ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2021

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 291

Advanced Methods for Evaluationof the Performance of ComplexDrug Delivery System

CAROLINE ALVEBRATT

ISSN 1651-6192ISBN 978-91-513-1078-7urn:nbn:se:uu:diva-425586

Dissertation presented at Uppsala University to be publicly examined in A1:111a, BiomedicalCenter, Husargatan 3, Uppsala, Friday, 15 January 2021 at 09:15 for the degree of Doctor ofPhilosophy (Faculty of Pharmacy). The examination will be conducted in English. Facultyexaminer: Professor Caitriona O'Driscoll (University College Cork).

AbstractAlvebratt, C. 2021. Advanced Methods for Evaluation of the Performance of ComplexDrug Delivery System. Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 291. 68 pp. Uppsala: Acta Universitatis Upsaliensis.ISBN 978-91-513-1078-7.

Low oral bioavailability of drugs originating from poor aqueous solubility is a common issuein drug development. Various enabling formulations have been presented to circumvent thislimitation, many making use of supersaturation. In these, the drug is delivered to the gastro-intestinal lumen in a high energy state e.g. in amorphous form or a liquid lipid vehicle.Concentrations surpassing the equilibrium solubility of the crystalline drug are achieved, whichfacilitate increased absorption for dissolution-rate limited compounds. Meanwhile the use ofthe enabling formulation can be beneficial to increase the bioavailability of poorly water-soluble drugs, in vitro evaluation of these systems remain challenging. Limited methods havealso evaluated several different types of enabling formulation in the same experimental setup.The overall aim of this thesis was therefore to develop assays to study the performance ofvarious complex drug delivery systems. In the first part, a small scale dissolution apparatus,the µDiss Profiler, was used to study drug release from drug-loaded mesoporous magnesiumcarbonate (MMC). A protective filter was developed to minimize particle interference on theUV-measurements, enabling studies of supersaturation from the amorphous carrier. In thesecond paper, lipids were adsorbed onto the MMC. A modified in vitro lipolysis setup wasestablished and the samples were analyzed with nuclear magnetic resonance spectroscopy. Astability study of the lipid-loaded MMC was also performed. The methods developed in thefirst two projects provided an insight to events occurring in the intestinal lumen. The intestinalabsorption has however been shown to be a complex interplay between dissolution-digestionand permeation. In the final two projects, two devices comprising of a donor (luminal) chamberand a receiver (serosal) chamber were studied (the µFLUX and the enabling absorption, ENA,device). The two chambers were separated by a semipermeable membrane (cell-based and/orphospholipid-based). A wide range of enabling formulations were evaluated in the two assays.As the exposure in the donor correlated poorly with the exposure in the receiver compartment,this emphasizes the importance of in vitro methods taking both the dissolution-digestion andpermeation into account. The ENA results also predicted the in vivo performance in rats well. Toconclude, several models have been established in the thesis to study the in vitro performanceof enabling formulations, which will be valuable for screening of appropriate drug deliverysystems.

Keywords: amorphous solid dispersion, coadministration, digestion, dissolution, drugabsorption, enabling formulation, in vitro assay, lipid-based formulation, mesoporous carrier,supersaturation

Caroline Alvebratt, Department of Pharmacy, Box 580, Uppsala University, SE-75123Uppsala, Sweden.

© Caroline Alvebratt 2021

ISSN 1651-6192ISBN 978-91-513-1078-7urn:nbn:se:uu:diva-425586 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-425586)

Till Farfar, Mormor, Farmor och Morfar

”Just smile and wave boys, smile, and wave” Skipper the Penguin, Madagascar

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals.

I Alvebratt, C., Cheung, O., Strømme M., Bergström C.A.S.

(2018) A modified in situ method to determine release from a complex drug carrier in particle-rich suspensions. The AAPS PharmSciTech, 19(7):2859–2865

II Alvebratt, C., Dening, T.J., Åhlén, M., Cheung, O., Strømme M., Gogoll, A., Prestidge, C.P., Bergström C.A.S. (2020) In vitro Performance and Chemical Stability of Lipid-Based Formula-tions Encapsulated in a Mesoporous Magnesium Carbonate Car-rier. Pharmaceutics, 12:426

III Alvebratt, C.*, Keemink, J.*, Edueng, K., Cheung, O., Strømme M., Bergström C.A.S. (2020) An in vitro dissolution–digestion–permeation assay for the study of advanced drug delivery sys-tems. European Journal of Pharmaceutics and Biopharmaceu-tics, 149:21-29

IV Alvebratt, C., Karlén, F., Åhlén, M., Edueng, K., Dubbelboer, I., Bergström, C.A.S. Two is better than one: benefits of combin-ing supersaturating and solubilizing effects. (In manuscript)

Reprints were made with permission from the respective publishers.

*The authors contributed equally to this work.

Other contributions

I Andersson, S.B.E., Alvebratt, C., Bevernage, J., Bonneau, D., da Costa Mathews, C., Dattani, R., Edueng, K., He, Y., Holm, R., Madsen, C., Müller, T., Muenster, U., Müllertz, A., Ojala, K., Rades, T., Sieger, P., Bergström, C.A.S. (2016) Interlaboratory validation of small-scale solubility and dissolution measure-ments of poorly water-soluble drugs. Journal of Pharmaceutical Sciences, 105(9):2864–2872

II Andersson, S.B.E., Alvebratt, C., Bergström, C.A.S. (2017) Controlled suspensions enable rapid determinations of intrinsic dissolution rate and apparent solubility of poorly water-soluble compounds. Pharmaceutical Research, 34(9):1805-1816

III Yang, J., Alvebratt, C., Zhang, P., Zardán Gómez de la Torre, T., Strømme, M., Bergström, C.A.S., Welch, K. (2017) En-hanced release of poorly water-soluble drugs from synergy be-tween mesoporous magnesium carbonate and polymers. Interna-tional Journal of Pharmaceutics, 525:183-190

IV Yang, J., Alvebratt, C., Lu, X., Bergström, C.A.S., Strømme, M., Welch K. (2018) Amorphous magnesium carbonate nanopar-ticles with strong stabilizing capability for amorphous ibuprofen. International Journal of Pharmaceutics, 548(1): 515-521

Contents

Introduction ................................................................................................... 11 The Biopharmaceutics Classification System .......................................... 12 Enabling formulations .............................................................................. 14

Amorphous solid dispersions ............................................................... 15 Carrier-based delivery ......................................................................... 15 Lipid-based formulations ..................................................................... 16

Selection of appropriate enabling formulations ....................................... 17 In vitro methods to evaluate the performance of enabling formulations .. 18

Dissolution assays ................................................................................ 19 In vitro lipolysis ................................................................................... 19 Permeation models............................................................................... 20 Combined dissolution-permeation models .......................................... 20

Limitations of current methods ................................................................ 21

Aims of the thesis.......................................................................................... 23

Methods ........................................................................................................ 24 Selection of model compounds ................................................................ 24 Selection of enabling formulations ........................................................... 24 Preparation of enabling formulations ....................................................... 25

Micronized drug .................................................................................. 25 Drug-loading of MMC ......................................................................... 25 Lipid-based formulations ..................................................................... 25 Adsorption of lipids onto MMC .......................................................... 26 Amorphous solid dispersions ............................................................... 26

Solid state characterization techniques ..................................................... 26 Differential scanning calorimetry ........................................................ 26 Powder X-ray diffraction ..................................................................... 27 Scanning electron microscopy ............................................................. 27 Thermal gravimetric analysis .............................................................. 27 Nitrogen sorption analysis ................................................................... 27 Fourier-transform infrared spectroscopy ............................................. 28 Polarized light microscopy .................................................................. 28

In vitro release and digestion assays ........................................................ 28 The µDiss Profiler ............................................................................... 28 In vitro lipolysis and 1H NMR ............................................................. 29

Combined digestion-permeation assays ................................................... 29 Cell culture .......................................................................................... 29 The µFLUX system ............................................................................. 30 Enabling absorption devices ................................................................ 31

In vivo pharmacokinetic study .................................................................. 33

Results and discussion .................................................................................. 34 In situ measurements in particle rich systems (Paper I) ........................... 34

Design of protective filters .................................................................. 35 Release studies from drug-loaded carrier ............................................ 35 Detection of chemical degradation ...................................................... 37 Limitations with protective filters ....................................................... 37

In vitro evaluation of solid lipid based formulations (Paper II) ............... 38 Modified lipolysis with 1H NMR analysis ........................................... 38 Chemical stability of lipid-loaded MMC ............................................. 39

Dissolution–digestion–permeation assays (Papers III-IV) ....................... 40 In vitro performance of enabling formulations .................................... 41 In vitro-in vivo correlation of the ENA assay ...................................... 46 Strengths and limitations with the digestion-permeation assays ......... 47

Conclusions ................................................................................................... 49

Future perspective ......................................................................................... 51

Populärvetenskaplig sammanfattning ........................................................... 53

Acknowledgements ....................................................................................... 55

References ..................................................................................................... 58

Abbreviations

ASD Amorphous solid dispersion AUC Area under the curve A/V Area-to-volume ratio BCS Biopharmaceutics classification system BSA Bovine serum albumin CAP-MMC Captex-loaded mesoporous magnesium carbonate CV Coefficient of variation DSC Differential scanning calorimetry DG Diglyceride EMA The European Medicines Agency ENA Enabling absorption FA Fatty acids FaSSIF Fasted state simulated intestinal fluid FDA The U.S. Food and Drug Administration FFA Free fatty acid FTIR Fourier-transform infrared spectroscopy GI Gastro-intestinal HCl Hydrochloric acid 1H NMR Proton nuclear magnetic resonance spectroscopy HBSS Hank’s balanced salt buffer HPLC High-performance liquid chromatography HPMC Hydroxypropyl methylcellulose HPMCAS Hydroxypropyl methylcellulose acetate succinate I.V. Intravenous IVIVC In vitro in vivo correlation LBF Lipid based formulation LC Long chain LFCS Lipid Formulation Classification System LiDo Lecithin-in-dodecane logP Partition coefficient MC Medium chain MG Monoglyceride MMC Mesoporous magnesium carbonate MS/MS Tandem mass spectrometry NaOH Sodium hydroxide PAMPA Parallel Artificial Membrane Permeability Assay

PK Pharmacokinetic pKa Aqueous dissociation constant PLM Polarized light microscopy PVP Polyvinylpyrrolidone PVPA Phospholipid Vesicle-based Permeation Assay PWSD Poorly water-soluble drugs Sapp Apparent solubility SD Standard deviation SEM Scanning electron microscopy SR Supersaturation ratio Tg Glass transition temperature TG Triglyceride TGA Thermal gravimetric analysis Tm Melting point UPLC Ultra-performance liquid chromatography UV Ultraviolet

11

Introduction

The formulation of pharmaceutical compounds as pills dates back to 1550 B.C., and oral drug delivery remains the preferred administration route today. In 2018, 50% of the 59 new drugs registered at U.S. Food and Drug Admin-istration (FDA) were intended for oral use.1 The non-invasive nature and ease of handling of oral formulations typically results in higher patient compliance and lower cost compared to e.g., parenteral injections.1,2 Other advantages in-clude good long-term storage stability and creation of sustained and controlled release.3,4 While the oral administration offers many attractive features, for-mulation design for oral drug delivery systems is far from simple. The milieu in the gastro-intestinal (GI) tract offers many challenges that affect the bioa-vailability of a compound. When the drug is ingested, it will encounter diges-tive enzymes, pH shifts, mucus layers, the GI epithelium and first pass metab-olism.1,3 This hostile environment limits successful oral administration of macromolecules such as proteins and peptides.1,2,5

In order to be absorbed, the drug first has to dissolve in the GI fluids (Figure 1). Parameters that affect the dissolution process are summarized in the Noyes-Whitney’s equation/Nernst–Brunner equation (Equation 1):6,7

= ∗ ∗ ( − )ℎ where dm/dt is the amount of drug dissolved over time (e.g. µg/min), D is the diffusion coefficient (e.g., cm2/min), A is the total surface area of the particles exposed to the solvent (e.g., cm2), Cs is the equilibrium solubility of the com-pound in the given media at a specific temperature (e.g., µg/mL), C is the concentration of the drug in the bulk at a given time (e.g., µg/mL), and h is the thickness of the diffusion layer around the particles (e.g., cm). From the equation, it can be derived that the dissolution rate is proportional to the total surface area and solubility. Alterations in these properties, by e.g., size reduc-tion of the drug particles or addition of co-solvents to the dissolution medium, will directly impact the dissolution rate. Once dissolved, the free molecules have to pass through the GI epithelium.4 Depending on the physicochemical properties of the compound, this occurs through passive diffusion (transcellu-lar or paracellular), and/or via passive or active carrier-mediated transport.8,9

12

Commonly discussed physicochemical properties that affect permeation in-clude the molecular weight, the lipophilicity, number of hydrogen donor/ac-ceptors and the compound’s conformational flexibility, as indicated by the number of rotatable bonds.10–12

Figure 1. The absorption process after ingestion of a tablet. The tablet first has to disintegrate into primary particles (1). The primary particles then dissolve at the mo-lecular level (2), and the free molecules are able to permeate the GI epithelium, reach-ing the blood (3). The drug will be transported to the liver and be subjected to first pass metabolism prior to reaching the systemic circulation.

