterial disease (PAD) is an important medical problem caused by a blockage

of blood flow to the limbs. The prevalence of PAD increases with age and those newly

diagnosed are six-times more likely to die within the next 10 years; this is because the diagnosis

of PAD often coincides with underlying advanced atherosclerosis or Diabetes Mellitus.

Symptomatic PAD can cause pain when walking, and in severe cases, lead to non-healing ulcers

or limb amputation. Treatment options often involve surgical revascularization, such as surgical

bypass or percutaneous angioplasty. However, many patients are unable to withstand surgery

due to the presence of surgically-prohibitive co-morbidities, thereby leaving them without

treatment.

Therapeutic angiogenesis offers an alternative treatment option for PAD, which relies

on the delivery of biologics or stem cells to promote host blood vessel reperfusion. In particular,

stem cell-based therapeutics may offer advantages to biologics by having the potential to

respond to environmental cues and secrete a multitude of paracrine signals. Despite this

potential, many stem cell therapeutics remain deficient, which is partly related to poor stem cell

engraftment and insufficient paracrine signaling. Gene therapy has the potential to overcome

these issues by strategically overexpressing therapeutic factors involved in stem cell mediated

therapeutic angiogenesis, such as angiogenic growth factors, cell survival factors, or cell homing

factors. Indeed, various genes encoding angiogenic factors along with various cell types have

been studied showing promising results.

The goal of this thesis is to develop a clinically translatable platform for treating PAD

using autologous stem cells genetically engineered with non-viral vectors. To achieve such a

platform several criteria were applied, including: 1) selection of a readily and abundantly

available autologous stem cell source; 2) use of a safe and efficient method for gene delivery to

primary cells in vivo; and, 3) the proper selection of therapeutic factors for genetic upregulation.

To meet these criteria, we developed two different strategies for achieving a genetically

engineered stem-cell based therapy for PAD. These were: 1) ex vivo engineered stem cells; and

2) endogenous progenitor cell recruitment and programming in situ.

With regards to the first approach, we achieve non-viral engineered stem cells ex vivo

using the following: 1) adipose-derived stem cells (ADSCs), which are a readily available

autologous stem cell source; and 2) biodegradable polymeric vectors, which are a non-viral

transfection agent. Using this platform, we assessed several genes for therapeutic enhancement.

1

In particular, we assessed CXC chemokine receptor 4 (CXCR4) for its role in stem cell homing

to ischemia. We also assessed vascular endothelial growth factor (VEGF) and hepatocyte

growth factor (HGF); where the latter is explored as an angiogenic factor that has shown the

most clinical potential to date.

With regards to the second approach, we attempt to achieve non-viral engineered stem

cells in situ using: 1) endogenous progenitor cells, which are recruited cell populations to sites

of ischemia under the action of chemokines; and 2) plasmid DNA, for therapeutic enhancement

of the recruited populations. This strategy is meant to circumvent the need for isolating stem

cells, which otherwise requires cumbersome and expensive manipulation in the laboratory, and

subsequent transplantation back to the patient. Effectively, this approach eliminates the need to

perform multi-step procedures by achieving genetically engineered stem cells wholly in situ.

We attempted to achieve this approach by controlled release of stromal derived factor-1α (SDF-

1α), a chemokine involved in cell homing to ischemia, and subsequent plasmid DNA delivery

encoding therapeutic factors.

These two approaches are described over 7 chapters (Chapters 2-8), which are

organized as follows:

Chapter 2 describes the main background for using stem cells and biomaterials for

therapeutic angiogenesis

Chapter 3 focuses on the optimization of ADSC transfection using poly(β-amino esters)

(PBAE), a biodegradable polymeric vector. Several aspects are explored for their effects on

ADSC transfection including: cell source and isolation; PBAE type; cell culture conditions; and,

plasmid design and gene.

Chapter 4 assesses the effects of paracrine release from non-virally engineered ADSCs

on endothelial cell (EC) behavior in vitro.

Chapter 5 explores the effects of CXCR4-overexpressing ADSCs on promoting cell

survival and engraftment, and subsequent therapeutic response in vivo. Further study on the

phenotypic behavior of CXCR4-overexpressing ADSCs cultured under hypoxia is also

performed.

Chapter 6 demonstrates the development of ADSCs co-overexpressing CXCR4 and

hepatocyte growth factor (HGF) using a bicistronic plasmid designed to enable dual expression

in the same ADSC. The therapeutic effects of these cells were assessed in an aged murine model

of hindlimb ischemia induced with severe ischemia to better reflect severe PAD.

2

Chapter 7 shows the potential of endogenous cell recruitment and programming as an

alternative to stem cell-based therapeutics. The effects of controlled SDF-1α release on cell

recruitment and in situ transgene expression are demonstrated. Timing of DNA delivery post-

surgery is further assessed. Therapeutic potential is validated in a murine model of hindlimb

ischemia using HGF.

Chapter 8 is about the development of monodisperse, biodegradable polyethylene

glycol (PEG) microspheres for tuning the release of biologics using a microfluidic platform.

The goal of this platform is to enable controlled delivery of protein or DNA, which may be used

to achieve sequential release of protein and DNA for recruitment and programming strategy.

Chapter 9 summarizes the results of this thesis and also discusses limitations and future

directions.

3

2. Background

Therapeutic Angiogenesis for Treating Cardiovascular Diseases

Lorenzo Deveza1, 2, Jeffrey Choi3, Fan Yang2,4 ,*

1 School of Medicine, Stanford University, Stanford, CA, 94305, USA; 2 Department of Bioengineering,

Stanford University, Stanford, CA, 94305, USA; 3 Department of Biology, Stanford University, Stanford,

CA, 94305, USA; 4 Department of Orthopaedic Surgery, Stanford University, Stanford, CA, 94305, USA

Abstract

Cardiovascular disease is the leading cause of death worldwide and is often associated with

partial or full occlusion of the blood vessel network in the affected organs. Restoring blood

supply is critical for the successful treatment of cardiovascular diseases. Therapeutic

angiogenesis provides a valuable tool for treating cardiovascular diseases by stimulating the

growth of new blood vessels from pre-existing vessels. In this chapter, we discuss strategies

developed for therapeutic angiogenesis using single or combinations of biological signals, cells

and polymeric biomaterials. Recent progress in exploiting genetically engineered stem cells and

endogenous cell homing mechanisms for therapeutic angiogenesis is also discussed.

2.1 Clinical significance

Cardiovascular disease (CVD) represents a global medical and economic problem with

high morbidity and mortality rates. The World Health Organization (WHO) has listed CVD as

the number one cause of death worldwide. In the United States alone, CVD accounted for

32.8% of the approximately 2.5 million deaths in 2008 [1]. The prevalence of CVD is similarly

staggering, with an estimated 82.6 million American adults (1 in 3) having one or more types

of CVD. Of these, 40.4 million are estimated to be older than 60 years of age with the average

annual rates of first time CVD events increasing with age [1]. For those that have experienced

one CVD event, CVD has been listed as one of the 15 leading conditions that cause functional

disability, thereby affecting quality of life and the ability to work.

Depending on the organs affected, CVD can be classified into coronary artery disease,

cerebrovascular disease, peripheral arterial disease and aortic (thoracic or abdominal)

atherosclerosis. CVD is generally characterized by narrowing or occlusion of the blood supply

of these vascular beds, and is most commonly caused by atherosclerosis [2, 3]. Treatment

options for CVD generally aim to re-establish blood flow through the affected vascular beds,

4

and are administered based on the severity of the disease [2, 4]. For early-stage disease,

management focuses on lifestyle modifications to reduce the number of modifiable risk factors.

As the disease progresses, pharmacological or surgical interventions may be needed to increase

the blood flow through the affected tissue or to reduce the energy requirements.

Pharmacological therapy acts to decrease oxygen demand by applying drugs to decrease the

heart rate; or to increase the blood supply by applying drugs that cause vascular smooth muscle

dilation. In cases of acute disease or full vascular occlusion, vascular stents may be used to

expand vessels when one or a few vessels are affected, while surgical bypass is necessary when

multiple vascular beds are occluded [4]. Despite the set of currently available treatment options

for patients with CVD, there is a subset of patients with advanced disease for whom surgical

revascularization is not an option due to the existence of various co-morbidities that prohibit

them from undergoing surgical procedures.

Driven by the clinical demands, therapeutic angiogenesis aims to stimulate and augment

the growth of new blood vessels from pre-existing vessels in order to re-supply blood flow to

affected ischemic tissues. In this chapter, strategies for therapeutic angiogenesis are discussed,

including direct delivery of angiogenic growth factors and the delivery of cells to ischemic

tissues. Recent progress on therapeutic angiogenesis utilizing polymeric biomaterials, combined

stem cell and gene therapy and regulation of endogenous stem cell homing are also discussed.

2.2 Biology of Angiogenesis

There are several mechanisms by which blood vessel formation occurs including

vasculogenesis, angiogenesis, and arteriogenesis [5, 6]. Each of these processes is interrelated

and leads to the formation of the vasculature of the body. The earliest blood vessel formation

in a developing embryo arises via vasculogenesis, in which endothelial progenitor cells coalesce

to form solid cords. These initially lumenless cords then transform into patent vessels in the

process of tubulogenesis [6]. Unlike the de novo blood vessel formation process associated with

vasculogenesis, angiogenesis is defined as sprouting new blood vessels from pre-existing blood

vessel networks and is important for expanding the vascular bed initially formed via

vasculogenesis. Arteriogenesis is the maturation of arterio-arteriolar anastomoses by the

recruitment and coating of pre-formed vessels with pericytes or vascular smooth muscle cells.

Arteriogenesis results in completely developed, functional arteries [7]. In the post-natal period,

angiogenesis and arteriogenesis play the major role in re-vascularizing under-supplied tissues

[5].

5

Under normal physiological conditions, there is a steady state in which quiescent

endothelial cells are maintained by autocrine signaling including vascular endothelial growth

factor (VEGF), NOTCH, angiopoietin-1 and fibroblast growth factor (FGF) [5, 6]. This steady

state may become disrupted under conditions of low oxygen, inflammation, wound healing, or

within a tumor. Such hypoxic conditions can be sensed by endothelial cells and other stromal

cells, which express oxygen sensors and hypoxia-inducible factors, such as prolyl hydroxylase

domain 2 (PHD2) and hypoxia-inducible factor 2α (HIF-2α) [5, 6]. When hypoxia is sensed,

angiogenic signals are released, such as VEGF, angiopoietin-2, and FGF [5, 6]. Initially, these

signals cause pericytes to detach from the vessel walls and endothelial cells to loosen their cell-

cell junctions, which enable the blood vessels to dilate. This increases vascular permeability

and allows plasma proteins to extravasate and form a provisional extracellular matrix (ECM)

through which ECs can migrate. Simultaneously, proteases liberate angiogenic molecules from

the ECM, which further potentiate the process. The migrating ECs differentiate to become

guiding tip cells or proliferating stalk cells. The tip cells lead the direction of sprouting, while

the stalk cells proliferate to elongate the sprouting vessels. Blood flow is initiated when two

growing vessels meet and fuse. Maturation later ensues as pericytes are stimulated to cover

endothelial cells under the action of platelet-derived growth factor-B (PDGF-B), angiopoietin-

1 (ANG-1), and transforming growth factor-β (TGF-β) [5]. It is known that bone-marrow

derived endothelial progenitor cells (EPCs) also participate in angiogenesis and become

incorporated into the vascular wall, but their importance is not well understood [8].

2.3 Therapeutic angiogenesis

Therapeutic angiogenesis aims to induce, augment and control the host angiogenic

response in order to re-vascularize ischemic tissues, and often involves delivery of growth

factors or stem/progenitor cells. Growth factors may be delivered in the form of proteins or

genes encoding target proteins. The premise behind this approach is to apply well-studied

growth factors (VEGF, FGF) in ischemic tissues to guide angiogenic cellular and tissue

behavior. Cell therapy may similarly act to induce the angiogenic response by the release of

paracrine signals. Delivered cells may also act by becoming incorporated into the growing

vascular supply, thereby acting as building blocks to form new blood vessels. Both mechanisms

likely occur upon delivery of cells to ischemic tissues. Transmyocardial laser revascularization

offers another strategy for therapeutic angiogenesis, which is believed to stimulate host

angiogenic response by injuring the ischemic myocardium in specified locations [9].

6

2.3.1 Growth Factor Therapy

Various growth factors have been applied for therapeutic angiogenesis including

VEGF, bFGF, hepatocyte growth factor (HGF), PDGF, ANG-1 and insulin-like growth factor

(IGF-1). Among these, VEGF and bFGF are the most well-studied and have reached human

clinical trials. VEGF is the most important regulator of physiological angiogenesis during

growth, healing and in response to hypoxia [5, 10]. VEGF is upregulated 30-fold by hypoxia-

inducible transcription factor, which is more than any other inducible angiogenic factor.

However, when administered alone, VEGF may lead to the formation of leaky, unstable

capillaries. PDGF-B can help stabilize nascent blood vessels by recruiting mesenchymal

progenitors, and co-delivery of VEGF and PDGF has been shown to lead to early formation of

mature vessels [11]. bFGF is among the first discovered angiogenic factors to have both

angiogenic and arteriogenic properties, and FGF receptors are expressed on both endothelial

cells and smooth muscle cells, which may facilitate formation of a mature blood vessel network

[5, 10]. The HGF family induces potent angiogenic responses by binding to the c-MET receptor,

which is expressed on endothelial cells, vascular smooth muscle cells and hematopoietic stem

cells. HGF is known to have mitogenic, angiogenic, anti-apoptotic, and anti-fibrotic activities

in various cells [10].

To date, there have been several pre-clinical and clinical studies evaluating the safety

and efficacy of growth factors for inducing angiogenesis in ischemic heart disease and

peripheral arterial disease (PAD). Early phase I and II human clinical trials with VEGF, bFGF

and HGF were promising, but larger randomized placebo-controlled trials failed to demonstrate

significant benefits in the approved end-points. Preclinical data demonstrate that the delivery

of bFGF or VEGF eventually results in unstable vessel growth that resembles immature tumor

vasculature. The limited clinical response to growth factor therapy is likely due to a number of

factors. Growth factor delivery is generally limited by their rapid diffusion, poor biostability

and short half-lives in vivo. Furthermore, it often requires supraphysiological doses or multiple

injections, which could lead to excessive uncontrolled vascular formation in undesired

locations.

3.2. Cell-based therapy

Cell-based therapy is the most extensively studied approach so far for therapeutic

angiogenesis and may contribute via either directly participating in new vessel formation or

7

secreting paracrine signals. Excellent reviews have been published previously with detailed

discussion on this topic [12, 13]. In this review, we have chosen to focus on more recently

developed platforms for therapeutic angiogenesis including biomaterials-based approach,

combined stem cells and gene therapy and exploiting stem cell homing for angiogenesis. As a

therapeutic approach, cells have the potential to act as hypoxia-responsive paracrine release

vehicles by virtue of their diverse cytokine contents. As cells migrate and respond to the

environment, they may also lead to more dynamic paracrine release. The most extensively

studied cell populations for therapeutic angiogenesis are endothelial progenitor cells (EPC) and

bone marrow-mesenchymal stem cells (BMMSCs). EPCs have been found to mobilize in

response to ischemia and become incorporated into the growing vasculature [8, 14]. However,

it should be recognized that various studies have employed different protocols for isolating this

population of cells, and this has resulted in inconsistent expression profiles among different

studies, which consists of true EPCs and other cell types, such as hematopoietic stem cells [12].

For this reason, this cell population has been reported and identified according to the method of

cell isolation, e.g., “granulocyte colony stimulating factor (G-CSF) mobilized peripheral blood

cells,” or “bone marrow aspirate mononuclear cells” (BMNCs). These cell populations can

differentiate into endothelial cells, pericytes, or smooth muscle cells, thereby acting as building

blocks for angiogenesis, but may also potentiate vascular growth through paracrine

mechanisms. BMMSCs also have the potential for differentiation into vascular support cells,

but they more likely contribute to angiogenesis by the release of proangiogenic factors. Adipose-

derived stem cells (ADSC) exhibit similar differentiation potential and paracrine release

characteristics as BMMSCs, and is currently being evaluated in preclinical trials for therapeutic

angiogenesis. Compared to BMMSC, ADSCs are more abundant and easier to isolate, yet more

studies are still needed to fully assess their potential for therapeutic angiogenesis.

The main advantages of autologous stem cell therapy relate to immunogenicity and

ethical issues. Because these cells will be transplanted back into the same patient from which

they are harvested from, there should be no immune rejection, such as that seen in allogeneic

implants. Also, since these progenitor cells are isolated from adult tissues and are not derived

from fetal tissues, they face no major identified ethical concerns. Furthermore, as adult stem

cells, their differentiation potential is relatively limited with lesser risk for aberrant tissue

formation and tumorigenesis. On the other hand, cell-based therapies still face several

limitations. Upon isolation, the cell population is often heterogeneous, which may lead to varied

responses. To obtain the large cell numbers needed for transplantation, ex vivo cell expansion

8

is often required, which leads to regulatory concerns and increased cost and time, thereby

increasing the barriers for clinical translation. Furthermore, the cell engraftment efficiency is

typically very low upon transplantation, and may induce inflammatory responses. Still, cell-

based therapies have met with success in the clinic. In a meta-analysis of cell therapy for PAD,

it was found that autologous cell delivery (BMMSCs and G-CSF mobilized peripheral blood

cells) led to improved indices of ischemia and subjective symptoms (pain-free walking) [15].

However, the study also pointed out limitations of their analysis, such as publication bias and

the inclusion of unblinded study results. Further work is needed to determine the efficacy of

autologous stem cell therapy.

2.4 Polymeric Biomaterials for therapeutic angiogenesis

Polymeric biomaterials may serve as drug delivery depots for controlled release of

biological signals in a temporal and spatial-controlled manner. In the ideal situation, such

systems may release growth factor within a physiological range over the course of several days

to weeks, thereby overcoming the issues of short protein half-lives and rapid diffusion from the

site. Polymeric biomaterials have been applied in therapeutic angiogenesis to delay the release

of single growth factors, to release multiple biologics or to achieve environment-responsive

release of biologics in response to changes in pH or cell-mediated matrix metalloproteinase

activities.

2.4.1 Commonly utilized polymers for therapeutic angiogenesis

Polymers for therapeutic angiogenesis include synthetic-based polymers, such as

poly(lactic-co-glycolic acid) (PLGA) and poly(ethylene glycol) (PEG), as well as naturally-

derived polymers, such as alginate or gelatin. PLGA was among the first to be studied for drug

release and many initial systems for therapeutic angiogenesis utilized PLGA as a base polymer

for the controlled release of VEGF [16-18]. PLGA is a favorable polymer due to its

biodegradability and ability to control release, but when degraded it increases the local acidity,

which could cause undesirable inflammation. PEG is a biocompatible polymer that is inert by

nature, with easily tunable biochemical and mechanical properties [19]. PEG has been utilized

to create cell-responsive hydrogels for VEGF release [20]. Alginate and gelatin are composed

of natural polymers that can interact well with cells in vivo and may also help to control drug

release. However, natural-derived materials suffer from batch-to-batch variance, and they are

less tunable in chemical properties compared to synthetic polymers.

9

2.4.2 Strategies for modulating growth factor release using biomaterials

To achieve effective therapeutic angiogenesis, it would be desirable to have a sustained

release of biological signals over time. However, most drug delivery systems have an initial

“burst-release,” where the large-majority of loaded growth factors is released in the first few

hours. While the mechanisms underlying “burst-release” have not yet been entirely elucidated,

several parameters are believed to play a critical role in regulating matrix-controlled drug

release including processing conditions, surface characteristics, and geometry [21]. Several

strategies have been developed to delay such burst release by modulating physical interactions,

ionic/biochemical affinity interactions, and covalent binding between the loaded biologics and

the polymeric biomaterials (Figure 1).

2.4.2.1 Physical interactions. Physical interactions or physical entrapment can delay protein

release based on the size of the drug relative to the pore size of the scaffolds. Controlled release

of VEGF from PLGA scaffolds improved angiogenesis compared to free-VEGF administration

in a myocardial ischemia model or hindlimb ischemia model [17, 18]. These reports

demonstrate the benefit of prolonged release of VEGF over bolus growth factor injection.

Biological signals can also be loaded into composite scaffolds using both PLGA microspheres

containing alginate hydrogel, which also demonstrated statistically significant increases in

vessel density compared to recombinant human-VEGF (rh-VEGF) injection alone in vivo [22].

2.4.2.2 Ionic/biochemical interactions. Proteins can form hydrogen or ionic bonds with

macromolecules such as heparin sulfate. Such interactions can be utilized to slow down the

protein release by incorporating heparin, heparan sulfate or hyaluronic acid into the drug-

releasing scaffolds [23-27]. bFGF specifically binds to heparin and heparan sulfate, and loading

bFGF into heparin conjugated PLGA nanospheres led to linear release up to 15 days [23]. This

release can be further prolonged by encapsulating these PLGA nanospheres into fibrin hydrogels

[23]. Sustained release of bFGF resulted in significant improvements in capillary density and

cell proliferation in a murine hindlimb ischemia model, [23]. Enhanced angiogenesis was also

observed using heparin-containing chitosan hydrogels or fibrous matrices to release bFGF in

vivo, and release can be tuned by varying heparin concentration in the scaffold [24, 27]. Heparin

binding alginate hydrogels have also been developed for prolonged release of HGF, which

resulted in significant increases in blood flow and capillary density compared to bolus HGF

10

delivery [26]. These studies show that release of growth factor may be prolonged by exploiting

heparin-binding interactions and this has potential for improving growth factor approaches for

therapeutic angiogenesis.

2.4.2.3 Covalent linking. Lastly, some biologics can be covalently bound to the polymeric

matrix via a hydrolytically degradable or MMP-degradable linker [20, 28, 29]. It has been

demonstrated that matrices incorporating covalently attached VEGF via a MMP-degradable

linker can lead to the formation of a more controlled and stable vasculature [20, 29]. However,

covalent linkage requires chemical modification of the biologic which can affect its biological

activity. Also, the chemistry of the linker must be carefully designed to prevent unwanted

immune responses [19].

2.4.3 Co-delivery of multiple biologics.

Angiogenesis is a dynamic process that is regulated by complex biological signals.

Polymeric biomaterials offer the potential to simultaneously or sequentially deliver multiple

biologics better mimicking the complexity of angiogenesis [11, 30-35]. For example, VEGF is

a potent angiogenic factor, but when delivered alone is not sufficient for developing a mature

and stable vascular network. Dual delivery of VEGF and PDGF using PLGA or alginate

hydrogel resulted in the formation of larger and more mature vessels and improved blood flow,

thereby suggesting the potential benefits of co-delivering synergistic growth factors (Figure 2)

[11] [30]. Biodegradable hydrogels such as hyaluronan (HA) has also been examined to achieve

sequential delivery of VEGF and keratinocyte growth factor (KGF) [33]. When transplanted

into the ear pinna of a mouse they found significant increases in microvascular density in the

dual-delivery group from the HA hydrogel compared to dual growth factor bolus injections, or

single growth factor delivery from HA hydrogels [33]. Gelatin-based hydrogel has also been

shown as a promising depot for co-delivery of bFGF and G-CSF to enhance therapeutic

angiogenesis [31]. bFGF is a mitogen and chemoattractant of both fibroblasts and endothelial

cells, while G-CSF acts as a potent mobilizer of hematopoietic stem cells and bone marrow

stromal cells. Dual release of bFGF and G-CSF from a gelatin hydrogel enabled increased

reperfusion and capillary density by 2 weeks, which was sustained up to 8 weeks. In contrast,

bFGF alone led to a reduction in these parameters by 8 weeks [31]. Bolus delivery of these

factors led to improvements in capillary density, but no effects on blood reperfusion, which

further supports the benefit of controlled release [31]. Dextran-based hydrogels capable of

11

releasing multiple angiogenic factors (VEGF, SDF-1, IGF, Ang I) has been recently developed

[35]. Significant increases in vessel number and size were observed when all four growth

factors were delivered in a subcutaneous model compared to single or dual growth factor

delivery, again suggesting the potential synergistic benefit of delivering multiple biological

signals for therapeutic angiogenesis [35].

2.4.4 Temporal controlled release.

Angiogenic response to growth factors is time-dependent and the ability to control the

temporal release of growth factors is highly desirable to achieve maximal benefits. In support

of this concept, Silva and colleagues demonstrated that a gradual reduction in VEGF

concentration led to increases in vessel sprouts in an in vitro sprout assay compared to

consistently high VEGF concentration or increasing VEGF concentration over time (Figure 3)

[36]. Tengood and colleagues developed a cellulose hollow fiber capable of sequential release

of VEGF followed by sphingosine-1-phosphate (S1P) [34]. VEGF initiates angiogenesis by

increasing vascular permeability and recruits endothelial cells, while S1P acts to stabilize

intracellular junctions and decrease the permeability of endothelial cells but simultaneously

inhibits endothelial cells migration [34]. If delivered together, these two signals may inhibit

each other. Their results showed that a significant increase in angiogenic response was only

observed when VEGF and S1P were delivered sequentially. In contrast, low angiogenic

response was observed when the two factors were delivered simultaneously or if S1P was

delivered first. These results demonstrate that growth factors can act in temporal manner

depending on the stage of angiogenesis.

2.4.5 Environmental-responsive hydrogels for releasing biologics.

To aid releasing angiogenic growth factors specifically in ischemic tissues, pH and

temperature responsive hydrogels have been developed. pH-responsive heparin functionalized

chitosan/poly(γ-glutamic acid) nanoparticles have been developed to gel and enable sustained

release of bFGF under low pH, and then disintegrate and release heparin under normal

physiological pH (Figure 4) [37]. This design aimed to utilize bFGF to induce angiogenesis,

while releasing heparin as an anti-coagulant [37]. The efficacy of such a system for in vivo

therapeutic angiogenesis remains unknown and requires further testing. Another study reported

the development of a pH and temperature sensitive hydrogel capable of gelation at pH of 6.8

and at 37°C [38]. When transplanted in a rat model of myocardial ischemia, it led to statistically

12

significant improvements in blood flow, capillary density and arteriole density compared to

bolus injection and polymer only controls, and these parameters correlated with improvements

in myocardial functions [38]. These results demonstrate the feasibility and promise of applying

environmental-responsive polymers to achieve better control over growth factor release.

2.4.6 Biomimetic polymers to modulate cell response.

Biomimetic scaffolds containing cell-responsive domains can promote therapeutic

angiogenesis. For example, matrices can be designed to interact with cells via integrin binding

sites to facilitate cell migration and cell-cell interactions. Injecting RGD-modified alginate

hydrogels into a chronic rat myocardial infarction model led to significant improvements in

cardiac function 70 days post-infarction, with commensurate increases in arteriole density

compared to non-RGD containing gel control [39]. In another study, an MMP-sensitive PEG

hydrogel containing RGD and covalently linked VEGF was evaluated for therapeutic

angiogenesis in vivo (Figure 5) [20]. The cell-responsive hydrogel led to angiogenesis

specifically within the hydrogel, while soluble VEGF release led to increased vascular density

in the surrounding tissue [20]. These systems demonstrate the importance of cell-matrix

interactions and potential of exploring biomimetic polymers to modulate cell responses for

therapeutic angiogenesis.

2.5 Combined Stem Cell and Gene-based Therapy

The therapeutic potential of stem cells can be further enhanced by combining with gene

therapy. In this approach, stem cells can be genetically modified prior to transplantation in such

a way that particular cellular processes are strategically exploited (Figure 6). In particular,

researchers have focused on addressing the challenges associated with applying cells alone, such

as insufficient paracrine release, poor cell survival upon engraftment, and lack of cell homing.

Gene delivery can be used to over-express desired therapeutic factors to induce a biological

response. In contrast to growth factor delivery, gene delivery may also be utilized to up-regulate

expression of intracellular transcription factors or cell surface receptors. These strategies offer

the advantage of controlling cell behavior at the intracellular signaling level.

2.5.1 Viral vs. Nonviral

Both viral and non-viral vectors have been employed for gene delivery for therapeutic

angiogenesis, with most studies utilizing a viral-based approach due to its high efficiency [40-

13

50]. However, clinical translation of viral-based gene delivery is limited due to safety concerns

such as immunogenicity and insertional mutagenesis. Non-viral methods provide a potentially

safer alternative, and rely on physical methods such as electroporation and nucleofection, or

using cationic lipids or polymers [51]. However, most non-viral based gene delivery methods

suffer from low transfection efficiency and often high cytoxicity, and few studies for therapeutic

angiogenesis have applied such methods [52-54]. All of the following genetic approaches

described here utilized a viral-approach, except those described in the “Non-viral Approaches”

section.

2.5.2 Stem cells overexpressing growth factors

The efficacy of stem cell-based therapy may be augmented by overexpressing desirable

paracrine signals that promote angiogenesis. Overexpressing VEGF reduced the number of

EPCs needed by 30-fold to achieve a similar therapeutic effect as unmodified EPCs alone in

promoting angiogenesis and limb salvage [42]. By enhancing paracrine effects, fewer cells

would be necessary to achieve therapeutic effects. Similarly, virally transduced BMMSCs

overexpressing VEGF or ANG I led to significantly enhanced angiogenesis and improved

cardiac function in a rat myocardial infarction model compared to non-transfected controls [43,

44]. HGF and bFGF are known to stimulate both endothelial cells and smooth muscle cells, and

may stimulate both the initiation and maturation of angiogenesis. Virally transduced MSCs or

ADSCs overexpressing HGF resulted in improvement in cardiac performance and increasing

number of capillaries [41, 50]. The improvement in cardiac function can be detected as early as

day 7 and continued over time, supporting the benefits of HGF overexpression on both early

and late stages of angiogenesis. EPC overexpressing FGF1 also led to significant increases in

mature vessel density by α-smooth muscle actin staining in vivo [40]. Overall, these studies

demonstrate the potential of stem cells overexpressing angiogenic factors for accelerating blood

vessel growth and tissue functions.

2.5.3 Genetic modification to enhance cell survival

To enhance cell survival post-transplantion, ex vivo genetic modification of stem cells

with cell survival factors has been studied [55, 56]. Akt is a serine threonine kinase that is

known to be involved in several cellular processes, such as glucose metabolism and cell

migration. Foremost, among these processes is that it is a cell survival factor. Based on these

effects, Mangi and colleagues sought to enhance survival of MSCs upon transplantation by

14

overexpressing Akt1 by a retroviral method [55]. In comparison to GFP transfected controls,

they found significant increases (over two-fold) in c-kit+ cells (transplanted cells) and decreases

in c-kit+ cell apoptosis 24 hours post-transplantation, and this led to significant decreases in

fibrosis and increases in cardiac function [55]. Their results also demonstrated improvements

in cardiac function utilizing half the number of Akt-positive MSCs as LacZ controls (2.5x105

cells vs 5.0x105 cells) further supporting their hypothesis. Another report sought to build on

these results by positing an interesting approach to co-overexpress a cell-survival factor with a

growth factor. In this study, Ang1 and Akt co-overexpressing MSCs were developed utilizing

adenovirus and then were transplanted into a rat myocardial infarction model. They

demonstrated that Akt overexpression leads to enhanced cell engraftment as long as 3 months

compared to non-transfected controls with resulting increases in mature vessel density and

improvements in heart function [56]. However, it is unclear what the benefits of dual over-

expression were, since the study lacked single Akt or Ang1 controls [56]. More work will need

to be done to determine how these effects might synergize.

2.5.4 Genetic modification to enhance stem cell homing

Efficient cell homing to ischemic tissue is critical for achieving effective angiogenesis

as the therapeutic effects are largely dependent upon the engraftment efficiency of transplanted

cells. Current methods rely mostly on intramuscular or intravascular cell injection, which often

results in poor cell engraftment at the site of ischemia. Gene therapy has been applied to modify

stem cells to overexpress cell homing factors to enhance cell targeting to ischemic tissues [45,

47-49]. The hypoxia-inducible factor-1 (HIF-1)/stromal-derived factor-1 (SDF-1)/CXC

receptor 4 (CXCR4) is a well-known axis that enables host cells to home to sites of ischemia

[57]. Based on this understanding, both CXCR4 and SDF-1 have been studied for induction of

transplanted cell homing and host cell recruitment [45, 47-49]. Two studies have reported the

use of virally transduced CXCR4 overexpressing MSCs for therapeutic application in rat models

of myocardial ischemia [45, 49]. The method of delivery was IV injection for both studies and

each found significant increases in cell engraftment into the ischemic myocardium compared to

non-transfected controls. However, these studies reported conflicting results with regards to

increases in vessel density due to CXCR4 transduction in MSCs [45, 49]. This may be due to

differences in negative controls with one demonstrating significant increases compared to

EGFP-adenoviral transfected controls and one showing no difference compared to non-

transfected controls. In either case, CXCR4-overexpressing MSCs led to improvements in

15

cardiac function, thereby demonstrating that such increased cell homing may provide

therapeutic effects by another mechanism. Tang and colleagues took another approach seeking

to upregulate SDF-1 in MSCs for the purpose of recruiting host cells for therapeutic

vascularization [47, 48]. In their study, they chose to overexpress both VEGF and SDF-1 to

achieve the effect of initiation and maturation of new blood vessels in a rat model of myocardial

infarction [48]. Interestingly, they demonstrated that overexpression of SDF-1 leads to genetic

upregulation of the survival factor, Akt [47, 48]. They found that SDF-1 overexpression alone

was able to mediate significant increases in vascular density by vWF and α-smooth muscle actin

staining, but combination therapy led to significantly higher increases with corresponding

improvements in cardiac function [48]. These approaches may become more interesting as an

understanding of host cell recruitment for injury becomes better understood.

2.5.5 Non-viral approaches

Non-viral gene transfection approaches are important for clinical translation. However,

few studies have reported the use of non-viral vectors for ex vivo gene modification in

therapeutic angiogenesis. Many of the current approaches for non-viral gene transfer are

unsatisfactory due to low transfection efficiency or high cytotoxicty. To overcome some of

these issues, Yang and colleagues developed VEGF over-expressing MSCs utilizing novel non-

viral biodegradable polymeric nanoparticles called poly(β-amino esters) that were able to

achieve 4-5 times the transfection efficiency of the leading commercially available reagent,

Lipofectamine 2000, with minimal cytoxicity [54]. When transplanted into a murine model of

hindlimb ischemia, they found enhancement in angiogenesis and limb salvage compared to GFP

transfected controls, thereby demonstrating the clinical potential of non-viral engineered stem

cell-based therapy for therapeutic angiogenesis (Figure 6) [54]. In another study, a novel

plasmid for a hypoxia inducible VEGF was transfected into MSCs utilizing a non-degradable

polymer (linear poly-ethyliminine), which also demonstrated improved efficiency and

cytoxicity characteristics compared to Lipofectamine. This plasmid was designed to enable

better control over VEGF release, and led to improved cell engraftment, capillary density and

increased cardiac functions in a rat myocardial ischemia model. These results demonstrate the

promise of applying non-viral based gene delivery for therapeutic efficacy, which may provide

a safer alternative than the viral approach.

16

2.6 Exploiting stem cell homing for therapeutic angiogenesis

Stem cell therapy holds great potential for therapeutic angiogenesis, but in general,

clinical translation of stem cell-based approaches for therapeutic angiogenesis has been slow

due to several inherent limitations. First, most methods require lengthy procedures which

include cell isolation, in vitro cell modification, and subsequent transplantation back into the

patient. Such multi-step procedures lead to high cost and significant regulatory concerns.

Second, the vast majority (>90%) of transplanted cells die and few manage to migrate to the

ischemic site. Third, ex vivo stem cell modifications may be dependent on culture conditions

and be transient, and may thus lose effectiveness upon transplantation in vivo.