The Biopharmaceutics Classification System In the Biopharmaceutics Classification System (BCS), the fundamental pa-rameters for the rate and extent of drug absorption are identified as the disso-lution and GI permeability of the drug. Compounds are divided into four clas-ses based on their solubility and permeability: Class I is high solubility and high permeability, Class II is low solubility and high permeability, Class III is high solubility and low permeability, and Class IV is low solubility and low permeability. The framework is applied in the regulatory frameworks of the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) to justify biowaivers for immediate release formulations, and thereby reduce the need for costly in vivo bioequivalence studies.13,14 For compounds to be classified as highly soluble, the drug needs to be soluble in aqueous buff-ers within a wide pH-range (pH 1 to 6.8). This criterion may, however, be too

13

restrictive for e.g., weak acids. Many weak acids are classified as Class II de-spite being completely soluble in the environment in the small intestine. As the absorption there is not dissolution-rate limited, the in vivo drug perfor-mance is more similar to a Class I compound i.e., it is fully absorbed.15,16 The GI tract also contains a number of solubilizing agents, affecting the solubility of lipophilic drugs. The dissolution media used in the in vitro assay therefore need to accurately reflect the in vivo situation.15 Due to these limitations, sub-classes of BCS Class II have been suggested based on the physicochemical properties of the drug.16

Apart from aiding decisions around biowaivers, the BCS classification may serve as a foundation for selecting suitable formulation strategies of drugs. Although strategies to affect the solubility are plentiful, the possibility to af-fect the permeability—and hence increase the bioavailability of Class III and partially Class IV compounds—is limited and the main focus is directed to-wards the use of various permeation enhancers.17

Solubility-limited drugs Computational chemistry and high-throughput screening have changed the landscape of drug development, enabling rapid screening of novel targets. En-dogenous ligands of these targets often are lipophilic, and as a result, the av-erage lipophilicity of novel drug compounds has increased, typically resulting in lowered aqueous solubility.11,18 Between 70 and 90% of all drugs under development are believed to suffer from poor water solubility.17,18 Many for-mulation strategies for dissolution- and/or solubility-limited Class II com-pounds have therefore been developed (Figure 2).17,19–22 Common approaches to increase the dissolution rate include size reduction of the primary particles e.g., micronization and nanocrystals. Other ways to improve the solubility and dissolution of a drug are to alter the crystalline structure of the material by salt formation and co-crystals.17,23 Many drugs also possess polymorphs. Due to their differences in the crystalline lattice and/or molecular conformations, the polymorphs differ in physical properties which can affect their solubility and dissolution pattern.24,25 A vast number of enabling formulations have been de-veloped to increase the bioavailability of poorly water-soluble drugs (PWSD) (Figure 2). 11,23,26

14

Enabling formulations Many enabling formulations make use of the creation of supersaturation. Su-persaturation occurs when the drug is introduced to the medium in a higher energy state (compared to the stable crystalline form), and is achieved by add-ing the drug in e.g., an amorphous form or in a solution (Figure 3). By increas-ing the drug concentration in the lumen, the driving force for absorption will increase, resulting in a higher bioavailability.27 However, supersaturation is an inherently thermodynamically unstable state—and depending on the crystal-lization tendency of the drug—precipitation will occur. The precipitation may be prolonged or completely inhibited by the addition of a stabilizing agent.28

Figure 2: Commonly used strategies to improve the dissolution of poorly water-sol-uble drugs.

15

Figure 3. Dissolution profiles of crystalline and supersaturated systems. Supersatura-tion (green line) is achieved when the drug is presented in a higher energy state than the thermodynamically stable crystalline form (blue line). The supersaturation is, however, thermodynamically unstable, typically transient, and the drug will precipi-tate (green line). The supersaturation may be prolonged or completely stabilized by the addition of a stabilizing agent e.g., a polymer (yellow line).

Common enabling formulations include amorphous solid dispersions (ASDs), carrier-based delivery systems and lipid-based formulations (LBFs) (Figure 2).11,27–30

Amorphous solid dispersions An amorphous material lacks long-range order, and is therefore in a higher energy state than its crystalline counterpart.31,32 This makes the material ther-modynamically unstable, and it will eventually re-crystallize to the stable crystalline form.32,33 To kinetically stabilize the system, the drug can be dis-persed in a carrier matrix e.g., a polymer, creating an ASD.32–35 Addition of the polymer lowers the molecular mobility of the drug, decreasing the free energy of the system, and ultimately leading to increased storage stability.32 The physical stability of the ASD is influenced by the miscibility of the drug in the polymer and intermolecular drug–polymer interactions.36 There are sev-eral examples of ASD formulations that have led to improved treatment re-gimes. One example is vemurafenib, used to treat late-stage melanoma pa-tients. In the end of the Phase I trial, the highest dose regime administered to the patient daily was 1600 mg vemurafenib (equivalent to 32 capsules). The drug was then formulated in a stable ASD, increasing the drug exposure five-fold (Zelboraf). Subsequently, the dosing regime could be reduced to 8 cap-sules per day.37

Carrier-based delivery In 1992, Mobile Oil Company introduced a highly ordered mesoporous silica, which has since been extensively investigated for several applications, includ-ing carrier of drugs.36,38,39 Other mesoporous carriers (e.g., mesoporous alumi-num oxide and mesoporous carbon) have also been evaluated.40,41 Another

16

promising application of mesoporous material is for combined diagnostic and therapeutic, i.e., theranostic, purposes.42

Mesoporous materials are defined as having a pore size of 2-50 nm,36 how-ever, the shape and morphology of the pores in the carrier varies. Once the drug is adsorbed onto the carrier, it is stabilized in a non-crystalline state due to the decrease in Gibbs free energy. Nucleation and crystal growth are further hindered due to spatial constraints.36 This means that factors such as pore size/volume, surface area, and surface chemistry of the carrier affect the fea-sibility of using a given carrier for a specific drug.36,43–45

In 2013, a mesoporous magnesium carbonate (MMC) was introduced.46 The pores form spontaneously during the synthesis, which eliminates the need to remove (either chemically or by heating) a surfactant template commonly used for e.g. synthesis of mesoporous silica.39,47 The MMC displayed a high surface area (~800 m2/g) and a high tortuosity (~15). The pore size can be tuned (3-20 nm) by varying the energy input during synthesis.46,48 As with other mesoporous carriers, MMC are able to stabilize drugs in a non-crystal-line state.49–51

Lipid-based formulations Around 2 to 4 % of marketed drugs are formulated in solubilizing formulation such as LBFs.52 LBFs consists of mixtures of oils/lipids, surfactants (water-soluble/-insoluble), and co-solvents.21,53,54 Based on the ratio of each compo-nent, LBFs are classified according to the Lipid Formulation Classification System (LFCS). The classes range from type I, consisting solely of lipids, to type IV which only contains surfactants and co-solvents. Type II and III con-tain different ratios of all three components.55,56 The drug is pre-dissolved in the LBF, and the liquid or semi-solid formulation is filled into soft or hard capsules.21 Once ingested, the LBF is dispersed in the GI fluids and the lipids (tri-/diglycerides) are digested by gastric and intestinal lipases into mono-glycerides (MGs) and free fatty acids (FFAs) (Figure 4). The lipolytic com-ponents are mixed with e.g., endogenous bile salts/phospholipids/cholesterol, forming vesicular and micellar structures. In addition to the free dissolved drug in the intestinal fluid, the drug is solubilized in the colloidal structures. The total dissolved content in the intestinal lumen is increased and thereby enable increased fraction absorbed of PWSD.21,54,57

17

Figure 4. Schematic of in vivo digestion of triglycerides in the GI tract. The ester bond in the triglyceride is hydrolyzed, forming a diglyceride and a free fatty acid. The di-glyceride is then further digested into a monoglyderide and a second free fatty acid. The final digestion products, i.e., one monoglycide and two free fatty acids, are incor-porated into mixed colloidal structures together with endogenous compounds in the GI tract.58,59 Complete digestion may also take place, whereby the monoglyceride is further digested into glycerol and a free fatty acid.59

Solid lipid-based formulations Liquid LBFs, however, are associated with comparatively high manufacturing costs and physicochemical instability of the final dosage form. To address these concerns, attempts have been made to create solid formulations incor-porating the liquid LBFs.60–63 A common strategy is to use chemically inert carriers upon which the LBF is adsorbed.64 These are typically referred to as solid LBFs; however, it should be noted that it is not the LBF itself that is solidified but an integrated part of a solid system, due to the properties of the carrier. The formation of a solid system eases handling of the formulations and the powder can be filled into capsules or compressed into tablets.62,65,66 Different techniques can be used to adsorb the lipid, including spray drying, freeze drying, and rotary evaporation.61,62 With mesoporous carriers, adsorp-tion can also be achieved by simply mixing the carrier and the LBF.62,64 Prom-ising results have been seen for silica and silicate-based solidifying carriers. Other materials, such as mesoporous carbon, have also been investigated.61,62

Selection of appropriate enabling formulations The benefits of incorporating a drug into a specific enabling formulation de-pend on the physicochemical properties of the compound, the desired dose, and the route of administration.67–69 Lipophilic drugs are good candidates for LBFs, but the melting point (Tm) of the drug needs to be taken into consider-ation.21,70 A high Tm typically means low lipid solubility, making the com-pound less suitable for LBFs.21,71 Molecular properties such as polar surface area, number of nitrogen atoms and size/shape, are also useful predictors of the solubility of the drug in lipid excipients.70 Good candidates for LBFs type I-IIIA (i.e., lipid-rich LBFs) are neutral or basic drugs that have few polar groups and a Tm <150°C. For drugs with a higher Tm, LBFs type IIIB-IV with higher ratios of co-solvent and surfactants are more suitable.72

18

The properties of the crystalline solid state of the drug also impact its suit-ability for incorporation as an amorphous formulation. Commonly investi-gated properties for drugs in ASDs are the Tm, the glass transition temperature (Tg), and the ratio between the two. The Tm is a strong indicator of the ther-modynamic driving force for crystallization, where a higher Tm is more likely to cause crystallization. The Tg relates to the kinetic barrier for molecular dif-fusion in the amorphous material. A good candidate for an ASD would hence have a comparatively low Tm and a high Tg.37 Apart from being used to eval-uate the glass-forming ability of the compound, the Tm/Tg ratio is also used to select drug-loading of the ASD.73

For mesoporous drug delivery systems, the success of drug-loading de-pends not only on the ability of the drug to form an amorphous system but also on the inherent properties of the carrier. 36,43,44 Apart from properties re-lating to the porosity of a carrier, the surface chemistry of the material is also critical for drug adsorption onto carriers. By modifying its surface chemistry, the carrier can be tailored to improve drug-load and stability of the system.43,45 Surface functionalization has been explored for a wide range of materials, in-cluding mesoporous silica and MMC.39,74,75 Furthermore, to evade negative storage effects on solid-LBFs, pre-coating a silicate carrier with PVP can be used.76

In vitro methods to evaluate the performance of enabling formulations The inherent properties of the enabling formulations create different demands on assays that test their in vitro performance. Typically, the dissolution and release from amorphous formulations are studied in relatively simple assays in which the formulations are dispersed in a relevant aqueous medium.77,78 More complex methods incorporating digestive enzymes are usually used for LBFs.57,79 Commonly employed methods to study enabling formulations are briefly described below (Figure 5).