One strategy to overcome the aforementioned limitations is to exploit natural cell

homing mechanisms to attract host progenitor cells to sites of ischemic injury. In the normal

physiological state, bone marrow-derived populations are recruited to sites of hypoxia to

participate in the angiogenic process [8, 58]. However, during states of cardiovascular disease,

host cells often have reduced functionality and a compromised cell niche [58]. Various

researchers have proposed methods to increase cell homing to ischemic tissue for therapeutic

angiogenesis and for tissue engineering [58, 59]. The number of circulating EPCs from the

bone marrow has been shown to increase in response to ischemia [8], which led to improved

corneal neovascularization when ischemia was induced in the mouse hindlimb, and applying

cell mobilization factor GM-CSF led to further increases in angiogenesis. These results

demonstrate that ischemia induced at a distant site can have an effect on angiogenesis

throughout the body by cell mobilization [8]. The number of EPCs recruited to the site of

ischemia has been shown to be correlated to the level of hypoxia [57]. Increasing hypoxia led

to increasing SDF-1 expression and vessel density, which was attenuated using antibodies to

block the SDF-1/CXCR4 pathway [57]. Further study revealed that G-CSF induced

mobilization of bone marrow-derived cells by disrupting the SDF-1/CXCR4 interactions in the

bone marrow niche, with gradual decrease in SDF-1 signaling and increased CXCR4 expression

over 5 days [60]. When G-CSF mobilized cells were transplanted into a bone marrow transplant

model, cell engraftment was inhibited by blocking SDF-1 or CXCR4 pathway with antibodies.

In a later study, it was shown that HSCs, EPCs and MSCs could be differentially mobilized by

application of different combinations of G-CSF, CXCR4 antagonists and VEGF [61].

The potential of cell homing strategies for treating cardiovascular disease has been

demonstrated by the application of G-CSF in a mouse model of myocardial ischemia. In a study

performed by Orlic and colleagues, G-CSF was applied once daily for 5 days prior to inducing

17

myocardial infarction, followed by 3 more days of G-CSF injection post-myocardial infarction

[62]. This protocol resulted in over 70% mouse survival in comparison to about 20% mouse

survival in the untreated group [62]. This treatment also led to increased left ventricular wall

thickness and enhanced ejection fraction [62]. However, in practice, this type of protocol is

unrealistic since myocardial infarction gives few early warning signs. Despite some positive

results in animal models, clinical trials with G-CSF for acute myocardial infarction have been

inconclusive. One potential cause for the unsatisfactory results may be the lack of homing

efficiency of G-CSF mobilized cells to sites of injury due to the rapid cleavage of SDF-1 by

CD26/dipeptidylpeptidase-IV (DPP-IV) [63]. By applying both G-CSF and DPP-IV inhibitor

(Diprotin-A), researchers have shown significant increases in homing to the infarcted

myocardium, which correlated with improvements in capillary density and cardiac function and

mouse survival compared to G-CSF treatment alone (Figure 7) [63]. Since these factors are

administered at the time of injury, this therapy may be more realistic for clinical intervention.

The field of exploiting stem cell homing for therapeutic angiogenesis is young, and represents

a promising future direction for therapeutic angiogenesis. For future work, polymeric scaffolds

that are capable of in situ recruiting and programming endogenous stem cells in ischemic sites

would provide valuable tools for reducing the translational barrier for therapeutic angiogenesis.

2.7 Conclusions

Therapeutic angiogenesis is an important interventional approach for treating a broad

range of cardiovascular diseases, which represents a global challenge with high morbidity.

Conventional strategies involve delivering growth factors or stem cells, or a combination of

both. Polymeric biomaterials have the potential to enhance therapeutic angiogenesis by

prolonging the release of angiogenic signals, and co-delivery of synergistic signals may further

accelerate the formation of functional and mature blood vessel networks. Engineering

biomimetic scaffolds that can release biological signals in an environmental-responsive manner

would facilitate targeted angiogenesis, while mitigating the undesirable side effects caused by

uncontrolled diffusion. Controlled release systems with precise temporal and spatial control

will be of great value for therapeutic angiogenesis by mimicking the dynamic biological release

profiles characteristic of normal angiogenesis processes. Combined stem cell and gene-based

therapy may further improve the efficacy of using autologous cells alone by overexpressing

therapeutic factors or cell homing factors, and its translation relies largely on safe and effective

methods to deliver genes into stem cells. Finally, strategies exploiting cell homing may provide

18

a solution to reduce the translational barrier of cell-based therapeutic angiogenesis by reducing

the time and cost associated with ex vivo cell expansion and manipulation.

References

1. Roger, V.L., et al., Heart disease and stroke statistics--2012 update: a report from the

American Heart Association. Circulation, 2012. 125(1): p. e2-e220.

2. Libby, P. and P. Theroux, Pathophysiology of coronary artery disease. Circulation,

2005. 111(25): p. 3481-8.

3. Jawad, E. and R. Arora, Chronic stable angina pectoris. Dis Mon, 2008. 54(9): p. 671-

89.

4. McFalls, E.O., et al., Coronary-artery revascularization before elective major vascular

surgery. N Engl J Med, 2004. 351(27): p. 2795-804.

5. Carmeliet, P. and R.K. Jain, Molecular mechanisms and clinical applications of

angiogenesis. Nature, 2011. 473(7347): p. 298-307.

6. Xu, K. and O. Cleaver, Tubulogenesis during blood vessel formation. Semin Cell Dev

Biol, 2011. 22(9): p. 993-1004.

7. Heil, M., et al., Arteriogenesis versus angiogenesis: similarities and differences. J Cell

Mol Med, 2006. 10(1): p. 45-55.

8. Takahashi, T., et al., Ischemia- and cytokine-induced mobilization of bone marrow-

derived endothelial progenitor cells for neovascularization. Nat Med, 1999. 5(4): p.

434-8.

9. Horvath, K.A., Transmyocardial laser revascularization. J Card Surg, 2008. 23(3): p.

266-76.

10. Losordo, D.W. and S. Dimmeler, Therapeutic angiogenesis and vasculogenesis for

ischemic disease. Part I: angiogenic cytokines. Circulation, 2004. 109(21): p. 2487-91.

11. Richardson, T.P., et al., Polymeric system for dual growth factor delivery. Nat

Biotechnol, 2001. 19(11): p. 1029-34.

12. Leeper, N.J., A.L. Hunter, and J.P. Cooke, Stem cell therapy for vascular regeneration:

adult, embryonic, and induced pluripotent stem cells. Circulation, 2010. 122(5): p. 517-

26.

13. Rafii, S. and D. Lyden, Therapeutic stem and progenitor cell transplantation for organ

vascularization and regeneration. Nat Med, 2003. 9(6): p. 702-12.

19

14. Asahara, T., et al., VEGF contributes to postnatal neovascularization by mobilizing

bone marrow-derived endothelial progenitor cells. EMBO J, 1999. 18(14): p. 3964-72.

15. Fadini, G.P., C. Agostini, and A. Avogaro, Autologous stem cell therapy for peripheral

arterial disease meta-analysis and systematic review of the literature. Atherosclerosis,

2010. 209(1): p. 10-7.

16. Cleland, J.L., et al., Development of poly-(D,L-lactide--coglycolide) microsphere

formulations containing recombinant human vascular endothelial growth factor to

promote local angiogenesis. J Control Release, 2001. 72(1-3): p. 13-24.

17. Formiga, F.R., et al., Sustained release of VEGF through PLGA microparticles

improves vasculogenesis and tissue remodeling in an acute myocardial ischemia-

reperfusion model. J Control Release, 2010. 147(1): p. 30-7.

18. Golub, J.S., et al., Sustained VEGF delivery via PLGA nanoparticles promotes vascular

growth. Am J Physiol Heart Circ Physiol, 2010. 298(6): p. H1959-65.

19. Lin, C.C. and K.S. Anseth, PEG hydrogels for the controlled release of biomolecules in

regenerative medicine. Pharm Res, 2009. 26(3): p. 631-43.

20. Zisch, A.H., et al., Cell-demanded release of VEGF from synthetic, biointeractive cell

ingrowth matrices for vascularized tissue growth. FASEB J, 2003. 17(15): p. 2260-2.

21. Huang, X. and C.S. Brazel, On the importance and mechanisms of burst release in

matrix-controlled drug delivery systems. J Control Release, 2001. 73(2-3): p. 121-36.

22. Lee, J. and K.Y. Lee, Local and sustained vascular endothelial growth factor delivery

for angiogenesis using an injectable system. Pharm Res, 2009. 26(7): p. 1739-44.

23. Jeon, O., et al., Long-term and zero-order release of basic fibroblast growth factor from

heparin-conjugated poly(L-lactide-co-glycolide) nanospheres and fibrin gel.

Biomaterials, 2006. 27(8): p. 1598-607.

24. Kim, M.S., et al., Development of functional fibrous matrices for the controlled release

of basic fibroblast growth factor to improve therapeutic angiogenesis. Tissue Eng Part

A, 2010. 16(10): p. 2999-3010.

25. Pike, D.B., et al., Heparin-regulated release of growth factors in vitro and angiogenic

response in vivo to implanted hyaluronan hydrogels containing VEGF and bFGF.

Biomaterials, 2006. 27(30): p. 5242-51.

26. Ruvinov, E., J. Leor, and S. Cohen, The effects of controlled HGF delivery from an

affinity-binding alginate biomaterial on angiogenesis and blood perfusion in a hindlimb

ischemia model. Biomaterials, 2010. 31(16): p. 4573-82.

20

27. Fujita, M., et al., Therapeutic angiogenesis induced by controlled release of fibroblast

growth factor-2 from injectable chitosan/non-anticoagulant heparin hydrogel in a rat

hindlimb ischemia model. Wound Repair Regen, 2007. 15(1): p. 58-65.

28. DuBose, J.W., C. Cutshall, and A.T. Metters, Controlled release of tethered molecules

via engineered hydrogel degradation: model development and validation. J Biomed

Mater Res A, 2005. 74(1): p. 104-16.

29. Ehrbar, M., et al., Cell-demanded liberation of VEGF121 from fibrin implants induces

local and controlled blood vessel growth. Circ Res, 2004. 94(8): p. 1124-32.

30. Sun, Q., et al., Sustained release of multiple growth factors from injectable polymeric

system as a novel therapeutic approach towards angiogenesis. Pharm Res, 2010. 27(2):

p. 264-71.

31. Layman, H., et al., Co-delivery of FGF-2 and G-CSF from gelatin-based hydrogels as

angiogenic therapy in a murine critical limb ischemic model. Acta Biomater, 2009. 5(1):

p. 230-9.

32. Patil, S.D., F. Papadmitrakopoulos, and D.J. Burgess, Concurrent delivery of

dexamethasone and VEGF for localized inflammation control and angiogenesis. J

Control Release, 2007. 117(1): p. 68-79.

33. Peattie, R.A., et al., Dual growth factor-induced angiogenesis in vivo using hyaluronan

hydrogel implants. Biomaterials, 2006. 27(9): p. 1868-75.

34. Tengood, J.E., et al., Sequential delivery of vascular endothelial growth factor and

sphingosine 1-phosphate for angiogenesis. Biomaterials, 2010. 31(30): p. 7805-12.

35. Sun, G., et al., Functional neovascularization of biodegradable dextran hydrogels with

multiple angiogenic growth factors. Biomaterials, 2011. 32(1): p. 95-106.

36. Silva, E.A. and D.J. Mooney, Effects of VEGF temporal and spatial presentation on

angiogenesis. Biomaterials, 2010. 31(6): p. 1235-41.

37. Tang, D.W., et al., Heparinized chitosan/poly(gamma-glutamic acid) nanoparticles for

multi-functional delivery of fibroblast growth factor and heparin. Biomaterials, 2010.

31(35): p. 9320-32.

38. Garbern, J.C., et al., Delivery of basic fibroblast growth factor with a pH-responsive,

injectable hydrogel to improve angiogenesis in infarcted myocardium. Biomaterials,

2011. 32(9): p. 2407-16.

39. Yu, J., et al., The effect of injected RGD modified alginate on angiogenesis and left

ventricular function in a chronic rat infarct model. Biomaterials, 2009. 30(5): p. 751-6.

21

40. Chen, S.Y., et al., Autologous transplantation of EPCs encoding FGF1 gene promotes

neovascularization in a porcine model of chronic myocardial ischemia. Int J Cardiol,

2009. 135(2): p. 223-32.

41. Duan, H.F., et al., Treatment of myocardial ischemia with bone marrow-derived

mesenchymal stem cells overexpressing hepatocyte growth factor. Mol Ther, 2003.

8(3): p. 467-74.

42. Iwaguro, H., et al., Endothelial progenitor cell vascular endothelial growth factor gene

transfer for vascular regeneration. Circulation, 2002. 105(6): p. 732-8.

43. Matsumoto, R., et al., Vascular endothelial growth factor-expressing mesenchymal

stem cell transplantation for the treatment of acute myocardial infarction. Arterioscler

Thromb Vasc Biol, 2005. 25(6): p. 1168-73.

44. Sun, L., et al., Mesenchymal stem cells modified with angiopoietin-1 improve

remodeling in a rat model of acute myocardial infarction. Biochem Biophys Res

Commun, 2007. 357(3): p. 779-84.

45. Cheng, Z., et al., Targeted migration of mesenchymal stem cells modified with CXCR4

gene to infarcted myocardium improves cardiac performance. Mol Ther, 2008. 16(3):

p. 571-9.

46. Jabbarzadeh, E., et al., Induction of angiogenesis in tissue-engineered scaffolds

designed for bone repair: a combined gene therapy-cell transplantation approach. Proc

Natl Acad Sci U S A, 2008. 105(32): p. 11099-104.

47. Tang, J., et al., Mesenchymal stem cells over-expressing SDF-1 promote angiogenesis

and improve heart function in experimental myocardial infarction in rats. Eur J

Cardiothorac Surg, 2009. 36(4): p. 644-50.

48. Tang, J., et al., Combination of chemokine and angiogenic factor genes and

mesenchymal stem cells could enhance angiogenesis and improve cardiac function after

acute myocardial infarction in rats. Mol Cell Biochem, 2010. 339(1-2): p. 107-18.

49. Zhang, D., et al., Over-expression of CXCR4 on mesenchymal stem cells augments

myoangiogenesis in the infarcted myocardium. J Mol Cell Cardiol, 2008. 44(2): p. 281-

92.

50. Zhu, X.Y., et al., Transplantation of adipose-derived stem cells overexpressing hHGF

into cardiac tissue. Biochem Biophys Res Commun, 2009. 379(4): p. 1084-90.

51. Park, H.J., F. Yang, and S.W. Cho, Nonviral delivery of genetic medicine for

therapeutic angiogenesis. Adv Drug Deliv Rev, 2012. 64(1): p. 40-52.

22

52. Das, H., et al., Stem cell therapy with overexpressed VEGF and PDGF genes improves

cardiac function in a rat infarct model. PLoS One, 2009. 4(10): p. e7325.

53. Kim, S.H., et al., Hypoxia-inducible vascular endothelial growth factor-engineered

mesenchymal stem cells prevent myocardial ischemic injury. Mol Ther, 2011. 19(4): p.

741-50.

54. Yang, F., et al., Genetic engineering of human stem cells for enhanced angiogenesis

using biodegradable polymeric nanoparticles. Proc Natl Acad Sci U S A, 2010. 107(8):

p. 3317-22.

55. Mangi, A.A., et al., Mesenchymal stem cells modified with Akt prevent remodeling and

restore performance of infarcted hearts. Nat Med, 2003. 9(9): p. 1195-201.

56. Shujia, J., et al., Stable therapeutic effects of mesenchymal stem cell-based multiple

gene delivery for cardiac repair. Cardiovasc Res, 2008. 77(3): p. 525-33.

57. Ceradini, D.J., et al., Progenitor cell trafficking is regulated by hypoxic gradients

through HIF-1 induction of SDF-1. Nat Med, 2004. 10(8): p. 858-64.

58. Dimmeler, S., Regulation of bone marrow-derived vascular progenitor cell mobilization

and maintenance. Arterioscler Thromb Vasc Biol, 2010. 30(6): p. 1088-93.

59. Chen, F.M., et al., Homing of endogenous stem/progenitor cells for in situ tissue

regeneration: Promises, strategies, and translational perspectives. Biomaterials, 2011.

32(12): p. 3189-209.

60. Petit, I., et al., G-CSF induces stem cell mobilization by decreasing bone marrow SDF-

1 and up-regulating CXCR4. Nat Immunol, 2002. 3(7): p. 687-94.

61. Pitchford, S.C., et al., Differential mobilization of subsets of progenitor cells from the

bone marrow. Cell Stem Cell, 2009. 4(1): p. 62-72.

62. Orlic, D., et al., Mobilized bone marrow cells repair the infarcted heart, improving

function and survival. Proc Natl Acad Sci U S A, 2001. 98(18): p. 10344-9.

63. Zaruba, M.M., et al., Synergy between CD26/DPP-IV inhibition and G-CSF improves

cardiac function after acute myocardial infarction. Cell Stem Cell, 2009. 4(4): p. 313-

23.

23

Figure 2.1 Methods for incorporating biological signals into scaffolds for controlled release. A)

Biologics encapsulated directly via physical entrapment. B) Biologics covalently tethered to

polymer chains. C) Biologics loaded into microspheres and then encapsulated into hydrogel

scaffolds. Reproduced with permission from ref [19].

24

Figure 2.2 A) Polymeric scaffolds for dual delivery of VEGF and PDGF. (B) In vitro release

kinetics of VEGF from scaffolds fabricated from PLGA (85:15, lactide:glycolide), measured

using 125I-labeled tracers. (C) In vitro release kinetics of PDGF pre-encapsulated in PLGA

microspheres ( 75:25, intrinsic viscosity = 0.69 dl/g; 75:25, intrinsic viscosity = 0.2 dl/g),

before scaffold fabrication. D) Dual delivery led to early maturation of blood vessels as

demonstrated by α-smooth muscle actin staining at 2 weeks. E) Representative micro-CT

images after 5 weeks intramuscular injections of alginate (Blank), alginate containing VEGF165

(VEGF), or alginate containing VEGF165 combined with PLGA microspheres containing

PDGF-BB (VEGF/PDGF).(A-D) Reproduced with permission from ref [11] (E) Reproduced

with permission from ref [30].

25

Figure 2.3 Effects of VEGF temporal presentation on angiogenesis. Varying amounts of VEGF

is applied to human microvascular endothelial cells (HMVEC) over five days in a sprouting

assay. Each color bar represents different temporal presentations of VEGF. Red bars represent

the same concentration of VEGF applied daily over five days; blue bars represent decreasing

VEGF dose over five days; green bars represent gradual decrease in VEGF concentration; and,

brown bars represent increasing concentrations of VEGF applied. These are represented in the

lower graphs. As shown in the larger graph, decreasing the VEGF dose from an initial high

concentraion (blue bars) induced a greater number of endothelial cells sprouts, as compared to

constant VEGF doses (50 ng/ml day) (red bars), increasing VEGF doses (brown bars), or a

gradual VEGF dose decrease over time (green bars). Reproduced with permission from ref [36].

26

Figure 2.4 Environmental-responsive controlled release system for sustained bFGF delivery in

ischemic tissue triggered by pH. TEM micrography of heparinized chitosan/poly(gamma-

glutamic acid) (HP-CS/g-PGA) nanoparticles at a distinct pH value: (A) pH 6.0 (after 2 h), (B)

pH 7.4 (after 10 min), (C) pH 7.4 (after 30 min), (D) pH 7.4 (after 2 h). (E) Release of bFGF or

heparin from the smart nanoparticles, depending on the environmental pH variation.

Reproduced with permission from ref [37].

27

Figure 2.5 Cell-triggered VEGF delivery using biomimetic hydrogels. A) Schematic of MMP-

degradable PEG hydrogels containing RGD peptide and covalently linked VEGF. B) Cell-

triggered VEGF delivery led to targeted angiogenesis within the hydrogel region, whereas

soluble VEGF treatment resulted in diffusive and uncontrolled angiogenesis. Reproduced with

permission from ref [20].

28

Figure 2.6 Combined stem cell and gene therapy approach for therapeutic angiogenesis (A).

Transplantation of genetically modified stem cells into ischemic tissues led to enhanced

paracrine secretion and angiogenesis in situ. Reproduced with permission from ref [51] (B).

VEGF-overexpressing MSCs using biodegradable polymeric nanoparticles led to enhanced

angiogenesis and limb salvage in a murine hindlimb ischemia model (C). Reproduced with


29

Figure 2.7 Exploiting cell homing for treating acute myocardial infarction. A) Signaling

pathway related to stem cell homing to ischemic tissue. G-CSF promoted cell mobilization and

Diprotin A (Dip) inhibited SDF-1 protease, thereby maintaining the stability of cell homing

factor SDF-1 released from ischemic tissue. B) G-CSF and Dip dual treatment led to

significantly increased myocardial stem cells detected in the ischemic tissue. C) G-CSF and Dip

dual treatment increased mouse survival after myocardial infarction. Reproduced with


30

3. Overexpressing Angiogenic Growth Factors in Adipose-Derived Stem Cells using

Biodegradable Polymeric Nanoparticles

Lorenzo Deveza1, 2, Jeffrey Choi3, Fan Yang2,4 ,*

1 School of Medicine, Stanford University, Stanford, CA, 94305, USA; 2 Department of

Bioengineering, Stanford University, Stanford, CA, 94305, USA; 3 Department of Biology,

Stanford University, Stanford, CA, 94305, USA; 4 Department of Orthopaedic Surgery, Stanford

University, Stanford, CA, 94305, USA

Abstract

Stem cell-based therapies are promising candidates for treating ischemic cardiovascular

diseases by promoting angiogenic reperfusion of the affected tissues. Genetic modification of

stem cells may further promote their ability for ischemic tissue repair by strategically

overexpressing relevant therapeutic genes, and may also enable stem cells to present a more

consistent phenotype across various autologous sources. Amongst the available cell sources

adipose-derived stem cells (ADSCs) are a clinically abundant cell source that are easy to isolate

and show promise for ischemic tissue repair. However, non-viral transfection of ADSCs has

been difficult to achieve using commercially available polymeric vectors. In this study, we

assess and optimize non-viral transfection of ADSCs using optimized biodegradable polymeric

nanoparticles, known as poly(β-amino esters). We demonstrate that transfection varies across

multiple cell isolations of ADSCs, which can be partly rescued by optimizing culture conditions,

such as adding 10 ng/ml basic fibroblast growth factor (bFGF). Transfection of ADSCs with

therapeutic genes is then demonstrated using eight different plasmids encoding for three

different angiogenic factors, vascular endothelial growth factor (VEGF), hepatocyte growth

factor (HGF), and CXC chemokine receptor 4 (CXCR4). Paracrine release from transfected

ADSCs is then evaluated for promoting endothelial cell and smooth cell viability in vitro.

3.1 Introduction

Stem cells represent potential therapeutic options for tissue regeneration and repair,

which may act via two main mechanisms: 1) engraftment into the developing tissue; and, 2)

delivery of paracrine signals to promote repair. In particular, the second mechanism is

increasingly believed to be more likely in treating ischemic muscle and myocardium in a process

called therapeutic angiogenesis; as few cells are found to exhibit long-term cell engraftment and

31

a similar therapeutic response is achieved by delivering stem cell conditioned medium [1].

Despite the purported role of stem cells as delivery vehicles, stem cell therapies have had limited

clinical success, which may be due to issues of variability in autologous cell source and

insufficient paracrine release. Genetic engineering of stem cells may help to overcome these

issues by reinforcing the overexpression of desired therapeutic factors across multiple cell

sources, and in effect, make them more consistent. To achieve such an approach for clinical

usage, what is desired is a readily available cell source and a safe transfection system of relevant

therapeutic genes.

Adipose-derived stem cells (ADSCs) are a clinically abundant cell source that are easy

to isolate and have the potential to differentiate towards mesenchymal lineages [2]. Importantly,

ADSCs have been shown to secrete a number of angiogenic and anti-apoptotic growth factors

[3, 4] and have demonstrated the ability promote therapeutic angiogenesis in vivo [5]. Genetic

modification of ADSCs has also been shown to promote their therapeutic capabilities using

various angiogenic factors with beneficial effect, including vascular endothelial growth factor

(VEGF) [6], hepatocyte growth factor (HGF) [7], and CXC chemokine receptor 4 (CXCR4) [8].

While these approaches have shown promise for the therapeutic use of ADSCs, their clinical

usage remains hampered by the use of viruses to achieve gene transfection, due to issues of

immunogenicity and insertional mutagenesis.

There are several non-viral approaches for gene transfection, which may overcome the

concerns associated with viral-based vectors. Amongst such approaches, cationic polymers,

such as Lipofectamine (Lipo) and polyethylenimine (PEI) enable gene delivery by forming

complexes with plasmid DNA to generate nanoparticles, which protect plasmid DNA from

extracellular degradation and enable passage across cell membranes. However, in contrast to

viral-based vectors, many non-viral approaches suffer low transfection efficiency and high

cytotoxicity. Poly(β-amino esters) (PBAE) are a class of biodegradable cationic polymers for

gene delivery that were screened in a relatively high-throughput, combinatorial fashion for

achieving high transfection efficiency with low cytotoxicity in human cells [9, 10], and further

screened for transfection into stem cells [11]. Using PBAE, high transfection efficiencies were

also achieved in human ADSCs [11-13]. However, transfection was only explored in ADSCs

of one cell isolation in these studies. It is well known that transfection efficiency is cell type

and cell isolation specific and for clinical usage, it is important to understand how ADSCs of

various isolations might transfect using an optimal non-viral vector, such as PBAE.

32

In this study, we assess transfection of various ADSC cell isolations using PBAE

polymeric nanoparticles complexed with plasmids encoding for VEGF, HGF, and/or CXCR4.

We first assess two different PBAE polymers for transfection efficiency and cell cytoxicity. We

then optimize the culture conditions by exploring tranfection in the presence or absence of basic

fibroblast growth factor (bFGF) and fetal bovine serum (FBS). Lastly, ADSCs are transfected

with various angiogenic factors and assessed for mRNA expression, protein production, and

bioactivity of paracrine signals against human umbilical vein endothelial cells (HUVECs) and

human aortic smooth muscle cells (HASMCs).

3.2 Materials and Methods

3.2.1 Materials

Monomers were purchased from Sigma-Aldrich (St Louis, MO, USA), Scientific

Polymer Products Inc. (Ontario, NY, USA) and Molecular Biosciences (Boulder, CO, USA).

All chemicals were used as received. Anhydrous dimethyl sulfoxide was purchased from

(Sigma-Aldrich). A 25-mM sodium acetate buffer solution pH5.2 (NaAc buffer) was prepared

by diluting a 3-M stock (Molecular Biosciences). Plasmid DNAs pGluc and pcDNA3.1 was

obtained from Elim Biopharmaceuticals (Hayward, CA, USA). Plasmid DNA pBLAST-VEGF

was obtained from Aldeveron (Fargo, ND, USA). Plasmid DNA pCK-HGF was obtained from

Viromed (Seoul, Korea). Plasmid DNAs pCMV-VEGF, pCMV-CXCR4/VEGF, pCMV-

CXCR4, pCK-CXCR4/HGF were cloned in house.

3.2.2 Polymer Synthesis

Acrylate-terminated (C32-Ac and D32-Ac) polymer were synthesized by mixing

butanediol diacrylate (C) or hexanediol (D) with aminopentanol (32) in a 1.2:1.0

diacrylate/amine molar ratio at 90 ºC for 24 h to generate acrylate-terminated polymer chains

(Figure 1A, 1B). Amine-capped polymers (C32-122 and D32-122) were prepared by

subsequently reacting C32-Ac or D32-Ac with diamine monomers (122) in dimethyl sulfoxide

(Figure 1A, 1B). End-capping reactions were performed overnight at room temperature using a

1.6-fold molar excess of amine over acrylate end groups.

3.2.3 Cell Isolation and Culture

Human adipose-derived stem cells (hADSCs) were isolated from fat tissues that were

obtained from the abdominal fat of three separate female patients who had undergone a free flap

33

breast reconstruction surgery at Stanford University. All procedures were approved and guided

under Stanford IRB protocol. Mouse adipose-derived stem cells (mADSCs) were isolated from

the inguinal fat pad of FVB/NJ mice. Fat tissues were washed 2-3 times with Phosphate

buffered saline and digested at 37 °C for 30 min with 0.5 U/ml Blendzyme 3 (Roche

Diagnostics, Indianapolis, IN, USA). Enzyme activity was neutralized with DMEM, containing

10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA, USA), 1% Penicillin/Streptomycin.

Cells were then filtered through a 100 µm nylon mesh to remove cellular debris before seeding.

Following the initial 48 hours of incubation at 37 °C and 5% CO2, cells were washed with PBS

and maintained in growth medium containing DMEM, 10% FBS, 1% penicillin/ streptomycin

and 10ng/ml of the basic fibroblast growth factor (FGF) (PeproTech, Rocky Hill, NJ, USA).

Once 85-95% confluent, the cells were passaged by trypsin (0.05%) digestion and plated at a

density of 5000-6000 cells per cm2 for further expansion. Human umbilical vein endothelial

cells (HUVECs) and human aortic smooth muscle cells (HASMCs) were obtained from Lonza

(Basal, Switzerland) and expanded utilizing endothelial growth medium (EGM-2) or smooth

muscle growth media (Lonza).

3.2.4 Transfection

Cells were transfected with the various DNA plasmids shown in Table 1. Passage 3-4

ADSCs were seeded at 12 000 cells per well in clear 96 well plates or 300 000 cells per well in

clear 6 well plates and incubated over night prior to transfection. Cells were then transfected

utilizing poly (β-amino esters) (PBAE) C32-122 or D32-122. Briefly, polyplexes were formed

by mixing pDNA and PBAE polymers in 25 mM sodium acetate at a 20:1 polymer to DNA

weight ratio. Polyplexes were added to the cell culture after 10 minutes of complexation. Cells

were then incubated with polyplexes for four hours, at which time cell medium was replaced.

Effects of culture media on transfection was assessed by using different culture conditions

during four hour incubation, as well as after replacement. The effects of basic fibroblast growth

factor (bFGF) on transfection was assessed by transfecting cells with and without 10 ng/ml

bFGF. Effects of replacement media was assessed by varying the serum concentration (0%,

1%, 5%, 10%) and bFGF concentration (0 ng/ml, 1 ng/ml, 5 ng/ml, 10 ng/ml) in the media used

for four hour replacement.

34

3.2.5 Luciferase Expression Assay

Cells transfected with plasmid DNA encoding for Gaussia Luciferase (pGLUC) were

assessed by collecting the cell supernatant at 48 hours post-transfection and assaying for

secreted luciferase. Cell supernatant was diluted 10-fold in PBS and added to opaque 96-well

plates (Falcon). The substrate Luciferin (New England Biolabs) was diluted in buffer according

to the manufacturers instructions, and added at 1:1 ratio to the diluted cell supernatant.

Luminescence was immediately measured using a plate reader.

3.2.6 Cell Viability and Proliferation

To measure cytotoxicity and cell viability, cellular metabolic activity was measured

using the Cell Titer 96 Aqueous One Solution assay kit (Promega). Metabolic activity was

measured using optical absorbance on a Multi-label plate reader (Perkin Elmer Life Sciences,

Boston, MA, USA). Measurements of treated cells were converted to percentage viability in

comparison with untreated controls

3.2.7 RNA Isolation, Reverse Transcription, and quantitative PCR

To confirm successful transfection by PBAE in vitro, total RNA was harvested from

cells at 48 hours post-transfection using RNeasy Kit (Qiagen, Germantown, MD, USA). cDNA

was synthesized by reverse transcription using the Superscript First-Strand Synthesis System

(Invitrogen) and 1 μg of total RNA. Quantitative PCR was performed using Power SYBR Green

PCR Master Mix following the manufacturer's protocol. Primers used are shown in Table 2.

All samples were run for 40 PCR cycles on Applied Biosystems 7900 Real-Time PCR System

(Carlsbad, CA, USA). The relative expression level of target genes was determined by the

ΔΔCT method, where target gene expression was first normalized to an endogenous

housekeeping gene, GAPDH, followed by a second normalization to the gene expression level

measured in untransfected control cells.

3.2.8 VEGF and HGF ELISA

VEGF and HGF production were assessed by collecting conditioned medium and

measuring using an ELISA kit for human VEGF (Peprotech, Rocky Hill, NJ) and human HGF

(R&D Systems, Minneapolis, MN) according to the manufacturer’s protocol. The level of

secreted VEGF and HGF was expressed as the amount of VEGF per milliliter of medium

supernatant.

35

3.2.9 HUVEC and HASMC viability assay

HUVECs or HASMCs were plated in gelatin-coated 96 well plates (11 200 cells/well).

HUVECs and HASMCs were cultured in low serum medium (DMEM + 2% FBS + 1% P/S) for

24 hours prior to switching to conditioned medium. HUVEC and HASMC viability was

determined by MTS assay as described above.

3.2.10 Statistical Analysis

All experiments were performed in triplicate and repeated for consistency. The results

were reported as mean ± standard error of the mean. Statistical analysis was performed using

the one way analysis of variance with Dunnett’s multiple comparison test as the post-test to

compare all the experimental groups.

3.3 Results

3.3.1 Polymer Synthesis

Poly(β-amino esters) were synthesized by the conjugate addition of amine monomers

to diacrylates. For the synthesis of acrylate-terminated C32 (C32-Ac) or D32 (D32-Ac) ,

butanediol diacrylate (C) or hexanediol diacrylate (D) was mixed in a 1.2:1.0 molar ratio with

aminopentanol (32; Figure 1). End-modified polymers were synthesized by subsequently

reacting C32-Ac or D32-Ac individually with one diamine monomer (122) to yield end-

modified C32-122 (Figure 1A) or D32-122 (Figure 1B) polymers.

3.3.2 Assessment of transfection of ADSCs using C32-122 or D32-122

Two end-modified PBAE polymers (C32-122 and D32-122) were compared for

transfection into five different cell cultures: 3 human ADSCs from different patients, 1 mouse

ADSC cell source, and HEK293T cells as a control. Transfections were assessed for cell

viability and transfection efficiency. A polymer to DNA weight ratio of 20:1 and a DNA dose

of 5 µg per 70 000 cells was used for all transfections. PBAE-mediated transfection of all cells

caused a change in cell morphology in comparison to non-transfected controls (Figure 2A).

Control images demonstrate that normal ADSC morphology was spindle shaped, while

transfected ADSCs had a cobblestone pattern, suggesting some change in cell behavior. When

assessing for cell viability, there was no significant difference at 48 hours post-transfection for

all cells in comparison to controls. Transfection efficiency, as assessed using a Gaussia

36

Luciferase reporter plasmid, demonstrated that C32-122 led to higher overall transfection

compared to D32-122. D32-122 was unable to efficiently and consistently transfect hADSCs.

However, transfection by D32-122 was successful in HEK293T cells. We found D32-122

transfection to be variable based on the batch. Transfection efficiency also varied by cell type

and culture. Human ADSCs transfected less efficiently than mouse ADSCs and HEK293T cells.

Amongst the hADSC isolations, hADSC-I3 transfected the most efficiently, while transfection

in hADSC-I2 was low or undetectable.

3.3.3 Enhancing ADSC proliferation and transfection with bFGF

Considering the difference in transfection potential in each hADSC cell isolation and

known reliance of non-viral transfection on cell proliferation, the various hADSC cell isolations

were assessed for their ability to proliferate (Figure 3A, 3B, 3C). We found that transfection

efficiency of each hADSC cell isolation correlated with their relative ability to proliferate, as

demonstrated by BrDu assay (Figure 1A), MTS assay (Figure 1B), and Picogreen DNA content

assay (Figure 1C). Furthermore, hADSC proliferation could be enhanced in the presence of 10

ng/ml basic fibroblast growth factor (a common component of cell growth medium) (Figure 1A,

1B, 1C). This treatment improved transfection efficiency in two of the three hADSC cell

isolations (hADSC-I1 and hADSC-I3), as demonstrated by green fluorescence protein

transfection (Figure 3D) and luciferase assay (Figure 3E). Transfection efficiency in hADSC-

I2 remained low despite the proliferation increase by 10 ng/ml bFGF, suggesting that promoting

proliferation does not wholly explain the ability to transfect.