19

Figure 5. Illustration of commonly used in vitro methods to evaluate performance of different enabling formulations. Various types of dissolution baths, with or without in situ probes, are used for solid formulations such as ASDs whereas in vitro lipolysis is typically used to evaluate the performance of LBFs. In the dissolution baths and the in vitro lipolysis, samples are manually collected and immediately centrifuged. The phases are separated, diluted, and analyzed using e.g., HPLC-UV.

Dissolution assays Similar to what is used for traditional formulations such as tablets, common assays to study the performance of amorphous formulations (e.g. ASDs and mesoporous drug carriers) include USP Type II dissolution baths.77 Samples are collected at predetermined time points, and are centrifuged or filtered prior to analysis. The drug concentrations are then measured using chromatographic methods, e.g., high-performance liquid chromatography with a ultraviolet-spectrophotometer (HPLC-UV)80,81. Optionally, in situ techniques, e.g., UV probes, can be used to continuously measure the concentration in the vessel over time.82

In vitro lipolysis In vitro lipolysis of LBFs is commonly performed using a pH stat appa-ratus.11,21 In short, the LBF is dispersed in the lipolysis medium and pancreatic extract is added. The triglycerides (TGs) are digested, and as the FFAs form, the pH decreases. Sodium hydroxide (NaOH) is titrated to counteract the pH shift. Based on the amount NaOH added, the amount of FFA formed and the extent of digestion are calculated. To determine the drug release, samples are collected at predetermined time points during the digestion and centrifuged,

20

and the phases (oil, aqueous and solid material precipitated as pellet) are sep-arated. The amount of drug in each phase is determined using e.g., HPLC-UV.79

Permeation models The experimental setup and complexity of permeation assays depends greatly on the purpose of the study.83 Many different cell-free, cell-based, and tissue-based models have been developed, but Caco-2 cell monolayers remain the gold standard for drug permeation studies (Figure 6).84–86 The Caco-2 cell line is in vivo relevant, but associated with some drawbacks including lengthy cul-tivation times (21 days) and large inter- and intra-laboratory variability in transporter expression.83,84 Assays using artificial membranes have therefore been developed including the Parallel Artificial Membrane Permeability As-say (PAMPA), the Phospholipid Vesicle-based Permeation Assay (PVPA), and Permeapad.84,87,88 Although the experimental setup can vary, e.g., number of wells, plate size and membrane composition, some key elements are similar for the permeation assays. Setups such as Transwells or similar are commonly used, in which the apical compartment is separated from the basolateral one by a semipermeable membrane (Figure 6). The drug solution is added to the apical compartment and sampled from both the apical and basolateral sides of the membrane at predetermined time points.89 Similar methods are also used to study enabling formulations including ASDs and LBFs.90–93

Figure 6. Illustration of a typical permeation assay. A semipermeable membrane sep-arates the apical compartment (corresponding to the lumen) from the basolateral com-partment (corresponding to the blood). Depending on the purpose of the assay, the membrane can be an artificial one or a cell monolayer e.g., Caco-2 cells.

Combined dissolution-permeation models The standard dissolution/digestion methods described above may provide an understanding of processes occurring in the GI lumen, but in vivo absorption is a complex interplay between the solubility and permeation.94,95 Assays in-corporating both phenomena are therefore of value.96,97 This is of particular relevance for enabling formulations. The currently available combined disso-lution-permeation models vary in size and complexity (Figure 7). Assays

21

range from large-scale systems, e.g., the biphasic dissolution assay98 and the MacroFLUX system (Pion Inc, Billerica, USA)99, to smaller ones such as the BioFLUX (Pion Inc, Billerica, USA)100, and side-by-side diffusion chambers e.g., µFLUX (Pion Inc, Billerica, USA)101 and Side-bi-Side Cells (PermeGear Inc,, Hellertown, USA).102 In biphasic dissolutions, drug absorption is deter-mined as a result of the partitioning into the organic phase, e.g., octanol or decanol.103–105 In the Bio-/MacroFLUX and side-by-side diffusion cells, eval-uation of the absorption is more similar to the traditional Transwell experi-ments, i.e., the amount of drug permeation across the membrane into the re-ceiver chamber is determined.99,100,102 In the Macro-/Bio-/µFLUX, artificial absorption membranes are used.99,100,106 In the Side-bi-Side Cells, studies are typically performed with artificial membranes,102,107,108 but experimental set-ups incorporating cell-based absorption membranes have also been ex-plored.109

Figure 7. Examples of different combined dissolution and permeation models. Each model has its advantages and drawbacks. Between 500 and 1000 mL of dissolution media are used in the bi-phasic dissolution assays and the MacroFLUX, whereas much lower volumes (e.g., 6-7 ml) are used in the side-by-side diffusion cells.

Limitations of current methods All in vitro models have advantages and drawbacks. The compendia assays, e.g., USP type II-dissolution baths, require large amounts of formulation and dissolution media, which may limit its use in early drug development. 110 Man-ual sampling also makes it difficult to capture time-sensitive events such as supersaturation and precipitation. In situ UV probes will circumvent this, but the probes are sensitive to conditions that may interfere with the UV- meas-urement such as turbid or particle-rich systems.111 Although these simple mod-els may be of great value to in distinguishing differences in dissolution be-tween e.g., powder formulations, the in vivo relevance of the results may be

22

questioned as the assays do not incorporate any absorption.94 To increase the in vivo relevance when studying LBFs, digestive pancreatic enzymes are used in the in vitro lipolysis. Despite that, there is poor in vitro in vivo correlation (IVIVC) for in vitro lipolysis.112 Current permeation assays also suffer from some limitations. The static permeation models, e.g., Transwells, are efficient for studying drug permea-bility from solutions/stable suspensions, but sufficient stirring is necessary as the thickness of the unstirred water layer affects the permeability.84 The ex-perimental setup is however ill-suited for studies of solid powder because the possibility of adequate stirring in the donor compartment is limited. Further-more, lipid digestion in the static systems may be detrimental to e.g., cell membranes, limiting its use for studying drug release from LBFs.91 Regardless if dissolution or permeation is being studied, few assays are designed for dif-ferent enabling formulations in the same experimental setup.113 This remains true for the combined dissolution–permeation models in which the assays fo-cus on one specific enabling formulation, making head-to-head comparisons between different formulation types difficult.

Consequently, there is a need to develop new in vitro models for better prediction of the in vivo performance of enabling formulations. Ideally, the models should be designed for evaluation and comparison of different ena-bling formulations in the same experimental setup. A deeper understanding of the inherent properties affecting the performance of different enabling formu-lations is warranted. Further attention is also needed for combinations of for-mulation strategies to increase the absorption of PWSD. The work included in this thesis aimed to expand the current toolbox of in vitro assays, ultimately allowing head-to-head comparison of different enabling formulations.

23

Aims of the thesis

The overall aim of this thesis was to develop methods to study different ena-bling formulation in a resource- and time-efficient manner. In the first part of the thesis work (Papers I-II), emphasis was placed on studies of the perfor-mance of a drug-/lipid-loaded mesoporous carrier. Method development was focused on solving the challenges due to the inherent properties of the carrier. The scope of the thesis then expanded to include a wide range of enabling formulations (Papers III-IV). In Paper III, a small-scale and relatively fast screening assay was developed. Paper IV then investigated the most promising formulations found in paper III in a more complex in vitro model for which the data were correlated with the in vivo performance in rats. Specific aims of the thesis were to:

• Develop an in vitro method to evaluate the degree of supersaturation ob-

tainable with a drug-loaded mesoporous carrier. The feasibility of stabi-lizing the supersaturation with a polymer was also studied (Paper I).

• Design and produce a solid LBF using a mesoporous carrier for the solid-ification (Paper II).

• Establish an in vitro method to study the release and digestion of lipids from the solid LBF (Paper II).

• Use different, physiologically relevant dissolution–digestion–permeation models to study the performance of various enabling formulations, includ-ing the drug-loaded carrier and solid LBF (Paper III and IV).

• Evaluate the in vivo performance of the enabling formulations studied in the in vitro dissolution–digestion–permeation models, and to determine a potential in vitro-in vivo correlation (Paper IV).

24

Methods

Selection of model compounds Compounds with poor water solubility were studied in all four papers. In Pa-per I, the compounds (fenofibrate, hydrocortisone, ketoconazole, tamoxifen, and tolfenamic acid) were selected on the basis of their physicochemical prop-erties, e.g., charge and polarity. A wide range of compounds was selected to understand the feasibility of their incorporation into a mesoporous carrier and thereby increase the amount dissolved. In Papers III-IV, felodipine was se-lected as the model compound. Felodipine is a highly lipophilic, neutral drug (logP = 3.6).114 It is classified as a stable glass former (Class III according to the classification system established by Taylor et al.115), having a Tm of 147°C and a Tg of 45°C (Tm/Tg=1.32).116 Felodipine has very low solubility in aque-ous media (around 1 µg/mL)117, and would therefore benefit substantially from incorporation into several types of enabling formulations.

Selection of enabling formulations Depending on the physicochemical properties and the physical form of a com-pound, the benefits of a particular enabling formulation differ.19,68 There is therefore a need to increase the understanding of the potential of different types of drug delivery strategies, and to expand the assortment of formulations available. The different enabling formulation strategies used in this thesis are presented in Figure 8.

25

Figure 8. Schematic of the different enabling formulations in this thesis. Porous car-rier: Poorly water-soluble drugs adsorbed in a non-crystalline state onto a mesoporous magnesium carbonate (MMC) (Papers I and III). Adsorbed lipids: A single lipid and an LBF (Paper II), and drug-loaded LBFs (Paper III) adsorbed onto MMC. Liquid lipids: A single lipid (Paper II) and drug-loaded LBFs (Papers III and IV). ASD: Amorphous solid dispersions with varying drug:polymer ratios. Combined formula-tions: An ASD used in combination with a blank LBF.

Preparation of enabling formulations Micronized drug A planetary ball mill (Model PM 100 Retsch, Germany) was used to micronize the crystalline drug in Paper I. By reducing the particle size, the particles be-come more homogenous in size and the dissolution rate increases. The crys-talline drug was added to the milling bowl together with ten small steel beads (Ø 5 mm). The material was then milled at 600 rpm for 20 min.

Drug-loading of MMC In Papers I and III, the release from drug-loaded MMC with a pore size of ~5 nm was investigated. The drugs were adsorbed onto the MMC using solvent impregnation in which the crystalline drug is dissolved in ethanol, and the MMC added to the solution.49 The suspension was placed on a magnetic stirrer (500 rpm) for 24 h and then placed in an oven (70°C) for 24 h to evaporate the ethanol. A 1:10 w/w ratio drug:MMC was selected in Paper I, enabling studies of drug-release in particle-rich systems. In paper III, a higher drug-load (25% w/w) was used.

Lipid-based formulations The LBFs were prepared as previously described.91,114,118 In short, the excipi-ents were weighed into glass vials in predefined fractions (% w/w). The vials were vortexed and placed on a shaker for 24 h prior to use (37°C). For Papers II and IV, a type IIIB medium chain (IIIB-MC) LBF was selected. In Paper

26

III, two additional types of LBFs were also studied: one type IIIB long chain (IIIB-LC) and one type IV. The felodipine was incorporated into the LBFs by adding the required amounts of the drug to the LBF, and placing the vials on a shaker for 24 hours (37°C).119

Adsorption of lipids onto MMC Lipid/LBF-loaded MMC with a larger pore size of ~20 nm were investigated in Papers II and III. Lipids were added to the MMC via pipetting (1:1 w/w ratio), and the powder was stirred until all lipid had been adsorbed, giving a white free-flowing powder.64 In Paper II, the lipids were also adsorbed onto the MMC using solvent immersion and evaporation (similar to the procedures used for drug-loading), and by freeze drying.120,121

Amorphous solid dispersions The ASDs were produced in a Büchi Mini Spray Dryer B-290 (Switzerland). In Paper III, polyvinylpyrrolidone (PVP) was used to stabilize the amorphous drug (ASDPVP) and in Paper IV, hydroxypropyl methylcellulose acetate suc-cinate (HPMCAS) was selected (ASDHPMCAS). The drug was dissolved in the solvent, and the following spray drying conditions were used: inlet tempera-ture (55°C), aspiration rate (75%), and pump rate (4 mL/min).122 To remove any residual solvent, the ASDs was dried overnight. The ASDs were then stored in a vacuum desiccator containing silica beads until analysis.