3.3.4 Effects of post-transfection culture conditions

To determine the effects of culture media on ADSC transfection after incubating with

PBAE nanoparticles, ADSC-I3 was transfected and subsequently placed in media containing

varying amounts of FBS (0%, 1%, 5%, 10%) and bFGF (0 ng/ml, 1 ng/ml, 5 ng/ml, 10 ng/ml).

The amount of FBS had the largest effect on ADSC viability (Figure 4A) and transfection

(Figure 4B). The effects of bFGF on ADSC viability and transfection relied on the

concentration of FBS in the media. Interestingly, bFGF only affected viability and transfection

in the presence of 10% FBS, but did not lead to any significant effect otherwise (Figure 4A,

4B). In the presence of 10% FBS, as little as 1 ng/ml bFGF led to a significant increase in

transfection. Also, when bFGF was not added, there was no increase in transfection by

increasing FBS above 5%.

37

3.3.5 Upregulation of various angiogenic growth factors in ADSCs

To assess the ability to transfect ADSCs with various angiogenic factors, ADSCs were

transfected with seven different plasmids encoding functional genes and one backbone plasmid

(pcDNA3.1) (Table 1). All transfections were performed using C32-122 and in the presence of

10 ng/ml bFGF. Of the plasmids transfected, two encoded for vascular endothelial growth factor

(VEGF) which differed based on promoter, one encoded for CXC chemokine receptor 4

(CXCR4) only, one encoded for hepatocyte growth factor (HGF) only, one encoded for both

CXCR4 and VEGF, and one encoded for CXCR4 and HGF. The plasmids mainly differed in

promoters and size. Transfection with each of these plasmids were confirmed by RT-PCR

(Figures 5A, 5B, 5C) and ELISA (Figures 5D, 5E). As further confirmation of the above

luciferase data, human ADSCs transfected less well than mouse ADSCs and HEK293T cells.

Protein production from human ADSCs was low overall, and transfection led to increases that

were significant (p < 0.05). VEGF upregulation using all three plasmids encoding for VEGF

was comparable, suggesting little differences in transfection due to promoter and plasmid size

(Figure 5D). HGF upregulation using the two plasmids encoding for HGF was similarly

comparable (Figure 5E). The highest amounts of VEGF and HGF production occurred by

mouse ADSCs and HEK293T, reaching over 10 ng/ml and 50 ng/ml for VEGF and HGF,

respectively.

3.3.6 Effects of transfected ADSC paracrine release on HUVEC and HASMC viability

To determine whether growth factor production from ADSCs was bioactive,

conditioned media (CM) was collected from transfected ADSCs at 48 hours post-transfection

and applied to human umbilical vein endothelial cells (HUVECs) and human aortic smooth

muscle cells (HASMCs), then assessed for viability. CM from mADSCs and HEK293T cells

did increase HUVEC viability that was comparable to 10 ng/ml VEGF or 10 ng/ml HGF and

was significantly higher than transfection controls (pGLUC) and non-transfection controls (cells

only; Figure 6A). However, CM from hADSCs did not increase HUVEC (Figure 6A) or

HASMC (Figure 6B) viability as compared to basal media controls. No effects on HASMC

viability were observed due to VEGF and HGF protein controls, so as can be expected there

was no effect on HASMC viability due to CM from any transfected cell type (Figure 6B). The

ability of transfected mADSCs and HEK293T cells to elicit effects on HUVEC viability,

suggested that the effects of released growth factor was bioactive and concentration dependent.

38

3.4 Discussion

Herein, we demonstrated the transfection of various isolations of ADSC using PBAE

biodegradable nanoparticles with eight different plasmid DNAs encoding angiogenic factors or

reporter genes. We found that the PBAE polymer C32-122 achieved higher transfection in the

presence of 10 ng/ml bFGF, and further found that the culture conditions post-transfection also

affected transfection efficiencies and cell viabilities. Under these conditions, the transfection

of various angiogenic factors into the ADSCs using multiple plasmids was successfully

demonstrated. Notably, transfection efficiencies amongst various hADSC isolations varied, and

was overall lower than mADSCs and HEK293T cells. This was apparent when paracrine release

from transfected ADSCs were applied to HUVECs, where only mADSCs and HEK293T

paracrine release led to significant increases in viability.

Using the three different isolations of hADSCs, we determined that ADSCs are difficult

to transfect and further that their ability to transfect varied. At least one characteristic that

differed amongst these hADSCs was their proliferative potential, which correlated with their

ability to transfect. This is not surprising, as it is fairly well-established that non-viral

transfection relies on cell proliferation. [14] In lieu of this correlation, bFGF was used to

enhance ADSC proliferation, which also promoted ADSC transfection in two of three cell

isolations. Curiously, hADSC-I2 increased in proliferation in the presence of bFGF, but this

could not rescue their ability to transfect. This suggests that proliferation does not fully explain

transfection potential. Still, the idea of promoting proliferation to enhance transfection had not

yet been rigorously explored previously to our knowledge and does open up possibilities for

further exploration. Considering that two of three hADSCs responded well to bFGF suggests

that despite initial differences, select autologous hADSCs can be genetically modified to present

a similar phenotype.

Transfection into ADSCs was demonstrated using eight different plasmids, which

encoded for four different genes (VEGF, HGF, CXCR4, GLUC), that contained different

promoters, had different molecular weights, and were either singly expressing or dually

expressing. While each gene was able to be transfected into ADSCs, we found that there were

some differences in their level of expression. For example, pck-HGF led to much higher

increases in expression as compared to pBLAST-VEGF or pCDNA-VEGF. This is likely

plasmid specific, whereby pck-HGF is a highly engineered plasmid [15], whereas pBLAST-

39

VEGF and pCDNA-VEGF have not been fully optimized. This suggests that plasmid

development is another strategy to promote non-viral mediated transfection.

Considering that endothelial cell and smooth muscle cell viability are two cellular

aspects of angiogenesis, the effects of paracrine release from transfected cells were assessed

against HUVECs and HASMCs. The robust response of HUVECs by paracrine release from

transfected mADSC and HEK293T cells but not from the untransfected cells, suggested that

non-viral gene upregulation of VEGF and HGF can successfully promote angiogenic behavior.

However, considering the lower transfection efficiency of hADSCs, this effect was not observed

by transfected hADSCs, suggesting that the concentration of angiogenic growth factors were

too low. This could also be due to using conditioned medium, as opposed to a co-culture system,

which suffers from some limitations due to protein degradation and does not enable potential

cell-cell interactions. Further study may explored a co-culture system of transfected hADSCs

with HUVEC or HASMCs.

In conclusion, non-viral gene transfection of hADSCs can be achieved using PBAE

biodegradable nanoparticles and may promote more consistent behavior amongst various

hADSC cell isolations. Transfection of hADSCs is difficult, but could be improved by

polymeric vector, transfection conditions, and plasmid design. Overexpression of angiogenic

factors can promote endothelial cell survival in vitro. Overall, PBAE-mediated transfection of

hADSCs using angiogenic factors may be a strategy for promoting consistent hADSC behavior

towards therapeutic angiogenesis.

References

1. Ranganath, S.H., et al., Harnessing the mesenchymal stem cell secretome for the

treatment of cardiovascular disease. Cell Stem Cell, 2012. 10(3): p. 244-58.

2. Zuk, P.A., et al., Multilineage cells from human adipose tissue: implications for cell-

based therapies. Tissue Eng, 2001. 7(2): p. 211-28.

3. Thangarajah, H., et al., IFATS collection: Adipose stromal cells adopt a proangiogenic

phenotype under the influence of hypoxia. Stem Cells, 2009. 27(1): p. 266-74.

4. Rehman, J., et al., Secretion of angiogenic and antiapoptotic factors by human adipose

stromal cells. Circulation, 2004. 109(10): p. 1292-8.

5. Miranville, A., et al., Improvement of postnatal neovascularization by human adipose

tissue-derived stem cells. Circulation, 2004. 110(3): p. 349-55.

40

6. Jabbarzadeh, E., et al., Induction of angiogenesis in tissue-engineered scaffolds

designed for bone repair: a combined gene therapy-cell transplantation approach. Proc

Natl Acad Sci U S A, 2008. 105(32): p. 11099-104.

7. Zhu, X.Y., et al., Transplantation of adipose-derived stem cells overexpressing hHGF

into cardiac tissue. Biochem Biophys Res Commun, 2009. 379(4): p. 1084-90.

8. Cho, H.H., et al., Overexpression of CXCR4 increases migration and proliferation of

human adipose tissue stromal cells. Stem Cells Dev, 2006. 15(6): p. 853-64.

9. Green, J.J., et al., Combinatorial Modification of Degradable Polymers Enables

Transfection of Human Cells Comparable to Adenovirus. Advanced Materials, 2007.

19(19): p. 2836-2842.

10. Zugates, G.T., et al., Rapid optimization of gene delivery by parallel end-modification

of poly(beta-amino ester)s. Mol Ther, 2007. 15(7): p. 1306-12.

11. Yang, F., et al., Gene delivery to human adult and embryonic cell-derived stem cells

using biodegradable nanoparticulate polymeric vectors. Gene Ther, 2009. 16(4): p. 533-

46.

12. Nauta, A., et al., Adipose-derived stromal cells overexpressing vascular endothelial

growth factor accelerate mouse excisional wound healing. Mol Ther, 2013. 21(2): p.

445-55.

13. Deveza, L., et al., Paracrine release from nonviral engineered adipose-derived stem cells

promotes endothelial cell survival and migration in vitro. Stem Cells Dev, 2013. 22(3):

p. 483-91.

14. Brunner, S., et al., Cell cycle dependence of gene transfer by lipoplex, polyplex and

recombinant adenovirus. Gene Ther, 2000. 7(5): p. 401-7.

15. Pyun, W.B., et al., Naked DNA expressing two isoforms of hepatocyte growth factor

induces collateral artery augmentation in a rabbit model of limb ischemia. Gene Ther,

2010. 17(12): p. 1442-52.

41

Figures

Figure 3.1. Synthesis of end-modified poly(β-amino esters) (PBAE). (A) Shows the synthesis

of C32-Ac by mixing 1,4-butanediacrylate (C) with 1-aminopentanol in 1.2:1 molar ratio, then

shows the synthesis of C32-122 by the addition of 122 in excess molar ratio to C32-Ac. (B)

Shows the synthesis of D32-122, which is synthesized in the same manner as C32-122, except

using 1,6 hexanediacrylate (D).

42

Figure 3.2. Effects of polymer on transfection and viability of human adipose-derived stem

cells (hADSCs), mouse adipose-derived stem cells (mADSCs), and human embryonic kidney

cells (HEK293T). (A) Images demonstrating the morphology and observable health of the cells

at 48 hours post-transfection with either C32-122 or D32-122. (B) MTS cell viability assay of

transfected cells at 48 hours. (C) Luciferase release from transfected cells.

43

Figure 3.3. Enhancement of hADSC proliferation and transfection in the presence of basic

fibroblast growth factor. (A) BrDu cell proliferation assay, (B) MTS cell viability assay, and

(C) Picogreen DNA content assay, on various isolations of hADSC at 48 hours with and without

10 ng/ml bFGF. (D) Fluorescence images of various isolations of ADSCs transfected with

EGFP with and without 10 ng/ml bFGF. (E) Luciferase release from various isolations of

ADSCs transfected with Gaussia Luciferase with and without 10 ng/ml bFGF.

44

Figure 3.4. Effects of culture conditions post-transfection. (A) MTS cell viability assay and

(B) luciferase release from hADSC-I3 transfected with Gaussia Luciferase and then cultured

with various concentrations of fetal bovine serum (FBS; 0%, 1%, 5%, 10%) and various

concentrations of bFGF (0 ng/ml, 1 ng/ml 5 ng/ml 10 ng/ml).

45

Figure 3.5. Transfection of cells with different plasmids encoding for multiple genes, including

vascular endothelial growth factor (VEGF), hepatocyte growth factor (HGF), CXC chemokine

receptor 4 (CXCR4). (A-C) RT-PCR mRNA expression assessment of transfected cells for

VEGF (A), HGF (B), and CXCR4 (C). (D-E) ELISA protein release assessment of transfected

cells for (D) VEGF, and (E) HGF.

46

Figure 3.6. Effects of transfected cell paracrine release on human umbilical vein endothelial

cell (HUVEC) and human aortic smooth muscle cell (HASMC) viability. Conditioned medium

was collected from cells transfected with different plasmids at 48 hours and then applied to

HUVEC and HASMCs. (A) HUVEC and (B) HASMC cell viability was then assessed at 48

hours by MTS assay.

47

Table 3.1: List of plasmids used for transfection

Plasmid ID Genes encoded

pBlast-VEGF Vascular Endothelial Growth Factor 165

pCMV-VEGF Vascular Endothelial Growth Factor 165

pCK-HGF Hepatocyte Growth Factor 723 & 728

pCMV-CXCR4 CXC Chemokine Receptor 4

pCMV-CXCR4/VEGF CXC Chemokine Receptor 4 + Vascular Endothelial Growth Factor 165

pCK-CXCR4/HGF CXC Chemokine Receptor 4 + Hepatocyte Growth Factor 723 & 728

pGLUC Gaussia Luciferase

pCDNA-3.1 Blank Vector

Table 3.2: List of primers used for quantitative real time-polymerase chain reaction

Gene

Name Species Primer Sequence

hGAPDH Human Forward (5'-ACAGTCAGCCGCATCTTCTT-3')

Reverse (5'-CGACCAAATCCGTTGACTC-3')

mGAPDH Mouse Forward (5'-AACGACCCCTTCATTGAC-3')

Reverse (5'-TCCACGACATACTCAGCAC-3')

hCXCR4 Human Forward (5’- TTCTACCCCAATGACTTGTG-3’)

Reverse (5’- ATGTAGTAAGGCAGCCAACA-3’)

hVEGF Human Forward (5’-TACCTCCACCATGCCAAGTG-3’)

Reverse (5’-TGATGATTCDTGCCCTCCTCC-3’)

hHGF Human Forward (5'-CAATAGCATGTCAGGTGGAG-3')

Reverse (5'-CTGTGTTCGTGTGGTATCAT-3')

48

4. Paracrine Release from Non-viral Engineered Adipose-Derived Stem Cells Promotes

Endothelial Cell Survival and Migration in vitro

Lorenzo Deveza1,2, Jeffrey Choi3, Galym Imanbayev 4, Fan Yang PhD2,5,*

1School of Medicine, Stanford University, Stanford, CA; 2 Department of Bioengineering,

Stanford University, Stanford, CA; 3Department of Biology, Stanford University, Stanford,

CA; 4Department of Economics, Stanford University, Stanford, CA; 5Department of

Orthopaedic Surgery, Stanford University, Stanford, CA

Abstract

Stem cells hold great potential for therapeutic angiogenesis due to their ability to directly

contribute to new vessel formation or secrete paracrine signals. Adipose-derived stem cells

(ADSCs) are particularly attractive autologous cell sources for therapeutic angiogenesis due to

their ease of isolation and relative abundance. Gene therapy may be used to further enhance

the therapeutic efficacy of ADSCs by over-expressing desired therapeutic factors. Here we

developed VEGF overexpressing ADSCs utilizing poly(β-amino esters) (PBAE), a

biodegradable polymeric nanoparticle, and examined the effects of paracrine release from

non-viral modified ADSCs on the angiogenic potential of human umbilical vein endothelial

cells (HUVEC) in vitro. PBAE polymeric vectors deliver DNA into ADSCs with high

efficiency with low cytotoxicity, and led to over 3-fold in VEGF production by ADSCs

compared to Lipofectamine 2000. Paracrine release from PBAE/VEGF transfected ADSCs

enhanced HUVECs viability and decreased HUVEC apoptosis under hypoxia. Furthermore,

paracrine release from PBAE/VEGF transfected ADSCs significantly enhanced HUVEC

migration and tube formation, two critical cellular processes for effective angiogenesis. Our

results demonstrate that genetically engineered ADSCs using biodegradable polymeric vectors

may provide a promising autologous cell source for therapeutic angiogenesis in treating

cardiovascular diseases.

4.1 Introduction

Therapeutic angiogenesis is a promising treatment approach for patients suffering

from ischemic cardiovascular disease. This approach involves methods of regenerative

medicine to augment the host angiogenic response. Angiogenesis is the physiological process

of blood vessel growth from pre-existing vessels and is the body’s natural mechanism for

49

restoring blood perfusion to sites of ischemia. Stem cell-based therapies hold great promise for

therapeutic angiogenesis by either acting as cellular building blocks for new blood vessels or

by secreting paracrine signals to promote blood vessel growth [1-2]. The most well studied

cell sources for therapeutic angiogenesis include endothelial progenitor cells (EPCs) and bone

marrow-derived mesenchymal stem cells (BMMSCs) [3]. Despite the therapeutic efficacy of

these cell sources [3-7], their broad clinical application is largely limited due to requiring an

invasive cell isolation procedure with low cell yield, and the need for ex vivo cell expansion

prior to transplanting back to the patient.

Adipose-derived stem cells (ADSCs) represent a promising alternative cell source due

to their relative abundance, ease of isolation, and demonstrated potential for therapeutic

angiogenesis [8-9]. Indeed, when directly compared to BMMSCs, ADSCs were reported to

possess higher angiogenic potency demonstrating improved blood perfusion and a reduction in

infarct size two weeks post-transplantation in a mouse model of hindlimb ischemia [10].

Among the mechanisms by which stem cells contribute to angiogenesis, paracrine signaling is

regarded is being a key contributor [5,11]. In support of this, it was shown that direct

injection of conditioned medium (CM) from BMMSC versus BMMSC transplantation into a

rat model of myocardial ischemia led to similar reductions in infarct size and improvements in

ventricular function by 72 hours [12]. A broad spectrum of angiogenic cytokines were found

to be present in the conditioned medium of BMMSCs including vascular endothelial growth

factor (VEGF), fibroblast growth factor-2 (FGF-2), interleukin-6, and monocyte chemotactic

protein-1 (MCP-1) [6]. Similarly, ADSCs have been shown to release a number of angiogenic

and anti-apoptotic paracrine factors [13], which lead to enhanced new blood vessel formation

in vivo [13-18]. Despite the promise of stem cell-based therapies for therapeutic angiogenesis,

the efficacy of using stem cells alone remains limited due to the insufficient production of

therapeutic factors in vivo, as well as low cell engraftment and survival in ischemic tissues.

Gene therapy has the potential to improve stem cell-mediated therapeutic

angiogenesis by increasing the paracrine release of specific angiogenic factors. In fact, it has

been shown that transplantation of virally transduced BMMSCs overexpressing VEGF into rat

models of myocardial ischemia lead to enhanced blood vessel density and heart function in

comparison to MSCs alone [19]. Furthermore, it was demonstrated that overexpressing VEGF

reduced the number EPCs needed by 30-fold to achieve a similar therapeutic effect as

unmodified EPCs in promoting angiogenesis and limb salvage [20]. However, most studies so

far have relied on viral vectors for efficient gene transfer, which are limited by potential risks

50

of mutagenesis, carcinogenesis, and immunogenicity. In the pursuit of a safer alternative

approach, strategies utilizing non-viral vectors for therapeutic angiogenesis are beginning to

emerge [21]. Conventional non-viral delivery systems involve using positively charged

polymers which can condense negatively charged DNA into nanoparticles for cellular uptake.

Polyethylenimine and Lipofectamine are the most widely used polymer-based transfection

reagents for gene delivery, but their clinical application is limited due to non-degradability,

poor transfection efficiency and high toxicity [21]. Poly(β-amino esters) (PBAE) are a family

of hydrolytically biodegradable polymers, which can deliver DNA into human stem cells with

high efficiency and low toxicity [11,22]. We have previously reported that MSCs modified

using PBAE/VEGF nanoparticles lead to substantially enhanced angiogenesis and limb

salvage in a mouse hindlimb ischemia model [22]. However, the efficacy of non-viral

engineered ADSCs for therapeutic angiogenesis remains largely unknown.

In this study, we assess the effects of paracrine release from ADSCs modified

utilizing PBAE/VEGF nanoparticles on human umbilical vein endothelial cell (HUVEC)

behavior in vitro. We first developed biodegradable PBAE/VEGF nanoparticles that can

transfect ADSCs with high efficiency for VEGF overexpression. Conditioned medium from

transfected ADSCs were then applied to HUVECs to examine the effects of such paracrine

release on HUVEC viability and apoptosis under low oxygen (1% O2), HUVEC migration and

tube formation.


4.2.1 Cell isolation and culture

Fat tissue was obtained from the abdominal fat of a female patient who had undergone

a free flap breast reconstruction surgery at Stanford University. All procedures were approved

and guided under Stanford IRB protocol. The fat tissue was washed 2-3 times with Phosphate

buffered saline and digested at 37 °C for 30 min with 0.5 U/ml Blendzyme 3 (Roche

Diagnostics, Indianapolis, IN, USA). Enzyme activity was neutralized with DMEM,

containing 10% fetal bovine serum (FBS, Invitrogen, Carlsbad, CA, USA), 1%

Penicillin/Streptomycin. Cells were then filtered through a 100 µm nylon mesh to remove

cellular debris before seeding. Following the initial 48 hours of incubation at 37 °C and 5%

CO2, cells were washed with PBS and maintained in growth medium containing DMEM, 10%

FBS, 1% penicillin/ streptomycin and 10ng/ml of the basic fibroblast growth factor (FGF)

(PeproTech, Rocky Hill, NJ, USA). Once 85-95% confluent, the cells were passaged by

51

trypsin (0.05%) digestion and plated at a density of 5000-6000 cells per cm2 for further

expansion. Human umbilical vein endothelial cells (HUVECs) were obtained from Lonza

(Basal, Switzerland) and expanded utilizing endothelial growth medium (EGM-2) (Lonza).

Passage 4 HUVECs were utilized for all experiments.

4.2.2 Transfection

Human adipose-derived stem cells (ADSCs) were transfected with DNA plasmids

encoding vascular endothelial growth factor (pVEGF) (Aldevron, Fargo, ND, USA) or

enhanced green fluorescence protein (pEGFP) (ELIM Biopharm, Hayward, CA, USA).

Passage 4 ADSCs were seeded at 65 000 cells per well in clear 24-well plates or 10 400 cells

per well in clear 96 well plates and incubated over night prior to transfection. Cells were then

transfected utilizing poly (β-amino esters) (PBAE) C32-122 as previously reported [11].

Briefly, polyplexes were formed by mixing pDNA and PBAE polymers in 25 mM sodium

acetate. Polyplexes were added to the cell culture after 10 minutes of complexation. Cells

were then incubated with polyplexes for four hours, at which time cell medium was replaced.

Transfection parameters were optimized by varying pDNA loading dose (1 µg, 2 µg, 3 µg, 4.5

µg, 6 µg) and GFP expression was examined at 48 hours. Lipofectamine 2000 (Invitrogen)

was used as a positive control and non-transfected cells were included as a negative control.

GFP expression by transfected ADSCs was measured at 48 hours post-transfection utilizing

fluorescence-activated cell sorting on a BD Facscalibur (BD Biosciences, San Jose, CA,

USA). Dead cells were excluded from the analysis by staining with propidium iodide and

10,000 live cells per sample were included for analyses. The percentage of GFP positive cells

was quantified using Flowjo 7.6 software (Tree Star, Inc. Ashland, OR, USA). Cell viability

was determined 48 hours post-transfection using Cell Titer 96 Aqueous One Solution Cell

Proliferation Assay (MTS) (Promega, Madison, WI, USA), and the value was normalized

using non-treated cells as a control.

4.2.3 Preparation of conditioned medium

Conditioned medium (CM) was prepared by culturing ADSCs in 24 well plates (65

000/well) for 48 hours in serum-free endothelial basal medium (EBM) (Lonza). Three groups

were examined including: 1) PBAE-VEGF transfected ADSCs, 2) Lipofectamine-VEGF

transfected ADSCs and 3) non-transfected ADSCs. CM was collected at 48 hours and stored

at -80 °C. To examine the effects of hypoxia on VEGF release, ADSCs were cultured in

52

either low oxygen (1% O2) or normoxia (20% O2). CM from both groups was collected at 48

hours, and CM from ADSCs cultured under normoxia was also collected on days 4, 6, and 8 to

quantify VEGF release kinetics. VEGF in the supernatant was determined by enzyme linked

immunosorbent assay (ELISA) (Peprotech).

4.2.4 HUVEC survival under hypoxia

Passage 5 HUVECs were plated in either 96 well plates (11 200 cells/well) for MTS

assay or in gelatin-coated 8-well chamber slides (28 000 cells/well) for TUNEL staining.

HUVECs were cultured in serum-free medium for 24 hours prior to switching to conditioned

medium, when they were placed in a low oxygen chamber (1% O2) for 48 hours. HUVEC

viability was determined by MTS assay as described above. HUVEC apoptosis was

determined utilizing the DeadEnd Fluorometric TUNEL System (Promega), and mounted with

Vectashield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA, USA) for

counterstain. Fluorescent microscopic images were analyzed utilizing ImageJ software (NIH,

USA). Apoptosis was analyzed by determining the ratio of TUNEL positive cells to DAPI

positive cells.

4.2.5 HUVEC Migration

HUVEC migration was assessed in a modified boyden chamber (E&K Scientific,

Santa Clara, CA, USA). The upper wells were pre-treated with 50 µl of 10 µg/ml fibronectin

(Sigma-Aldrich, St. Louis, MO) and incubated at 4oC for 24 hours prior to HUVEC plating.

HUVECs were seeded onto the upper chamber (20 000 cells/well) and EBM was added to the

bottom chamber. Cells were allowed to settle for 24 hours at 37oC. EBM in the bottom wells

was then replaced with hADSC-CM in quadruplicate and cells were allowed to migrate for 6

hours. The cells were stained with Hemacolor (Merck, Darmstadt, Germany) and cell number

was quantified by microscopy per high powered field. Results were reported as a ratio of

migrated cells over total number of cells.

4.2.6 Tube Formation Assay

Tube formation was assessed by culturing HUVECs on BD Matrigel (BD

Biosciences). Prior to HUVEC seeding, 96-well microtiter plates were coated with 80 µl of

Matrigel per well and incubated for 1 hour. Passage 5 HUVECs were suspended in hADSC-

CM or serum-free EBM and plated in triplicate at 12 000 cells per well. The formation of

53

tube-like structures was examined at each hour up to 16 hours. Representative images were

taken and analyzed utilizing ImageJ software (NIH, USA). Tube formation was analyzed by

quantifying the number of branches per high powered field.


All experiments were performed in triplicate and repeated for consistency. The results

were reported as mean ± standard error of the mean. Statistical analysis was performed using

the one way analysis of variance with Dunnett’s multiple comparison test as the post-test to

compare all the experimental groups with Lipofectamine 2000 as a positive control and non-

transfected cells as negative controls.

4.3 Results

4.3.1 Gene delivery efficiency into ADSCs using PBAE nanoparticles

Transfection efficiency of poly(β-amino esters) (PBAE) was first examined in human

adipose-derived stem cells (ADSCs) using a leading PBAE polymer C32-122 [11] and a

reporter DNA encoding enhanced-green fluorescence protein (pEGFP). Fluorescence

microscopy and flow cytometry confirmed gene delivery into ADSCs 48 hours post-

transfection (Figure 1). Increasing DNA doses from 1 µg to 2 µg significantly increased the

transfection efficiency from 6% to 26%, and further increase in DNA doses up to 6 µg did not

significantly change the transfection efficiency. The PBAE-mediated transfection also showed

4-fold higher efficiency compared to the positive control transfected using Lipofectamine

2000 (5.27%±0.7%) (Figure 2a). Meanwhile, all groups transfected with PBAE demonstrated

much higher cell viability compared to the Lipofectamine 2000 transfected group (88.6±6%

vs. 70.4±4%, p < 0.05) (Figure 2b). Varying DNA dose led to no statistically significant

change in cell viability among PBAE-transfected groups.

4.3.2 VEGF release from PBAE/VEGF transfected ADSCs

The effect of DNA dose on VEGF release from PBAE/VEGF transfected ADSCs was

examined by VEGF ELISA 48 hours post-transfection (Figure 3). ADSCs transfected using

PBAE with low DNA doses (1-3 µg DNA) or lipofectamine 2000 did not show significant

increase in VEGF production compared to non-transfected ADSC control. Increasing

PBAE/DNA dose to 4.5 µg led to enhanced VEGF production (5290±563 pg/106 cells) and

further increase was observed when DNA dose increased to 6 µg (6308±1810 pg/106 cells),

54

which was over 3-fold higher than Lipofectamine control (1726.25±46.25 pg/106 cells, p <

0.05) and non-transfected ADSC control (1188±133.75 pg/106 cells, p < 0.05) (Figure 3a).

Accumulated VEGF release overtime showed that the optimized PBAE transfected group

produced over 3-fold higher amount of VEGF than that of non-transfected ADSC controls,

while Lipofectamine only slightly increased VEGF production.

4.3.3 Effect of hypoxia on VEGF release and ADSC proliferation

To assess the effects of ischemia on ADSC fate, ADSCs were cultured in a hypoxia

chamber (1% O2) for 48 hours to mimic ischemic conditions in vivo. Outcomes were analyzed

for VEGF release and ADSC cell viability (Figure 4). Hypoxia treatment enhanced VEGF

release in non-transfected ADSC controls by over 200% when compared to culture under

normoxia. PBAE/VEGF transfection led to further increase in VEGF production, with the

highest VEGF concentrations observed for the 4.5 µg and the 6 µg DNA doses (Figure 4a).

The lowest cell number was observed in the Lipofectamine transfected group (~70%). Groups

transfected using PBAE demonstrated high viability (85%-95%) overall. Hypoxia treatment

led to further increase in cell number in all PBAE transfected groups except the 6 µg dose.

This increase is likely a combination of improved cell survival and increased cell proliferation.

All PBAE-transfected groups increased in cell number under hypoxia in comparison to the

normoxia-cultured, non-transfected ADSC controls (Figure 4b).

4.3.4 Effects of paracrine release on HUVEC viability and apoptosis under hypoxia

Conditioned medium from ADSCs (transfected or control) were collected and added

to human umbilical vein endothelial cell (HUVEC) culture under hypoxia to evaluate the

effects of paracrine signals from ADSCs on HUVEC viability and apoptosis (Figure 5).

Conditioned medium from all ADSC groups led to an increase in HUVEC viability

(87.3±3.9%) compared with EBM (69.6±5.1%, p < 0.05) (Figure 5a). However, no significant

difference in HUVEC viability was observed between groups treated with conditioned

medium from VEGF-overexpressing ADSCs or untransfected ADSCs. The effects of

paracrine release from ADSCs on HUVEC apoptosis was analyzed using TUNEL assay.

Conditioned medium from PBAE-transfected ADSCs led to a significant decrease in HUVEC

apoptosis (13.7±2.6%) compared to Lipo (33.2±2.4%, p < 0.05) and non-transfected controls

(39±5.3%, p < 0.05).

55

4.3.5 Effects of paracrine release on HUVEC migration and tube formation

We lastly assessed the effects of VEGF-overexpression on endothelial cell migration

and tube formation, two key steps involved in angiogenesis. HUVEC migration towards

ADSC paracrine release was assessed in a modified Boyden chamber (Figure 6). Conditioned

medium from non-transfected ADSC increased HUVEC migration (12.7±1%, p < 0.05)

compared to EBM (6.4±2%, p < 0.05), and conditioned medium from PBAE/VEGF-

transfected ADSCs led to further increase in HUVEC migration (16.8±1%). Similarly,

HUVEC formed homogeneous and interconnected tubular structures when exposed to

conditioned medium from PBAE/VEGF transfected ADSCs (49±4 branches per high powered

field) (Figure 7), whereas significantly fewer tubular structures were observed in HUVECs

exposed to conditioned medium from Lipo/VEGF-transfected ADSCs (31.7±1 branches per

high powered field, p < 0.05) or ADSCs alone (32.0±2 branches per high powered field, p <

0.05).

4.4 Discussion

Here, we demonstrated that poly(β-amino esters) (PBAE) are an effective non-viral,

polymeric vector for gene delivery to adipose-derived stem cells (ADSCs). VEGF-

overexpressing ADSCs utilizing biodegradable nanoparticles enhanced angiogenesis-related

cellular behavior, including endothelial cell survival under hypoxia, migration and tube

formation. Although a previous study has shown beneficial effects of VEGF-overexpressing

ADSCs on angiogenesis using a viral-based approach [23], the broad clinical translation of

viral-mediated gene delivery remains limited by safety concerns. Non-viral based gene

delivery is potentially safer, but often suffers from low transfection efficiency and high

toxicity. Our results showed that PBAE-nanoparticles are promising non-viral polymeric

vectors that could transfect ADSCs with 4-5 times higher transfection efficiency than

Lipofectamine 2000, a commercially available transfection reagent, and with higher cell

viability (Figure 2). Polyethylenimine (PEI) [24], nucleofection, and electroporation [25-26]

are other commonly used non-viral gene delivery methods, which may achieve higher

transfection efficiency than Lipofectamine 2000, but these suffer from high toxicity. PBAE

nanoparticles are able to achieve high transfection at little expense to cell viability, which

makes them a more attractive option to over-express desired therapeutic genes. PBAE-

mediated VEGF overexpression is transient (Figure 3), which may limit any negative effects

due to constitutive release of VEGF, such as excessive and aberrant vascularity. This

56

demonstrates that PBAE nanoparticles could be utilized to strategically overexpress known

therapeutic factors, thereby manipulating stem cells to augment their natural angiogenic

responses.

Given that ADSCs will be exposed to hypoxia when transplanted into ischemic

tissues, it is important to understand ADSC behavior under hypoxia to determine their

response post-transplantation. ADSCs are known to release several paracrine factors with

angiogenic effects [13,27]. Hypoxia has been shown to enhance paracrine release of

angiogenic factors by ADSCs or BMMSCs, which may induce therapeutic benefit both in

vitro and in vivo [27-30]. Paracrine release from ADSCs cultured under hypoxia were shown

to improve endothelial cell (EC) survival and such effects were attenuated by addition of anti-

VEGF blocking antibodies [30]. We examined the effects of hypoxia on ADSC VEGF release

and cell viability, and how such cellular responses changed upon up-regulating VEGF by

PBAE. Consistent with previous reports, our results showed that hypoxia led to an increase in

both ADSC VEGF release and cell proliferation compared with normoxia culture. VEGF-

overexpression further improved VEGF release by ADSCs due to hypoxia, demonstrating that

gene therapy and hypoxia can act synergistically. We also found that hypoxia enhanced

ADSC viability, which has been shown in several prior reports utilizing different assays

[13,27,30]. Endogenous ADSCs have been found to proliferate in situ due to ischemia at a

distant site, suggesting that ADSCs respond favorably to low oxygen tension [27]. These

findings support ADSCs as an excellent cell source for treating ischemic disease, due to their

ability to survive under hypoxia and enhance paracrine release of angiogenic factors. Though

most previous work has focused on the use of bone marrow-derived mesenchymal stem cells

(BMMSCs) for angiogenesis [6-7,12,22,31], ADSCs represent a more attractive alternative

cell source due to their relative abundance and ease of isolation. Both cell sources have been

reported to improve angiogenesis in vivo, but their paracrine release profiles are different.

While both cell types secrete VEGF, BMMSCs secrete more fibroblast growth factor (bFGF)

[6] and ADSCs more prominently secrete hepatocyte growth factor (HGF) and transforming

growth factor-β (TGF-β) [13].

Angiogenesis is the formation of new blood vessels from pre-existing vessels, which

involves endothelial cell proliferation, migration, and new tube formation. To determine the

effects of paracrine release from VEGF-overexpressing ADSCs on angiogenesis, we applied

conditioned medium from PBAE/VEGF transfected ADSCs to human umbilical vein

endothelial cells (HUVEC) and examined HUVEC cellular processes related to angiogenesis.

57

Prior studies assessed endothelial cell survival under low-serum (2-5%) containing medium

[4,7,28]. In our study, endothelial cells were cultured in serum-free EBM, which represents a

more challenging culture condition for endothelial cells. Our results confirmed hypoxia

exposure had detrimental effects on HUVECs, with decreased cell viability and markedly

increased cell apoptosis (Figure 5). Consistent with previous reports, here we found that

paracrine release from ADSCs alone enhanced HUVEC viability,[13,16] HUVEC migration

[6,30], and HUVEC tube formation [28,30]. PBAE/VEGF overexpression led to a marked

decrease in HUVEC apoptosis, with improvements in HUVEC migration and more

homogeneous tube formation. These results confirmed the beneficial effects of VEGF-

overexpression on HUVEC behavior. Addition of anti-VEGF blocking antibodies was

previously reported to reduce endothelial cell migration [30], again suggesting the functional

importance of VEGF. While previous study has also shown protective effects on endothelial

cells using VEGF supplementation, the dose of VEGF applied was two orders of magnitude

higher than that secreted by cells alone (4-10 ng/ml versus 0.08-0.1 ng/ml VEGF from cells

alone) [6], which may lead to undesirable excessive vascular formation. It is also worth noting

that Lipofectamine/VEGF transfection did not improve tube formation or HUVEC apoptosis

compared to ADSCs alone, which is likely the result of the low transfection efficiency using

Lipofectamine compared with PBAE nanoparticles (Figure 1). Overall, these results show

that PBAE-mediated non-viral gene delivery provides an effective strategy for overexpressing

therapeutic factors to induce HUVEC cellular processes related to angiogenesis.

4.5 Conclusion

In conclusion, we have shown that non-virally modified ADSCs using biodegradable

polymeric vectors have beneficial effects on HUVEC behavior in vitro. ADSCs are promising

autologous cell sources that can serve as delivery vehicles for therapeutic paracrine signals.

Biodegradable poly(β-amino esters) nanoparticles can deliver DNA into ADSCs with high

efficiency and low cytotoxicty. Paracrine release from VEGF-overexpressing ADSC improved

HUVEC survival, migration and tube formation, and markedly reduced HUVEC apoptosis.

The results of this study suggest that non-virally engineered ADSCs may be a promising

autologous cell source for therapeutic angiogenesis in cardiovascular disease.

58

Acknowledgements

The authors would like to thank American Heart Association National Scientist Development

Grant (10SDG2600001), Stanford Bio-X Interdisciplinary Initiative Program, Donald E. and

Delia B. Baxter Foundation and Stanford Medical Scholars Research Program for funding.

References

1. Potente M, H Gerhardt and P Carmeliet. (2011). Basic and therapeutic aspects of

angiogenesis. Cell 146:873-87.

2. Losordo DW and S Dimmeler. (2004). Therapeutic angiogenesis and vasculogenesis

for ischemic disease. Part I: angiogenic cytokines. Circulation 109:2487-91.

3. Losordo DW and S Dimmeler. (2004). Therapeutic angiogenesis and vasculogenesis

for ischemic disease: part II: cell-based therapies. Circulation 109:2692-7.

4. Gruber R, B Kandler, P Holzmann, M Vogele-Kadletz, U Losert, MB Fischer and G

Watzek. (2005). Bone marrow stromal cells can provide a local environment that

favors migration and formation of tubular structures of endothelial cells. Tissue Eng

11:896-903.

5. Kamihata H, H Matsubara, T Nishiue, S Fujiyama, Y Tsutsumi, R Ozono, H Masaki,

Y Mori, O Iba, E Tateishi, A Kosaki, S Shintani, T Murohara, T Imaizumi and T

Iwasaka. (2001). Implantation of bone marrow mononuclear cells into ischemic

myocardium enhances collateral perfusion and regional function via side supply of

angioblasts, angiogenic ligands, and cytokines. Circulation 104:1046-52.

6. Kinnaird T, E Stabile, MS Burnett, CW Lee, S Barr, S Fuchs and SE Epstein. (2004).

Marrow-derived stromal cells express genes encoding a broad spectrum of

arteriogenic cytokines and promote in vitro and in vivo arteriogenesis through

paracrine mechanisms. Circ Res 94:678-85.

7. Potapova IA, GR Gaudette, PR Brink, RB Robinson, MR Rosen, IS Cohen and SV

Doronin. (2007). Mesenchymal Stem Cells Support Migration, Extracellular Matrix

Invasion, Proliferation, and Survival of Endothelial Cells In Vitro. STEM CELLS

25:1761-1768.

8. Zuk PA, M Zhu, H Mizuno, J Huang, JW Futrell, AJ Katz, P Benhaim, HP Lorenz

and MH Hedrick. (2001). Multilineage cells from human adipose tissue: implications

for cell-based therapies. Tissue Eng 7:211-28.

59

9. Schäffler A and C Büchler. (2007). Concise Review: Adipose Tissue-Derived Stromal

Cells—Basic and Clinical Implications for Novel Cell-Based Therapies. STEM

CELLS 25:818-827.

10. Kim YJ, HK Kim, HH Cho, YC Bae, KT Suh and JS Jung. (2007). Direct Comparison

of Human Mesenchymal Stem Cells Derived from Adipose Tissues and Bone Marrow

in Mediating Neovascularization in Response to Vascular Ischemia. Cellular

Physiology and Biochemistry 20:867-876.

11. Yang F, JJ Green, T Dinio, L Keung, SW Cho, H Park, R Langer and DG Anderson.

(2009). Gene delivery to human adult and embryonic cell-derived stem cells using

biodegradable nanoparticulate polymeric vectors. Gene Ther 16:533-46.

12. Gnecchi M, H He, OD Liang, LG Melo, F Morello, H Mu, N Noiseux, L Zhang, RE

Pratt, JS Ingwall and VJ Dzau. (2005). Paracrine action accounts for marked

protection of ischemic heart by Akt-modified mesenchymal stem cells. Nat Med

11:367-8.

13. Rehman J, D Traktuev, J Li, S Merfeld-Clauss, CJ Temm-Grove, JE Bovenkerk, CL

Pell, BH Johnstone, RV Considine and KL March. (2004). Secretion of angiogenic

and antiapoptotic factors by human adipose stromal cells. Circulation 109:1292-8.

14. Cao Y, Z Sun, L Liao, Y Meng, Q Han and RC Zhao. (2005). Human adipose tissue-

derived stem cells differentiate into endothelial cells in vitro and improve postnatal

neovascularization in vivo. Biochemical and Biophysical Research Communications

332:370-379.

15. Kondo K, S Shintani, R Shibata, H Murakami, R Murakami, M Imaizumi, Y

Kitagawa and T Murohara. (2009). Implantation of Adipose-Derived Regenerative

Cells Enhances Ischemia-Induced Angiogenesis. Arteriosclerosis, Thrombosis, and

Vascular Biology 29:61-66.

16. Miranville A, C Heeschen, C Sengenes, CA Curat, R Busse and A Bouloumie. (2004).

Improvement of postnatal neovascularization by human adipose tissue-derived stem

cells. Circulation 110:349-55.

17. Moon MH, SY Kim, YJ Kim, SJ Kim, JB Lee, YC Bae, SM Sung and JS Jung.

(2006). Human adipose tissue-derived mesenchymal stem cells improve postnatal

neovascularization in a mouse model of hindlimb ischemia. Cell Physiol Biochem

17:279-90.

60

18. Rubina K, N Kalinina, A Efimenko, T Lopatina, V Melikhova, Z Tsokolaeva, V

Sysoeva, V Tkachuk and Y Parfyonova. (2009). Adipose stromal cells stimulate

angiogenesis via promoting progenitor cell differentiation, secretion of angiogenic

factors, and enhancing vessel maturation. Tissue Eng Part A 15:2039-50.

19. Matsumoto R, T Omura, M Yoshiyama, T Hayashi, S Inamoto, KR Koh, K Ohta, Y

Izumi, Y Nakamura, K Akioka, Y Kitaura, K Takeuchi and J Yoshikawa. (2005).

Vascular endothelial growth factor-expressing mesenchymal stem cell transplantation

for the treatment of acute myocardial infarction. Arterioscler Thromb Vasc Biol

25:1168-73.

20. Iwaguro H, J Yamaguchi, C Kalka, S Murasawa, H Masuda, S Hayashi, M Silver, T

Li, JM Isner and T Asahara. (2002). Endothelial progenitor cell vascular endothelial

growth factor gene transfer for vascular regeneration. Circulation 105:732-8.

21. Park HJ, F Yang and SW Cho. (2012). Nonviral delivery of genetic medicine for

therapeutic angiogenesis. Adv Drug Deliv Rev 64:40-52.

22. Yang F, S-W Cho, SM Son, SR Bogatyrev, D Singh, JJ Green, Y Mei, S Park, SH

Bhang, B-S Kim, R Langer and DG Anderson. (2010). Genetic engineering of human

stem cells for enhanced angiogenesis using biodegradable polymeric nanoparticles.

Proceedings of the National Academy of Sciences 107:3317-3322.

23. Jabbarzadeh E, T Starnes, YM Khan, T Jiang, AJ Wirtel, M Deng, Q Lv, LS Nair, SB

Doty and CT Laurencin. (2008). Induction of angiogenesis in tissue-engineered

scaffolds designed for bone repair: a combined gene therapy-cell transplantation

approach. Proc Natl Acad Sci U S A 105:11099-104.

24. Ahn HH, JH Lee, KS Kim, JY Lee, MS Kim, G Khang, IW Lee and HB Lee. (2008).

Polyethyleneimine-mediated gene delivery into human adipose derived stem cells.

Biomaterials 29:2415-2422.

25. Lopatina T, N Kalinina and E Parfyonova. (2009). Nonviral Transfection of Adipose

Tissue Stromal Cells: An Experimental Study. Bulletin of Experimental Biology and

Medicine 147:509-512.

26. Zaragosi L-E, N Billon, G Ailhaud and C Dani. (2007). Nucleofection Is a Valuable

Transfection Method for Transient and Stable Transgene Expression in Adipose

Tissue-Derived Stem Cells. STEM CELLS 25:790-797.

27. Thangarajah H, IN Vial, E Chang, S El-Ftesi, M Januszyk, EI Chang, J Paterno, E

Neofytou, MT Longaker and GC Gurtner. (2009). IFATS Collection: Adipose Stromal

61

Cells Adopt a Proangiogenic Phenotype Under the Influence of Hypoxia. STEM

CELLS 27:266-274.

28. Hung S-C, RR Pochampally, S-C Chen, S-C Hsu and DJ Prockop. (2007). Angiogenic

Effects of Human Multipotent Stromal Cell Conditioned Medium Activate the PI3K-

Akt Pathway in Hypoxic Endothelial Cells to Inhibit Apoptosis, Increase Survival,

and Stimulate Angiogenesis. STEM CELLS 25:2363-2370.

29. Kinnaird T, E Stabile, MS Burnett, M Shou, CW Lee, S Barr, S Fuchs and SE

Epstein. (2004). Local delivery of marrow-derived stromal cells augments collateral

perfusion through paracrine mechanisms. Circulation 109:1543-9.

30. Nakagami H, K Maeda, R Morishita, S Iguchi, T Nishikawa, Y Takami, Y Kikuchi, Y

Saito, K Tamai, T Ogihara and Y Kaneda. (2005). Novel Autologous Cell Therapy in

Ischemic Limb Disease Through Growth Factor Secretion by Cultured Adipose

Tissue–Derived Stromal Cells. Arteriosclerosis, Thrombosis, and Vascular Biology

25:2542-2547.

31. Burdon TJ, A Paul, N Noiseux, S Prakash and D Shum-Tim. (2011). Bone marrow

stem cell derived paracrine factors for regenerative medicine: current perspectives and

therapeutic potential. Bone Marrow Res 2011:207326.

62

Figures

Figure 4.1. Efficiency of gene delivery to human adipose-derived stem cells (hADSC)

utilizing poly(β-amino esters) (PBAE) nanoparticles. Under leading transfection conditions,

PBAE (a) delivered eGFP DNA into hADSCs with 4-times the efficiency of Lipofectamine

2000 (b). Fluorescence microscopy showed markedly increased GFP signals in hADSCs

transfected using PBAE nanoparticles (d) compared to cells transfected using Lipofectamine

(e). Scale bar: 500 µm.

63

Figure 4.2 Effects of nanoparticle dose on transfection efficiency and cell viability. (a)

Poly(β-amino esters) (PBAE) led to significantly increased transfection efficiency in human

adipose-derived stem cells (hADSCs) using doses as low as 2 µg/well compared to

Lipofectamine 2000 (Lipo) (# p < 0.001 compared to Lipo). (b) Cells transfected using PBAE

at all doses demonstrated high cell viability, which were about 1 fold higher compared to the

cells transfected using Lipo (# p < 0.01 versus Lipo, * p < 0.05 versus Cells only).

64

Figure 4.3. VEGF release from human adipose-derived stem cells (hADSCs) after

transfection using polymeric nanoparticles. (a) VEGF release from PBAE-modified hADSCs

increased with DNA dose and reached a maximum at DNA dose of 6 µg, which was

significantly higher than groups transfected using Lipofectamine 2000 (Lipo) (# p < 0.05) and

cells alone (* p < 0.05). (b) Accumulated VEGF release showed continuously enhanced

VEGF release from cells transfected using PBAE/VEGF nanoparticles compared to Lipo

control.

65

Figure 4.4 Effects of hypoxia on VEGF production and cell viability of non-viral engineered

ADSCs. (a) Hypoxia (1% O2 for 48 hrs) enhanced VEGF release from non-viral engineered

human adipose-derived stem cells in a dose-dependent manner. (b) Hypoxia treatment led to

an increase in ADSC viability in most PBAE transfected groups, except in the groups

transfected with 6 µg PBAE/DNA or Lipofectamine 2000 (Lipo).

66

Figure 4.5. Effects of paracrine signals from non-viral engineered ADSCs on endothelial cell

viability and apoptosis. (a) Paracrine release from hADSCs significantly improved HUVEC

viability under hypoxia (* p < 0.05 versus EBM), with no significant difference in cell

viability due to VEGF overexpression. (b) Paracrine release from hADSCs markedly

decreased HUVEC apoptosis under hypoxia, and PBAE/VEGF transfected group led to the

greatest decrease in cell apoptosis (* p < 0.05 versus EBM, # p < 0.05 versus Lipo

VEGF/hADSCS-CM). (c-f) Representative images of TUNEL stained endothelial cells

undergoing apoptosis. Scale bar 200 µm.

67

Figure 4.6. Paracrine release from PBAE/VEGF engineered human adipose-derived stem

cells (hADSCs) enhanced endothelial cell migration. (a-c) Representative images of migrated

endothelial cells in response to conditioned medium from hADSCs or EBM (negative

control). (d). Percentage of migrated endothelial cells towards conditioned medium (* p < 0.05

versus EBM only, # p < 0.05 versus hADSC-CM only).

68

Figure 4.7. Paracrine release from PBAE/VEGF engineered human adipose-derived stem

cells (hADSCs) enhanced endothelial cell tube formation. (a-d). Conditioned medium from

PBAE/VEGF transfected hADSCs led to formation of interconnected tubular structures by

endothelial cells (scale bar: 500 µm). (e) The number of branch points formed by endothelial

cells in response to conditioned medium from all groups. (* p < 0.05 versus EBM, # p < 0.05

compared to Lipo).

69

70

5. CXCR4-overexpressing Stem Cells Enhance Tissue Regeneration

in a Mouse Model of Hindlimb Ischemia

Lorenzo Deveza1,2, Jeffrey Choi3, Jerry Lee4, Ngan Huang5, John Cooke4, Fan Yang1,6*

1Department of Bioengineering, Stanford University, Stanford, CA; 2Stanford University School

of Medicine, Stanford, CA; 3Department of Biology, Stanford University, Stanford, CA;

4Division of Cardiovascular Medicine, Stanford University, Stanford, CA; 5Department of

Cardiovascular Surgery, Stanford University, Stanford, CA;6Department of Orthopaedic

Surgery, Stanford University, Stanford, CA.

Abstract

Peripheral arterial disease (PAD) is an important disease caused by a blockage of blood flow to

the limbs. Treatment often involves surgical revascularization, but many patients possess co-

morbidities which prohibit their treatment. Stem cell-based therapeutics offer an alternative,

which may promote angiogenesis in ischemic tissues. However, current approaches have met

with variable clinical success, which might be attributed to insufficient paracrine signals or poor

cellular engraftment into regenerating tissues. To overcome these limitations, we developed

CXC chemokine receptor 4 (CXCR4) overexpressing adipose-derived stem cells (ADSCs)

using biodegradable nanoparticles. CXCR4 is a well-known receptor involved in stem cell

targeting to sites of ischemia, which we hypothesized would promote ADSCs as drug delivery

vehicles to ischemia. CXCR4-overexpressing ADSCs led to a robust healing response in murine

models of hindlimb ischemia, leading to 100% limb salvage that was associated with early

reperfusion and muscle repair. CXCR4- overexpressing ADSCs also improved cell engraftment

and survival over 2 weeks compared to transfection controls. Assessement of CXCR4-ADSCs

in vitro revealed that CXCR4-overexpression induced a pro-angiogenic and anti-inflammatory

phenotype in ADSCs. These results suggest that non-virally engineered ADSCs overexpressing

CXCR4 are a promising strategy for treating PAD.

5.1 Introduction

Peripheral arterial disease (PAD) represents a large, yet underrepresented medical

problem worldwide. It is caused by a reduction in blood flow to the limbs, and usually signifies

advanced underlying atherosclerotic disease [1, 2]. Poor blood flow by PAD can cause

significant leg pain when walking, and in severe cases lead to non-healing ulcers or limb

71

amputation. Treating PAD generally involves surgical revascularization [1, 2]. Unfortunately,

many patients suffering from PAD are unable to withstand surgery. Therapeutic angiogenesis

using stem cells offers a valuable alternative for treating PAD by secreting paracrine signals to

promote blood vessel growth in situ [3]. However, stem cell-based approaches have met with

variable clinical success, which might be attributed to poor cell engraftment and an insufficient

paracrine response.

Combination stem cell-gene therapy can overcome many of these limitations by

overexpressing genes to enhance relevant cellular processes, such as cell migration and

angiogenic growth factor release. Previously, we demonstrated that vascular endothelial growth

factor (VEGF)-transfected bone marrow-derived mesenchymal stem cells (BM-MSCs) could

improve therapeutic responses in a murine model of hindlimb ischemia [4]. Notably,

transfection was achieved utilizing an optimized biodegradable polymeric nanoparticle for non-

viral gene transfection, namely poly(β-amino ester) (PBAE), which is an important

characteristic for clinical translation [5]. However, the use of VEGF and BM-MSCs may not

be ideal, as VEGF can lead to leaky vasculature and promote inflammation, while BM-MSCs

are difficult to obtain from the patient and suffer low yields and donor site morbidity. Further

developments of combined stem cell-gene therapy for clinical translation include: 1) the

identification of therapeutic gene targets that can promote cell engraftment and enhance

paracrine signaling; as well as, 2) the proper selection of a reliably abundant autologous cell

population.

CXC chemokine receptor 4 (CXCR4) is a G-protein coupled receptor known to be

involved in cell homing towards ischemia, and may potentially be involved in other cellular

processes, such as cell survival and growth factor release [6]. Physiologically, CXCR4 is

expressed on progenitor and inflammatory cells, which localize to hypoxic tissues to participate

in physiologic revascularization and tissue repair [7]. The ligand for CXCR4 is stromal derived

factor-1α (SDF-1α), which is secreted under the action of the hypoxia inducible factor-1α

transcription factor and is mainly expressed in hypoxic tissues [8-10]. Due to its physiologic

involvement in hypoxic signaling, CXCR4 is a promising homing factor to overexpress in stem

cells to promote their role as drug delivery vehicles to ischemia.

With regards to cell source, adipose derived stem cells (ADSCs) represent a relatively

abundant, autologous cell population with similar lineage differentiation potential to BM-

MSCs. Clinically, ADSCs are easier to isolate and have high yields. For therapeutic purposes,

ADSCs have shown benefits in models of hindlimb ischemia [11, 12] and myocardial ischemia

72

[13], and are known to secrete a number of proangiogenic and anti-apoptotic cytokines [14].

Previously, we engineered ADSCs using PBAE nanoparticles to overexpress VEGF and found

that their paracrine release could enhance endothelial cell processes of angiogenesis in vitro

[15].

In the present study, we aimed to evaluate the potential of CXCR4-overexpressing

ADSCs using PBAE nanoparticles for enhancing therapeutic angiogenesis in a murine model

of hindlimb ischemia (figure 1A). We hypothesized that CXCR4 overexpression would enable

cell survival and localization to the ischemic environment. Therapeutic efficiency of genetically

engineered ADSCs was assessed by evaluating cell survival and engraftment, angiogenic and

inflammatory response, and muscle morphology.


5.2.1 Materials

Monomers for synthesizing poly(β-amino esters)were purchased from Sigma–Aldrich,

Scientific Polymer Products, and Molecular Biosciences. Anhydrous dimethyl sulfoxide

(DMSO) was purchased from Sigma–Aldrich. A 25 mM sodium acetate (NaOAc) buffer

solution (pH 5.2) was prepared by diluting a 3 M stock solution (Sigma-Aldrich, St Louis, MO).

Plasmids (pEGFP-N1 and pcDNA3.1(+)-CXCR4) were purchased from Elim

Biopharmaceuticals and pVEGF165 plasmid was obtained from Aldevron.

5.2.2 Poly(β-Amino Esters) Synthesis

PBAE polymers were synthesized as previously described [4]. Briefly, acrylate-

terminated C32 polymer (C32-Ac) was prepared by mixing butanediol (C) and aminopentanol

(32) monomers using a 1.2:1.0 diacrylate/amine monomer molar ratio at 90 °C for 24 h.

Subsequently, amine-terminated C32 polymers were generated by reacting C32-Ac with

diamine monomer (122) in DMSO. End-capping reactions were performed overnight at room

temperature by using a 1.6-fold molar excess of amine over acrylate end groups.

5.2.3 ADSC isolation and culture

ADSCs were collected by harvesting the inguinal fat pads from 5 to 6 weeks old

luciferase-positive/GFP-positive transgenic male mice (Joseph Wu, MD, PhD laboratory,

Stanford University) and FVB wild-type male mice (purchased from Charles River, Willington,

MA). Following serial betadine washes, the fat pads were finely minced in cold, sterile PBS,

73

followed by digestion in 0.075% Collagenase II (Sigma-Aldrich) at 37 °C for 30 minutes in a

shaking water bath. Digestion was neutralized with Dulbecco’s modified Eagle’s medium

(Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Invitrogen) and 1%

penicillin-streptomycin (Sigma-Aldrich). Neutralized cells were centrifuged at 4 °C and 1,000

rpm for 5 minutes, and floating adipocytes and media were removed before pelleted ADSCs

from the stomal-vascular portion were resuspended in culture media, passed through a 100 μm

Falcon nylon cell strainer (Becton Dickinson, Bedford, MA) and plated on a 10 cm dish. Cells

were incubated at 37 °C and 5% atmospheric CO2. Twenty-four hours after isolation, the cells

were washed in PBS to remove debris and nonadherent cells, and fresh medium was added to

the cells. Medium was changed every 2 days, and adherent spindle-shaped cells were grown to

subconfluence, passaged using standard methods of trypsinization, and propagated through

passage 2.

5.2.4 Transfection

ADSCs were transfected with CXCR4 plasmid, VEGF plasmid, or control plasmid

(EGFP) in DMEM containing 10% FBS using optimized PBAE transfection conditions [15].

PBAE nanoparticles were formulated by mixing C32-122 PBAE polymers with plasmids at 30:1

polymer weight to DNA weight ratio in 25 mM NaAc and incubuating for 10 minutes. PBAE

nanoparticles were then added to ADSCs cultured in DMEM containing 10% FBS. After 4 h

incubation, the medium was replaced with fresh growth medium. For transfection

characterization assays, ADSCs were transfected at 3.0 x 105 cells per well in 6 well plates. For

therapeutic animal studies, ADSCs were transfected in T150 tissue culture flasks at 3.5 x 106

cells.

5.2.5 Flow cytometric analysis of ADSC transfection efficiency and CXCR4 cell surface

expression

To verify transfection efficiency by PBAE nanoparticles, GFP expression by

transfected WT ADSCs was measured at 48 h post transfection utilizing fluorescence-activated

cell sorting (FACS) on a BD FACScalibur (BD Biosciences, San Jose, CA). Dead cells were

excluded from the analysis by staining with propidium iodide and 10,000 live cells per sample

were included for the analysis. The percentage of GFP-positive cells was quantified using

Flowjo 7.6 software (Tree Star, Inc.,Ashland, OR).

74

5.2.6 Quantitative real-time polymerase chain reaction (RT-PCR) in vitro

To confirm successful transfection of human CXCR4 and human VEGF by PBAE in

vitro, total RNA was harvested from ADSCs at 48 hours post-transfection using RNeasy Kit

(Qiagen, Germantown, MD, USA). cDNA was synthesized by reverse transcription using the

Superscript First-Strand Synthesis System (Invitrogen) and 1 μg of total RNA. Quantitative

PCR was performed using Power SYBR Green PCR Master Mix following the manufacturer's

protocol and running all samples for 40 PCR cycles on Applied Biosystems 7900 Real-Time

PCR System (Carlsbad, CA, USA). The relative expression level of target genes was

determined by the ΔΔCT method, where target gene expression was first normalized to an

endogenous housekeeping gene, GAPDH, followed by a second normalization to the gene

expression level measured in untransfected control cells. Genetic upregulation of mouse

angiogenic and inflammatory genes was confirmed by isolating total RNA from ADSCs

cultured under 1% O2 conditions for 48 hours then proceeding through cDNA synthesis and RT-

PCR. Measurement of mRNA extracted from ischemic muscle was performed by digesting

muscle utilizing TRIzol Reagent (Invitrogen) and proceeding through RNA extraction, cDNA

synthesis, and RT-PCR as described above. The utilized primers are listed in Table S1.

5.2.7 Measurement of VEGF production by Enzyme-Linked Immunosorbent Assay

(ELISA)

VEGF production from ADSCs was assessed by collecting conditioned medium and

measured using an ELISA kit for human VEGF (Peprotech, Rocky Hill, NJ) according to the

manufacturer’s protocol. VEGF production was measured on days 2, 4, 6, 8. Accumulated

VEGF release was calculated by quantifying the total amount of VEGF released from all

previous time points. The level of secreted VEGF in vitro was expressed as the amount of

VEGF per milliliter of medium supernatant.

5.2.8 Determination of CXCR4 upregulation by Western Blot

CXCR4 protein production was determined by Western blot method. Total protein was

isolated utilizing RIPA buffer (Thermo Scientific, Waltham, MA). Proteins were then separated

by gel electrophoresis utilizing NuPage Novex Bis-Tris Mini Gels (Invitrogen) at 200 V for 35

minutes. Separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane

(Invitrogen) utilizing the XCell II Blot Module (Invitrogen) at 30 V for 1 hour. To assess for

CXCR4, blotted PVDF membranes were stained anti-human CXCR4 (Abcam, Cambridge,

75

MA), utilizing a HRP-labeled secondary antibody (Abcam). Staining for anti-β actin (Abcam)

served as a loading control. X-ray films were developed utilizing the chemiluminescent reagent,

SuperSignal West Pico (Thermo Scientific).

5.2.9 Cell viability/proliferation assay

Cell viability or proliferation following transfection was evaluated at 48 hours post-

transfection using the Cell Titer 96 Aqueous One Solution (MTS) assay kit (Promega,

Fitchburg, WI) to measure metabolic activity. Results were expressed as a percentage of optical

absorbance values obtained for nontreated cells.

5.2.10 Transplantation of stem cells in mouse model of hindlimb ischemia

All experiments were performed in accordance with the Stanford University Animal

Care and Use Committee Guidelines and approved APLAC protocols. Hindlimb ischemia was

induced in female FVB mice age 10-12 weeks (Charles River) as previously described [16].

Briefly, the femoral artery was ligated at two sites (distal to the external iliac artery and proximal

to the popliteal artery) through a skin incision with 5-0 silk suture (Ethicon). The femoral artery

was excised between the points of ligation. After arterial dissection, cells (1.0 x 106 cells per

injection) were suspended in 100 µL of PBS and injected intramuscularly into two sites (the

adductor muscle region in the medial thigh and the gastrocnemius) using a 29-gauge tuberculin

syringes. Four experimental groups (n = 10 per group) were examined as follows: (i) CXCR4-

ADSC, (ii) CXCR4/VEGF-ADSC, (iii) CXCR4/VEGF-ADSC, and (iv) GFP-ADSC

transfection control. Physiological status of ischemic limbs was followed up to 4 weeks after

treatment. All of the animals were killed at the 4-week time point, and limbs of the ischemic

sides were retrieved for analyses.

5.2.11 Bioluminescent Imaging (BLI)

ADSC survival was monitored post-injection by BLI over the course of two weeks. BLI

was performed on the day of surgery, and on days 1, 3, 4, 7, 10, 14. Animals were anesthetized

by 2–3% inhaled isoflurane and injected intraperitoneally with the reporter probe D-Luciferin

at 100 mg/kg body weight. IVIS Spectrum system (Xenogen, Alameda, CA) was used to image

the animals while under 2% inhaled isoflurane anesthesia. Each animal was scanned until the

peak signal was reached, recording radiance of regions of interest in photons. Radiance was

quantified in photons per second per centimeter squared per steridian.

76

5.2.12 In vivo assessment of clinical limb salvage and limb reperfusion

Ischemic limbs were evaluated by direct observation of clinical limb status and by

measurement of blood reperfusion. Limbs were assessed every week up to 4 weeks. Grading

of limb ischemia was based on the severity of limb ischemia utilizing following categories: 1)

limb salvage; 2) toe necrosis; 3) foot necrosis; 4) foot loss; and 5) limb loss. Blood reperfusion

in ischemic limbs was measured using a PeriScan PIM3 laser Doppler system (Perimed AB,

Sweden), as previously described [16]. Briefly, each animal was pre-warmebd to a body

temperature of 36°C. ROIs were drawn around the entire hindlimbs and the level of perfusion

in ischemic and normal hindlimbs were quantified using the mean pixel value within each ROI.

Relative changes in blood flow were expressed as the ratio of the ischemic limb to the normal

limbs.

5.2.13 Histology, Immunofluorescence, Capillary Densitometry, and Macrophage

Quantification

Mice were euthanized at 4 days and 28 days to retrieve muscle tissue from ischemic

limbs for histological analysis. Specimens were embedded in optimal cutting temperature

(O.C.T.) for cryosectioning. Sections were cut at 8 µm thickness. General tissue morphology

was assessed with Masson’s Trichrome Staining. Immunofluorescence was performed to assess

morphologic features of angiogenesis and inflammation. Transplanted GFP-positive/luciferase-

positive cells were detected in day 28 sections with anti-GFP (Invitrogen). Capillaries were

detected with anti-CD31 (BD Biosciences, San Jose, CA). Macrophages were detected with

anti-CD68 (AbD Serotec, Oxford, UK). Sections were mounted with Vectashield Mounting

Medium with 4’,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA)

for counterstain. Sections were imaged using a fluorescent microscope. The number of CD31-

positive vessels was quantified to estimate capillary density per high powered field (hpf) using

ImageJ software (NIH). Quantitative analysis of macrophages was performed by determining

the number of DAPI-positive cells that overlap CD68-positive signals and reported as a

percentage of total DAPI-positive cells utilizing MATLAB software (MathWorks, Natick, MA).


GraphPad Prism (Graphpad Software, San Diego, CA, USA) was used to perform all

statistical analyses. One-way analysis of variance with Tukey's multiple comparison test as a

77

post-hoc test was performed to compare all experimental groups and determine statistical

significance. All quantitative data were expressed as mean ± standard error of the mean.

5.3 Results

5.3.1 Successful overexpression of CXCR4 and VEGF in ADSCs utilizing biodegradable

polymeric nanoparticles

To generate CXCR4- and/or VEGF- overexpressing ADSCs, PBAE nanoparticles were

applied to ADSCs ex vivo (figure 1A). Genetic overexpression was then verified by RT-PCR

and ELISA or Western blot. Initially, transfection was performed using WT mouse ADSCs

with the reporter gene enhanced green fluorescent protein (EGFP), which verified that PBAE

nanoparticles improved transfection efficiency (19.9±2.0%) compared to Lipo controls

(5.4±0.8%, p < 0.05) (figure S1A). For all remaining experiments, transfections were carried

out in luciferase positive/GFP positive (luc+/GFP+) mouse ADSCs; these cells enabled live cell

tracking during animal experiments.

RT-PCR and Western blot confirmed successful CXCR4-overexpression in ADSCs

using PBAE nanoparticles (figure 1B, 1C). As CXCR4 is a transmembrane receptor, cell

surface expression was quantified by flow cytometry (FACS). Cell surface expression of

CXCR4 was statistically greater than GFP-transfected controls (figure 1D); however, the

increased cell surface expression was subtle (1.41±0.02% vs 0.9±0.11%, p < 0.05) suggesting

that CXCR4 is predominantly intracellular as supported by previous studies [17]. To understand

the kinetics of CXCR4 overexpression, CXCR4 mRNA levels were quantified over 10 days.

This demonstrated that CXCR4 expression peaked at 48 hours post-transfection and remained

high over 8 days (figure 1E). VEGF-overexpression was similarly confirmed by RT-PCR

(figure 1F) and VEGF ELISA (figure 1G). As with CXCR4 expression, VEGF release peaked

at 48 hours and was significantly higher in VEGF-transfected ADSCs (2.41±0.19 ng/ml) and

CXCR4/VEGF- co-transfected ADSCs (0.69±0.09 ng/ml) compared to GFP-transfected

controls (0.25±0.4 ng/ml, p < 0.05) (figure 1G). By 10 days post-transfection, accumulated

VEGF release was over three-times greater than non-transfected controls (figure 1H). Finally,

it was demonstrated that cell viability due to PBAE-mediated transfection was not significantly

different compared to non-transfected controls.

78

5.3.2 Transplanted ADSCs overexpressing CXCR4 exhibited prolonged cell survival

To assess the effects of non-viral engineered, CXCR4- overexpressing ADSCs on cell

survival, GFP+/luc+ ADSCs were injected into the medial muscles of a murine model of

hindlimb ischemia and monitored by bioluminescence imaging (BLI) (figure 2A, 2B). GFP-

transfected ADSCs were utilized as a control for PBAE-mediated transfection. BLI

demonstrated that cell engraftment of CXCR4-, VEGF-, and both CXCR4/VEGF- transfected

ADSCs was significantly greater than that of GFP-transfected controls (13.00±6% combined

for CXCR4, VEGF and CXCR4/VEGF groups vs 6.11±3% for GFP group, p < 0.05) (figure

S2A). For all groups, BLI signal reached a peak at day 4, indicating that the transplanted cells

were able to recover and proliferate (figure 2B). At day 7, BLI signal for all groups began to

decline, but CXCR4- and CXCR4/VEGF- transfected groups remained relatively high

(11.20±8.8% combined for CXCR4 and CXCR4/VEGF) compared to the VEGF- and GFP-

transfected groups (2.96±2.6% combined for VEGF and GFP, p < 0.05) (figure 2B, S2B). By

day 10, BLI signal was only detected in CXCR4 and CXCR4/VEGF groups (figure 2A, 2B),

thereby indicating improved cell survival due to CXCR4-overexpression. To confirm genetic

upregulation of CXCR4 or VEGF in transplanted ADSCs, ischemic tissue sections were

harvested on day 4 and analyzed by RT-PCR. Genetic upregulation of CXCR4 and/or VEGF

was confirmed in their respective groups (figure 2C), which verified the cell survival of

transplanted ADSCs to day 4.