Solid state characterization techniques Differential scanning calorimetry Conventional DSC (Q2000 Differential Scanning Calorimeter; TA Instrument Co.,USA) was used to determine the Tm of the crystalline compounds in Pa-pers I, III and IV. DSC was also used to ensure absence of crystalline material directly after manufacturing of the ASDHPMCAS in Paper IV, and to study the storage stability. The drug/ASD was added to an aluminum pan which was then sealed with an aluminum lid with pinholes. The samples were equili-brated at 0°C, then heated at 10°C/min to 30-50°C above the expected Tm. The samples were purged with nitrogen at 50 mL/min. To increase the sensitivity and to confirm the absence of crystallinity, modulated DSCs (mDSCs) were used for the drug-loaded MMC and the ASDPVP (Papers I and III). The formu-lations were weighed into the aluminum pans and the samples equilibrated at 0°C. The samples were heated to 20–57°C above the Tm of the drug at 1°C/min with a modulation of ± 0.5 °C every 60 s.122

27

Powder X-ray diffraction To further confirm the absence of crystalline material in the ASDHPMCAS, pow-der X-ray diffraction (PXRD) was used in Paper IV. A Bruker D8 Advanced XRD TWIN-TWIN instrument (Bruker, Bremen, Germany) with a Bragg-Brentano setup was used to collect the PXRD diffractograms. Silicon zero background holders were used. The samples were exposed to CuKα radiation (λ=0.154 Å; 45 kV and 40 mA), and the diffraction pattern was collected for 2θ at 10° to 80°. A step size of 0.02° and a measuring time of 2 s per step were used.48

Scanning electron microscopy The surface morphology of the particles in Papers II (lipid-loaded MMC) and IV (ASDHPMCAS) was studied using a high-resolution scanning electron mi-croscopy (SEM, Zeiss Merlin, Oberkochen, Germany).123,124 The samples were mounted on aluminum stubs with double-sided adhesive tape. To avoid charging effects, the samples were sputter-coated with a uniform layer of gold/palladium. An accelerating voltage of 5-8kV was used for imaging.

Thermal gravimetric analysis The thermal behaviors of MMC and lipid-loaded MMC were evaluated by thermal gravimetric analysis (TGA; Mettler Toledo TGA/SDTA8511e bal-ance, Greifensee, Switzerland). The samples were added to aluminum cruci-bles and pretreated at 200°C under an air flow of 40 mL/min for 10 min to remove adsorbed water, and cooled to room temperature. They were then heated at a rate of 10°C to 800°C (air flow of 40 mL/min). The lipid content in the lipid-loaded MMC was calculated by comparing the weight loss ob-served in the TGA profiles of MMC and lipid-MMC (Paper II).50

Nitrogen sorption analysis The average pore size, pore volume, and surface area of the MMC were eval-uated using nitrogen sorption analysis (Micromeritics ASAP 2020 surface area analyzer, Micromeritics, Norcross, GA, USA).47,48 The nitrogen adsorp-tion/desorption isotherm of the samples was recorded at -195.15°C. The spe-cific surface area was determined using the Brunauer-Emmett-Teller (BET) equation and sample pore size was determined using the Density Functional Theory (DFT, N2 slit pore model48) with the MicroActive software. The N2 equilibrium uptake at a relative pressure (p/p0) of 0.98 was used to calculate the total pore volume.

28

Fourier-transform infrared spectroscopy Fourier-transform infrared spectroscopy (FTIR) was used to study interactions between lipid and MMC in Paper II. Each compound displays a unique spec-trum based on the chemical bonds within the molecules. Changes in the spec-trum (e.g., the appearance of a new covalent bond or a peak shift/deformation) indicate a strong interaction between the lipid an the carrier.123 The spectra were acquired with a Bruker ALPHA II coupled with a single-reflection, dia-mond-attenuated total reflection module (Platinum ATR, Billerica, 146 MA, USA). Background scans (n=24) were collected prior to each measurement, and 24 scans were recorded for each spectrum (400-4000 cm-1), at a resolution of 4 cm-1.

Polarized light microscopy Polarized light microscopy (PLM) was used in Papers III and IV to study drug precipitation during dissolution/digestion. PLM is useful for studying crystal-line structures, e.g., polymorphs and solvates, by relying on their birefrin-gence.125 Crystalline particles (precipitate/undissolved drug) appear as glow-ing, bright spots, whereas the absences of ordered structures in amorphous precipitate results in black images. The images were recorded with an Olym-pus BX51 microscope (Olympus, Japan) equipped with cross-polarizing fil-ters.72 Crystalline felodipine, the polymers, and the ASD were examined as references.

In vitro release and digestion assays The µDiss Profiler A small-scale dissolution apparatus, the µDiss Profiler (Figure 9), was used to determine apparent solubility and release from drug-loaded MMC. The assay is fairly simple, as no manual sampling is required. The concentrations in the dissolution vials are monitored using in situ UV-absorbance probes117, allow-ing measurement of the concentration up to every two seconds. This makes the µDiss particularly useful for studies of supersaturation and precipitation events.122,126 In Paper I, a fixed amount of sample was weighed into the vials, and the measurements started upon addition of 15 mL fasted state simulated intestinal fluid version 1 (FaSSIF-V1). UV-spectra were then collected con-tinuously. Analysis was performed using the second derivative of the spectra and a single wavelength.127,128 However, in situ UV detection is sensitive to interferences caused by e.g., high amounts of particle in the medium, resulting in extensive light scattering. To circumvent this issue, a protective filter of net nylon was developed in Paper I. The nylon filters were placed on the in situ UV probes to protect them from the accumulation of particles.

29

Figure 9. Schematic illustration of the µDiss Profiler. The system is equipped with six in situ UV probes for real-time monitoring of drug dissolution. The temperature in the vials (10-20 mL dissolution media) is kept constant using a heating block. The vial holder also allows magnetic stirring in each vial.117,129

In vitro lipolysis and 1H NMR A modified in vitro lipolysis was developed to study the release of lipid from lipid-loaded MMC in Paper II. As described in the introduction, release and digestion is usually evaluated using titration with NaOH.21,79 The pH stat ap-paratus is also used to study release from solid LBFs that contain e.g., silica and smectite clays.120,121,123,130 However, the standard lipolysis assay was not suitable for studies of lipid-loaded MMC due to the pH increase caused by the dissolution of the carrier itself. Titration with hydrochloric acid (HCl) was therefore made during the digestion of the lipid-loaded MMC to keep a con-stant pH of 6.5, and thereby maintain the enzyme activity of the pancreatic extract. Samples were collected at predetermined time points during the mod-ified in vitro lipolysis and the formation of FFA and other digestion products were evaluated using proton nuclear magnetic resonance spectroscopy (1H NMR).123,131 The lipolytic components in the sample were extracted with di-chloromethane, and the organic phase was collected. The solvent was evapo-rated under vacuum and the lipid film was redissolved in deuterated chloro-form. The samples were transferred to an NMR-tube, and a 1H NMR spectrum was acquired for each sample. Based on the spectra, the relative molar per-centage of each lipolytic component were calculated.59,131,132

Combined digestion-permeation assays

Cell culture Caco-2 cell monolayers were used in the two assays presented below. The cells (American Type Culture Collection, VA, USA) were cultivated at 37°C in an atmosphere of 10% CO2 and 90% air, as described by Hubatsch et al.89

30

The cells, passages 95-105, were seeded at a density of 170,000 cells/cm2 on polycarbonate filter supports (0.45 μm pore size, 75 mm diameter; Transwell Costar, Sigma-Aldrich).119 The growth medium consisted of Dulbecco’s mod-ified Eagle’s medium supplemented with 10% fetal calf serum, 1% nonessen-tial amino acids, penicillin (100 U/mL), and streptomycin (100 μg/mL). The monolayers were used 21-26 days after seeding.

The µFLUX system The µDiss Profiler system was used in combination with the µFLUX system, to simultaneously evaluate dissolution and permeation of the drug compound. The µFLUX consists of two side-by-side horizontal chambers which are sep-arated by a semipermeable membrane (Figure 10). An in situ UV-probe from the µDiss Profiler was placed in each chamber to continuously determine the concentration.101,133,134 Typically, artificial membranes are used in the µFLUX101,106, but in Paper III, a Caco-2 cell monolayer was used as the ab-sorption membrane to increase the in vivo relevance of the assay.

Different enabling formulations (LBFs, solid-LBFs, an ASD, and a drug-loaded carrier) were evaluated, which all posed different challenges to the sys-tem. The protective nylon filters developed in Paper I were used to minimize particle interference with the UV-measurements from the drug-/LBF-carriers. Furthermore, as LBFs are digested in the GI-tract, digestive enzymes had to be incorporated into the assay. An immobilized lipase, Novozyme 435 (lipase B from Candida antarctica adsorbed on macroporous acrylic polymer resin135), was used, as it is compatible with Caco-2 cells.91 The immobilized lipase is also more resilient to pH changes than pancreatic extract, thereby maintaining the enzyme activity over a broad pH range (Strem Chemicals Inc.,136).

During the assay, the pH changes as a result of the lipid digestion. Because the µFLUX-system does not allow for titration, a buffer with high buffering capacity was required in the donor chamber. The more commonly used tris-maleate buffer was hence replaced with a buffer containing Bis-tris 137. The buffer was supplemented with 3mM taurocholate and 0.75 mM phospholipid (FaSSIF-V1 powder). HBSS (pH 7.4) was used in the receiver compartment and supplemented with 0.2% d-α-tocopheryl polyethylene glycol 1000 suc-cinate (TPGS), to ensure sink conditions.

The formulations were weighed into the donor chamber (equivalent to 1.75 mg/mL drug) and the in situ measurements started upon addition of 15 mL dissolution/receiver media. After a 10-min dispersion phase, the immobilized lipase was added. The UV spectra were analyzed using the second derivative of the spectrum and a narrow wavelength range. As some interference with the UV-measurements was observed for the donor chamber, despite the pres-ence of the protective nylon filters, samples were also collected from the donor compartment and then analyzed with HPLC-UV. The membrane integrity of

31

cell monolayer was evaluated using a hydrophilic permeation marker, Lucifer yellow .138 Lucifer yellow was added to the donor chamber, and samples were continuously collected from the receiver chamber. The Lucifer yellow con-centrations were then determined in a UV-plate reader (Tecan, Austria).119

Figure 10. The µFLUX setup. The two vials are separated by a semi-permeable mem-brane. A Caco-2 cell monolayer was used as absorption membrane. The formulations were added to the donor chambers, and the digestion was performed with an immobi-lized lipase. To counteract the pH shift caused by the digestion of lipids, a buffer with high buffering capacity was used in the donor compartment. To reduce particle inter-ference with the UV reading, a protective nylon filter was placed on the probe. Mem-brane integrity was monitored with Lucifer yellow.

Enabling absorption devices The digestion–release and permeation of drugs from enabling formulations was also studied in a larger scale system, the enabling absorption device (ENA).118,119 It consists of vertical donor- and receiver-chambers, separated by a semipermeable membrane (Figure 11). In the ENA device, the membrane area-to-volume ratio (A/V) is more similar to the in vivo situation than in other release-permeation devices, e.g., the µFLUX.119 Permeation across both a Caco-2 monolayer and an artificial semipermeable membrane (lecithin-in-do-decane, LiDo) were evaluated in Paper IV. The experimental setup varied de-pending on the absorption membrane used.

When the Caco-2 cell membrane was used in the ENA, the assay was per-formed similar to the traditional in vitro lipolysis. The donor medium was a tris-maleate buffer supplemented with 3mM taurocholate and 0.2 mM phos-pholipid (FaSSIF-V2). HBSS (pH 7.4) supplemented with 4% w/w bovine se-rum albumin (BSA) was used in the receiver chamber. Due to the pH differ-ence between the two chambers, pH increase in the donor compartment can be used as a marker for membrane integrity. To counteract the pH decrease

32

caused by the digestion of the lipids the medium in the donor compartment was titrated with 0.6M NaOH.

In the ENA setup with the artificial LiDo membrane, high-buffering Bis-tris buffer was used in the donor compartment because the LiDo affects the pH electrode. Lucifer yellow was added to the donor compartment, to monitor membrane integrity by measuring of the amount permeated to the receiver. As for the ENA with the cell setup, HBSS supplemented with BSA was used in the receiver chamber with the LiDo setup.