5.3.3 Improved therapeutic efficacy due to CXCR4-overexpressing ADSCs

Therapeutic efficacy was monitored by observation of clinical limb outcome and

functional blood reperfusion. Importantly, CXCR4-overexpression led to 100% clinical limb

salvage (n = 8), which was the best therapeutic response of all groups (figure 2D, 2I). Co-

transfection with CXCR4/VEGF also led to improved clinical limb salvage compared to VEGF-

and GFP-transfected groups, albeit with two mice suffering toe necrosis and one mouse

suffering foot necrosis. Interestingly, VEGF-transfection led to 50% limb salvage, which was

greater than controls, but also led to the highest amount of limb loss. With regards to functional

blood reperfusion, laser Doppler perfusion imaging (LDPI) was performed weekly (figure 2E,

2G). LDPI measurements performed on the day of surgery confirmed successful induction of

hindlimb ischemia (figure 2G). By day 7, the CXCR4-overexpressing groups demonstrated a

robust reperfusion response, including CXCR4- and CXCR4/VEGF- groups, compared to

VEGF- and GFP- transfected groups (p < 0.05) (figure 2G, S3). Thereafter, relative reperfusion

79

levels were not statistically different from each other, indicating that the host regenerative

response may have taken over.

5.3.4 Improvement of muscle regeneration in ischemic hindlimbs due to CXCR4-

overexpressing ADSCs

Muscle histology was performed to determine the health of the ischemic hindlimbs at 4

weeks post-ischemic induction (figure 3A). Sections were stained with Masson’s Trichrome

and assessed for muscle regeneration or damage (fibrosis and atrophy). Several regenerating

muscle fibers were observed in the CXCR4- and CXCR4/VEGF- groups, as indicated by the

presence of centrally located nuclei. This suggested that these tissues had reached late stage

muscle repair. In contrast, there were few regenerating muscle fibers noted in the VEGF-

groups; instead, frank fibrosis was observed throughout. Similarly, much fibrosis and muscle

atrophy was observed in GFP- groups. Considering that this was a 4 week time point, it is likely

that these tissues had undergone chronic fibrosis.

5.3.5 Effects of CXCR4- and/or VEGF- overexpressing ADSCs on angiogenic and

inflammatory response

As angiogenesis and inflammation are important processes in muscle repair, ischemic

tissue sections were immunostained for capillaries (anti-CD31/PECAM) and macrophages

(anti-CD68). Transplanted cells were immunostained with anti-GFP antibody to determine the

location of their long-term engraftment (figure 3B). In day 28 tissue sections, the few remaining

transplanted ADSCs identified had become incorporated into nascent capillaries (figure 3B).

Their engraftment location suggested a potential role for these transplanted cells in the

angiogenic process. To gauge the magnitude of the angiogenic response, capillaries were

quantified by high powered field (hpf) at 4 weeks. As expected, VEGF-overexpressing ADSCs

led to the largest increase in vessel density (48.9±10 vessels per hpf) compared to GFP-

transfected controls (42.7±2 vessels per hpf, p < 0.05) (figure 3D). Co-transfection with

CXCR4/VEGF also resulted in increased vessel density, but this was not statistically greater

than GFP-controls. Interestingly, vessel density in the CXCR4- group was not statistically

different compared to GFP- controls.

Because the initial inflammatory response generally peaks around days 3-7, day 4 tissue

sections were stained for macrophages, the expected dominant inflammatory cell population at

this time point. This analysis revealed a large inflammatory response in ischemic muscles

80

transplanted with VEGF- overexpressing ADSCs (figure 3C). Indeed, both VEGF- and

CXCR4/VEGF- groups led to a statistically significant increase in CD68 macrophages

compared to GFP- controls (p < 0.05) (figure 3E). CXCR4- overexpressing ADSCs led to an

apparent increase in CD68 macrophages, but this was not statistically greater than GFP-

controls. RT-PCR for inflammatory cytokines, including IL-1β, TNF-α, IL-6, and IL-10, was

performed on day 4 tissue sections to complement immunostaining assessment for

inflammation. Consistent with CD68 staining, the VEGF group led to significant increases in

pro-inflammatory cytokines IL-1β (p < 0.001) (figure 3F) and TNF-α (p < 0.001) compared to

controls (figure 3G), while co-transfection by CXCR4/VEGF led to significant increases in

TNF-α only (p < 0.001). IL-6 was upregulated in both CXCR4- and CXCR4/VEGF- groups

compared to VEGF- and GFP- groups (p < 0.001), indicating that CXCR4 had a role in

upregulating this cytokine. The anti-inflammatory cytokine IL-10 followed similar trends to

that of TNF-α, suggesting that its upregulation may have occurred in response to the increase in

pro-inflammatory cytokines (figure 3I).

5.3.6 CXCR4-overexpression in ADSCs promotes a pro-angiogenic and anti-inflammatory

phenotype in vitro hypoxia culture

To better understand how CXCR4-overexpression might be promoting therapeutic

response into ischemic tissues, transfected ADSCs were cultured under hypoxia in vitro and

assessed for genetic markers of angiogenesis and inflammation. Interestingly, CXCR4-

overexpression led to genetic upregulation of angiogenic factors compared to GFP- and cells

only controls, including mouse VEGF (p < 0.05), basic fibroblast growth factor (bFGF) (p <

0.05), and hepatocyte growth factor (HGF) (p < 0.05) (figure 4A). With regards to inflammatory

cytokines, CXCR4-overexpression led to significant decreases in IL-1β (p < 0.05) and IL-6 (p

< 0.05) compared to GFP- and cells only controls (figure 4B). These trends were not observed

for CXCR4/VEGF- or VEGF- transfected groups.

Paracrine release from transfected ADSCs was assessed against human umbilical vein

endothelial cells (HUVECs) and THP-1 macrophages in assays of HUVEC proliferation and

THP-1 migration in vitro. Both CXCR4-overexpressing ADSCs and VEGF-overexpressing

ADSCs increased HUVEC proliferation compared to controls (p < 0.05), which was not seen

by co-transfection with CXCR4-VEGF. THP-1 migration was significantly enhanced by

CXCR4-VEGF and VEGF-transfected ADSCs compared to controls (p < 0.05), which was not

81

observed by CXCR4-ADSCs. These results indicated that CXCR4-overexpression might be

promoting a pro-angiogenic and anti-inflammatory in response to hypoxia (figure 4C).

5.4 Discussion

PAD remains a significant problem in the U.S. that affects roughly 12% of U.S. adults.

For many patients, treatment remains lacking due to the presence of co-morbidities that prohibit

them from undergoing surgical revascularization. Therapeutic angiogenesis using stem cells

has potential for treating PAD, but is hindered by poor cell engraftment and insufficient

paracrine signaling. Genetic engineering can improve stem cell therapy by enhancing cellular

processes related to cell survival, migration, and paracrine signal release [3]. One interesting

target known to mediate these processes is CXCR4, which is involved in host stem cell homing

and survival in relation to ischemia and injury [6]. In this study, we found that ADSCs

engineered to overexpress CXCR4 using biodegradable polymers enhanced cell engraftment

and survival, which subsequently led to a quicker reperfusion response and complete limb

salvage. Because this approach used a readily available stem cell population and a non-viral

transfection agent, it represents a potentially clinically translatable method for treating PAD.

As one of the main goals for treating PAD is the prevention of ischemic tissue damage,

the most interesting result in our study was that CXCR4-overexpressing ADSCs led to complete

limb salvage in all hindlimb ischemia murine models studied (100% limb survival). In

comparison, VEGF-overexpressing ADSCs resulted in less than 60% limb survival. Muscle

histology further demonstrated that CXCR4-overexpression improved healing response by

indicating wide-spread muscle regeneration that was not observed in the VEGF- and GFP-

overexpressing groups. Together, these findings clearly demonstrate that CXCR4-

overexpressing ADSCs have a beneficial effect and promote tissue salvage. Furthermore,

because our findings for VEGF-overexpressing ADSCs were similar to results obtained in prior

studies using VEGF-transfected EPCs [18] and VEGF-transfected MSCs [4], it is reasonable to

compare our results to prior work.

In considering the mechanism of repair, CXCR4-overexpressing ADSCs likely acted

by promoting an early angiogenic response and moderating inflammation. We found that

CXCR4-overexpressing ADSCs led to a robust reperfusion response by day 7 which then

aligned with the host healing response by day 14 (figure 2G). Histological sections at day 28

revealed a low relative number of capillaries compared to VEGF-overexpressing ADSCs, which

suggested that the tissue repair process had returned to baseline. This phenomenon can be

82

explained by the natural response to ischemic injury, where the blood supply increases above

normal levels during healing and then gradually returns to pre-ischemic levels upon complete

repair [19]. The other aspect of tissue repair is the early inflammatory phase, which is a

necessary process that can become damaging to the tissue if excessive. We found that CXCR4-

ADSCs tended to moderate inflammation, while VEGF-ADSCs led to an exuberant and likely

damaging inflammatory response as indicated by the increase in inflammatory cytokines (IL-1

and TNF-α) and CD68+ tissue macrophages. These findings were further supported in vitro,

where CXCR4-ADSCs cultured under hypoxia promoted endothelial cell proliferation and

decreased THP-1 macrophage migration.

Cell engraftment and survival remain challenges for stem cell-based therapeutics. Our

study suggests that CXCR4-overexpresson may help both of these processes. On BLI, we found

that CXCR4-overexpressing ADSCs, including CXCR4- and CXCR4/VEGF- ADSCs,

demonstrated high BLI signal compared to controls at 24 hours, a trend which persisted over 14

days (figure 2B). As CXCR4 is a G-protein coupled receptor known to activate multiple

intracellular signaling cascades leading to cell migration and cell survival [6], these findings

suggest a role for utilizing CXCR4 to overcome issues related to cell engraftment and survival.

The improvements in cell engraftment and survival likely contributed to the early angiogenic

response achieved by CXCR4- overexpressing ADSCs, which then promoted tissue salvage.

Two important characteristics for successful clinical translation of combined stem cell-

gene therapy include: 1) an available and abundant cell population and 2) the use of a non-viral

gene transfection vector. With regards to cell population, ADSCs represent a relatively

abundant, autologous cell source that are easy to isolate. While this is clear, the therapeutic

potential for PAD is still being explored. Indeed, multiple studies have demonstrated their

therapeutic effects in murine models of a hindlimb ischemia [11, 12, 14, 20]. We further studied

the effects of VEGF-overexpressing ADSCs on endothelial cells in vitro, finding enhancement

in cellular process of angiogenesis, in a prior report [15]. Carrying this forward here, we have

demonstrated that CXCR4-overexpressing ADSCs support angiogenesis and tissue repair in a

murine model of hindlimb ischemia. Overall, these findings, in combination with their ready

availability, support ADSCs as a clinically applicable therapeutic stem cell source.

With regards to non-viral gene transfection vectors, such vectors are clinically preferred

because they may overcome many of the issues related to viral vectors, such as immunogenicity

and insertional mutagenesis. However, many non-viral vectors suffer low transfection

efficiency and high cytotoxicity. In our platform, we used a biodegradable polymer for non-

83

viral gene transfection, poly(β-amino esters) (PBAE), which had been previously optimized

based on combinatorial polymer synthesis and high-throughput screening [21, 22]. We further

screened these polymers for transfection in stem cells, where we identified PBAE C32-122 for

efficiently transfecting multiple stem cell types with low cytotoxicity, including embryonic stem

cells, BM-MSCs, and ADSCs [5]. [20]. In the present study, PBAE C32-122 was used to

achieve successful overexpression of CXCR4 and VEGF. Protein expression of each gene was

clearly demonstrated by Western blot or ELISA for CXCR4 or VEGF, respectively.

Furthermore, genetic upregulation by PBAE C32-122 is transient, suggesting that the plasmid

DNA is not integrated into the genome and would not lead to insertional mutagenesis. Thus,

the use of PBAE C32-122 for transfection into ADSCs constitutes a platform that may meet

clinical standards for combined stem cell-gene therapy.

While our study shows promise for non-viral engineered CXCR4-overexpressing

ADSCs for treating PAD, some limitations of the study include: 1) the use of small rodent model

to characterize therapeutic response and 2) the reliance on ex vivo genetic engineering. Use of

a large animal of a similar size to humans would better represent ischemic tissue damage and

repair because of the dependence of diffusion distance for oxygen and nutrient delivery.

Therefore, further study in large animals is warranted that may discover beneficial modifications

to the present platform for human therapy. The next limitation relates to using a multi-step

procedure to transfect stem cells, whereby ADSCs would be isolated in one procedure, then ex

vivo expanded and transfected, and then delivered back to the patient in a second procedure.

The process would increase cost and regulatory concerns, and may perhaps cause variability in

ADSC phenotype. To overcome this issue, platforms based on a one-step approach (i.e.

enabling cell isolation and transfection in the same procedure) would overcome many of the

concerns related to regulatory and cost, and may thus be more readily applied in the clinic.

In summary, this study suggests that CXCR4-overexpressing ADSCs using

biodegradable polymers can promote therapeutic angiogenesis and treat PAD. CXCR4

overexpression supports cell engraftment and survival and promotes an appropriate healing

response to ischemic injury. This approach has clinical potential for the engineering and

regeneration of many other tissue types, as well as in treating other types of ischemic diseases

such as myocardial infarction and cerebral ischemia.

84

Acknowledgments: Thank others for any contributions. Funding: This work was funded by

the American Heart Association (10SDG2600001) supported by Stanford Medical Scholars

Research funding and Stanford Medical Scientist Training Program.

Supplementary Materials

Fig. S5.1. Improvement of transfection efficiency and low cytotoxicity by PBAE nanoparticles.

Fig. S5.2. Cell engraftment of transfected ADSCs.

Fig. S5.3. Reperfusion of the ischemic hindlimb at one week post-procedure.

Table S5.1: List of primers used for quantitative real time-polymerase chain reaction

References:

1. Selvin, E. and T.P. Erlinger, Prevalence of and risk factors for peripheral arterial disease

in the united states results from the national health and nutrition examination survey,

1999–2000. Circulation, 2004. 110(6): p. 738-743.

2. Criqui, M.H., et al., Mortality over a period of 10 years in patients with peripheral

arterial disease. New England Journal of Medicine, 1992. 326(6): p. 381-386.

3. Fischbach, M.A., J.A. Bluestone, and W.A. Lim, Cell-based therapeutics: the next pillar

of medicine. Sci Transl Med, 2013. 5(179): p. 179ps7.



p. 3317-22.



46.

6. Busillo, J.M. and J.L. Benovic, Regulation of CXCR4 signaling. Biochim Biophys

Acta, 2007. 1768(4): p. 952-63.



434-8.



9. Jin, D.K., et al., Cytokine-mediated deployment of SDF-1 induces revascularization

through recruitment of CXCR4+ hemangiocytes. Nat Med, 2006. 12(5): p. 557-67.

85

10. Petit, I., D. Jin, and S. Rafii, The SDF-1–CXCR4 signaling pathway: a molecular hub

modulating neo-angiogenesis. Trends in Immunology, 2007. 28(7): p. 299-307.



12. Kim, Y., et al., Direct comparison of human mesenchymal stem cells derived from

adipose tissues and bone marrow in mediating neovascularization in response to

vascular ischemia. Cell Physiol Biochem, 2007. 20(6): p. 867-76.

13. Cai, L., et al., IFATS collection: Human adipose tissue-derived stem cells induce

angiogenesis and nerve sprouting following myocardial infarction, in conjunction with

potent preservation of cardiac function. Stem Cells, 2009. 27(1): p. 230-7.





p. 483-91.

16. Niiyama, H., et al., Murine model of hindlimb ischemia. J Vis Exp, 2009(23).

17. Schantz, J.T., H. Chim, and M. Whiteman, Cell guidance in tissue engineering: SDF-1

mediates site-directed homing of mesenchymal stem cells within three-dimensional

polycaprolactone scaffolds. Tissue Eng, 2007. 13(11): p. 2615-24.

18. Iwaguro, H., et al., Endothelial progenitor cell vascular endothelial growth factor gene

transfer for vascular regeneration. Circulation, 2002. 105(6): p. 732-8.

19. Hong, G., et al., Near-Infrared II Fluorescence for Imaging Hindlimb Vessel

Regeneration with Dynamic Tissue Perfusion Measurement. Circulation:

Cardiovascular Imaging, 2014.

20. Cao, Y., et al., Human adipose tissue-derived stem cells differentiate into endothelial

cells in vitro and improve postnatal neovascularization in vivo. Biochem Biophys Res

Commun, 2005. 332(2): p. 370-9.

21. Green, J.J., et al., Combinatorial modification of degradable polymers enables

transfection of human cells comparable to adenovirus. Advanced Materials, 2007.

19(19): p. 2836-+.



86

Figures:

Figure 5.1. Genetic upregulation of CXCR4 and/or VEGF in ADSCs by PBAE nanoparticles.

(A) Schematic showing experimental design. Briefly, ADSCs were transfected ex vivo by first

generating PBAE nanoparticles utilizing plasmid DNA encoding CXCR4 or VEGF and then

applying to cells. Gene transfected ADSCs were transplanted into murine models of hindlimb

ischemia. CXCR4-overexpression was verified by RT-PCR (B) and by western blot (C) at 48

hours post-transfection. (D) FACS quantification of CXCR4 cell surface expression. (E)

Kinetics of CXCR4 mRNA expression over 10 days. VEGF overexpression was verified by RT-

PCR (F) and ELISA (G) at 48 hours. (H) VEGF release over time quantifying release every 48

hours. (I) Effects of PBAE transfection on ADSC cell viability. (* p < 0.05, ** p < 0.01

compared with GFP-ADSC controls)

87

Figure 5.2. Improvement of cell fate and therapeutic efficacy due to CXCR4-overexpressing

ADSCs. (A) Representative BLI images at day 10. (B) Percentage of initial BLI signal

remaining over the first 14 days post-procedure. (C) RT-PCR analysis of harvested tissue at 4

days post-procedure measuring human CXCR4 mRNA and human VEGF mRNA (relative to

limbs transplanted with CXCR4-ADSC and limbs transplanted with VEGF-ADSC,

respectively). Representative images of clinical limb outcome (D) and laser doppler perfusion

imaging (E) at day 14 post-procedure. Ischemic limbs are indicated by red circles.

Quantification of clinical limb outcome at day 21 (F) and relative reperfusion over 21 days (G).

(* p < 0.05 compared to GFP-ADSC controls)

88

Figure 5.3. CXCR4-overexpressing ADSCs enhanced muscle regeneration and mediated host

angiogenic and inflammatory response. (A) Representative tissue sections at 28 days post-

procedure stained with Masson’s Trichrome. (B) Immunostaining of tissue sections at 28 days

post-procedure for capillaries (anti-CD31: red), transplanted cells (anti-GFP: green), and

cellular counterstain (Dapi: blue). (C) Immunostaining of tissue sections at 4 days post-

procedure for macrophages (anti-CD68: red) and cellular counterstain (Dapi: blue). (D)

Quantification of CD31-positive capillaries per high powered field (hpf). (E) Percentage of

CD68-positive macrophages to all cells. RT-PCR analysis on harvested tissue at day 4 for

inflammatory cytokines, including IL-1β (F), TNF-α (G), IL-6 (H), and IL-10 (I). (* p < 0.05,

** p < 0.01, *** p < 0.001 compared to GFP-ADSC controls; # p < 0.05 compared to VEGF-

ADSCs)

89

Figure 5.4. CXCR4-overexpression in ADSCs promotes a pro-angiogenic and anti-

inflammatory phenotype under hypoxia. Transfected ADSCs were cultured under hypoxia and

assessed at 48 hours by RT-PCR for markers of angiogenesis (A) and inflammation (B) in

comparison to normoxia culture. (C) Paracrine release from transfected ADSCs applied to

HUVECs and assessed for MTS viability. (D) Paracrine release from transfected ADSCs

assessed for THP-1 macrophage migration. (E) Proposed model for CXCR4- mediated

therapeutic response to tissue ischemia. SDF-1α is secreted from sites of ischemia, which

activates CXCR4 on transfected ADSCs to promote an ischemia specific therapeutic response.

90

Supplementary Materials:

Figure S5.1. Improvement of transfection efficiency and low cytotoxicity by PBAE

nanoparticles. WT ADSCs were transfected with the reporter GFP utilizing PBAE nanoparticles

or Lipo as a control. (A) FACS quanitification of GFP-positive cells. (B) Cell viability of

transfected cells. (*** p < 0.001).

Figure S5.2. Cell engraftment of transfected ADSCs. Percentage of BLI signal remaining from

transplanted cells measured on the day of transplantation at 24 hours (A) and 7 days (B). (* p <

0.05).

91

Figure S5.3. Reperfusion of the ischemic hindlimb at one week post-procedure. Relative

reperfusion as assessed by LDPI at one week. (* p < 0.05)

92

Table S5.1: List of primers used for quantitative real time-polymerase chain reaction

Gene

Name

Species Primer Sequence

GAPDH Mouse Forward (5'-AACGACCCCTTCATTGAC-3')

Reverse (5'-TCCACGACATACTCAGCAC-3')

CXCR4 Human Forward (5’- TTCTACCCCAATGACTTGTG-3’)

Reverse (5’- ATGTAGTAAGGCAGCCAACA-3’)

VEGF Human Forward (5’-TACCTCCACCATGCCAAGTG-3’)

Reverse (5’-TGATGATTCDTGCCCTCCTCC-3’)

VEGF Mouse

Forward (5'-

GCCCTGGAGTGCGTGCCCACGTCAGAGAGCA-3')

Reverse (5'-TGGCGATTTAGCAGCAGATA-3')

bFGF Mouse Forward (5'-GCCAGCGGCATCACCTCGCT-3')

Reverse (5'-TATGGCCTTCTGTCCAGGTCCCGT-3')

HGF Mouse Forward (5'-TTGGCCATGAATTTGACCTC-3')

Reverse (5'-ACATCAGTCTCATTCACAGC-3')

IL-1β Mouse Forward (5'-CAACCAACAAGTGATATTCTCCAT G-3')

Reverse (5'-GATCCACACTCTCCAGCTGCA-3')

TNF-α Mouse Forward (5'-CATCTTCTCAAAATTCGAGTGACAA-3')

Reverse (5'-TGGGAGTAGACAAGGTACAACCC-3')

IL-6 Mouse Forward (5'-TGGCTAAGGACCAAGACCATCCAA-3')

Reverse (5'-AACGCACTAGGTTTGCCGAGTAGA-3')

IL-10 Mouse Forward (5'-GGTTGCCAAGCCTTATCGGA-3')

Reverse (5'-ACCTGCTCCACTGCCTTGCT-3')

6. Adipose-Derived Stem Cells Co-expressing Hepatocyte Growth Factor and CXCR4

Promotes Ischemic Tissue Repair

Jeffrey Choi1*, Lorenzo Deveza2,3*, Sungwon Lim2, Fan Yang2,4

1Department of Biology, Stanford University, Stanford, CA; 2Department of Bioengineering,

Stanford University, Stanford, CA; 3Stanford University School of Medicine, Stanford, CA; 4Department of Orthopaedic Surgery, Stanford University, Stanford, CA.

Abstract

Stem cells hold great potential for therapeutic angiogenesis due to their ability to respond to

environmental cues and secrete beneficial growth factors. However, stem cell therapies have

been largely unsuccessful in clinical trials, which is in part related to issues of poor cell

engraftment and cell homing, as well as insufficient paracrine signaling. To overcome these

issues, we engineered adipose-derived stem cells (ADSCs), a clinically abundant cell source, to

dually express CXC chemokine receptor 4 (CXCR4) and hepatocyte growth factor (HGF);

based on the involvement of CXCR4 in stem cell homing to sites of ischemia, as well as the

potent angiogenic effects of HGF. To achieve this, we developed a bicistronic plasmid encoding

both genes and then assembled polymeric nanoparticles using poly(β-amino esters) (PBAE), a

non-viral transfection vector, prior to delivery to ADSCs ex vivo. Engineered syngeneic ADSCs

were transplanted into aged mice surgically induced with severe hindlimb ischemia to evaluate

their therapeutic efficacy. HGF overexpression resulted in increased overall angiogenesis in the

ischemic hindlimb based on blood reperfusion and density of vascular structures. CXCR4

overexpression prolonged ADSC retention in vivo and led to improved muscle integrity with a

muted inflammatory response. Notably, ADSCs co-expressing HGF and CXCR4 not only

exhibited increased cell survival, but also quenched early inflammation and promoted functional

angiogenesis, leading to improved limb salvage. These results demonstrate that non-viral

transfer of multiple genes via a single plasmid to ADSCs can tailor their behavior for treating

ischemic skeletal muscle, and perpetuate the larger goal of engineering cell-based therapeutic

tools that can be controlled and tuned for different diseases.

6.1 Introduction

Peripheral arterial disease (PAD) represents an important, yet underrepresented medical

problem, caused by decreased blood flow to the limbs. It may signify severe underlying

93

atherosclerosis and often occurs with advanced Diabetes Mellitus. Treatment options for PAD

resemble those of other ischemic diseases, including surgical revascularization via bypass

surgery or percutaneous angioplasty. However, because of the presence of surgically

prohibitive co-morbidities, many patients are without treatment.

Stem cell-based therapeutics are an alternative treatment option for patients unable to

withstand surgery [1]. It involves transplanting autologous stem cells to ischemic regions to

induce host-blood vessel growth in a process known as therapeutic angiogenesis, and likely acts

via secreting beneficial paracrine factors [2]. Multiple stem cell sources have shown therapeutic

effect in pre-clinical animal models [3-6], but overall success in human clinical trials has been

limited [7]. This lack of response in human trials might be attributed to insufficient paracrine

signaling, poor stem cell engraftment, or an inappropriate response to the disease state [2, 8].

Ex vivo genetic manipulation of stem cells may improve their behavior by enhancing

those cellular process involved in therapeutic angiogenesis [8]. Prior strategies to enhance stem

cell-therapeutics have included genetic modification with angiogenic growth factors [9, 10],

survival factors [11], and homing factors [12]. However, as stem cells have complex

intracellular signaling pathways and respond to environmental stimuli, single genetic

modifications with paracrine signals might underappreciate the potential for stem cell

therapeutics. Stem cells could be engineered to overexpress appropriate cellular receptors to

respond to environmental cues, as well as overexpress paracrine signals to induce a therapeutic

response. In particular, CXC chemokine receptor 4 (CXCR4) is a receptor involved in stem cell

homing and is known to localize cells to sites of ischemia in response to stromal derived factor-

1α (SDF-1α) [13]. Hepatocyte growth factor (HGF) is a potent angiogenic growth factor that

has shown great promise in human clinical trials for treating clinical limb ischemia [14, 15].

HGF is also known to have activity against both endothelial cells and smooth muscle cells, and

may thereby stimulate both angiogenic initiation and maturation. Dual overexpression of

CXCR4 and HGF may thus promote stem cell-based therapeutics by localizing stem cells to

sites of ischemia, where they may appropriately deliver HGF to induce an angiogenic response.

Methods to achieve genetic overexpression of two or more genes, such as CXCR4 and

HGF, typically use as many plasmids to transfect cells ex vivo. While this approach can be

satisfactory, there is no guarantee that the same cell is transfected with both genes; and, dually

transfected cells may vary in the relative expression of each gene due to variations in the number

of plasmid copies or by the promoter regulating each gene. An alternative approach is to

recombine both genes into a single plasmid and design the genes to be regulated by a single

94

promoter. Engineering both CXCR4 and HGF into a single bicistronic plasmid could thus

ensure that both genes are transfected into the same cell and expressed under a single promoter.

Clinically, a successful approach for combined stem cell/gene therapy would be safe

and broadly applicable. To meet this criteria, our approach has been to utilize non-virally

transfected adipose-derived stem cells (ADSCs) [16, 17]. Poly(β-amino esters) (PBAE) are a

class of non-viral gene transfection vectors that have been optimized for efficiently transfecting

stem cells for therapeutic angiogenesis, including ADSCs [9, 18]. ADSCs are an easy to isolate

and clinically abundant cell source [19] that naturally possesses angiogenic potential [6, 20],

making them excellent candidates for treating ischemic disease. Indeed, we have demonstrated

that ADSCs can be efficiently engineered using PBAE and that paracrine release from VEGF-

transfected ADSCs improves therapeutic response both in vitro [16] and in vivo [17].

In this study, we genetically engineered ADSCs with a bicistronic plasmid encoding

both HGF and CXCR4 using PBAE biodegradable polymers, and then assessed their ability for

therapeutic efficacy in an aged murine model of hindlimb ischemia. Two types of dually

expressing plasmids for HGF and CXCR4 were designed and assessed for genetic upregulation;

using an optimized HGF gene sequence that expresses two isoforms of HGF [21]. PBAE-

mediated transfection of ADSCs was then optimized using the bicistronic plasmid. Therapeutic

efficacy was assessed in retired breeder mice induced with hindlimb ischemia to simulate a

difficult healing situation in an immunocompetent mouse (many previous studies have focused

on younger mice and/or nude mice). Outcomes analysis in vivo included evaluating cell survival

and engraftment, angiogenic and inflammatory response, and muscle morphology.


6.2.1 Materials

Monomers for synthesizing poly(β-amino esters)were purchased from Sigma–Aldrich,

Scientific Polymer Products, and Molecular Biosciences. Anhydrous dimethyl sulfoxide

(DMSO) was purchased from Sigma–Aldrich. A 25 mM sodium acetate (NaOAc) buffer

solution (pH 5.2) was prepared by diluting a 3 M stock solution (Sigma-Aldrich, St Louis, MO).

Plasmids (pEGFP-N1 and pcDNA3.1(+)-CXCR4) were purchased from Elim

Biopharmaceuticals and pck-HGF plasmid was obtained from Viromed (Seoul, Korea).

95

6.2.2 Poly(β-Amino Esters) Synthesis

PBAE polymers were synthesized as previously described [9]. Briefly, acrylate-

terminated D32 polymer (C32-Ac) was prepared by mixing hexanediol (D) and aminopentanol

(32) monomers using a 1.2:1.0 diacrylate/amine monomer molar ratio at 90 °C for 24 h.

Subsequently, amine-terminated D32 polymers were generated by reacting D32-Ac with

diamine monomer (122) in DMSO. End-capping reactions were performed overnight at room

temperature by using a 1.6-fold molar excess of amine over acrylate end groups.

6.2.3 ADSC isolation and culture

ADSCs were collected by harvesting the inguinal fat pads from 5 to 6 weeks old

luciferase-positive/GFP-positive transgenic male mice (Joseph Wu, MD, PhD laboratory,

Stanford University) and FVB wild-type male mice (purchased from Charles River, Willington,

MA). Following serial betadine washes, the fat pads were finely minced in cold, sterile PBS,

followed by digestion in 0.075% Collagenase II (Sigma-Aldrich) at 37 °C for 30 minutes in a

shaking water bath. Digestion was neutralized with Dulbecco’s modified Eagle’s medium

(Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Invitrogen) and 1%

penicillin-streptomycin (Sigma-Aldrich). Neutralized cells were centrifuged at 4 °C and 1,000

rpm for 5 minutes, and floating adipocytes and media were removed before pelleted ADSCs

from the stomal-vascular portion were resuspended in culture media, passed through a 100 μm

Falcon nylon cell strainer (Becton Dickinson, Bedford, MA) and plated on a 10 cm dish. Cells

were incubated at 37 °C and 5% atmospheric CO2. Twenty-four hours after isolation, the cells

were washed in PBS to remove debris and nonadherent cells, and fresh medium was added to

the cells. Medium was changed every 2 days, and adherent spindle-shaped cells were grown to

subconfluence, passaged using standard methods of trypsinization, and propagated through

passage 2.

6.2.4 Transfection

ADSCs were transfected with pCK-HGF-X7, pCK-HGF-X7-IRES-CXCR4,

pcDNA3.1-CXCR4, or the pcDNA3.1 empty plasmid as a negative control for transfection.

ADSCs were seeded at 65 000 cells per well in clear 24-well plates and incubated over night at

37 °C. Cells were then transfected utilizing novel end-modified PBAE polymers (D32-122).

Transfection efficiency was optimized by varying vector parameters including polymer/DNA

weight ratios (10:1, 12.5:1, and 15:1) and plasmid DNA (pDNA) loading dose (4, 5, 6, and 7 μg

96

per well in a 24-well plate). Polyplexes were formed by mixing pDNA with PBAE polymers in

25 mM sodium acetate buffer and waiting 10 minutes for complexes to form. Polyplexes were

added to the cells cultured in growth medium containing 10% serum. Cells were then incubated

with polyplexes for 4 hours, at which time complexes were removed and replaced with fresh

medium.

6.2.5 Quantitative real-time polymerase chain reaction (qRT-PCR)

To evaluate transfection efficiency, total RNA was harvested from cells using RNeasy

Kit (Qiagen, Germantown, MD, USA). cDNA was synthesized by reverse transcription using

the Superscript First-Strand Synthesis System (Invitrogen) and 1 μg of total RNA. Quantitative

PCR was performed using Power SYBR Green PCR Master Mix () following the manufacturer's

protocol and running all samples for 40 PCR cycles on Applied Biosystems 7900 Real-Time

PCR System (Carlsbad, CA, USA). The expression of the transfected genes – HGF and CXCR4

– was examined. Relative expression level of target genes was determined by the comparative

CT method, where target gene expression was first normalized to an endogenous housekeeping

gene, GAPDH, followed by a second normalization to the gene expression level measured in

untransfected control cells.

6.2.6 Cell Viability/Proliferation Assay

To evaluate possible cytotoxic effects of transfection, cell viability or proliferation was

evaluated 48 hours post-transfection using the Cell Titer 96 Aqueous One Solution assay kit

(Promega) to measure metabolic activity by optical absorbance on a plate reader. Measurements

of cells transfected with polymer-plasmid polyplexes were converted to relative absorbance

values in comparison to untransfected control cells.

6.2.7 Mouse Hindlimb Ischemia Model and Cell Transplantation

FVB female retired breeder mice were housed five per cage in a 12-hour light/dark

cycle and provided ad libitum with standard food and water. Aged mice were used to better

distinguish the effect of therapeutic intervention from the inherent robust recovery of young

mice. Hindlimb ischemia was induced as previously described (28). Briefly, the femoral artery

and its branches were ligated through a skin incision with 5-0 silk suture (Ethicon). The external

iliac artery and all the above arteries were then ligated. The femoral artery was excised from its

proximal origin as a branch of the external iliac artery to the distal point where it bifurcates into

97

the saphenous and popliteal arteries. Laser Doppler imaging of blood perfusion was used to

confirm induction of hindlimb ischemia post-operatively.

After inducing hindlimb ischemia, cells (1.0 x 106 cells per injection) were suspended

in 100 μL of PBS and injected intramuscularly into two sites of the gracilis muscle in the medial

thigh by using 27-gauge tuberculin syringes. Five experimental groups (n=5 per group) were

examined as follows: (i) PBS, (ii) ADSC-pcDNA3.1, (iii) ADSC-HGF, (iv) ADSC-CXCR4,

and (v) ADSC-HGF-IRES-CXCR4. All experiments were performed in accordance with the

Stanford University Animal Care and Use Committee Guidelines and approved APLAC

protocols.

6.2.8 Bioluminescence Imaging

ADSC survival was monitored post-injection with bioluminescence imaging

technology. Bioluminescence imaging (BLI) was performed on the day of surgery, day 1, day

4, day 7, day 10, and day 14. Animals were anesthetized by 2–3% inhaled isoflurane and injected

intraperitoneally with D-Luciferin reporter probe at 100 mg/kg body weight. IVIS Spectrum

system (Xenogen, Alameda, CA) was used to image the animals while under 2% inhaled

isoflurane anesthesia. Each animal was scanned until the peak signal was reached, and using

Living Image 4.3.1 software (Caliper Life Sciences), radiance of regions of interest (ROIs) in

photons was recorded. Radiance was quantified in photons per second per centimeter squared

per steridian.

6.2.9 In Vivo Assessment of Ischemic Damage and Limb Perfusion

Physiological status of ischemic limbs was followed up to 3 weeks after treatment.

Hindlimb blood perfusion was assessed using a PeriScan PIM3 laser Doppler system (Perimed

AB, Sweden)on day 0, 1, 4, 7, 14, and 21, as previously described (30). Each animal was pre-

warmed to a body temperature of 37°C to avoid data variations raised from confounding factors.

ROIs were drawn around the entire hindlimbs, and the level of perfusion in ischemic and normal

hindlimbs were quantified using the mean pixel value within each ROI. Relative changes in

blood flow were expressed as the ratio of the left (ischemic) over the right (normal) laser

Doppler blood perfusion.

98

6.2.10 Histology and Immunofluorescence staining

Mice were euthanized on day 7 and day 21 post-surgery and cell transplantation to

retrieve muscle tissues from ischemic sides and normal controls for analyses. Hindlimbs were

embedded in optimal cutting temperature compound and snap frozen for cryosectioning. Lateral

cross-sections of the muscle tissues were collected at 8 μm thickness. Tissue sections were

stained with hematoxylin and eosin to assess muscle integrity in ischemic regions. Muscle

tissue from normal limbs with no surgical alterations was used as a positive control.

For immunofluorescence staining, tissue sections were fixed in 4% paraformaldehyde

for 10 minutes, then blocked with 1% BSA in PBS and permeabilized with 0.1% Triton X-100

in PBS. To detect capillaries in ischemic regions, tissue sections were immunofluorescently

stained with rat monoclonal antibody to CD31 overnight at 4°C, followed by secondary

antibody (Alexa Fluor 594 goat anti-rat) incubation for 1 hour at room temperature. Nuclei

were counterstained with DAPI mounting medium (Vectashield) and images were taken with a

Zeiss fluorescence microscope.





significance. All quantitative data were expressed as mean ± standard error of the mean.

6.3 Results

6.3.1 PBAE-mediated Gene Transfer of bicistronic plasmid co-encoding CXCR4 and

HGF to ADSCs

Non-virally engineered ADSCs were prepared by generating biodegradable polymeric

nanoparticles using poly(β-amino esters) (PBAE) and plasmid DNA, and then delivering the

resulting nanoparticles to ADSCs ex vivo as shown in Figure 1A. To ensure co-delivery of

CXCR4 and HGF to ADSCs, a plasmid was designed to encode for both genes (Figure 1B).

Two types of designs were initially evaluated: a bi-directional plasmid and a bicistronic plasmid

(pCK-HGF-IRES-CXCR4). Based on a preliminary evaluation, the bicistronic was chosen

because it led to higher genetic expression of both HGF and CXCR4 in HEK293T cells (Figure

S1). The PBAE polymer was synthesized as described in the methods (Figure 1C). Transfection

with the bicistronic plasmid was optimized in ADSCs using PBAE by varying polymer to DNA

99

weight ratio (Figure S2) and subsequently exploring DNA dose (Figure 1D, 1E). A polymer to

DNA weight ratio of 10:1 and DNA dose of 7 µg led to the highest gene expression and was

used for all subsequent experiments. The co-expressing bicistronic plasmid led to similar

genetic upregulation of HGF (Figure 1D) and CXCR4 (Figure 1E) as compared to single

expressing plasmids for the corresponding genes. ADSC viability was not significantly affected

by DNA dose as compared to non-transfected controls (Figure 1F).

6.3.2 Validation of Paracrine Release from Transfected ADSCs In Vitro

Paracrine release from transfected ADSCs was evaluated by HGF ELISA and by

assessing bioactivity against human umbilical vein endothelial cells (HUVECs) in vitro. HGF

ELISA demonstrated that ADSCs transfected with either the bicistronic HGF-CXCR4 plasmid

or the single HGF plasmid significantly increased HGF release compared to ADSCs transfected

with CXCR4 or the backbone plasmid (vehicle, p < 0.05; Figure 2A). The HUVEC viability

assay further confirmed that the paracrine release was bioactive, where conditioned medium

from HGF/CXCR4-, HGF- and CXCR4- overexpressing ADSCs led to significant increases in

HUVEC viability compared to vehicle and basal media controls (p < 0.05; Figure 2B). Notably,

CXCR4- overexpressing ADSCs led to similar increases in HUVEC viability as the co-

transfected HGF/CXCR4-ADSCs and the singly transfect HGF- ADSCs, suggesting additional

effects on ADSC paracrine release by CXCR4-overexpression. These data demonstrated that

transfected ADSCs exhibited potential therapeutic behavior.

6.3.3 Confirmation of HGF and CXCR4 Overexpression In Vivo

ADSCs transfected with HGF- and/or CXCR4- ex vivo were transplanted into the

adductor and gastrocnemius muscles of a murine model of hindlimb ischemia. To confirm that

HGF- and/or CXCR4 were upregulated in ADSCs in vivo, the ischemic hindlimb was harvested

at 4 days post-transplantation and assessed for human CXCR4 and human HGF mRNA

expression. Both hHGF and hCXCR4 were overexpressed in the corresponding ischemic

tissues after 4 days in vivo, which confirmed that the transplanted ADSCs were still expressing

these genes (Figure 3A, 3B). This also gave evidence that many transplanted cells remained

viable after 4 days post-transplantation.

100

6.3.4 Enhanced Cell Survival in the Ischemic Hindlimb

For all in vivo studies, ADSCs that constitutively expressed GFP and Luciferase

(GFP+/Luc+) were used, which enabled us to assess ADSC survival and engraftment in vivo by

bioluminescence imaging (BLI). At day 0, BLI demonstrated no significant differences in BLI

signal (data not shown), indicating that each animal had received a nearly equivalent number of

cells at the outset. By Day 1, much BLI signal was lost with roughly 10% of the original BLI

signal remaining; however, higher signal was observed in mice transplanted with HGF and/or

CXCR4- overexpressing ADSCs, compared to ADSCs transfected with the vehicle control

(Figure 3C). On kinetic read, BLI signal either increased or remained stable from day 1 to day

4, but dropped by day 7 and was undetectable by day 14. Only ADSCs transfected with CXCR4

(i.e. HGF/CXCR4 and CXCR4 only) were able to recover from day 1 to day 4. This indicated

that CXCR4 promoted cell survival post-transplantation compared to vehicle transfection,

whether by single transfection or co-transfection with HGF (p < 0.05; Figure 3D).

6.3.5 Increased Angiogenic Response in the Aged Mouse

Angiogenesis and reperfusion in the murine hindlimb ischemia model was assessed by

Laser Doppler Reperfusion Imaging (LDPI), as well as by immunofluorescence staining of

CD31 (capillaries) and alpha-smooth muscle actin (α-SMA; arterioles) (Figure 4). LDPI

confirmed that ischemic induction was successfully achieved (Day 0 images; Figure 4A), with

progressive reperfusion that occurred over several weeks. The largest changes in reperfusion

occurred between days 7 and 14 (Figure 4B). At day 7, a trend emerged where the HGF

transfected ADSCs led to the highest reperfusion response, followed by the HGF/CXCR4-

ADSCs and CXCR4-ADSCs, then finally, the vehicle-transfected ADSCs (Figure 4B). By day

14, reperfusion by HGF/CXCR4-ADSCs had caught up with the HGF-ADSCs, indicating that

both groups led to a robust reperfusion response (Figure 4B). These findings were further

supported by immunohistochemical staining at day 14, where CD31 capillary densitometry

demonstrated that HGF-ADSCs and HGF/CXCR4-ADSCs resulted in a significantly greater

number of capillaries than vehicle controls (p < 0.05; Figure 4C, 4E). Arteriole assessement

also demonstrated that these groups led to a statistically significant increase in the number of

arterioles compared to vehicle controls (p < 0.05; Figure 4D, 4F). Overall, these data

demonstrated that HGF was associated with an increased angiogenic response, which could be

achieved with either the single plasmid or the dual-expressing plasmid.

101

6.3.6 Promotion of Ischemic Limb Health

Limb health was assessed by gross limb status and muscle histology. Hindlimb

ischemia models in the aged mouse were found to be very severe in controls receiving only

PBS, which led to significant limb loss and necrosis (data not shown). Transplantation of

ADSCs that overexpressed either HGF or CXCR4 (or both) was able to rescue limb necrosis,

and led to complete limb salvage for many mice (Figure 5A, 5D). The healthiest limb status

was achieved by both HGF/CXCR4-ADSCs and CXCR4 only-ADSCs (Figure 5A). Day 4 limb

status measurements demonstrated that HGF/CXCR4-ADSCs and HGF-ADSCs caused some

early toe necrosis, which was less prevalent with CXCR4-ADSCs, but that both had the same

amount of limb salvage and toe necrosis by the final time point. Some foot necrosis was

observed by HGF-ADSC transplantation (Figure 5D), but was overall better than vehicle-ADSC

controls. Masson’s Trichrome staining of day 7 tissue samples was consistent with limb status

measurements, demonstrating good overall health of the muscle by HGF and/or CXCR4-

overexpressing ADSCs, but not by vehicle-ADSC controls (Figure 5B). Large areas of muscle

necrosis and inflammation were observed in vehicle-ADSC controls, which was not observed

in the other groups. Inflammatory assessement was further performed by immunostaining for

CD68+ macrophages on day 7 tissue samples (Figure 5C). As with Masson’s Trichrome

staining, much inflammation was observed in vehicle-ADSC controls. Notably, more

inflammation was observed in HGF-ADSC transplanted tissues than HGF/CXCR4-ADSC or

CXCR4-ADSC transplanted tissues. Overall, HGF and/or CXCR4-ADSCs were able to rescue

ischemic limb necrosis in the aged mouse model, with a less robust response by HGF-ADSCs.

6.4 Discussion

Herein, we developed HGF and CXCR4 overexpressing ADSCs using a bicistronic

plasmid for ischemic tissue repair. To develop this platform, a clinically applicable strategy

was employed by using PBAE biodegradable polymers for non-viral gene transfection, and an

abundant, autologous cell type with angiogenic potential. We hypothesized that dual

overexpression would synergize to enhance the behavior of ADSCs as delivery vehicles for

HGF paracrine release, which were evaluated in an aged murine model of hindlimb ischemia.

Overall, we found that dual HGF and CXCR4 overexpressing ADSCs enhanced ischemic tissue

repair by promoting stem cell survival and angiogenic response, as well as by attenuating

inflammation. Given the complex nature of stem cells, this approach demonstrates the ability

102

for exploiting multiple cellular processes involved in the therapeutic response to promote tissue

repair.

Two cellular processes involved in stem cell-based therapies for ischemic tissue repair

are: 1) stem cell engraftment and homing; and 2) angiogenic growth factor secretion. We found

that dual overexpression of CXCR4 and HGF could enhance both of these processes in ADSCs,

while single overexpression only promoted one process or the other. With regards to stem cell

engraftment and homing, BLI and RT-PCR data demonstrated that CXCR4-overexpression in

ADSCs could enhance this process, whether single or dual expression with HGF. This is further

supported by prior reports demonstrating that CXCR4 is required for stem cell engraftment in

the bone marrow [22] and promotes stem cell engraftment in the ischemic myocardium [23]. In

our own work, we have previously found that CXCR4-overexpressing ADSCs could promote

stem cell survival in the ischemic hindlimb in a similar manner as that demonstrated here. There

is some indication that dual overexpression by HGF and CXCR4 might even improve cell

survival in comparison to CXCR4-only overexpressing ADSCs, but the increase was not

significant (Figure 3D). Indeed, promoting stem cell survival may be important for long-term

therapeutic effect, as both CXCR4-ADSCs and HGF/CXCR4-ADSCs led to similarly improved

limb salvage.

With regards to the angiogenic response, HGF/CXCR4-ADSCs and HGF-ADSCs

promoted revascularization while CXCR4-ADSCs did not, as demonstrated by LDPI, CD31

capillary densitometry, and α-SMA arteriole assessment. This important difference between

CXCR4-ADSCs and HGF/CXCR4-ADSCs should be noted because a healthy angiogenic

response is important in promoting the health of the neuromuscular tissues. In fact, the effect

of HGF on the angiogenic response is well-known [15, 21]. Also, ADSCs are known to

naturally secrete HGF, in addition to a number of angiogenic growth factors, including vascular

endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF). While many of

these factors are upregulated in ADSCs by hypoxic pre-conditioning, HGF is less responsive

[20], which makes it an excellent candidate for genetic upregulation. Curiously, as HGF is also

known for its anti-inflammatory properties, we found that HGF overexpression could not

attenuate inflammation to levels observed by HGF/CXCR4- overexpression. This indicated that

CXCR4 may promote limb salvage by attenuating a damaging inflammatory response; which

generally acts to remove necrotic tissue and debris and could lead to widespread tissue loss, as

that observed by vehicle-ADSCs. Taken together, these data supported HGF and CXCR4

103

overexpression as acting synergistically for promoting long term tissue salvage and repair in

vivo.

The plasmid designed in this study was important in enabling therapeutic upregulation

of both HGF and CXCR4. After exploring both a bi-directional plasmid and a bicistronic

plasmid, we found that the bicistronic plasmid led to higher gene expression of both CXCR4

and HGF. This plasmid was organized with CXCR4 and HGF in sequence under a single CMV

promoter and separated by an internal ribosomal entry site (IRES). Further engineering on the

plasmid may be achieved by focusing on the genetic sequence of each respective gene. Here,

we have used a HGF gene sequence that has already been optimized for dual expression of two

HGF isoforms by alternative splicing [21]. Previously, the alternatively spliced HGF led to

enhanced therapeutic response when injected IM into a rabbit hindlimb ischemia model, in

comparison to single isoforms [21]. The potential for using this dually expressing plasmid was

further demonstrated in our study, where we found that HGF only-ADSCs led to a good

therapeutic response in a severe mouse model of hindlimb ischemia in comparison to vehicle-

ADSC controls. Designing a bicistronic plasmid that co-expressed HGF and CXCR4 using this

gene sequence, effectively enabled the overexpression of three different proteins, one isoform

of CXCR4 and two isoforms of HGF.

Two important criteria for successful clinical translation of genetically engineered stem

cells include: 1) the use of a non-viral gene transfection vector; and 2) an available and abundant

cell population. Non-viral gene transfection vectors may overcome many of the issues

associated with viral vectors, such as immunogenicity and insertional mutagenesis, but tend to

suffer low transfection efficiency and high cytotoxicity. Poly(β-amino esters) (PBAE) were

developed and optimized in response to these shortcomings by combinatorial polymer synthesis

and high-throughput screening [24, 25]. Indeed, we were able to successfully tranfect HGF and

CXCR4 into ADSCs using PBAE with minimal effects on cytotoxicity (Figure 1). An

additional benefit by PBAE transfection is the transient genetic upregulation, suggesting that

the plasmid DNA is not integrated into the genome and would not lead to insertional

mutagenesis.

In terms of cell population, ADSCs represent a relatively abundant, autologous cell

source that are easy to isolate. While this is clear, the therapeutic potential for PAD is still being

explored. Indeed, multiple studies have demonstrated their therapeutic effects in murine models

of a hindlimb ischemia [6, 20, 26, 27]. We further studied the effects of VEGF-overexpressing

ADSCs in vitro and CXCR4-overexpressing ADSCs in vivo [16, 17]. Carrying this forward

104

here, we have demonstrated that HGF/CXCR4-overexpressing ADSCs may further promote

ischemic tissue repair. Overall, these findings, in combination with their ready availability,

support ADSCs as a clinically applicable therapeutic stem cell source.

In summary, this study demonstrates that HGF/CXCR4-overexpressing ADSCs can be

developed using a bicistronic plasmid and biodegradable polymeric vectors. HGF and CXCR4

may synergize to promote ischemic tissue repair, by enhancing stem cell survival, angiogenic

response and ability to attenuate inflammation. Similar strategies may be employed to enhance

other cellular processes for therapeutic response, which may broaden the area of cellular-based

engineering. Overall, this approach has clinical potential for the engineering and regeneration

of many other tissue types, as well as in treating other types of ischemic diseases such as

myocardial infarction and cerebral ischemia.

Supplementary Materials

Fig. S6.1. Plasmid design optimization

Fig. S6.2. Optimization of bicistronic plasmid in ADSCs using poly(β-amino esters) (PBAE)

References


ischemic disease: part II: cell-based therapies. Circulation, 2004. 109(22): p. 2692-7.



3. Kalka, C., et al., Transplantation of ex vivo expanded endothelial progenitor cells for

therapeutic neovascularization. Proceedings of the National Academy of Sciences,

2000. 97(7): p. 3422-3427.

4. Katare, R., et al., Clinical-Grade Human Neural Stem Cells Promote Reparative

Neovascularization in Mouse Models of Hindlimb Ischemia. Arteriosclerosis,

Thrombosis, and Vascular Biology, 2014. 34(2): p. 408-418.

5. Kinnaird, T., et al., Local Delivery of Marrow-Derived Stromal Cells Augments

Collateral Perfusion Through Paracrine Mechanisms. Circulation, 2004. 109(12): p.

1543-1549.



105

7. Fadini, G.P., C. Agostini, and A. Avogaro, Autologous stem cell therapy for peripheral

arterial disease: Meta-analysis and systematic review of the literature. Atherosclerosis,

2010. 209(1): p. 10-17.





p. 3317-22.

10. Iwaguro, H., et al., Endothelial Progenitor Cell Vascular Endothelial Growth Factor

Gene Transfer for Vascular Regeneration. Circulation, 2002. 105(6): p. 732-738.

11. Mangi, A.A., et al., Mesenchymal stem cells modified with Akt prevent remodeling and

restore performance of infarcted hearts. Nat Med, 2003. 9(9): p. 1195-201.

12. Yu, J.-X., et al., Combination of stromal-derived factor-1α and vascular endothelial

growth factor gene-modified endothelial progenitor cells is more effective for ischemic

neovascularization. Journal of Vascular Surgery, 2009. 50(3): p. 608-616.



14. Morishita, R., et al., Phase I/IIa clinical trial of therapeutic angiogenesis using

hepatocyte growth factor gene transfer to treat critical limb ischemia. Arterioscler


15. Shigematsu, H., et al., Randomized, double-blind, placebo-controlled clinical trial of

hepatocyte growth factor plasmid for critical limb ischemia. Gene Ther, 2010. 17(9): p.

1152-61.



p. 483-91.



445-55.



46.

106

19. Zuk, P.A., et al., Multilineage cells from human adipose tissue: implications for cell-

based therapies. Tissue Eng, 2001. 7(2): p. 211-28.





2010. 17(12): p. 1442-52.

22. Peled, A., et al., Dependence of Human Stem Cell Engraftment and Repopulation of

NOD/SCID Mice on CXCR4. Science, 1999. 283(5403): p. 845-848.

23. Tang, Y.L., et al., Hypoxic Preconditioning Enhances the Benefit of Cardiac Progenitor

Cell Therapy for Treatment of Myocardial Infarction by Inducing CXCR4 Expression.

Circulation Research, 2009. 104(10): p. 1209-1216.

24. Green, J.J., et al., Combinatorial modification of degradable polymers enables

transfection of human cells comparable to adenovirus. Advanced Materials, 2007.

19(19): p. 2836-+.



26. Cao, Y., et al., Human adipose tissue-derived stem cells differentiate into endothelial

cells in vitro and improve postnatal neovascularization in vivo. Biochem Biophys Res

Commun, 2005. 332(2): p. 370-9.

27. Kim, Y., et al., Direct comparison of human mesenchymal stem cells derived from

adipose tissues and bone marrow in mediating neovascularization in response to

vascular ischemia. Cell Physiol Biochem, 2007. 20(6): p. 867-76.

107

Figures

Figure 6.1. Non-viral gene delivery to ADSCs via DNA polyplexes composed of a

biodegradable poly(β-amino ester) and pCK-HGF-IRES-CXCR4 plasmid. (A) Schematic of

plasmid DNA complexation with poly(β-amino esters) (PBAE) to form biodegradable

nanoparticles, which were used for transient genetic modification of adipose-derived stem cells

(ADSCs) ex vivo for the purpose of therapeutic angiogenesis, as assessed in a murine hindlimb

ischemia model. (B) Plasmid map of pCK-HGF-IRES-CXCR4, a bicistronic vector encoding

both human HGF and human CXCR4. (C) Synthesis of D32-122 polymer via a two-step

procedure, in which acrylate-terminated D32 (D32-Ac) polymer was prepared by 1.2:1.0

diacrylate:amine polymerization, followed by end modification of D32-Ac with an amine group

(122). (D) Human HGF and (E) CXCR4 gene expression analysis by qRT-PCR of ADSCs

transfected by D32-122-based polyplexes at 10:1 polymer/DNA weight ratio and varying DNA

loading doses at 48 hours post-transfection (* p < 0.05, ** p < 0.01, *** p < 0.001; n = 3). (F)

Cell viability of transfected ADSCs at 48 hours post-transfection, as assessed by MTS assay (no

statistical significance was observed).

108

Figure 6.2. Production of functional therapeutic protein from genetically modified ADSCs.

(A) HGF ELISA of conditioned media harvested from genetically modified ADSCs 48 hours

post-transfection (n = 3). (B) Viability of human umbilical cord vein endothelial cells

(HUVECs) cultured in transfected ADSC-conditioned medium as assessed by MTS assay (* p

< 0.05 vs basal media control, n = 4).

109

Figure 6.3. Survival of genetically modified ADSCs post-transplantation in vivo. (A) Human

HGF and (B) human CXCR4 gene expression analysis by qRT-PCR of total mRNA extracted

from gastrocnemius tissue lysate harvested at 48 hours post-transplantation (n = 2). (C)

Representative images from bioluminescence imaging (BLI) to noninvasively track survival of

ADSCs transfected with HGF-CXCR4, HGF, CXCR4, or vehicle control by PBAE polymer

over two weeks. (D) Quantification of BLI signal observed in mice treated with ADSCs at day

4 post-transplantation (* p < 0.05, ** p < 0.01, *** p < 0.001; n = 4-5).

110

Figure 6.4. Angiogenic response in ischemic hindlimbs post-transplantation in vivo. (A)

Representative laser Doppler blood reperfusion images of treatment groups at day 0, 7, and 14.

(B) Quantification of mean blood perfusion ratio of the ischemic foot over the control foot of

ADSC-treated mice (* p < 0.05 vs Vehicle, # p < 0.05 vs. CXCR4, n = 4-5). (C) CD31

immunofluorescent staining of ischemic muscle sections 21 days post-transplantation. Scale

bar: 200 µm. (D) Double immunofluorescent staining of α-SMA and laminin in ischemic muscle

sections 21 days post-transplantation. Scale bar: 200 µm. (E) Quantification of CD31+

capillaries and (F) α-SMA+ vessels in the adductor muscle of the ischemic hindlimb at day 21

(* p < 0.05, *** p < 0.001 compared to vehicle control; # p < 0.05, ## p < 0.01 compared to

CXCR4).

111

Figure 6.5. Analysis of limb status and muscle health. (A) Representative photographs of

ADSC-treated ischemic hindlimbs at day 7 and day 14 after surgery and cell transplantation.

(B) Masson’s Trichrome staining of adductor muscle tissue sections from the ischemic hindlimb

at day 7 post-transplantation. (C) Immunofluorescent staining of CD68 to assess inflammation

in adductor muscle sections at day 7 post-transplantation. Scale bar: 200 µm. (D) Physiological

status of hindlimbs after surgery and cell transplantation at day 1, 4, and 7, as rated on three

levels of severity: limb salvage (limb integrity and morphology is similar to control limb of the

same animal), toe necrosis, and foot necrosis (HGF-CXCR4, HGF, CXCR4: n=5; Vehicle: n=4)

112

Figure S6.1 – Plasmid Design of bi-directional and bicistronic vectors. Plasmid maps of bi-

directional (A) and bicistronic (B) plasmids. Transfection of each in HEK 293T cells showing

RT-PCR for CXCR4 (C) and HGF (D).

113

Figure S6.2 – Optimization of D32-122 transfection by polymer to DNA weight ratio. RT-

PCR for HGF (A) and CXCR4 (B). MTS cell viability at 48 hours post transfection.

114

7. Recruiting and Programming Endogenous Progenitor Cells in situ

for Therapeutic Angiogenesis

Lorenzo Deveza1,2, Jeffrey Choi3, Jothikritika Ashoken4, Fan Yang2,5

1Department of Bioengineering, Stanford University, Stanford, CA; 2Stanford University

School of Medicine, Stanford, CA; 3Department of Biology, Stanford University, Stanford, CA; 4Department of Biological Sciences, San Jose State University, San Jose, CA; 5 Department of

Orthopaedic Surgery, Stanford University, Stanford, CA.

Abstract

Stem cell therapy holds great promise for therapeutic angiogenesis via secreting a broad

spectrum of paracrine signals to promote new vessel growth. However, current methods often

require cumbersome and expensive procedures of isolating and modifying cells in vitro.

Endogeneous progenitor cells are known to possess the capacity to migrate towards ischemia in

situ under the action of stromal derived factor-1α (SDF-1α). Herein, we assess the potential for

sequential delivery of SDF-1α and DNA for recruiting and programming endogeneous

progenitor cells in situ for therapeutic angiogenesis. We further examine the effects of timing

of plasmid DNA delivery on transfection efficiency in situ. The outcomes of recruiting and

transfecting endogenous cells were evaluated using a murine model of hindlimb ischemia. We

found that controlled SDF-1α release enhanced the recruitment of CXCR4+ and c-kit+

endogenous cells by FACS, which led to improvements in the level and duration of luciferase

transgene expression. Timing of DNA delivery was important, which demonstrated only DNA

delivery at 7 days post-surgery led to enhanced transgene expression. Lastly, we found that the

therapeutic effects on cell recruitment were improved by HGF DNA delivery, which led to

improvements on the angiogenic response and ischemic muscle repair. Overall, this study

demonstrates that the strategy of recruitment and programming may offer an alternative

strategy for promoting angiogenesis by circumventing the need for isolating and

manipulating cells ex vivo.

7.1 Introduction

Peripheral arterial disease (PAD) represents an important medical problem caused by

decreased blood flow to the limbs. Treating PAD usually involves surgical revascularization,

but many patients suffering PAD carry surgically prohibitive co-morbidities. For such patients,

115

stem cell-based therapeutics are a promising, alternative treatment option [1]. This approach

involves stem cell injection into ischemic regions to promote host blood vessel growth in a

process known as therapeutic angiogenesis [2, 3]. Further enhancement on stem cell-based

therapeutics can be achieved by ex vivo genetic manipulation, which has shown promise by

strategically promoting cellular processes involved in angiogenic induction [2-4], such as

angiogenic growth factor release [5-7].

While these approaches have shown potential, clinical translation of stem cell-based

therapeutics has been slow due to several inherent limitations. Firstly, implementing most cell-

based therapies requires multi-step procedures which include: 1) autologous cell isolation; 2) ex

vivo cell expansion and transfection; and 3) transplantation back to the patient. In practice, this

requires multiple facilities with good laboratory practices (GLP), as well as multiple skilled

professionals, thereby leading to high cost and significant regulatory concerns. Secondly, the

technical aspects can be challenging, as ex vivo genetic manipulation of stem cells can be

inefficient and cytotoxic, and the majority of the cells die during re-injection.

One strategy to overcome these limitations is to exploit natural cell homing mechanisms

to attract endogeneous progenitor cells to sites of ischemic injury in order to replenish the stem

cell pool [8]. This is a process that has been shown to occur naturally, whereby bone marrow-

derived cell populations are recruited to sites of hypoxia to participate in the reparative process

[9-11]. Many researchers have proposed methods to augment systemic cell homing to ischemic

tissue as a way of supporting the local niche for therapeutic angiogenesis and tissue engineering

[8, 9]. The potential of using cell homing strategies for treating cardiovascular disease has been

successfully demonstrated using mobilization agents and chemokines in mouse models of

ischemic heart disease [12-14]. However, during states of cardiovascular disease, host cells

often have reduced functionality and a compromised cell niche that affects their ability for

therapeutic response [9].

In a similar manner that genetic manipulation of stem cells ex vivo can improve their

therapeutic effects, endogenous progenitor cell therapeutics might be enhanced by genetic

modification. Intramuscular (IM) injection of plasmid DNA to promote therapeutic transgene

expression has long been a strategy for therapeutic angiogenesis [15]. Indeed, multiple clinical

trials are currently being conducted to evaluate the potential of gene therapy for ischemic heart

disease and peripheral arterial disease; and are evaluating growth factors such as: basic

fibroblast growth factor (bFGF), vascular endothelial growth factor (VEGF) and hepatocyte

growth factor (HGF). In terms of the cells transfected in situ, plasmid DNA mediated delivery

116

to skeletal muscle has been shown to preferentially transfect regenerating muscle fibers [16, 17]

and non-inflammatory, infiltrating mononuclear cells [18]. In the setting of ischemic injury,

both of these cells increase in number; which are contributed by multiple sources, including

interstitial muscle satellite cells [19] and circulating progenitor cells [19, 20]. As this is the

case, it is very likely then that recruited endogenous progenitor cells can be transfected by IM

plasmid DNA delivery during ischemia.

Based on these findings, we hypothesize that endogenous progenitor cells could be

recruited to sites of ischemic injury and subsequently transfected to promote therapeutic

angiogenesis. Specifically, we propose to develop a biomaterials depot for sequential delivery

of stromal derived factor-1α (SDF-1α) and plasmid DNA for recruiting and programming

endogeneous progenitor cells in situ. The platform is evaluated in a murine model of hindlimb

ischemia. We initially evaluate cell recruitment by FACS and assess in situ transfection by

bioluminescence imaging (BLI). We further examine the effects of timing of plasmid DNA

delivery on transfection efficiency in situ. We lastly evaluate the potential for therapeutic

efficacy using a plasmid encoding for HGF.


7.2.1 Materials

Fibrinogen and thrombin were purchased from Sigma–Aldrich (St Louis, MO).

Recombinant murine stromal derived factor-1α (SDF-1α) was purchased from Peprotech

(Rocky Hill, NJ). Plasmids (pEGFP-N1 and pCMV-LUC) were purchased from Elim

Biopharmaceuticals; pCK-HGF plasmid was obtained from Viromed (Seoul, Korea).

7.2.2 Preparation of SDF-1α Encapsulated Fibrin Gels

SDF-1α encapsulated fibrin gels were prepared by premixing two solutions: 1)

Fibrinogen (50 mg/ml) + SDF-1α (5 µg/ml); and 2) Thrombin (2 U/ml). The two solutions were

then mixed together to create a solution with final concentrations of 25 mg/ml Fibrinogen, 2.5

µg/ml SDF-1α, and 1 U/ml Thrombin. The resulting solution (200 µl) was injected into mice

prior to polymerization. SDF-1α release from varying concentration Fibrin gels (10 mg/ml, 25

mg/ml, and 50 mg/ml) was assessed in vitro by preparing gels in a 50 µl gel mold and incubating

in a 48 well plate at 4 ºC. SDF-1α was released in a collection media containing 1 mg/ml BSA

and quantified by SDF-1α ELISA (Peprotech).

117

7.2.3 Induction of hindlimb ischemia and therapeutic injections

All experiments were performed in accordance with the Stanford University Animal

Care and Use Committee Guidelines and approved APLAC protocols. Hindlimb ischemia was

induced in female C57/Bl mice aged 7-9 weeks (Jackson Laboratory) as previously described

[21]. Briefly, the femoral artery was ligated at two sites (distal to the external iliac artery and

proximal to the popliteal artery) with 5-0 silk suture (Ethicon). The femoral artery was excised

between the points of ligation. One day after arterial dissection, 200 µL of Fibrin/SDF-1α

solution (25 mg/ml Fibrin, 500 ng SDF-1α) was injected intramuscularly into two sites (the

adductor muscle region and the gastrocnemius) using a 27-gauge tuberculin syringes. After a

specified time delay post-Fibrin/SDF-1α injection (0 days, 2 days, 6 days), 10 µg plasmid DNA

(pEGFP-N1, pCMV-LUC, or pCK-HGF) was injected into the same muscles. These time points

effectively represented 1 day, 3 days, and 7 days post-surgery.

7.2.4 Tissue harvest and FACS assessement of recruited cell populations

To assess recruited cell populations, cells were isolated from hindlimbs and analyzed

by flow cytometry (FACS). Cells were isolated by harvesting hindlimbs and then removing and

mincing all limb muscles. Minced limb muscles were then digested using 0.1 mg/ml

Collagenase type II (Invitrogen, Carlsbad, CA) on a shaker at 37 ºC for 45 min. Digestion was

then continued using 1 mg/ml Collagenase/Dispase (Roche Life Sciences, Branford, CT) on

shaker at 37ºC for another 90 min. The resulting digested tissue was then serially passed through

three cell strainers (100 µm, 70 µm, 40 µm) to remove debris and non-digested tissue lysate.

Cells were then centrifuged to remove the supernatant. Upon re-suspending the cells were split

into separate round bottom tubes for FACS analysis. Cell staining was performed by incubating

cells with one or three antibodies: CD45-PerCP, CD34-FITC, c-kit-PE, CXCR4-PE (all from

BD Pharmingen, San Jose, CA). Triple color staining was performed for either: 1) CD34, CD45,

CXCR4; or 2) CD34, CD45, c-kit. Single color staining or unstained cells were used as controls.

After incubating for 30 minutes, antibodies were removed and replaced with 300 µl Hank’s

Balanced Salt Solution (HBSS, Life Technologies, Carlsbad, CA). Stained cells were subjected

to flow cytometry using a LSR II.UV flow cytometer. Post-FACS analysis was performed using

FlowJo v7.6.5 software.

118

7.2.5 Bioluminescence Imaging

In vivo transfection with pCMV-LUC reporter plasmid was assessed by

bioluminescence imaging (BLI). Animals were anesthetized by 2–3% inhaled isoflurane and

injected intraperitoneally with the reporter probe D-Luciferin at 100 mg/kg body weight. IVIS

Spectrum system (Xenogen, Alameda, CA) was used to image the animals while under 2%

inhaled isoflurane anesthesia. Each animal was scanned until the peak signal was reached,

recording radiance of regions of interest in photons. Radiance was quantified in photons per

second per centimeter squared per steridian.

7.2.6 Protein extraction, ELISA and Western Blot

Protein was extracted from the gastrocnemius muscle of the ischemic hindlimb, which

was harvested at 5 or 14 days post-surgery. Harvested muscle was snap-frozen by collecting

into an 1.5 ml Eppendorf tube and subsequently submerging into liquid nitrogen. Samples were

then stored at -80 ºC. Protein was then extracted by directly submerging the muscle into liquid

nitrogen and then grinding it using a mortar and pestle. At least 50-80 mg of the ground muscle

was then collected into a new Eppendorf, into which 500 µl of RIPA Lysis and Extraction buffer

(Thermoscientific, Waltham, VA) containing 1x Halt Protease and Phosphatase Inhibitor

Cocktail (Thermoscientific) was added. The tissue was vortexed and pipetted to further break

up the muscle. The tissue was then centrifuged, and the supernatant was collected and stored

for protein analysis. Total protein content was measured using Bio-Rad Protein assay (Bio-Rad

Laboratories, Hercules, CA). HGF ELISA was performed by diluting the sample 2x and then

assaying according to the manufacturer’s protocol (R&D Systems, Minneapolis, MN). Western

Blot was performed by loading 6 µg of protein per sample and separating proteins using sodium

dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE, Invitrogen). Proteins were

then transferred to polyvinylidine fluoride (PVDF) (Invitrogen) membranes overnight. PVDF

membranes were stained with antibodies for CD31, alpha smooth muscle actin (α-SMA) or beta

actin (β-actin) as a loading control (all from Abcam, Cambridge, England). Horse radish

peroxidase (HRP) conjugated secondary antibodies (Abcam) were used for labeling.