Regardless of the membrane used in the ENA, an immobilized lipase was used for digestion. Samples were withdrawn at predefined time points from the donor and receiver compartments. The samples were centrifuged to sepa-rate the different phases. The BSA in the receiver samples was precipitated with acetonitrile. Donor samples were analyzed with HPLC-UV and receiver samples with ultra-performance liquid chromatography-tandem mass spec-trometry (UPLC-MS/MS).

Figure 11. The enabling absorption (ENA) device. The two vertical chambers are sep-arated by a semipermeable membrane (artificial or cell-based) and both the donor (representing the GI-lumen) and receiver chambers (representing the blood) are sam-pled. This makes it possible to determine both the degree of supersaturation achieved e.g., by the enabling formulations in the donor compartment and the simultaneous flux of drug across the membrane.

33

In vivo pharmacokinetic study In Paper IV the in vivo performance of an ASD, an LBF, and a combination of the two were evaluated in rats. The pharmacokinetic (PK) study was carried out by XenoGesis (Nottingham, UK) in their animal research facility, Saretius Ltd. (Reading, UK). The in vivo protocol was reviewed and approved by the Reading University Animal Welfare and Ethical Review Body (AWERB) (UK home office license: P513DA7FD 19b 2). Male Sprague Dawley rats (256-360 g on the day of dosing) were divided into five groups. One received intravenous (I.V.) dosing (n=4) and four received different combinations of oral formulations (n=6 per group, Figure 12). Food was removed 16 hours prior and 4 hours after dosing of the rats with oral formulations. Water was available ad libitum. Gelatin capsules were prepared with the solid formula-tions (grinded crystalline drug and an ASD) and the drug-loaded LBF. The drug-loaded LBF and the ASD were administered by an oral gavage, followed by administration of 1 mL sterile water. Additionally, the ASD-capsules were administered with blank LBF dispersed in water. The capsules containing crystalline drug were also administered with the LBF dispersion. Plasma sam-ples were collected at predefined time points and bioanalysis was performed by UPLC-MS/MS (Thermo TSQ Quantiva with a Thermo Vanquish UPLC system, Waltham, MA, USA) at XenoGesis.

Figure 12. Overview of the in vivo pharmacokinetic study. Male Sprague Dawley (n=6 per group; four groups) received gelatin capsules containing grinded crystalline drug (Cryst drug), a drug-loaded LBF, or an ASD. The drug-loaded LBF was admin-istered with water whereas the crystalline drug was administered in combination with a dispersion of blank LBF in sterile water. The ASD capsules were administered both with water and together with the LBF dispersion.

34

Results and discussion

In situ measurements in particle rich systems (Paper I) In Paper I, we aimed to develop a method to determine release from a drug carrier, MMC, making use of the µDiss Profiler (Figure 9). To our knowledge, all previous release studies of drug-loaded MMC had been performed in Type II dissolution baths, which requires large amounts of formulation and dissolu-tion media.49,51,139 A first attempt to study the supersaturation created from the drug-loaded MMC in the µDiss Profiler revealed that large amounts of parti-cles in the assay interfered with the UV readouts.50 Paper I therefore focused on modifying the existing experimental setup. To increase the complexity of the system being evaluated for supersaturation degree and stability, a polymer, hydroxypropyl methylcellulose (HPMC; Methocel E4M Premium CR) was added to the dissolution media. Five drugs were selected to cover a wide range of different physicochemical properties (Table 1), and loaded onto the MMC.

Table 1. Physicochemical properties of drugs loaded onto the mesoporous magne-sium carbonate. Charge pKaa Tm (°C)a logDpH6.5b Sapp FaSSIFc

Fenofibrate neutral n/a 80.5 ± 0.02 5.2 16.4 ± 0.3 Hydrocortisone neutral n/a 221.4 ± 0.35 1.4 454.3 ± 32.1 Ketoconazole base 6.2;4.2 147.5 ± 0.22 3.5 30.0 ± 0.8 Tamoxifen base 8.5 97.1 ± 0.21 4.6 152.9 ± 10.8

(62.2 ± 2.3)d (32.1 ± 0.7)e

Tolfenamic acid acid 3.7 212.2 ± 0.20 2.4 66.7 ± 0.8 (317.6 ± 4.3)d

aMelting point (Tm) experimentally determined with DSC, bDistribution coefficient at pH 6.5 (logDpH6.5) predicted using ADMET predictor, cApparent solubility (Sapp) in FaSSIF V1 pH 6.5 experimentally determined in the µDiss Profiler, dSapp in FaSSIF V1 pH 7.5, eSapp in FaSSIF V1 pH 8.5, n/a: not applicable.

35

Design of protective filters

We hypothesized that particle interference with the UV-measurements could be reduced by placing a protective filter on the in situ probes. Many materials and dimensions were evaluated for their ability to min-imize the particle interference without compromising the concentration measurements, e.g., by generation of a lag phase. A hydrophilic net ny-lon filter (Merck Millipore) with a pore size of 11 µm was selected, and sheets were cut and welded, forming “tubes”. The protective filters were placed on the in situ UV-probes, significantly reducing the UV-interference (Figure 13). The presence of filters enabled release-meas-urements in systems with MMC-concentrations of up to 8 mg/mL.

Release studies from drug-loaded carrier As expected, the adsorption of the drugs onto the MMC increased the disso-lution for all compounds, compared to the crystalline drug. This was attributed to the non-crystalline state of the drug arising from the spatial constraints within the pores. In Figure 14, the degree of superstaturation obtained by the drug-loaded carriers is shown as the supersaturation ratio (SR)(i.e., the con-centration determined at a specific time point divided by the apparent solubil-ity, Sapp, of the drug).

The pH in the dissolution media increased from pH 6.5 to pH 8.3–9.4 dur-ing the assays involving the MMC, due to partial dissolution of the carrier. For two of the compounds—hydrocortisone (a neutral compound) and keto-conazole (which based on its physicochemical properties performs similar to a neutral compound in the concerned pH range)—this pH drift would not af-fect the solubility significantly. For both drugs, supersaturation was achieved

Figure 13. Impact on dissolution-/release profiles after addition of filters on the in situUV-probes. Significant particle interference was seen during the dissolution assaywhen 2 mg/mL MMC was used in combination with tolfenamic acid (TOL). Whenthe protective filter was added to probes, concentrations of drug-loaded MMC of upto 8 mg/mL could be studied. HYDRO: Hydrocortisone.

36

(Figure 14a, b) and was stabilized by the polymer present in the dissolution media. The supersaturation could be maintained for hydrocortisone through-out the assay, whereas it was a transient event for ketoconazole. This was also reflected in the area under the curve (AUC) of the dissolution profiles. These differences in precipitation kinetics were accredited to the amount of drug-loaded MMC used in relation to the apparent solubility of the compound. For systems with a higher degree of supersaturation, i.e. a higher free concentra-tion of the drug, the probability of nucleation and subsequently precipitation is higher.122 Due to the large difference in Sapp of the compounds (Table 1), drug-loaded MMC equivalent to 1.6 x Sapp was added for hydrocortisone and 12 × Sapp for ketoconazole. Although precipitation was not completely inhib-ited by the polymer, it was slower for ketoconazole in the presence of the HPMC. This suggests that the HPMC mainly prevented crystal growth rather than nucleation.27,50,140

In contrast, the performance of the two ionizable drugs, tamoxifen and tolf-enamic acid, was strongly influenced by the pH shift. An increased pH reduces the solubility of basic compounds, and indeed the solubility of tamoxifen was ~5-fold lower at pH 8.5 than pH 6.5 (Table 1). It could therefore be assumed that the transient high concentrations achieved during approximately 50 min were more likely a pH effect, rather than a true supersaturation (Figure 14c). For tolfenamic acid, an increased pH is beneficial and increases the apparent solubility. All runs (regardless if the drug was in a crystalline state or loaded onto the MMC) resulted in final concentrations similar to the equilibrium sol-ubility at pH 7.5 (Table 1, Figure 14d). The concentrations obtained were also in line with the total amount of drug added to the run (360 µg/ml). Thus, no supersaturation was achieved for tolfenamic acid (Figure 14d).

37

Figure 14. Supersaturation ratio (SR; i.e., bulk concentrations divided by the apparent solubility, Sapp) of drug-loaded MMC with and without HPMC. For comparison, measurements were also made for crystalline drug combined with pure MMC. For hydrocortisone and ketoconazole, the calculation of the SR was done using Sapp in FaSSIF pH 6.5. Due to the pH increase caused by the dissolution of MMC, the SR values for tamoxifen and tolfenamic acid were calculated using the Sapp in FaSSIF pH 8.5 and 7.5, respectively.

Detection of chemical degradation A change in the UV spectra of the fenofibrate-loaded MMC was detected dur-ing the release experiments. It was therefore not possible to obtain release profiles in the µDiss despite the protective filters. This was suspected to be caused by degradation of the prodrug within the MMC. Hence, UPLC-MS/MS was used to analyze a solution of fenofibrate released from the MMC. The mass observed was consistent with that of fenofibric acid, the degradation product of fenofibrate, potentially formed by basic ester hydrolysis within the pores of the MMC.141

Limitations with protective filters The protective nylon filters allowed study of the supersaturation and precipi-tation events arising from a carrier-based drug delivery system. The pore size (11 µm) was suitable for the MMC, but it may be a limitation for systems with smaller particles, e.g., nanosuspensions. As only highly concentrated, super-saturated systems were evaluated, issues relating to adsorption of drugs onto the nylon were not taken into account. These issues would need to be further studied if the filters are to be used for diluted systems/studies with low drug concentrations. To better understand the applicability of the filters for studies

38

of enabling formulations, they were applied to more complex systems in the digestion–permeation assay developed in Paper III.

In vitro evaluation of solid lipid based formulations (Paper II) All previous studies of MMC had been directed towards release of adsorbed drugs. However, the possibility of tailoring MMCs pore size arose in 2016,48 providing an opportunity to use the carrier to adsorb lipids. The MMC with the larger pore size (~20 nm) was in the same range as a mesoporous alumina silicate, Neusilin US2, which has been successfully used for lipid adsorp-tion.64,76 The solid LBFs can then be filled into capsules or compressed into tablets, as had previously been done with mesoporous alumina silicate.65

Modified lipolysis with 1H NMR analysis As the standard lipolysis setup is dependent on pH changes during lipid diges-tion, its use for pH-altering formulations is limited. In Paper I, we had ob-served a pH increase during partial dissolution of MMC. Hence, a modified lipolysis assay was needed to study digestion of lipid-loaded MMC. However, as the enzymatic activity of the porcine extract is influenced by pH, the pH needed to be maintained throughout the assay. 1H NMR had been previously used in digestion of TG, and was considered to be a viable option to study the lipid-loaded MMC. To ease the analysis, the MMC was loaded with a single lipid (Captex 355) (CAP-MMC) instead of a more complex LBF. The impact on lipid release as a result of the preparation method (physical adsorption, solvent immersion, or freeze drying) was also evaluated.

The MMC dissolved rapidly in the modified lipolysis condition, enabling complete release of the TG. Complete release would be highly advantageous, as retention of the LBF-excipients has been seen when other mesoporous car-riers are evaluated.62,64,76,142 The incomplete release from a mesoporous alu-mina silicate also translates into poor in vivo performance.142 The in vitro dis-solution of the carrier makes MMC an attractive material for adsorption of lipids. It should, however, be noted that the 1H-NMR method only provides relative molar quantities, and hence is not a quantitative measurement. Fur-thermore, to what extent the dissolution of the MMC appears in vivo remains to be established.

The TG was extensively digested during the 60-min lipolysis assay, similar to what has been previously observed.143 The digestion of the TG and subse-quent formation of FFA from the CAP-MMC formulations were similar to those of the crude emulsion of the lipid. The similarity of the digestion profiles was accredited to the dissolution of the MMC (Figure 15).

39

Figure 15. Relative molar percentage of triglycerides and free fatty acids as deter-mined using 1H NMR spectroscopy (n ≥ 3) during in vitro lipolysis of a crude emul-sion of Captex (CAP C.E.) and CAP-MMC loaded using physical adsorption (P.A.), solvent immersion (S.I.), and freeze-drying (F.D.). No differences in lipid digestion (rate and extent) were seen between the crude emulsion and the lipid-loaded MMC.

The preparation method of the CAP-MMC did not appear to affect the lipid release, with the exception of the freeze-dried one. An increase in TG, and subsequently a decrease in FFAs, was observed for the freeze-dried material after 5 minutes. This may either be an artefact of the method, or a result of size reduction during manufacturing.

The relative di-/monoglyceride (DG/MG) content was very low for all for-mulations, in contrast to previous studies showing significantly higher amounts.123,143 This would indicate complete digestion of the TG into free fatty acids and one glycerol, which is known to occur to a greater extent in vitro compared to in vivo.59 However, due to the sample preparation used, the glyc-erol content could not be determined. The absence of DG/MG-peaks may also be an artefact of the parameter settings for the 1H NMR measurements, which were based on a previously developed in-house protocol for lipolysis samples, and not optimized for our samples. Meanwhile only minor differences were seen for all formulations during 60 min digestion, the TG content in the CAP-MMC samples at 0 min (i.e., before addition of the digestive enzyme) was low. The TG content in the three undigested CAP-MMC preparations was be-tween 85.1% ± 4.4% and 87.7% ± 2.7%. This is in contrast to the crude emul-sion of the Captex, for which the TG content was 95.6% ± 0.9%. To evaluate this discrepancy, the impact of storage of CAP-MMC was evaluated.

Chemical stability of lipid-loaded MMC For the second part of the study—which was directed towards lipid stability after being loaded onto the carrier—a more sensitive 1H NMR with optimized parameters was used. 1H NMR has been previously used to study degradation in e.g., edible oils144,145, but to the best of our knowledge, this was the first time it had been used to study storage stability of a lipid-loaded mesoporous carrier. Samples with CAP-MMC and CAP were prepared, and stored in a stability chamber at dry condition (<5% RH/25°C). Samples were taken monthly during 3 months and analyzed with 1H NMR. A significant change in

40

the TG content was observed for CAP-MMC, whereas Captex itself remained stable during the storage (Figure 16). The change in CAP-MMC was more pronounced during the first month, with only minor changes at two and three months. This would suggest that the degradation is a surface reaction on the MMC, as the contact area between the lipid and MMC would not have changed during the storage. The TG, like the pro-drug studied in Paper I, con-tains ester bonds which are susceptible to hydrolysis under alkaline condi-tions. The presence of small amounts of water may also expedite the conver-sion.

We hypothesized that since the MMC was not dried before lipid-loading, the degradation could have been a result of trace amounts of water in the po-rous matrix of the MMC. To study the impact of adsorbed water, further stud-ies were performed with MMC degassed prior to loading. However, changes in the molar percentage of TG were seen for the degassed CAP-MMC, indi-cating that the chemical degradation of the TG is not only a result of surface bound water on the MMC.

In conclusion, MMC adsorbed single lipids as well as a more complex LBF, and the loading could be performed with several relatively simple methods. However, the issue of the chemical stability needs to be addressed if the ma-terial is to be used for solidifications of lipids. The methods developed herein will be useful in future studies of, e.g., other mesoporous carriers and lipid-formulations for which pH-shifts may be a concern.

Figure 16. Molar percentage of lipid components during three months' storage (Cap-tex and CAP-MMC). For degassed CAP-MMC, the stability was only evaluated after one month’s storage. FFA: Free fatty acids, MG: Monoglycerides, DG: Diglycerides and TG: Triglycerides. CAP-MMC: Captex-loaded mesoporous magnesium car-bonate.

Dissolution–digestion–permeation assays (Papers III-IV) The assays developed in Papers I and II had focused on a single drug-delivery system. Furthermore, they focused solely on dissolution/release processes oc-curring in the GI-lumen, and lacks an absorption compartment. Thus, the last two papers investigated methods that incorporated both a donor and a receiver

41

milieu. In Paper III, the performance of drug-/LBF-loaded MMC was evalu-ated, together with three LBFs and one ASD, in the µFLUX system (Table 2). With the exception of the MMC systems, the formulations were commonly used in-house formulations and hence not optimized to improve their perfor-mance. On the basis of insights from Paper III, one LBF (IIIB-MC) and one ASD were selected for further evaluation in Paper IV.

Table 2. Composition of enabling formulations evaluated in the µFLUX system and the ENA device. Formulation Component 1a Component 2a Component 3a Component 4a

MMC FLD (25) MMC (75) - - LBF IIIB-MCb Cremophor EL (50) Carbitol (25) Captex 355 (12.5) Capmul MCM

EP (12.5) LBF IIIB-LCb Carbitol (50) Tween 85 (45) Soybean oil (5) - LBF IVb Cremophor EL (50) Carbitol (50) - - LBF-MMCb LBF (50)c MMC (50) - - ASDPVPd FLD (10) PVP (90) - - ASDHPMCASd FLD (25) HPMCAS (75) - - LBFBlank IIIB-MC (100) - - - aWeight fraction stated in parentheses (w/w) bDrug load: 70 mg FLD/g LBF cDrug load: 140 mg FLD/g LBF dPrepared via spray drying. LBF IIIB-MC; IIIB-LC; IV: Lipid-based formula-tions type IIIB medium chain; type IIIB long chain; type IV. MMC: Mesoporous magnesium carbonate. FLD: Felodipine. ASD: Amorphous solid dispersion. PVP: Polyvinylpyrrolidone. HPMCAS: Hydroxypropyl methylcellulose acetate succinate. In Paper IV, the ENA device (Figure 11) was used to study the in vitro per-formance of the formulations. For the first time, enabling formulations other than LBFs were evaluated in the ENA. The combined use of two enabling formulation strategies were also studied. We hypothesized that the solubiliz-ing properties of the LBF could be combined with the dissolution-enhancing properties of an ASD to increase drug absorption. The performance of the IIIB-MC, the ASD, and the ASD combined with the blank LBF were evalu-ated. As the PVP used in the ASD did not manage to maintain supersaturation of felodipine (Paper III), the PVP was replaced with HPMCAS in Paper IV. Apart from facilitating a stable supersaturation, the drug load could be in-creased to 25:75 w/w felodipine:HPMCAS (Table 2). The performance of crystalline drug combined with the blank LBF was evaluated as reference.

In vitro performance of enabling formulations The donor concentrations determined in the two assays varied significantly for the different formulations. Differences were also observed in mass transfer of the drug across the Caco-2 monolayer (Figure 17).

42

Figure 17. Felodipine concentrations in the donor (a, c, e, g) and receiver compart-ments (b, d, f, h) in the (a-f) µFLUX and the (g, h) ENA device for the enabling for-mulations (n≥3). The enzyme was added after a 10-min dispersion (here indicated by the dotted line).

43

Amorphous formulations For the amorphous formulations in the µFLUX-system, higher degrees of su-persaturation were achieved for the ASDPVP than the felodipine-loaded MMC (Figure 17a). The PVP used in the ASD did not, however, inhibit precipitation, and concentrations similar to the apparent solubility were reached after ap-proximately 20 min. The higher initial donor concentration also translated into a higher mass transfer for the ASDPVP. Furthermore, absorption appeared to be biphasic, with a higher mass transfer during the first 20 min. This phenom-enon has been previously reported by Tsinman et al., who speculated that it is a result of the supersaturation in the donor compartment.101 As the drug pre-cipitates, the amount of free drug available for absorption is reduced, thus de-creasing the driving force for absorption and leading to lower mass transfer.

However, the precise extent of supersaturation achieved for the ASDPVP could not be properly determined in the µFLUX due to turbidity in the donor compartment, which prompted manual sampling and analysis using HPLC-UV instead of making use of the in situ UV-probe. Apart from reducing the possibility to closely monitor the degree of superstaturation occurring the first minutes of dissolution, the manual sampling may induce precipitation.

The dissolution profile of the ASD stabilized with HPMCAS and evaluated in the ENA device differed from that observed in the µFLUX of the PVP-stabilized ASD. In the ENA device with the cell setup, the dissolution of the ASDHPMCAS was slow (Figure 17g). Even lower fractions of dissolved drug were seen for the ENA setup with LiDo-membrane, in which no titration was available. The low fractions dissolved were attributed to the poor solubility of the polymer, which has limited solubility at pH values ≤6.73 The dissolution of the HPMCAS caused a decrease in pH (from 6.5 to 6.1 by the end of the assay for ASDHPMCAS in the ENA with the LiDo setup). Regardless if titration were available or not to maintain the pH at 6.5, the pH change was believed to impact the dissolution of the polymer since the microclimate around the ASDHPMCAS particles is likely more acidic than that in the bulk.

Regardless of the type membrane used in the ENA device, high mass trans-fer of felodipine was observed in the receiver compartment for the ASDHPM-

CAS. This may be accounted for by liquid-liquid phase separation during the dissolution of the polymer, forming a drug-rich phase with high free-drug con-centrations and a drug-lean aqueous phase.146,147 The two phases are in a dy-namic equilibrium. As the drugs in the aqueous phase are absorbed, they are replaced with drugs from the drug-rich phase thus facilitating an increased absorption.

Lipid-based formulations In the µFLUX, the concentration profiles varied for the three LBFs (Figure 17c). The LBF type IIIB-MC reached the highest concentrations in the donor compartment and no precipitation occurred. High concentrations were also

44

achieved for the LBF type IV, but the concentration started to decrease after 10 min as a result of precipitation. The high donor concentrations achieved were most likely a result of solubilization and increased solvation capacity due to the presence of lipids, rather than a true supersaturation. For the long-chain LBF type IIIB, very low donor concentrations were observed. Turbidity, most likely caused by lipid droplets with a droplet size large than 0.25µm,56 also interfered with the UV-measurements. Thus, manual sampling followed by HPLC-UV analysis was necessary for the IIIB-LC donor samples.

High donor concentrations, however, did not translate into high mass trans-fer of the drug across the cell membrane for the LBFs. Thus, similar concen-tration profiles of the felodipine were observed for all LBFs in the receiver compartment (Figure 17d). The origin of this discrepancy (i.e., high-donor and low-receiver concentrations) may be due to events in the donor compartment. As the LBFs are dispersed in the dissolution medium, colloidal structures form. The lipophilic drug partitions into the solubilizing structures, thus re-ducing the amount of free drug available for absorption in the aqueous phase.119

High donor concentrations and comparatively low mass transfer were also observed for the IIIB-MC in the ENA device (Figure 17g, h). The underlying mechanism is believed to be the same as in the µFLUX, i.e., entrapment of felodipine in the colloidal structures in the donor compartment.

Solid lipid-based formulations For the LBF-loaded MMCs studied in the µFLUX, concentrations in the donor compartment were also significantly different for the three formulations (Fig-ure 17e). Here, the highest concentration was obtained for the IV-MMC, fol-lowed by IIIB-MC-MMC. For IV-MMC the drug precipitated, rapidly reduc-ing the concentrations in the donor chamber. The donor concentrations for the IIIB-LC-MMC were low, and turbidity interfered with the in situ UV-meas-urements. This made manual sampling and HPLC-UV analysis necessary.

In line with what was observed for the liquid LBFs, higher donor concen-trations of the LBF-MMCs did not translate into an elevated concentration in the receiver compartment. No differences in mass transfer were observed for the three LBF-MMC formulations (Figure 17f). The receiver profiles of the LBF-MMC were also similar to the ones obtained for the liquid-LBFs.

The adsorption of LBFs onto solid carriers is predominantly to ease manu-facturing and increase the storage stability of the final dosage form. As seen herein though, adsorption onto a carrier does not eliminate challenges associ-ated with in vitro studies of liquid LBFs (e.g., inability to distinguish between the aqueous concentration of free drug available for absorption and that parti-tioned into solubilizing complexes). Methods to address these concerns still need to be established in order to understand the full potential of solid LBFs.

45

Combined enabling formulations Simultaneous use of two enabling formulations strategies—amorphization and solubilization—was studied in the ENA device. As mentioned above, the dissolution of the ASDHPMCAS was slow. The concentrations obtained in the donor compartment were also low. When a blank LBF IIIB-MC was added, the rate and extent of dissolution increased and significantly higher donor-concentrations were obtained (Figure 17g). This was explained by the pres-ence of excipients in the LBF. HPMCAS contains hydrophobic regions,73 and the hydrophobic environment caused by the LBFs increases the solubility of the polymer when dissolved in presence of such lipophilic colloidal structures.

However, the elevated donor concentrations did not increase absorption. On the contrary, mass transfer decreased. The receiver concentrations were significantly lower than for the ASDHPMCAS alone. This phenomenon was be-lieved to be the same as that transpiring with the drug-loaded LBFs. After the ASDHPMCAS is dissolved, the free drug molecules partition into the colloidal structures formed from the excipients of the blank LBF. These are therefore not available for absorption.