Supersignal West Pico Chemiluminescent substrate (Thermoscientific) was applied to the

washed membranes, which were then assessed by X-ray film using CL-XPosure Film

(Thermoscientific).

119

7.2.7 Histology and Immunohistochemistry

Mice were euthanized at 14 days to retrieve muscle tissue from ischemic limbs for

histological analysis. Specimens were embedded in optimal cutting temperature (O.C.T.) for

cryosectioning. Sections were cut at 8 µm thickness. General tissue morphology was assessed

with Masson’s Trichrome Staining and immunofluorescence staining for Laminin (Abcam),

using a FITC-labeled secondary antibody. Angiogenesis was assessed by staining for capillaries

(CD31, Abcam) and arterioles (α-SMA, Abcam) Cells transfected with pEGFP-N1 were

assessed by immunostaining with anti-GFP (Abcam), using a FITC-labeled secondary antibody.

Sections were mounted with Vectashield Mounting Medium with 4’,6-diamidino-2-

phenylindole (DAPI) (Vector Laboratories, Burlingame, CA) for counterstain. Sections were

imaged using a fluorescent microscope. The number of CD31-positive vessels was quantified

to estimate capillary density per number of cell nuclei using MATLAB software (MathWorks,

Natick, MA).

7.2.9 Assessment of Limb Status

Ischemic limbs were evaluated by direct observation of limb status. Limbs were

assessed on days 1, 3, 7, 10, 14 post-ischemic induction. Grading of limb ischemia was based

on the severity of limb ischemia utilizing following categories: 1) limb salvage; 2) toe necrosis;

3) foot necrosis; 4) foot loss; and 5) limb loss. Limb salvage is plotted by showing the fraction

of each category per group (n = 5).





significance. Number samples per study was as follows: FACS analyses (n = 3); BLI

assessements (n = 3); therapeutic studies (n = 5). Each study was repeated to confirm results.

All quantitative data were expressed as mean ± standard error of the mean.

7.3 Results

7.3.1 Controlled Release of SDF-1α from Fibrin Gels

To protect and control the release of SDF-1α in vivo, SDF-1α was encapsulated into

injectable fibrin gels prior to delivery to the ischemic hindlimb (Figure 1A). The in vitro release

120

kinetics of SDF-1α from fibrin gels were studied to correlate timing of release with subsequent

cell recruitment and transfections. This demonstrated that the release was bi-phasic, showing a

burst release of nearly 70% on the first day, followed by a more gradual release over the next

3-4 days (Figure 1B). As the fibrin concentration did not significantly affect release, fibrin gels

composed of 25 mg/ml were used for all remaining experiments.

7.3.2 SDF-1α Enhances Cell Recruitment Post-ischemic Induction

The effect of SDF-1α treatment on in vivo cell recruitment was studied by injecting

SDF-1α/fibrin gels into ischemic hindlimbs at day 1 post-surgery and performing FACS

analysis on harvested tissues at days 1, 3, 7 (Figure 1C-1E). Recruited populations were

analyzed by staining for CD34, CD45, CXCR4 and c-kit. CD34 and CD45 were chosen as

general markers for progenitor and inflammatory cells, while CXCR4 and c-kit were chosen

because these populations have been shown to respond to SDF-1α (Figure 1C-1E). Overall,

ischemia was a strong stimulus for recruiting cells, which demonstrated increases in CXCR4+,

c-kit+, CD34-/CD45+/CXCR4+, and CD34-/CD45+/c-kit+ cell populations by days 3 and 7

compared to sham surgery controls (p < 0.05; Figure 1D, 1E). SDF-1α treatment further

improved recruitment of these populations compared to ischemia only and sham surgery

controls, which demonstrated increases at day 3 and 7; with a persistently high percentage of

CXCR4+ cells by day 7 (p < 0.05). We also observed an increase of CD45+ cells by SDF-1α

treatment at days 3 and 7 compared to ischemia only and sham surgery controls (p < 0.05). In

terms of cell recruitment kinetics, these data demonstrated that the peak cell recruitment was

around 3 days, but persisted towards 7 days with SDF-1α treatment. Overall, this indicated that

SDF-1α could enhance cell recruitment.

7.3.3 SDF-1α Promotes In Situ Transfection by Plasmid DNA Delivery

The effect of SDF-1α treatment on in situ plasmid DNA transfection was studied by

injecting SDF-1α/fibrin gels at day 1 post-surgery and then injecting plasmid DNA encoding

for either GFP or Luciferase at day 7 post-surgery (Figure 2A). Day 7 was chosen because cell

recruitment by SDF-1α treatment was enhanced compared to ischemia at this time point.

Analysis of GFP or Luciferase transfection was performed the day after plasmid DNA injection

by FACS or BLI assessment, respectively. Both FACS and BLI demonstrated increases in

transfection in situ by SDF-1α treatment compared to ischemia only (Figure 2B, 2C).

Specifically, FACS analysis demonstrated a ~2% increase in GFP+ cells by SDF-1α treatment

121

compared to ischemia only (p < 0.05; Figure 2B), while BLI data demonstrated a ~2-fold log

increase in luminescence by SDF-1α treatment compared to ischemia only (p < 0.05; Figure

2C). These data indicated that there were increases in both the number of cells transfected and

the amount of protein produced in ischemic hindlimbs treated with SDF-1α/fibrin gels, thereby

providing evidence that SDF-1α could promote plasmid DNA transfection in situ.

7.3.4 In Situ Transfected Cells are Largely Represented by the Recruited Cells

To assess whether the cell populations transfected were the same as those recruited, the

in situ transfected GFP+ populations were analyzed to determine their relative expression of

CD34, CD45, CXCR4, and c-kit (Figure 2D, 2E). We found that the transfected GFP+ cell

populations expressed many of the same markers as those assessed for cell recruitment (Figure

2D). Specifically, the GFP+ populations expressed the following: 50-60% CXCR4+; 50-60%

c-kit+; 30-40% CD34-/CD45+/CXCR4+, and 30-40% CD34-/CD45+/c-kit+ (Figure 2E). We

also determined that there was no significant difference in GFP+ transfected populations

whether SDF-1α was delivered or not (Figure 2E). Immunostaining was performed to determine

the relative location of transfected cells, which demonstrated that many of the transfected cells

were mononuclear cells that were located around regenerating muscle fibers (Figure 2F). This

was consistent with the FACS analysis, as c-kit+ or CXCR4+ cells represent mononuclear

inflammatory or progenitor cell populations. Overall, this provided evidence that the in situ

transfected cell populations were largely represented by the recruited cells.

7.3.5 Timing of Plasmid DNA Delivery Affects the Level and Duration of Transfection

The timing of plasmid DNA delivery was optimized by injecting plasmid DNA

encoding luciferase into ischemic hindlimbs at one of three different time points (days 1, 3, 7)

and then assessing for the level and duration of luciferase expression (Figure 3A). We found

that the timing of plasmid DNA delivery had a significant effect on luciferase expression, which

was lowest when delivered at day 1 and highest when delivered at day 7 (p < 0.05; Figure 3D,

3E). SDF-1α treatment was able to improve luciferase expression at each time point compared

to ischemia only (p < 0.05; Figure 3D, 3E), thereby further demonstrating the ability to use

SDF-1α to improve in situ transfection. In terms of the duration of transfections, we found that

plasmid DNA delivered at day 1 post-surgery only lasted one week, which was slightly

improved by SDF-1α treatment (Figure 3F). On the other hand, plasmid DNA that was

delivered at day 7 led to a robust and prolonged transfection response that persisted over 2 weeks

122

regardless of whether SDF-1α was delivered. Importantly, day 3 plasmid DNA injection was

strongly affected by SDF-1α treatment, where we found that the duration of luciferase

expression only persisted when SDF-1α was used (Figure 3F). Furthermore, the level of

luciferase expression from day 3 plasmid DNA injected hindlimbs treated with SDF-1α was

similar to the levels achieved by day 7 plasmid DNA injection. This suggested that SDF-1α

treatment could be used to enable earlier plasmid DNA injections for persistent transgene

expression.

7.3.6 Enhanced HGF Transfection In Situ by SDF-1α Delivery

To assess whether SDF-1α treatment could promote transgene expression of a

therapeutic growth factor, plasmid DNA encoding HGF (pHGF) was injected into mice at 3

days post-surgery and assessed for transfection on the fifth day (Figure 4A). Compared to

ischemia only, we found that SDF-1α treatment increased HGF release in ischemic hindlimbs

(p < 0.05; Figure 4B). Injecting pHGF alone at day 3 also led to increases in HGF release which

were significant compared to SDF-1α only and PBS controls (p < 0.05; Figure 4B). This

indicated that both treatment approaches, whether by timing plasmid DNA delivery or applying

SDF-1α were able to promote transgene expression of a therapeutic growth factor.

7.3.7 Angiogenic Improvement in the Ischemic Hindlimb

To assess whether promoting HGF transgene expression by SDF-1α treatment would

lead to an angiogenic response, day 14 tissue samples were analyzed for capillaries and

arterioles (Figure 5A). Compared to SDF-1α only and PBS controls, HGF transgene expression

led to an observable increase in both CD31+ capillaries (Figure 5B) and α-SMA+ arterioles

(Figure 5C) as demonstrated by immunofluorescence staining. CD31 quantification

demonstrated that the increase in capillaries due to HGF overexpression was statistically

significant compared to SDF-1α only and PBS controls (p < 0.05; Figure 5D). Western blotting

analysis of proteins extracted from muscle further confirmed that CD31 and α-SMA were

increased by HGF overexpression compared to SDF-1α only and PBS controls (Figure 5E). This

demonstrated the HGF transgene expression by SDF-1α treatment enhanced the angiogenic

response.

123

7.3.8 Promotion of Limb Health

The therapeutic effects of HGF transgene expression by SDF-1α pre-treatment were

assessed by evaluating limb status and by evaluating histological features of muscle health.

Limb status assessment demonstrated that the baseline response to ischemic induction was

moderate toe and foot necrosis, which presented around day 7 post-surgery and became

progressively worse over the next 7 days (PBS group; Figure 6B, 6E). Progressive limb damage

beyond day 7 was prevented by injecting SDF-1α (F/SDF; Figure 6E) or by HGF transgene

expression (F/SDF + HGF; Figure 6E), which were comparable. Notably, HGF transgene

expression led to improved muscle health and morphology compared to SDF-1α and PBS

controls, as demonstrated by Masson’s Trichrome staining and anti-Laminin

immunofluorescence staining (Figure 6C, 6D). This was clearly observable by anti-Laminin

staining, which demonstrated excellent overall muscle morphology with intact basement

membranes (Figure 6D). Overall, this suggested that HGF transgene expression by SDF-1α

could promote muscle health and lead to therapeutic effects.

7.4 Discussion

Herein, we demonstrated that scaffold-mediated release of SDF-1α followed by delayed

plasmid DNA injection offers a promising platform for recruiting and programming endogenous

progenitor cells in situ. Cell recruitment was strongly enhanced by ischemia and SDF-1α

delivery, which led to improvements in transgene expression in situ. The timing of plasmid

DNA delivery post-ischemia had a large effect on the level and duration of transgene expression,

where we found that SDF-1α had the potential to enable earlier transfection. As validation for

this approach, transfection with HGF was shown to promote therapeutic efficacy in a murine

model of hindlimb ischemia. These results demonstrate that such platforms may offer an

alternative strategy for promoting angiogenesis by circumventing the need for isolating and

manipulating cells ex vivo.

The strategy of recruiting and programming endogenous progenitor cells in situ relies

on two sequential processes: 1) initial cell homing; and 2) delayed transgene expression. In

many ways, the potential for using this approach was highlighted by the ability to improve

transgene expression by firstly inducing ischemia and secondly delivering DNA at a later time.

Here, the initial cell homing response is achieved by ischemia, which is a well-known stimulus

for recruiting inflammatory and progenitor cells [9]. This is further supported by our findings

that the CXCR4+ and c-kit+ populations increased over time due to ischemia compared to sham

124

surgery controls. As these populations peaked around the third to seventh day post-ischemic

induction, this suggested that it was the optimal time to deliver DNA for transfecting these cells.

Indeed, delayed transgene expression was achieved by delivering DNA at seven days post-

ischemic induction, which led to significant increases in transgene expression in situ. Further

analysis demonstrated that many of the transfected cells expressed CXCR4 and c-kit, suggesting

these were among the recruited populations. Overall, the finding that ischemic injury improves

transgene expression was consistent with other studies showing the same in skeletal muscle

injured by other mechanisms, such as bupivacaine injection [17] or cardiotoxin injection [16].

Each of these injuries may similarly lead to enhanced transfection by promotion of endogenous

cell recruitment and stem cell niche activation.

In our study, we exogenously delivered SDF-1α protein as a way for further promoting

cell recruitment and subsequent transfection. SDF-1α was chosen because it is one of the most

well-known chemokines for mediating cell homing to ischemia [22]. We found that

exogenously delivering SDF-1α one day post-ischemic induction was able to further promote

endogenous cell recruitment compared to with ischemia alone, demonstrating increases in the

CXCR4+ and c-kit+ populations. Indeed, these are known responders to SDF-1α [11, 13, 14,

23, 24]. These improvements were associated with better transfection of both luciferase and

HGF transfection, thereby supporting the strategy for promoting transfection by enhancing

endogenous cell recruitment using SDF-1α. This finding also opens many possibilities for

exploring other chemokines.

One of the major findings in this study was the effect of DNA delivery timing on the

ability for long-term transgene expression after inducing ischemia. In particular, we found that

delivering DNA too early after ischemic-injury led to short-term transgene expression, while

delivering DNA one week later enhanced transfection beyond that achieved in non-injured

muscles. Furthermore, pre-treatment of the ischemic muscle with SDF-1α enabled earlier

transgene expression (i.e. day 3 DNA delivery), which did not last when SDF-1α was not

delivered. These findings are pertinent for multiple reasons. Firstly, this suggests that there is

an optimal time to deliver DNA to treat acute injury (i.e. acute myocardial infarction or bone

fracture). Secondly, it demonstrates that DNA could be delivered earlier by SDF-1α treatment,

which may be important for therapeutic gene therapy when earlier treatment is warranted.

Thirdly, it suggests that injury may be a mechanism to enhance gene therapy, which may be

useful in therapeutic situations in which injury is purposely induced (i.e. transmyocardial laser

revascularization or microfractures for cartilage repair).

125

The therapeutic potential for genetically manipulating endogenous progenitor cells was

evaluated in a murine model of hindlimb ischemia. As with other studies, cell recruitment by

SDF-1α release could promote limb salvage. Further therapeutic effects were achieved by

programming endogenous cells with HGF, which is an angiogenic factor known to act on both

endothelial cells and smooth muscle cells; thereby promoting both capillaries and arterioles [25,

26]. Indeed, we found that this was the case, as HGF-modified endogenous cells increased both

types of vessels, which also led to therapeutic enhancement in the overall muscle health of the

ischemic hindlimb. This demonstrated that programming endogenous stem cells has promise

for treating peripheral arterial disease by therapeutic angiogenesis.

This study opens up the potential for many future studies and directions. Some of these

future directions, include: exploring other chemokines or therapeutic genes; developing dual

and sequential controlled release platform; and, studying the mechanism to better understand

the observations of this study. With regards to exploring other chemokines or therapeutic genes,

we chose to focus on SDF-1α, but many other proteins may achieve a similar enhancement on

transgene expression in situ, opening up the possibility for exploring these proteins. Also, the

field of gene therapy for inducing angiogenesis has demonstrated multiple factors for

upregulation, which could be explored upon initial cytokine delivery. With regards to a

controlled release platform, we achieved sequential release by controlled SDF-1α release and

delayed plasmid DNA injection, but a dual, sequential delivery platform would be more ideal

to truly enable a one-step procedure. Lastly, further study on mechanism would help us to more

specifically define which cell populations are transfected and why early DNA delivery post-

injury does not lead to prolonged transgene expression. This would enable us to design a better

strategy for achieving transgene expression in situ.

In conclusion, this study demonstrates the potential for using scaffold-mediated release

of SDF-1α and delayed DNA injection to program endogenous cells in situ. This strategy may

be useful for inducing therapeutic angiogenesis to treat ischemic diseases, such as PAD, which

can be an alternative to stem cell-based therapeutics. Additionally, it shows promise for other

tissue engineering approaches, such as wound or bone repair.

126

References


ischemic disease: part II: cell-based therapies. Circulation, 2004. 109(22): p. 2692-7.



3. Deveza, L., J. Choi, and F. Yang, Therapeutic angiogenesis for treating cardiovascular

diseases. Theranostics, 2012. 2(8): p. 801-14.





p. 483-91.



445-55.



p. 3317-22.

8. Chen, F.M., et al., Homing of endogenous stem/progenitor cells for in situ tissue

regeneration: Promises, strategies, and translational perspectives. Biomaterials, 2011.

32(12): p. 3189-209.

9. Dimmeler, S., Regulation of bone marrow-derived vascular progenitor cell mobilization

and maintenance. Arterioscler Thromb Vasc Biol, 2010. 30(6): p. 1088-93.



434-8.



12. Orlic, D., et al., Mobilized bone marrow cells repair the infarcted heart, improving

function and survival. Proc Natl Acad Sci U S A, 2001. 98(18): p. 10344-9.

13. Zaruba, M.M., et al., Synergy between CD26/DPP-IV inhibition and G-CSF improves

cardiac function after acute myocardial infarction. Cell Stem Cell, 2009. 4(4): p. 313-

23.

127

14. Zhang, G., et al., Controlled release of stromal cell-derived factor-1 alpha in situ

increases c-kit+ cell homing to the infarcted heart. Tissue Eng, 2007. 13(8): p. 2063-

71.


ischemic disease. Part I: angiogenic cytokines. Circulation, 2004. 109(21): p. 2487-91.

16. Davis, H.L., C.L. Millan, and S.C. Watkins, Immune-mediated destruction of

transfected muscle fibers after direct gene transfer with antigen-expressing plasmid

DNA. Gene Ther, 1997. 4(3): p. 181-8.

17. Vitadello, M., et al., Gene transfer in regenerating muscle. Hum Gene Ther, 1994. 5(1):

p. 11-8.

18. Ratanamart, J., C.G. Huggins, and J.A. Shaw, Transgene expression in mononuclear

muscle cells not infiltrating inflammatory cells following intramuscular plasmid gene

electrotransfer. J Gene Med, 2010. 12(4): p. 377-84.

19. Sherwood, R.I., et al., Determinants of skeletal muscle contributions from circulating

cells, bone marrow cells, and hematopoietic stem cells. Stem Cells, 2004. 22(7): p.

1292-304.

20. Cerletti, M., et al., Highly efficient, functional engraftment of skeletal muscle stem cells

in dystrophic muscles. Cell, 2008. 134(1): p. 37-47.

21. Niiyama, H., et al., Murine model of hindlimb ischemia. J Vis Exp, 2009(23).

22. Petit, I., D. Jin, and S. Rafii, The SDF-1-CXCR4 signaling pathway: a molecular hub

modulating neo-angiogenesis. Trends Immunol, 2007. 28(7): p. 299-307.

23. Askari, A.T., et al., Effect of stromal-cell-derived factor 1 on stem-cell homing and

tissue regeneration in ischaemic cardiomyopathy. The Lancet, 2003. 362(9385): p. 697-

703.

24. Jin, D.K., et al., Cytokine-mediated deployment of SDF-1 induces revascularization

through recruitment of CXCR4+ hemangiocytes. Nat Med, 2006. 12(5): p. 557-67.

25. Morishita, R., et al., Phase I/IIa clinical trial of therapeutic angiogenesis using

hepatocyte growth factor gene transfer to treat critical limb ischemia. Arterioscler




2010. 17(12): p. 1442-52.

128

Figures

Figure 7.1 SDF-1α Enhances Cell Recruitment Post-ischemic Induction. (A) Study design

showing injection of SDF-1α encapsulated in fibrin gels at one day post-ischemic induction,

followed by delayed plasmid DNA injection (10 µg). (B) Controlled release of SDF-1α from

synthesized fibrin gels in vitro. (C-E) FACS assessment of endogenous progenitor cell

recruitment to ischemic limbs at (C) day 1, (D) day 3, and (E) day 7 (* p < 0.05, n = 3).

129

Figure 7.2 SDF-1α Promotes In Situ Transfection of recruited endogenous cells (A) Study

design showing injection of SDF-1α encapsulated in fibrin gels at one day post-ischemic

induction, followed by day 7 plasmid DNA injection (10 µg EGFP or luciferase). (B) FACS

analysis of GFP+ transfected cells in harvested muscle at one day post-GFP injection. (C)

BLI assessment of luciferase expression in ischemic limbs at one day post-luciferase injection.

(D) FACS plots demonstrating further analysis and gating of GFP+ populations for CD34,

CD45, CXCR4 and c-kit. (E) Quantitative analysis of GFP+ populations. (F)

Immunostaining for GFP, Laminin to demonstrate location of transfected cells in situ. Dapi is

used as a counter-stain. (* p < 0.05, n = 3).

130

Figure 7.3 Timing of Plasmid DNA Delivery Affects the Level and Duration of Transfection.

(A-C) Schematic demonstrating study design for (A) day 1 luciferase injection, (B) day 3

luciferase injection, and (C) day 7 luciferase injection. (D) BLI images and (E) quantitative

lumuninescence in ischemic limbs one day post-luciferase injection, with and without SDF-1α.

(F) Kinetics of luciferase transgene expression when injected at specified time points post-

ischemic induction, with and without SDF-1α. (* p < 0.05, n = 3-5)

131

Figure 7.4 Enhanced HGF Transfection In Situ by SDF-1α Delivery. (A) Schematics

demonstrating timing of SDF-1α delivery and plasmid HGF injection. Mice were harvested

two-days post-HGF injection for analysis. (B) ELISA analysis of HGF release in harvested

muscle. (* p < 0.05, n = 3)

132

Figure 7.5 Angiogenic Improvement in the Ischemic Hindlimb by HGF overexpression. (A)

Schematic of therapeutic model study design demonstrating timing of SDF-1α and HGF

delivery. Tissue were harvested at day 14 post-surgery for analysis. (B) CD31

immunostaining and (C) α-SMA immunostaining for capillaries and arterioles, respectively.

(D) Capillary densitometry assessment represented as the ratio of CD31+ capillaries to DAPI+

cell nuclei (p < 0.05). (E) Representative western blot analysis of CD31 and α-SMA in

ischemic hindlimbs (n= 3; H: HGF + Fibrin/SDF; FS: Fibrin/SDF only; P: PBS treatment

only).

133

Figure 7.6 Promotion of limb salvage. (A) Schematic of therapeutic model study design

demonstrating timing of SDF-1α delivery and subsequent HGF injections. Tissue were

harvested at day 14 post-surgery for analysis. (B) Images showing foot and limb morphology

of ischemic limbs. (C) Masson’s Trichrome staining of harvested ischemic muscle. (D)

Laminin (green) staining on harvested muscle demonstrating muscle architecture and integrity,

DAPI (blue) was used as counterstain. (E) Quantitative limb status over two weeks,

demonstrating fraction of total mice with foot necrosis, toe necrosis, or limb salvage (n = 9).

134

135

8. Microfluidic Synthesis of Biodegradable Polyethylene-Glycol Microspheres for

Controlled Delivery of Growth Factors and DNA Nanoparticles

Lorenzo Deveza1,2, Jothikritika Ashoken3, Gloria Castaneda4,

Xinming Tong4, Li-Hsin Han4, Fan Yang1,4

1Department of Bioengineering, Stanford University, Stanford, CA; 2Stanford University School

of Medicine, Stanford, CA; 3Department of Biological Sciences, San Jose State University, San

Jose, CA; 4Department of Orthopaedic Surgery, Stanford University, Stanford, CA.

Abstract

Polymeric microparticles represent an injectable platform for controlled release of biologics.

The ability to synthesize uniform and homogeneous polymeric microparticles under mild

conditions is important for tuning release in a reliable manner. Microparticle diameter or

polymer character could then be varied to control release of the encapsulated biologics. In this

study, we report the development of monodisperse, biodegradable poly(ethylene glycol) (PEG)-

based microparticles for controlled release of growth factors or DNA nanoparticles using a

microfluidic platform. We found that simple changes in microfluidic design could speed the

rate of microparticle formation (>10-fold increase), as well as control the size of the

microparticles ranging from ~50-150 µm. By using thiol-ene crosslinkable PEG, microparticles

with homogeneous matrices were synthesized. Mesh size and degradation rate could be

controlled by varying PEG concentration from 7.5% to 15 % (W/V). Both microparticle

diameter and mesh size were varied to tune the release of basic fibroblast growth factor, which

retained its biological activity of stimulating cell proliferation when applied to human adipose-

derived stem cells. Controlled release of DNA nanoparticles encoding vascular endothelial

growth factor demonstrated the ability to transfect HEK293T cells. This platform may provide

a useful tool for synthesizing monodisperse microparticles as injectable carriers to achieve

coordinated growth factor or DNA nanoparticle release, which may be broadly useful for

guiding desirable cell fates and tissue regeneration in situ.

8.1 Introduction

In tissue regeneration and repair, soluble cues such as growth factors play an

important role in guiding cell fate, requiring well-orchestrated doses and timing. To

achieve such delivery characteristics, polymeric microparticles containing different

136

biologics, such as protein or DNA nanoparticles, with well-tuned release kinetics

represent a facile method for temporal release in vivo. While a number of factors are

known and employed to tune the release of biologics from polymeric microparticles, the

ability to generate homogeneous polymer matrices and the ability to synthesize uniform

microparticles are fundamental to reliably generate microparticles with desired release

kinetics. Once established, changes in polymer physical and chemical characteristics, as

well as changes in microparticle size and morphology may be exploited to tune growth

factor release.

Amongst the microfabrication approaches for synthesizing polymeric microparticles,

droplet microfluidics represents a relatively simple and consistent method for generating

uniform microparticles.[1] With this approach, polymeric microparticles containing multiple

reagents and biologics can be synthesized by adding polymer and biomolecules to the

discontinuous phase and polymerizing the resulting droplets.[2-7] Microparticle size and

morphology can be easily controlled by adjusting the input flow rates or the microchannel

dimensions.[6] Monodisperse drug releasing microparticles synthesized in this manner have

demonstrated better control over drug release than microparticles generated via simple water-

in-oil emulsion, where microparticle size has been shown to have an effect on drug release

characteristics.[4, 6] Of the systems previously studied, [2, 4-7] most focus on encapsulating

and releasing a small molecule. Those that have focused on releasing a protein have been mainly

polysaccharide-based microparticles and have generally used a model protein (i.e. BSA) that is

not assessed for bioactivity. As polysaccharide-based hydrogel systems lead to heterogeneous

networks, many monodisperse polysaccharide-based microparticles fail to achieve consistent

and prolonged release characteristics.

Polyethylene glycol (PEG) based hydrogel matrices are potentially more

homogeneous than other biodegradable hydrogels owing to being a synthetic polymer

with minimal biomolecule affinity and being easily modified chemically. As a drug

delivery platform, various approaches have been explored, such as physical entrapment,

inclusion of a protein binding molecule (i.e. heparin sulfate), and covalently linking the

protein to the gel. Amongst these, physical entrapment represents a low-cost and less-

complicated approach, which generally relies on encapsulating protein within a PEG

mesh diameter of comparable or smaller size and releasing the protein upon degradation

of the PEG network. PEG matrices that have these characteristics can be polymerized

via step-growth or condensation reactions using PEG precursors that have been

137

functionalized with a hydrolytically degradable moiety. Step-growth polymerized PEG

matrices can generate more homogeneous network structures, which may lead to more

consistent and reliable release kinetics and may have further in vivo benefit due to having

less renal-toxicity owing to smaller molecular weight degradation products. [8]

Furthermore, degradation rates can be tuned by simply varying PEG concentration,

which can effectively tune release kinetics.

In this study, we synthesized biodegradable PEG-based, monodisperse

microparticles for controlled release of growth factors or DNA nanoparticles using a

microfluidic device. We hypothesize that release of each from PEG microparticles could

be tuned by varying the polymer concentration in order to tune the mesh size and

degradability of the PEG matrix, as well as by using monodisperse microparticles with

controlled diameter. To study this, various microfluidic devices were designed for

rapidly synthesizing microparticles of different sizes. The effects of varying microsphere

size and PEG concentration on release kinetics were then assessed. Lastly, the biological

activity of released basic fibroblast growth factor (bFGF) was assessed using a cell

proliferation assay, while the ability for DNA nanoparticles to transfect were tested

against HEK293T cells.

8.2 Experimental Section

8.2.1 Materials

Polyethylene glycol (PEG, 8-arm, 10 kDa) was purchased from Jenkem

Technology USA (Allen, TX). Thioglycolic acid, norbornene acid, and L-cysteine were

purchased from Sigma-Aldrich (St. Louis, MO). Lithium arylphosphanate (LAP) was

synthesized as previously described. [9] Novec 7500 engineered fluid and Novec 7000

engineered fluid were purchased from 3M (St Paul, MN). Dupont Krytox 157 FSH was

purchased from Miller-Stephenson (Danbury, CT). The PEG-PFPE-PEG tri-block

surfactant was synthesized as previously described. [10] Fluorescein-conjugated bovine

serum albumin (FITC-BSA) was purchased from Life Technologies (Carlsbad, CA).

Recombinant basic fibroblast growth factor (bFGF) and bFGF ELISA kit were purchased

from Peprotech (Rocky Hill, NJ). Cell Titer Aqueous One Solution Cell Proliferation

Assay (MTS) was purchased from Promega (Madison, WI).

138

8.2.2 Microfluidic Device Fabrication

The microfluidic devices were fabricated using standard soft lithography and

micromolding techniques at the Stanford Microfluidics Foundry clean room. A

transparency mask was printed using a high-resolution printer (Fineline Imaging,

Colorado Springs, CO). Master molds were fabricated with one layer of negative SU8

photoresist (MicroChem, Newton, MA) on a silicon wafer (100 µm). The microfluidic

device was fabricated using PDMS (Sylgard 184, Midland, MI) micromolding against

the master mold. Sharpened needles (20 gauge) were used to punch inlets and outlets.

The molded PDMS layer was plasma bonded to a glass slide that was coated with an

additional layer of PDMS.

8.2.3 PEG functionalization

8-arm-PEG-Norbornene (8 arm PEG-NB) and 8-arm-PEG-Mercaptoacetic acid

(8 arm PEG-MAA) were synthesized as previously described.[11] Briefly, 8-arm PEG-

MAA was synthesized by reacting 8-arm PEG with mercaptoacetic acid in toluene. 5.0

g 8-arm PEG was dissolved in 150 mL toluene, followed by addition of 34 mg p-

toluenesulfonic acid. After adding 1.74 mL 3-mercaptopropionic, the solution was

refluxed overnight. The azeotropic mixture was collected periodically. After cooling the

solution to room temperature, the product was precipitated in ice-cold diethyl ether. 8-

arm PEG-NB was synthesized by reacting 8-arm PEG with norbornene acid following

the same procedure of 8-arm PEG MAA.

8.2.4 Microparticle Synthesis

Microparticles were synthesized using a T-junction droplet break-up microfluidic

device designed with various size input channels (50 μm, 100 μm, 150 μm, 200 μm).

The discontinuous phase consisted of 8 arm PEG-NB and 8 arm PEG-MAA in 1:1

stochiometric ratio and dissolved at various ratios (7.5 wt%, 10 wt%, 12.5 wt%, 15 wt%).

The final solutions incorporated 0.1% LAP photointiator in the discontinuous phase. The

continuous phase consisted of 2% PEG-PFPE-PEG in Novec 7500. The solutions were

run through the microfluidic channels at ~1.0 psi / ~1.0 psi (water phase / oil phase).

Synthesized spheres were collected in an oil bath with the same composition as the

continuous phase positioned under a broad spectrum, high intensity UV lamp. Novec

7500 was extracted by mixing with Novec 7000 (5 mL per wash) then removing the oil

139

phase after each wash cycle. This process was repeated five times. The residual Novec

7000 was allowed to evaporate in a water bath at 37 ºC. The resulting microparticles

were diluted in 2 mg/ml L-cysteine and then lyophilized. Prior to use, microparticles

were resolubilized in PBS at 2.5% (W/V).

8.2.5 Swelling Ratio and Mesh Size Determination

The equilibrium degree of swelling for each hydrogel was acquired in order to

calculate PEG mesh size. PEG hydrogels (100 μL) were polymerized under UV light and

their mass after swelling (MS) was measured. The hydrogels were then lyophilized and

their dry mass (MD) was measured. The mass swell ratio (QM) was determined by taking

the ratio of the swollen hydrogel mass to dry hydrogel mass. The mass swell ratio was

determined for hydrogels composed of various PEG wt% (7.5 wt%, 10 wt%, 12.5 wt%,

15 wt%).

Hydrogel mesh size (ξ) was calculated from the mass swell ratio using Flory-Rehner

calculations as previously described.[12] Firstly, the mass swell ratio was converted to a

volumetric swell ratio (QV) using Equation 1.

𝑄𝑉 = 1 +𝜌𝑝

𝜌𝑆(𝑄𝑀 − 1) (1)

where ρp is the density of the dry hydrogel (1.12 g/cm3 for PEG)3 and ρs is the density of the

solvent (1 g/cm3 for water).

Secondly, molecular weight between cross-links (MC) was calculated using Equation 2.

1

𝑀𝑐̅̅ ̅̅=

2

𝑀𝑛̅̅ ̅̅̅−

�̅�

𝑉1(ln(1−𝑣2)+𝑣2+𝜒1𝑣2

2)

𝑣2

13⁄ −

𝑣24

(2)

where 𝑀𝑛̅̅ ̅̅ is the number-average molecular weight of the un-cross-linked hydrogel (the

molecular weight of the polymer), V1 is the molar volume of the solvent (18 cm3/mol for

water), ν2 is the polymer volume fraction in the equilibrium swollen hydrogel (1/QV), �̅�

is the specific volume of the polymer (ρs / ρp), and χ1 is the polymer-solvent interaction

parameter (0.426 for PEG-water).[12]

Lastly, the mesh size was calculated using Equation 3.

𝜉 = 𝑣213⁄ 𝑙𝐶𝑛

12⁄ (2

𝑀𝐶̅̅ ̅̅ ̅

𝑀𝑟)12⁄

(3)

Where l is the average PEG bond length (0.146 nm),[12] Cn is the characteristic ratio of

PEG (4.0), and Mr is the molecular weight of the PEG repeat unit (44 for PEG).

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3050024/#R3

140

8.2.6 Microparticle Diameter Measurements

The average microparticle diameter was measured by image processing.

Microparticles were synthesized using varying size input channels (50 μm, 100 μm, 150

μm, 200 μm). Images were taken at 10x magnification. Images were analysed utilizing

MatLab Software (Natick, MA). Briefly, images were converted to black and white

based upon contiguous pixels. Microsphere area was then determined by quantifying the

number of pixels per microsphere and converting to micrometers. Sphere diameter was

then calculated. Ten images per group were assessed.

8.2.7 Protein Encapsulation and Release:

Protein encapsulation was assessed by adding FITC-labelled bovine serum

albumin (BSA) (0.1 mg/ml) or basic fibroblast growth factor (bFGF) (1.0 mg/ml) into

the discontinuous phase prior to microparticle synthesis. To assess protein encapsulation

efficiency, total BSA encapsulated was collected by dissolving the microparticles with

0.1 N NaOH and balancing the pH with 0.1 N HCL and phosphate buffered saline (PBS).

To determine the release kinetics of bFGF, microparticles were plated in the upper well

of a transwell plate (Corning, NY) with 1.0 mL PBS containing 1mg/ml BSA in the lower

well which comprised the collection medium. Released bFGF into the lower well was

collected at various time intervals over 10 days and replenished with fresh collection

medium. The amount of bFGF released was quantified by bFGF ELISA.