Donor–receiver correlations of the in vitro assays The µFLUX proved to be useful for the study of in vitro performance of ena-bling formulations, and the mass transfer in the receiver compartment could be monitored in situ. For the donor compartment, complementary HPLC-UV measurements were required for three of nine formulations. However, the do-nor–receiver correlation was poor, and a high donor concentration did not translate to a high mass transfer (Figure 18). Whereas no differences in mass transfer were observed for the drug-loaded MMC, the LBFs, or the LBF-MMCs, a significantly higher flux was obtained for the ASDPVP.

The underlying mechanism for these inconsistences can be explained by the inherent properties of the technique used for concentration determination in the µFLUX. The in situ-UV measurements are unable to discriminate be-tween free drug in the aqueous phase and that trapped in colloidal structures. Thus, the drug concentrations measured in the donor do not truly reflect the free drug available for absorption. Rather, they reflect the combination of drug molecularly dissolved in the aqueous phase and in the solubilizing complexes in the dispersion media.

46

Figure 18. Rank order in area under the curve (AUC) measurements during 40-min assays in the donor and receiver compartments of the µFLUX and the ENA device. Both setups used a Caco-2 cell monolayer. Formulations are listed in declining order of AUC. In general, the donor–receiver correlation was better for the ENA device than the µFLUX. However for both setups, the ASD-formulation displayed a low AUC in the donor compartments, but the highest receiver–AUC in relation to all other formu-lations. For the µFLUX there were no significant differences in the AUC in the re-ceiver for any formulation except ASD. IIIB-MC:IIIB-LC:IV; Lipid-based formula-tion type IIIB medium chain, IIIB long chain and type IV respectively, MMC: meso-porous magnesium carbonate, ASDPVP: Amorphous solid dispersion stabilized with PVP, ASDHPMCAS: ASD stabilized with HPMCAS, IIIB-MCblank: Blank, drug-free LBF type IIIB-MC,FLD: Felodipine, as received (µFLUX) or gently grinded (ENA).

The ENA device gave a better correlation between donor concentrations and mass transfer—regardless of membrane—compared to the µFLUX. However, consistent with the µFLUX-study, the highest mass transfer was obtained for the ASDHPMCAS, despite low donor concentrations (Figure 18). The rank order in the receiver compartments was the same for the ENA device and the LiDo setup. The donor compartment was also similar for the artificial LiDo mem-brane, except that dissolution was faster for the ASDHPMCAS+LBFblank. The do-nor concentration profiles for the drug-loaded LBF and the ASDHPM-

CAS+LBFblank were not, however, significantly different with the LiDo system. Similar to the µFLUX assay, the sample procedure used for the ENA experi-ments did not enable determination of the true aqueous concentration.

In vitro-in vivo correlation of the ENA assay With the exception of the ASDHPMCAS, the in vivo bioavailability and the AUC in the receiver compartment in the ENA device correlated well. Although the highest AUC was observed for the ASD in vitro, the bioavailability in vivo was not significantly higher than for the crystalline drug administered with a blank LBF-dispersion (Table 3). This was likely a result of the pH-dependent solubility of the HPMCAS, which has limited solubility in solvents with pH ≤6.73 The pH in the rat intestine is comparatively low, predominantly <6, with slightly higher pH in the proximal colon (6.2±0.4).148 The slow dissolution of

47

the ASD in vivo can also explain the high tmax (7h). Nevertheless, the low sol-ubility of the polymer may not pose an issue in humans, where the intestinal pH ranges between 6-8.149–151 Hence, the rodent model may not be optimal for evaluating pH-sensitive formulations. Another phenomenon that has been pre-viously observed for capsule–filled ASDs containing HPMCAS is formation of a solid plug upon dissolution, limiting the drug-release.152 The occurrence of such a “plug” during administration of the ASDHPMCAS-capsules evaluated in Paper IV merits further investigation.

Table 3. Pharmacokinetic parameters calculated from plasma profiles. The formula-tions were administered by oral gavage to the rats.

FLD+LBFblank LBFFLD ASDHPMCAS ASDHPMCAS+LBFblank

Cmax (ng/mL)a

51.6 ± 22.4 (43%)

253.2 ± 111.3 (44%)

73.3 ± 42.6 (58%)

120.4 ± 44.5 (37%)c

[1910]d tmax (h)b 0.5 - 6 (4) 0.25 - 0.5 (0.5) 4 - 8 (7) 2 - 6 (6)c [1]d AUC0-last (ng/mL*h)

182.5 ± 50.8 (28%)

360.0 ± 70.8 (20%)

258.8 ± 71.0 (28%)

406.2 ± 114.6 (28%)c

[4070]d F (%)a ≥6.7 ± 1.8

(27%) 13.8 ± 2.4 (17%)

≥8.5 ± 2.3 (27%)

≥13.7 ± 3.8 (28%)c

All parameters were generated using non-compartmental analysis in Phoenix 64 WinNonlin (Certara, Princeton, NJ, USA). aMean ± SD (CV %) bRange (median) cParameter excluding one rat whose overall exposure was significantly higher than for the other animals in this group. d

PK parameters of the rat with significantly higher overall exposure. eAUC calculated from time 0h until last observed data point. Due to the relatively low bioavailability of felodipine, only a few data points could be collected for the ASD and the AUCs are therefore presented as absolute AUC rather than AUC extrapolated to infinity. LBFFLD; Capsules with felodipine dissolved in the LBF administered with water, FLDGrinded+LBFblank; Capsules with grinded felodipine ad-ministered with an LBFblank emulsion, ASD; Capsules with ASD administered with water, ASD+ LBFblank; Capsules with ASD administered with an LBFblank emulsion.

The bioavailability and in vivo exposure of the ASDHPMCAS increased when co-administered with a blank LBF, consistent with the dissolution enhancement observed in vitro. Understanding this phenomena is of vital importance, as co-administration may be used as a strategy to overcome interindividual variabil-ities in the pH of the small intestine, and reduce impact of variabilities in bile secretion patterns.153

Strengths and limitations with the digestion-permeation assays Both methods used consider processes occurring in the GI lumen (dissolution, release, digestion) and the permeation across relevant membranes. However, neither the µFLUX nor the current sampling procedure of the ENA samples measures true aqueous concentration in the donor compartment. Rather, they

48

measure the total amount dissolved material i.e. the sum of the drug freely dissolved in the aqueous phase and partitioned into colloidal structures.

The experimental setups of the µFLUX and the ENA device had different strengths and weaknesses. The µFLUX-assay is comparatively rapid and is well suited for studies of supersaturation and precipitation, making it suitable for screening studies. The in situ probes bypasses, to some extent, the need for manual sampling. The setup can also be further simplified by using artificial membranes if not active transport is expected to play a key role in drug ab-sorption. Provided that the drug studied contain a sufficiently strong UV-chro-mophore, the sensitivity is also good, ~1 µg/mL. However, its usefulness is limited for formulations that interfere with UV-measurements (herein the ASDPVP and the LBF type IIIB-LC). To study the donor compartment for these formulations, manual sampling needs to be carried out, followed by, e.g., HPLC-UV analysis. This reduces the advantages with the method.

An important factor that has gained attention is the area-to-donor-volume ratio (A/V) of the absorption-surface, as it may impact the dissolution-perme-ation balance. A too low A/V may not capture the in vivo situtiuation.154 The A/V for the µFLUX is 0.1cm-1 and 0.47cm-1 for the ENA device. The donor volume in the ENA can be further reduced at least 3-fold, obtaining an A/V of 1.41cm-1,119 which is significantly closer to the in vivo condition (A/V esti-mated to 1.9−2.3 cm−1).155

The hydrodynamics are also different in the two devices, affecting the aqueous boundary layer adjacent to the absorptive membrane (as well as the boundary layer around particles during dissolution). Signs of insufficient stir-ring (i.e., powder accumulation of solid powders) were observed for both de-vices, so further studies are needed to evaluate the thickness of the unstirred water layer.

The ENA studies are more time-consuming and resource-demanding than those with the µFLUX. However, as a pH stat apparatus and analytical instru-ments like HPLC-UV are common in laboratories, the device can be easily used without major instrument purchases. Another advantage of the ENA setup is the inclusion of pH-monitoring and titration, which gives insight into many phenomena of in vivo relevance, e.g., pH-dependent dissolution of ex-cipients.

49

Conclusions

This thesis focused on identifying and developing techniques to study ena-bling formulations. As the landscape of drug development has shifted towards drugs with challenging physicochemical properties, formulations to address these challenges are warranted. Poor aqueous solubility is commonly men-tioned as the main hurdle for new chemical entities. Enabling formulations based on the principle of supersaturation have been successfully used for sev-eral compound to overcome poor solubility. However, in vitro assays for eval-uation of enabling formulations are scarce, especially assays that enable head-to-head comparisons of different formulation strategies. This thesis therefore aimed to expand the toolbox of in vitro assays for the evaluation of enabling formulations. Valuable insights into the in vitro performance of a wide range of formulation strategies were also gained. Specific conclusions of this thesis are:

• Special protective nylon filters can be applied to the µDiss setup to enable

release measurements in particle-rich suspensions and turbid dissolution media. Since no lag time in the measurements was observed, the method can be used to study time-sensitive events including supersaturation and precipitation (Paper I).

• Mesoporous magnesium carbonate (MMC) can be used to stabilize a range of drugs in their amorphous form. As the carrier partly dissolves in aqueous media—affecting the microclimate within the pores and around the carrier (e.g., by increasing the pH)—the benefits of MMC are com-pound dependent. Furthermore, the chemical stability of the drug may be compromised after loading onto the carrier (Paper I).

• MMC can be used to create solid lipid-formulations using several differ-ent loading techniques of varying complexity. Physical adsorption is the simplest. However, chemical instability was observed during storage of Captex-loaded MMC. To what extent this will affect the performance (both in vitro and in vivo) needs to be further evaluated if the MMC is to be used for lipid adsorption (Paper II).

• By modifying the in vitro lipolysis to counteract pH increase caused by MMC dissolution, the release and digestion of lipids from lipid-loaded

50

MMC could be studied. 1H NMR can be efficiently applied to analyze the lipolysis samples and to investigate the storage stability of lipid-loaded carriers. (Paper II)

• The µFLUX can be used to study the release-digestion and permeation of drugs from a wide range of enabling formulations if it is combined with filters to protect the in situ UV probes and immobilized enzyme. To in-crease the in vivo relevance, cell monolayers can also be used as the ab-sorption membrane (Paper III).

• The correlation was poor between the drug exposure in the donor and re-ceiver compartments in the µFLUX. The UV-measurements do not dis-criminate between the drug in the aqueous phase and the drug solubilized in, e.g., colloidal structures formed after dispersion of lipid-based formu-lations. Hence, the donor concentration does not reflect the free drug truly available for absorption. This strengthens the relevance of methods incor-porating both a dissolution and a receiver compartment, in order to capture the complex interplay of these phenomena (Paper III).

• The enabling absorption (ENA) device offers a new promising assay to study enabling formulations. Based on the transport routes of the drugs and the purpose of the study, both cell-based and artificial membranes can be used (Paper IV).

• The rank-order in in vivo bioavailability of the formulations correlated well with the in vitro exposure in the receiver compartment of the ENA, with one exception – the amorphous solid dispersion (ASD) (Paper IV).

• For the ASD, the flux was high in vitro but the in vivo bioavailability of the drug from the formulation was low. The origin of this discrepancy was explained by evaluation of the donor compartment of the ENA device, in which there was a pH-dependent solubility of the polymer used (Paper IV).

• The in vitro studies showed that the performance of a poorly soluble ASD

can be increased by combining it with a dispersion of a blank lipid-based formulation. This was also confirmed in an in vivo pharmacokinetic study (Paper IV).

51

Future perspective

The characteristics of drugs under development are continuously changing, and suitable formulations for the new drugs are warranted. Many currently marketed drugs also suffer from poor water solubility and would benefit from novel formulation strategies. In a move away from traditional small molecule drugs, treatment with biologics, e.g., proteins and peptides, is a therapeutic field being extensively explored. Although treatment with small peptides such as insulin has been around for a long time, the currently available biologics are almost exclusively administered parentally (i.e., intravenous, intramuscu-lar or subcutaneous). This route of administration is problematic. Apart from causing discomfort for the patient, it requires medically trained staff and/or training of the patient for proper administration. Oral formulations of the pro-teins would circumvent these issues and increase the patient tolerability of the dosage form. Studies of enabling formulations of peptides are currently widely investigated, and the field is likely to continue to expand in the future.