8.2.8 Cell Proliferation

The bioactivity of bFGF released from microparticles was determined by

assessing the mitogenic response of adipose-derived stem cells (ADSCs). ADSCs were

plated in a 96-well plate at a seeding density of 10,000 cells per well. After 24 hours,

the culture medium was replaced with 100 µl of collected released medium from various

time points, mixed with 50 µl of Dulbecco’s Modified Eagle Medium (DMEM,

Invitrogen, Carlsbad, CA) containing 2% fetal bovine serum (FBS, Invitrogen) per well,

and cultured for another 48 hours in a humidified chamber at 37 ºC. Cell proliferation

was assessed by MTS assay.

141

8.3 Results

8.3.1 Microfluidic Fabrication of PEG Microparticles

Three types of droplet microfluidic designs were evaluated for their ability to

efficiently synthesize monodisperse PEG microparticles in a repeatable and consistent

fashion (Figure 1). Essentially, the designs were composed of two main parts: 1) a

water-in-oil junction (i.e. either a t-junction or a flow-focusing design); and 2) an

extended output channel. This latter part was either a straight channel of the same

channel dimensions as the input channels, or an open channel that widens to roughly 7x

the dimension of the input channels (Figure 1A, 1B). We found that the volume of

spheres synthesized per hour increased from ~200 µl/hr to ~1400 µl/hr for the straight

channel and open channel designs, respectively (input channel width of 100 µm and input

pressures ~1.0 psi / ~1.0 psi, water/oil). To assess whether increasing the height of the

channel would further increase rates, we also compared microchannels of 100 µm height

with microchannels 150 µm height, but found the increased height led to jetting and could

not efficiently generate droplets. When comparing the t-junction to the flow-focusing

design (Figure 1B, 1C), we did not observe a significant difference in their ability to

efficiently synthesize monodisperse microparticles. However, due to the relative ease of

using and re-using the t-junction design, we chose to use the t-junction design with the

open extended output channel for the remaining experiments (Figure 1C).

The procedure for synthesizing monodisperse PEG-based microparticles is shown in

Figure 1D. The full description is provided in the experimental section. Briefly, end-

functionalized 8-arm PEG polymers were solubilized in water, which composed the

discontinuous phase. The generated droplets fell directly into an oil bath and were polymerized

under UV light. The synthesized microparticles underwent an oil-extraction procedure and then

were re-solubilized in an L-cysteine solution to prevent microparticle flocculation (i.e. covalent

bond formation between microparticles). The resulting solution was freeze-dried and then re-

solubilized in PBS prior to use.

8.3.2 Controlling microparticle diameter and mesh size

To control the microparticle diameter, four microfluidic devices were fabricated based

on the same design, which varied by input channel dimensions (i.e. 50 µm, 100 µm, 150 µm,

200 µm) (Figure 2). Using these four input channel dimensions and running the microfluidic

with input pressures of ~1.0 psi / ~1.0 psi (water pressure / oil pressure), microparticles with the

142

following average diameters were synthesized: ~42.4 µm, ~77.6 µm, ~120.1 µm, ~142.1 µm,

respectively (Figure 2C, 2D). As demonstrated in Figure 2C, the synthesized microparticles

were homogeneous and were easily controlled by varying input channel dimensions. Additional

tuning of microparticle diameter was achieved by varying the input pressures or flow rates.

Applying both of these strategies could lead to a broad range of microparticle size from ~25 μm

to ~200 μm.

To control PEG mesh size and degradation, the PEG concentration was varied in the

precursor solution ranging from 7.5% W/V to 15.0% W/V (Figure 3). Two types of end-

functionalized 8-arm PEG polymers with ester containing functional groups were used.

Polymerization of these two PEG polymers occurred via a thiol-ene reaction, which proceeds

as a step-growth reaction. As opposed to the radical chain polymerization scheme, step-growth

polymerization is expected to yield more homogeneous networks. We utilized this property to

synthesize PEG matrices with predicted mesh sizes that are on the order of protein

hydrodynamic radii (~2 nm). As demonstrated in Figure 3B, increasing PEG concentration

caused a progressive decrease in swell ratio, which corresponded with calculated mesh sizes of

~2.2 nm to ~0.9 nm (Figure 3B). To confirm degradation of the PEG matrices, swell ratio was

measured after incubating for 4 days in serum-free cell culture medium. At 4 days, the swell

ratio of each PEG hydrogel increased, which is consistent with a bulk degradation process and

would be expected for hydrolytically degradable hydrogels. At these swell ratios, the theoretical

mesh size for each group increased above 2 nm, except for that of 15 PEG wt%. Considering

the relative size of many growth factors, many growth factors would be free to diffuse through

these PEG matrices by this time point.

8.3.3 Protein encapsulation, release and bioactivity

To assess protein encapsulation, fluorescein-conjugated bovine serum albumin

(FITC-BSA) was encapsulated into PEG microparticles and confirmed by fluorescence

microscopy (Figure 4). It was found that the relative amount of FITC-BSA encapsulated

per microsphere did vary, which is demonstrated by the higher intensity of FITC in some

microparticles relative to the others. Protein encapsulation efficiency was then

quantified by digesting the PEG microspheres and measuring the amount of FITC-BSA

encapsulated compared to the amount used in the precursor solution. We found that

encapsulation efficiencies of varying concentration ranged from 19.7 ± 5.8 % to 23.2 ±

143

4.3% (Figure 4B). Furthermore, we found that varying microsphere size did not

significantly change protein encapsulation efficiency.

To assess growth factor release kinetics, we encapsulated basic fibroblast growth

factor (bFGF) into PEG microparticles with varying size or concentration and quantified

the accumulated release using ELISA (Figure 5). Microparticles fabricated using the 100

µm channel resulted in the highest level of protein release, with over 60% of encapsulated

growth factors released by day 10. In contrast, microparticles fabricated using 150 or 200

µm channels showed comparable slowed down protein release, with ~20% protein

released by day 10 (Fig. 5A). Decreasing PEG concentration generally resulted in more

rapid release, with 7.5% (W/V) PEG-based microspheres leading to ~70% accumulated

protein release by day 10 (Fig. 5B). In general, smaller microparticles and those

composed of lower PEG wt% released their contents most quickly. Of further note,

because of the type of degradation process (i.e. bulk degradation), it was interesting to

observe for any sudden boosts in release. Upon inspection, it was noted that around day

4, some of the release curves did experience a subtle increase in release rates. This

correlated with the increase in swell ratio by day 4 as shown in Figure 3. Still, much of

the bFGF was still retained in PEG matrices beyond this time point as suggested by the

release curves.

To confirm that the bioactivity of encapsulated bFGF was still retained upon

synthesis of PEG microparticles and subsequent release, the released bFGF was applied

to adipose-derived stem cells (ADSCs) and assayed for cell proliferation over 48 hours.

Compared to ADSCs with no bFGF treatment, ADSCs treated with the released

supernatant from all time points led to an increase in ADSC proliferation that was

comparable to positive controls treated with 10 ng/ml bFGF. The effect of timing on

bFGF release and ADSC proliferation was subtle, and might be due to the potency of

bFGF on ADSC. Consistent with the observation of a small increase in bFGF release

around day 4, we see that ADSC proliferation also increased around this time point. This

provided further evidence that the release of bFGF was delayed and remained bioactive

during processing and release.

8.3.4 DNA nanoparticle encapsulation and transfection

DNA nanoparticle encapsulated microparticles were synthesized by preparing

nanoparticles in a precursor solution containing various PEG wt % and flowing the

144

resulting solution through the microfluidic chip (Figure 6A). DNA nanoparticles were

prepared using D32-122 poly(β-amino ester) (PBAE) at a polymer to DNA weight ratio

of 20:1. Upon synthesis and re-solubilization, the resulting DNA nanoparticle

concentration was 70 µg per ml of microspheres.

The ability of encapsulated DNA nanoparticle to transfect was tested using DNA

nanoparticles encoding luciferase and vascular endothelial growth factor (VEGF).

Microparticles (200 µl or 14 µg DNA nanoparticles) were co-incubated with HEK293T

cells in transwell culture (Microparticles in the upper well and HEK293T cells in the

lower well). Medium was collected over 8 days and assayed for luciferase expression

(Figure 6B) and VEGF release (Figure 6C). Luciferase expression and VEGF release

correlated with PEG wt%, demonstrating more rapid increase in transfection by 7.5 wt%

PEG and slower transfection using 12.5 wt% PEG. This demonstrated that DNA

nanoparticles remained bioactive during encapsulation and release, and further that

release of DNA nanoparticles could be controlled by PEG wt%.

8.4 Discussion

Herein, we report the synthesis of biodegradable PEG-based, monodisperse

microparticles for controlled growth factor delivery and DNA nanoparticles with tunable release

kinetics using microfluidics. Due to the uniformity of the microparticles and the homogeneity

of the PEG matrices, this approach may serve as a reliable and consistent method for preparing

injectable growth factor delivery depots. Of particular interest is that controlling microparticle

properties was relatively simple and robust, and throughout the processing conditions the

encapsulated biologics remained bioactive (Figure 5C). Here, two microparticle properties were

controlled: 1) the mesh size of the hydrogel; and 2) the microparticle diameter. The first was

controlled by varying PEG concentration in the precursor solution; and, the second was

controlled by using microfluidic channels of differing dimensions. Varying these two

parameters considerably varied the rate of growth factor release and DNA nanoparticle release.

Given the bio-inertness of PEG and the mildness of the processing conditions, such a platform

may be useful for encapsulating and releasing a broad range of growth factors, as well as various

other biological molecules or cells.

Microparticle mesh size and degradation rate were easily controlled by varying PEG

concentration in the precursor solution, which correspondingly affected the rate of growth factor

release and DNA nanoparticle release. Controlling mesh size by varying polymer concentration

145

is a common strategy for controlling release, which has been demonstrated in many other studies

using bulk PEG hydrogels. Using this strategy in a microparticle formulation enables this

property to be used as an injectable formulation, in which separate microparticles could be

designed with different PEG concentrations. Indeed, we found that lower concentration PEG

wt% led to increased biologic release of both basic fibroblast growth factor and DNA

nanoparticles, while higher PEG wt% led to the opposite. Importantly, this led to differing rates

in transfection of HEK293T cells. These findings demonstrate the potential for generating

mixed formulations to control the release of multiple biologics, which otherwise cannot be

achieved by a bulk hydrogel.

PEG has many advantages to other polymers because it may be easily chemically

modified, which would provide another level of control over biologic encapsulation and release

[8]. Indeed, thiol-ene chemistry enables the easy incorporation of multiple types of functionality

by simply incorporating biomolecules containing a free thiol. For example, L-cysteine was

actually required in this experiment because of the spontaenous covalent crosslinking between

neighboring microparticles due to unreacted thiol groups. This piece of development also

demonstrated microparticles could be easily surface-modified by re-solubilizing the

microparticles with any other type of molecule that contained a thiol, such as a bio-functional

peptide. By the same token, functionality can be incorporated within the PEG microparticle by

adding a functional peptide or biomolecule containing a thiol to the precursor solution. By

taking advantage of this property, growth factor release could be further modulated by

incorporating protein binding biomolecules or even covalently linking the growth factor directly

to the microparticle.

Simple modifications to microfluidic design can be used to increase the rate of

microparticle synthesis and enable more robust control of size. To increase rate of microparticle

synthesis, the resistance of the output channel post-droplet formation was decreased by

expanding its width. This simple change led to increased rates of microparticle formation by

over ten-fold. To control microparticle size, multiple microfluidic chips were designed with

varying input channel dimensions. In comparison to adjusting input flow rates, we found this

to be a more robust approach. This is because relying on input flow rates requires more accuracy

between the water and oil channels. While accuracy can be easily achieved by a syringe pump,

a pneumatic controller might be a more efficient approach to use when running multiple

microfluidic devices in parallel, which would be subject to subtle changes in air pressure. We

demonstrated that such an approach was able to rapidly synthesize monodisperse microparticles

146

of varying sizes, which may support the use of the microfluidic device as a manufacturing

platform.

In summary, here we report a droplet microfluidics platform that allows synthesis of

biodegradable PEG-based microspheres for controlled release of proteins and DNA

nanoparticles. We further demonstrated our ability to synthesize monodisperse microspheres

by tuning the width of microfluidics channel width. PEG microsphere compositions can be

easily modulated by varying PEG concentration in the precursor solution. We demonstrated

gradual protein release from PEG-based microspheres, and the release kinetics can be further

turned by varying PEG composition and microsphere size. Given the bio-inertness of PEG, such

platform may be useful for encapsulating and releasing a broad range of growth factors or DNA

nanoparticles. Further chemical modification to the PEG composition may provide another level

of control over biologic encapsulation and release, and may be easily incorporated into the

present system. To achieve temporal release of multiple biologics to enhance tissue

regeneration, microparticles with different release kinetics or loaded with different cargos may

be easily combined to provide an injectable drug delivery system for treating a broad range of

diseases.

Acknowledgements

The authors would like to thank American Heart Association National Scientist

Development Grant (10SDG2600001) and Stanford Medical Scholars Research Program

for funding. The authors would also like to thank Stanford Microfluidics Foundry for

technical assistance with microchip fabrication.

References

1. Thorsen, T., et al., Dynamic pattern formation in a vesicle-generating microfluidic

device. Phys Rev Lett, 2001. 86(18): p. 4163-6.

2. De Geest, B.G., et al., Synthesis of monodisperse biodegradable microgels in

microfluidic devices. Langmuir, 2005. 21(23): p. 10275-9.

3. He, T., et al., A modified microfluidic chip for fabrication of paclitaxel-loaded poly(l-

lactic acid) microspheres. Microfluidics and Nanofluidics, 2011. 10(6): p. 1289-1298.

4. Huang, K.S., et al., Microfluidic controlling monodisperse microdroplet for 5-

fluorouracil loaded genipin-gelatin microcapsules. J Control Release, 2009. 137(1): p.

15-9.

147

5. Kesselman, L.R.B., et al., Synthesis of Monodisperse, Covalently Cross-Linked,

Degradable “Smart” Microgels Using Microfluidics. Small, 2012. 8(7): p. 1092-1098.

6. Xu, Q., et al., Preparation of monodisperse biodegradable polymer microparticles using

a microfluidic flow-focusing device for controlled drug delivery. Small, 2009. 5(13): p.

1575-81.

7. Yang, C.-H., K.-S. Huang, and J.-Y. Chang, Manufacturing monodisperse chitosan

microparticles containing ampicillin using a microchannel chip. Biomedical

Microdevices, 2007. 9(2): p. 253-259.

8. Lin, C.-C. and K. Anseth, PEG Hydrogels for the Controlled Release of Biomolecules

in Regenerative Medicine. Pharmaceutical Research, 2009. 26(3): p. 631-643.

9. Fairbanks, B.D., et al., Photoinitiated polymerization of PEG-diacrylate with lithium

phenyl-2,4,6-trimethylbenzoylphosphinate: polymerization rate and cytocompatibility.

Biomaterials, 2009. 30(35): p. 6702-7.

10. Holtze, C., et al., Biocompatible surfactants for water-in-fluorocarbon emulsions. Lab

Chip, 2008. 8(10): p. 1632-9.

11. Fairbanks, B.D., et al., A Versatile Synthetic Extracellular Matrix Mimic via Thiol-

Norbornene Photopolymerization. Advanced Materials, 2009. 21(48): p. 5005-5010.

12. Zustiak, S.P. and J.B. Leach, Hydrolytically degradable poly(ethylene glycol) hydrogel

scaffolds with tunable degradation and mechanical properties. Biomacromolecules,

2010. 11(5): p. 1348-57.

148

Figures

Figure 8.1. Schematics of microfluidic chip design for microsphere synthesis. (A-C) Design of

the three droplet microfluidic chips assessed. (A, C) Two t-junction designs with differing

output channel widths. (B) A flow-focusing device with an wide output channel. A snapshot

of microsphere formation in each microfluidic chip is shown on the right. (D) Schematic of

fabrication process of polyethylene glycol (PEG)-based microspheres. 8-arm PEG-norbornene

(PEG-NB) and 8-arm PEG-mercaptoacetic acid (PEG-MAA) were added to the continuous

phase in 1:1 stochiometric ratio. Upon exiting the microfluidics chip, the PEG microspheres

were collected in oil bath and photocrosslinked using UV light. Microspheres are further

processed by an oil extraction procedure followed by lyophilization. Image shows microsphere

formation after re-solubilizing.

149

Figure 8.2. Control of microsphere size by adjusting microchip in put channel dimensions (50

µm, 100 µm, 150 µm and 200 µm). (A, B) Schematic of the t-junction open-channel design and

corresponding dimensions. (C) Microspheres synthesized based on the four different input

dimensions. Scale Bar: 200 µm. (D) Averaged diameters of microspheres synthesized using

microfluidics chip with varying microchannel width. (* p < 0.05)

150

Figure 8.3. Effect of varying PEG concentration on microsphere property including swelling

ratio and microgel mesh size at days 0 and 4. Increasing PEG concentration led to decreased

mass swelling ratio (A) and decreased mesh size (B), which increased by day 4 indicating

hydrogel degradation. (*** p < 0.001)

151

Figure 8.4. Protein encapsulation in biodegradable PEG microspheres of differing size and

polymer composition. (A) Protein encapsulation was confirmed using fluorescein-conjugated

bovine serum albumin (FITC-BSA). Images demonstrate FITC-BSA encapsulation in two

different sized microparticles. (B) Protein encapsulation efficiency was about 19.7 ± 5.8 % to

23.2 ± 4.3%, and did not significantly vary in microparticles of different PEG composition.

152


varying size and PEG concentration. (A) Effects of varying microsphere size on bFGF release

over 10 days. (B) Effects of varying PEG concentration on bFGF release over 10 days. (C)

Bioactivity of released bFGF over 10 days was retained, as shown by the cell proliferation assay

using adipose-derived stem cell (ADSC). Cells without FGF (-) or with fresh FGF (+) were

included as controls. (* p < 0.05).

153

Figure 8.6. Transfection by DNA nanoparticles encapsulated and released from biodegradable

PEG microspheres of varying polymer weight percent (W/V). (A) Schematic illustrating the

formation of DNA nanoparticles in a PEG precursor solution prior to synthesis of microspheres.

(B) Transfection of HEK 293T cells by released DNA nanoparticles encoding luciferase from

PEG microspheres of varying polymer weight percent (W/V). (C) VEGF transfection of

HEK293T cells by released microspheres encapsulating DNA nanoparticles encoding VEGF.

9. Summary, Limitations, and Future Directions

9.1 Summary

Over the course of this thesis, we explored two different strategies to harness stem cells

as drug delivery vehicles for therapeutic angiogenesis. The first platform explored the potential

for genetically modifying adipose-derived stem cells (ADSCs) with angiogenic factors ex vivo

using poly(β-amino esters) (PBAE). The second platform explored the potential for sequential

release of chemokine and DNA to recruit and program endogenous progenitor cells in situ.

9.1.1 Summary of the non-viral engineered ADSCs approach

With regards to the ex vivo modified stem cell platform, we focused on optimizing the

following components: 1) cell source; 2) non-viral transfection agent; and, 3) selection of an

appropriate therapeutic gene. In terms of cell source, we focused on ADSCs due to their relative

abundance, ease of isolation and demonstrated angiogenic potential. In our studies, we found

that paracrine release from ADSCs alone could promote multiple endothelial cell processes of

angiogenesis in vitro. We further found that ADSCs behaved well under hypoxia, whereby their

proliferation potential and angiogenic growth factor production increase. Our results are

supported by many others that have also demonstrated the therapeutic potential of ADSCs,

which indicate that ADSCs are potentially useful for treating ischemic disease.

ADSCs were genetically modified using poly(β-amino esters) (PBAE), a non-viral

transfection vector. PBAE/DNA nanoparticles were optimized for transfection into ADSCs,

including human and mouse ADSCs of multiple cell isolations. We found that transfection was

dependent on the cell type and culture conditions, and could also be optimized by varying

transfection parameters such as polymer to DNA weight ratio and DNA dose. We also found

that ADSC proliferation correlated with their ability to transfect and that promoting proliferation

could enhance their transfection efficiency. Under the optimized conditions, PBAE led to high

transfection efficiency and low cytoxicity in ADSCs compared to the commercially available

gold standard, Lipofectamine 2000. These transfection conditions enabled us to transfect

multiple therapeutic genes on different plasmid constructs into ADSCs, which demonstrated the

versatility of using PBAE for genetic upregulation in ADSCs.

The main therapeutic genes we explored were vascular endothelial growth factor

(VEGF), CXC chemokine receptor 4 (CXCR4), and hepatocyte growth factor (HGF). Initially,

VEGF was chosen because it had been the most well-studied angiogenic growth factor. We

154

found that VEGF-overexpressing ADSCs could enhance the therapeutic effects of ADSCs

compared to ADSCs alone in vitro (chapter 4), but did not perform as well in vivo (chapter 5).

VEGF-overexpressing ADSCs led to enhanced inflammation, which proved to be quite

damaging and led to detrimental consequences on ischemic limb salvage. CXCR4 was explored

for its ability to enhance ADSCs as drug delivery vehicles to ischemia. When CXCR4 was co-

expressed with VEGF, we found that this improved ischemic limb salvage compared to VEGF-

overexpression alone in vivo, but overall the two genes did not seem to synergize. However,

CXCR4-overexpression alone was enough to enhance ischemic tissue repair, promote early

blood reperfusion, and attenuate some of the early inflammation in vivo. The therapeutic

potential of CXCR4-overexpressing ADSCs was confirmed in a second in vivo study in a more

severe animal model.

HGF was explored as an alternative angiogenic factor to VEGF (chapter 6). HGF-

overexpressing ADSCs promoted both endothelial cell and smooth muscle cell responses in

vitro and in vivo, which led to better therapeutic effects compared to VEGF-overexpressing

ADSCs. In this study, we also developed a plasmid containing both CXCR4 and HGF genes,

which was used to overexpress both in ADSCs in one-to-one correspondence. The combined

overexpression of CXCR4 and HGF promoted both ADSC survival and angiogenic potential.

9.1.2 Summary of the recruitment and programming endogenous cells approach

With regards to the recruitment and programming endogenous progenitor cells in situ

strategy, we sought to assess the feasibility for using endogenous progenitor cells as a cell source

for gene therapy. This approach was partly motivated by our studies on CXCR4-overexpressing

ADSCs and the overall objective of eliminating ex vivo cell culture. Recruitment and

programming endogenous progenitor cells was evaluated by considering each process

separately (chapter 7), including: 1) endogenous progenitor cell recruitment; and, 2) in situ

programming of endogenous progenitor cells. Endogenous cell recruitment was assessed by

FACS analysis, which demonstrated that ischemia was a potent stimulus for recruiting

endogenous cells; and further, that controlled release of stromal derived factor-1α (SDF-1α)

could increase recruitment compared to ischemia at each time point. Endogenous progenitor

cell programming in situ was assessed by bioluminescence imaging, which demonstrated that

in situ programming correlated with endogenous progenitor cell recruitment. Controlled SDF-

1α delivery further promoted in situ programming.

155

The duration of in situ transgene expression depended largely on the timing of DNA

delivery post-ischemic induction. Our main findings were that delivering DNA too early post-

ischemia without SDF-1α could not last, while delivering DNA later enabled prolonged

transgene expression. Controlled release of SDF-1α enabled earlier DNA delivery. This

approach was validated for therapeutic efficacy using HGF, which demonstrated improved limb

salvage and angiogenic response.

Development on a drug delivery platform was done by using a microfluidic platform to

generate biodegradable PEG-based microspheres with tunable release profiles (Chapter 8).

Microspheres synthesized using microfluidics were more uniform and homogeneous. With this

platform, we were able to achieve controlled release of bioactive growth factors and DNA

nanoparticles. The successful development of these microspheres enable the sequential delivery

of SDF-1α and DNA in a single injection and set the stage for further designing one-step

platforms for ADSC transfection.

9.2 Limitations

9.2.1 Limitations on the non-viral engineered ADSCs approach

Some of the limitations that should be noted in relation to the non-viral engineered

ADSC platform, include: variability of cell sources, animal model and mild genetic

overexpression. In terms of cell sources, we have found differences in ADSCs from different

cell sources, at different passage numbers, and over long-term storage. As an autologous cell

source, this should be noted, which suggests that there could be more work on isolating and

normalizing ADSCs from different sources. Some possibilities for differences include: different

anatomical locations of the fat pads, demographics of the patient (i.e. race, age, gender, etc), as

well as disease state of the patient. The cell isolation procedure is fairly basic, simply relying

on fat pad dissection, enzymatic digestion, and cell adherence to the tissue culture plate. This

process of negative/positive selection could be improved to help normalize the cell type. Other

methods for overcoming this limitation, include: genetic modifications ex vivo and use of

allogeneic cell sources. Genetic modifications may overcome some of the important

deficiencies per cell type by overexpressing relevant genes, which is the focus of this thesis.

Allogeneic cell sources, such as bone marrow-derived mesenchymal stem cells, are known to

exhibit low immunogenicity and are currently being used as an allogeneic source for cell

therapy.

156

In terms of the animal model, the models used in this study were generated by acutely

ligating the femoral artery in an otherwise healthy mouse. Since this is the case, many of these

mice possessed strong inherent regenerative capacity, which tended to override the healing

process beyond a certain time point. This could mask some of the therapeutic responses

afforded by the cell sources and could also lead to over-interpretation of the healing responses

caused by the gene modified cells, if additional measures are not also assessed. In an effort to

overcome these issues, retired breeder mice were used, which are older and possess less

regenerative potential. This model is a more difficult model to achieve a therapeutic response.

Other models of the general co-morbid state of patients with peripheral arterial disease (PAD)

include using mice with genetic backgrounds of being non-obese diabetic (NOD) or ApoE-

deficient mice.

With regards to transfection, we found that human ADSCs were a challenging cell

source to transfect in comparison to mouse ADSCs and human embryonic kidney cells

(HEK293T). Transfection efficiencies in human ADSCs generally reached 15-25%. Although

these were higher than that achieved by commercially available non-viral vectors, this generally

led to a mild genetic overexpression of therapeutic factors. To help overcome some of these

challenges, we have found that transfecting hADSCs in growth conditions could enhance their

transfection. Other methods to improve transfection efficiency include: further promoting cell

proliferation; synthetic modifications to the polymer; designing high-expressing plasmid

constructs; and, also better selection of ADSC sources. Promoting cell proliferation can

improve transfection because the nuclear membrane dissolves during cell proliferation, which

enables plasmid DNA to traffic from the cytoplasm to the nucleus. Modifications to the polymer

to incorporate nuclear localization signals are another strategy to enable trafficking from the

cytoplasm to the nucleus. Designing plasmids with high expressing promoters or as smaller

constructs, such as minicircle DNA, are methods to promote DNA entry and gene expression.

As far as cell sources, some of our recent ADSC isolations have enabled about 3x greater

transfection efficiencies (~60%) compared to those used in our earlier studies. This

demonstrates that certain ADSC populations have more transfection potential.

9.2.2 Limitations on the recruitment and programming endogenous cells approach

Some of the limitations of the recruitment and programming endogenous progenitor

cells approach include: 1) use of only one mouse strain; 2) reliance on acute injury, and, 3) lack

of control over gene delivery. The mouse strain used in our study was C57/Black, which was

157

chosen based on other similar studies of endogenous cell recruitment. Although we may

speculate that the results of our study extend to other mouse strains, as well as into other species,

further study is needed to demonstrate that this is the case. Importantly, endogenous progenitor

cell recruitment might vary according to disease state. This raises the question if similar

recruitment would be achieved in a more diseased animal model.

Another limitation in our study is that the hindlimb ischemia mouse model actually

simulated an acute ischemic episode rather than a chronic ischemic state. In this context, we

may conclude that acute ischemia is a potent stimulus for cell recruitment, which could be

further improved by SDF-1α delivery. It is unclear whether chronic ischemia would promote

endogenous progenitor cell recruitment to the same degree. Still, the acute ischemia model is

relevant in the clinical setting, because this reflects diseases such as acute myocardial infarction.

Also, causing an acute injury is often employed to promote tissue regeneration, such as

transmyocardial laser revascularization to promote angiogenesis or microfracture to promote

cartilage growth. These factors should be considered when further studying or designing

treatments based on the recruitment and programming approach.

In terms of the inability to control transgene expression by DNA delivery into tissues,

this is a concern because of the potential for unregulated gene expression, as well as the

possibility for causing an immune response. Several studies have suggested that plasmid DNA

mainly transfects regenerating muscle fibers, as well as non-inflammatory mononuclear cells.

Despite these findings, the precise fate of plasmid DNA is not entirely known. Further study

should be done to help determine which cell populations are transfected by plasmid DNA and

the mechanism of its removal. To help control gene expression, additional genes might be

engineered into the plasmid, which could be used for plasmid DNA regulation.

9.2.3 Limitations on the synthesis of biodegradable PEG microspheres using microfluidics

In the use of microfluidics to synthesize biodegradable polyethylene glycol (PEG)

microspheres using microfluidics, some of the limitations include: 1) ultraviolet (UV) light

initiated polymerization; and 2) requirement of high cross-linking efficiency. The use of UV

light to initiate polymerization is a potential limitation that might cause denaturation of the

encapsulated protein or DNA. While UV light is a standard approach in biomaterial research,

other methods that might be more biologic-friendly would be preferred, such as covalent cross-

linking. The requirement of high cross-linking polymers to effectively polymerize

microspheres limits versatility of the polymers used. For example, many designed biomaterials

158

might incorporate multiple signals or peptides that might not otherwise be available for use due

to their low cross-linking efficiency. Increasing concentration of the precursor solution may

overcome this issue.

9.3 Future Directions

There are many future directions that will help better understand the mechanism of each

approach, as well as further advance these approaches towards the clinic. These are discussed

separately for each platform.

9.3.1 Future directions of the non-viral engineered ADSCs approach

For the non-viral engineered ADSC platform, the directions that could be pursued

include: 1) understanding the therapeutic mechanism of CXCR4-overexpressing ADSCs; 2)

determining the fate of ADSCs upon transplantation; 3) using a biomaterial cell carrier for

delivery; and, 4) studying the therapeutic effects in a larger animal model.

Towards understanding the mechanism of CXCR4-overexpressing ADSCs, we have

begun to understand some of the changes in the growth factor and cytokine profile due to

CXCR4-overexpression, as well as the effects of paracrine release on endothelial cells and

macrophages. This may help to explain some of the therapeutic effects we see in vivo, but

further study is warranted to understand how CXCR4-overexpressing ADSCs lead to their

therapeutic effects. This may also enable us to identify and potentially exploit the relevant

process in future strategies.

When we transplant the ADSCs into hindlimbs of the mice, it is not completely clear

what happens to the fate of the cells with regards to location of engraftment and whether these

might persist over time. From bioluminescence imaging (BLI) and immunostaining, we have

an idea of the cell survival time and the location of cell engraftment. However, a direct

correlation of BLI signal and immunostained transplanted cells is inconsistent. For example,

despite the high BLI signal observed, immunostaining in the same limbs only identifies few

transplantated cells. This suggests that further study and potentially better experimental

methods would be important to understand the cell fate, which will also be important for clinical

translation.

To better improve cell survival upon transplantation, one approach is to include a

biomaterial cell carrier. A biomaterial might be able to protect the cells during injection and

may also provide a substrate to which the cells could attach in vivo. This direction could also

159

open up several avenues for tailoring the biomaterial to provide additional cues to the cells to

promote their therapeutic effects.

For clinical translation, it will be important to continue assessing the therapeutic effects

of non-viral engineered ADSCs in larger animal models. Performing studies in larger animal

models may better reflect the scale of ischemia in human patients and help to identify other

considerations not represented in the mouse.

9.3.2 Future directions of the recruitment and programming approach

For the recruitment and programming endogenous progenitor cells platform, future

studies would include: 1) deeper characterization of the recruited cell populations; 2)

assessement of inflammation; 3) modulating inflammation to promote endogenous progenitor

cell programming; and, 4) exploring additional cues for recruitment.

In our studies, we mainly focused on characterizing populations that are known to

respond to SDF-1α. However, there was a large increase in recruited cells due to ischemic

injury, and our current characterization only extended to a small proportion of these cells.

Overall, our observation was that the number of isolated cells increased nearly 10-fold due to

ischemia compared to sham controls. Further characterization on these cell sources may enable

us to know which cells are transfected and participate in therapeutic responses.

Inflammation is an important process that occurs post-ischemic injury that should be

studied in our experimental model. A proper understanding the dynamics of inflammation and

progenitor cell recruitment will help us to understand the effects of timing on gene delivery and

subsequent transfection. One finding was that early gene delivery did not last, which was likely

due to the inflammatory process. Considering that there was some transgene expression, it is

unlikely that the DNA was simply removed. More work should be done to understand how

inflammation might extinguish transgene expression in the transfected cells.

To potentially promote early gene delivery post-ischemic injury, one potential strategy

is to modulate inflammation. With a better understanding of the inflammatory cells that are

responsible for DNA removal, more targeted therapeutics might be designed to modulate the

relevant cell populations, while continuing to promote the recruitment of progenitor cells.

To further promote endogenous progenitor cell recruitment and activation, additional

recruitment cues could be studied. We found that ischemia was a potent stimulus for

endogenous cell recruitment, which could be further promoted by SDF-1α delivery. However,

there is a wide variety of stimuli that could be explored, including: homing factors, mobilization

160

factors, inflammatory factors, hypoxia stimuli, and inducing injury. Other homing factors

include: VEGF and sphingosine-1-phosphate (S1P). Inflammatory factors might include:

interleukin-1 (IL-1) and tumor necrosis factor alpha (TNF-α). Mobilization factors include:

granulocyte colony stimulating factor (G-CSF); granulocyte-macrophage colony stimulating

factor (GM-CSF), CXCR4 antagonists (AMD3100). Hypoxia mimicking agents include cobalt,

which promotes hypoxia inducible factor-1α (HIF-1α) stabilization. Also, other injurious

stimuli might initiate cell responses, such as transmyocardial laser revascularization or subtle

damaging agents.

9.3.3 Future directions of the biodegradable PEG microspheres using microfluidics

Moving forward with the biodegradable PEG microspheres for drug release, some of

the next steps include: 1) covalently cross-linking microspheres during microfluidic synthesis;

and, 2) application of the sequential delivery approach for biologics. Covalent cross-linking

during microfluidic synthesis could be done to obviate the need for UV light-based

polymerization. This would be a more mild synthesis condition for encapsulation of biologics

and even cells. Applying microspheres for temporal delivery of two or more biologics is now

possible based on the currently developed platform. Drug delivery systems should be designed

that may deliver protein and DNA to achieve recruitment and programming of endogenous

progenitor cells. Another possibility is the co-delivery of DNA-encapsulated microspheres and

ADSCs, which might enable transfection of ADSCs during injections.

9.4 Concluding remarks

For patients with PAD and other ischemic diseases, surgical revascularization will

remain the standard of care. Therapeutic angiogenesis will help treat the subset of patients that

are unable to withstand surgery. However, current clinical approaches for therapeutic

angiogenesis have led to variable results. To overcome these limitations, the work presented in

this thesis advances the strategies for both cell-based therapeutics and biologic-based

therapeutics. Overall, we have added new tools in the repertoire of therapeutic angiogenesis by

demonstrating the potential for non-viral engineered ADSCs with CXCR4 and/or HGF as

therapeutic targets; and also demonstrating the promise for recruiting and programming

endogenous progenitor cells in situ.

161

harnessing stem cells as drug delivery …vc043cd5324/deveza... · harnessing stem cells as drug...

Documents