Regardless if the drugs are small molecules or peptides, relevant in vitro assays are needed to predict the in vivo performance of the formulations. Cur-rent experimental setups for evaluation of orally administered formulations typically focus on one specific formulation, and are not broadly applicable to other formulations strategies. The two methods used in this thesis, the µFLUX and the enabling absorption device (ENA), are useful tools for the latter. Fur-ther studies need to be carried out to better understand the applicability of the assays, but the results are promising. Many other combined dissolution-per-meation assays are also under development as the importance of the complex interplay between the two phenomena is brought forward. Physiologically based biopharmaceutical modelling may also be a useful tool to better explain the in vivo performance of enabling formulations. Software needs to be adapted to include the features of the formulations, as current software limits the possibility to incorporate the formulation characteristics. As simulations thus may allow screening of different formulations, in vivo studies can be re-duced and thereby avoid animal experiments and high study costs.

The excipient selection in enabling formulations is also of vital importance. As shown in Paper IV, the polymer selection in an amorphous formulation strongly influences the in vivo performance of the formulation. The complex-ity of using polymers with pH-dependent solubility is also gaining more at-tention. Issues related to HPMCAS in e.g., rodents, are being reported, which

52

highlights the importance of considering species-specific gastro-intestinal characteristics when selecting a suitable animal model.

However, the development of new excipients is limited. The excipients commonly used have been so during a long time. Although the toxicity and physical characteristics of these excipients have been extensively studied, other concerns, e.g., pharmacological contributions of the excipients by recep-tor binding, are far less known. Furthermore, sex-specific effects have been observed for some excipients in vivo. These issues are gaining more attention and will likely continue to be of interest for the pharmaceutical industry.

To conclude, new formulation challenges are arising due to the expansion of drugs with poor aqueous solubility as well as biological drugs. Formulation strategies to overcome these limitations are therefore of vital importance, and new in vitro assays are needed to better capture the in vivo performance of the drug delivery systems. The excipient selection strongly influences the perfor-mance of the enabling formulations. Therefore a better understanding of the physiological impact of excipients is desirable.

53

Populärvetenskaplig sammanfattning

Redan 1550 före Kristus användes läkemedel i form av tabletter och kapslar, och ett oralt intag av läkemedel fortsätter vara den mest önskvärda läkeme-delsformuleringen. Genom att ta ett läkemedel oralt undviker patienten obe-hagliga nålstick och behöver inte ha hjälp av vårdpersonal för att ta läkemed-let. Transporten genom mag-och tarm kanalen är dock komplex och formule-ringen utsätts för många prövningar så som olika pH-värden och enzymer. För att ett läkemedel ska tas upp måste det först lösas upp, då endast enskilda fria molekyler kan tas sig igenom tarmmembranet och över till blodet. Upptaget över tarmen kommer också att påverkas av hur läkemedlet ser ut och bland annat kan nedbrytning pga enzymatisk aktivitet ske i tarmmembranet. Innan läkemedlet når den systemiska cirkulationen och sitt slutgiltiga mål genomgår det också första-passage metabolism i levern. Sammantaget kommer alla dessa processer att påverka hur mycket av läkemedlet patienten tagit som kommer att nå fram till sitt mål.

Många nya läkemedel har väldigt låg löslighet i tarmvätskorna, och kom-mer därför ha ett begränsat upptag när de sväljs. Detta är också sant för många läkemedel som redan finns på marknaden. Effekten blir att man måste ge höga doser av substansen, vilket ökar risken för biverkningar. För att kringgå den låga lösligheten pågår det därför mycket forskning runt framtagning av nya läkemedelsformuleringarna.

En strategi är att bryta sönder den kristallina strukturen som molekylerna sitter i, och skapa ett så kallat amorft system. I amorfa system sitter moleky-lerna i oordning, och det blir därför lätt för vattenmolekylerna att lösa upp läkemedlet. Detta är användbart för substanser som kräver mycket energi för att bryta söder bindningarna i kristallen, vilket kan ses genom att mäta smält-punkten hos substansen. En hög smältpunkt är ofta kopplat till en låg löslighet. Ett problem med amorfa system är dock att de är ostabila vid förvaring, och en stabiliserande substans som t.ex. en polymer kan behövas för att få en fy-sikaliskt stabil produkt.

En annan strategi är att använda sig av fetter och oljor, så kallade lipid-baserade formuleringar, i vilka läkemedlet löses upp. Detta gör att upplös-ningssteget i tarmen helt kringgås då substansen redan är upplöst vid admini-streringen. Väl i tarmen kommer sedan fetterna brytas ner, och tillsammans med kroppsegna gallsalter skapa strukturer där substansen kan hållas löst tills den tar upp genom tarmmembranet. Denna formuleringstyp är användbar för fettlösliga substanser. Många lipid-baserade formuleringar är dock kemiskt

54

instabila och både hjälpämnen och läkemedelssubstansen riskerar att brytas ner vid förvaring. Försök att göra om de vanligtvis flytande lipid-baserade formuleringarna till fast material, och därmed förbättra stabiliteten, har därför utforskats genom att adsorbera formuleringen på en porös bärare.

Denna avhandling har fokuserat på att ta fram metoder för att experimen-tellt på labb (s.k. in vitro) studera dessa olika formuleringsstrategier. Målet var att ta fram nya metoder för att kunna förutsäga hur läkemedelsformule-ringarna kommer att bete sig i kroppen (s.k. in vivo).

I det första projektet var målet att ta fram en metod för att titta på frisätt-ningen av olika läkemedel efter att de gjorts amorfa genom att de adsorberats på en bärarpartikel bestående av mesoporös magnesiumkarbonat. För att minska mängden formulering som behövdes vid experimenten använde vi ett litet upplösningsbad, med en UV-prob nerstoppad för att kunna studera fri-sättningen. Ett problem var dock att bärarpartiklarna störde mätningarna ef-tersom de ansamlades på proberna. Vi designade därför ett filter som kan sät-tas på utanpå proberna vilket skyddar dem från ansamling av material och därmed möjliggör kvantifiering av mängd frisatt läkemedel.

I det andra projektet undersökte vi om mesoporös magnesiumkarbonat kan användas för adsorption av fetter. En in vitro metod utvecklades där vi tog hänsyn till nedbrytningen av fetterna som sker i mag-tarmkanalen. Analyserna av proverna gjordes med hjälp av kärnmagnetisk resonans (1H NMR), vilket har visats vara användbart vid liknande studier. Stabiliteten hos de lipid-inlad-dade bärarna undersöktes också, där vi kunde se att fetterna till en viss del bröts ner vid förvaring av formuleringarna.

I de sista projekten var målet att utveckla och undersöka både frisättning och upptag av läkemedel från mer komplexa formuleringar samtidigt. Tidigare forskning har tydligt visat att både frisättningen och absorptionen är tätt sam-mankopplade i tarmarna, och det är därför viktigt att ta hänsyn till det även vid in vitro studier. Två metoder användes, den s.k. µFLUX och enabling-absorption (ENA)-utrustningen. Båda har en givarsida som ska efterlikna tar-men och en mottagarsida som ska likna blodet. Unikt för dessa studier är att vi tittade på många olika formuleringsstrategier i samma experimentella upp-sättning. I µFLUXen kunde vi se att koncentrationerna på givarsidan inte speglar det som mäts upp på mottagarsidan, vilket visar på vikten av att ha metoder som tar hänsyn till båda. För att studera hur formuleringarna fungerar in vivo administrerades de oralt till råtta. Råttornas plasmakoncentrationer jämfördes sedan med data från ENA:n. Resultaten från ENA:n stämde relativt bra överens med de observerade in vivo, med undantag av en amorf formule-ring. Detta förklarades med låg löslighet hos hjälpämnet i formuleringen i de artspecifika pH-värdena i råttans mag-tarm system.

Sammanfattningsvis har denna avhandling givit insikt i metoder som kan användas för att studera komplexa läkemedelsberedningar. Dessa kan för-hoppningsvis integreras i många nuvarande laboratorier, och ge en bättre pre-diktion av, och förståelse för, hur orala beredningar fungerar in vivo.

55

Acknowledgements

The work in this thesis was carried out at the Department of Pharmacy, Upp-sala University. The studies were financially supported by the Swedish Re-search Council and the Erling Person Family Foundation. The Australian Gov-ernment Department of Education, is acknowledged for financially supporting a research visit to the University of South Australia. Apotekare C D Carlssons stiftelse (Apotekarsociteten), IFs stiftelse inkl Källrotstipendiet (Apotekarso-citeten), and H F Sederholms stipendiestiftelse are also gratefully acknowl-edged for travel grants that allowed me to take part in several research confer-ences and visits.

This work would not have been possible without the help and support from many, many people. Thank you to all that has been involved during my stud-ies. I would like to direct some special thanks to the following: My scientific advisors: Min huvudhandledare, Prof. Christel Bergström: Du är en sann inspiration, och din osvikliga entusiasm för allt vad rör forskning är så motiverande. Trots alla förändringar som har skett under våra 5 år tillsammans har du varit en stabil och pålitlig mentor. Oavsett hur mycket annat du har haft på ditt bord har det alltid funnits tid för samtal. Tack för all handledning du givit mig både professionellt och privat, med målet att jag skulle bli en självständig forskare.

Min biträdande handledare, Prof. Maria Strømme: Det finns nog ingen som skickar så peppande mail som du! Tack för att du har underlättat detta avhandlingsarbete på alla sätt du kunnat, jag har alltid kunnat förlita mig på dina blicksnabba svar när de behövts.

Prof. Clive Prestidge: Thank you for giving me the unique opportunity to spend six months in your beautiful country, and for letting me be a part of your research group. It was a truly amazing experience. Thanks to all the other members in the Prestidge group for making me feel so welcome.

Prof. Adolf Gogoll: Thank you for all help with the NMR-studies, and for your patience with my (usually very confused) questions about the experi-ments.

To all the former and present members of the Drug delivery and disposition group and UDOPP: Thank you for being such nice colleagues and creating a great working environment.

56

Christine, Magnus and André: You immediately let me become a member of your group, and made my studies so enjoyable. Your extensive knowledge of mass spectrometry, BMC storage rooms, and mushrooms is nothing short of extraordinary.

Maria M: Dziękuję bardzo za twoją pomoc. Jesteś najlepsza!

To all the former and present members of the BLT-group, thank you for great discussions and being the best colleagues possible. Special thanks to Jenny and Oliver who read through this thesis, giving me valuable feedback.

Sara and Linda: Thank you for quickly taking me on and supporting me when I started my quest towards a PhD.

Khadijah: Your kindness knows no boundaries, and your patience always amazes me. Thank you for letting me visit your beautiful country and family, for always taking the time to help with anything and everything, and for all the yummy treats during these past years.

Janneke: Words are not enough to explain what you meant scientifically in my quest to become a researcher. Words are even more insufficient to describe your support as a friend during this time.

Julian and Madlen, my two favorite Germans: Thank you for a lot of laughs, uplifting words and support, and many glasses of wine.

Vicky: Remember that you are amazing and you’re doing a good job, even the days when you’re not feeling that way. Thank you for making sure that I got to relax on the beautiful beaches of Paros.

Ilse: My mentor with everything related to animals and modelling. Thank you for all research and non-research related discussions and for making sure I stayed sane these past years.

My current office mates, Jenny and Albin: Thank you for being the best possible office mates, and taking time to listen to my research issues, life is-sues, and everything in between.

Thanks to all my bachelor/master students that helped me with experiments throughout my studies. Special thanks to Filip, who was kind enough to stay on and help me finish up. I don´t know what I would have done without your reliable and skilled experimental work. Thanks to the members in the Nanotechnology and Functional Materials-group at Ångström who have helped me these past years. Special thanks to Michelle and Ocean, for all your help with solid-state analyzes and providing material for my research.

To my friends and family, whose patience and support are never failing:

My Aussie family: Dr. Dening: Thank you for taking me in, introducing me to you family (and trash TV :P), being an amazing roomie and remaining the most awesome of friends ever, even after I left the land down under. Leah:

57

thank you for company during long, cold evenings, Tim-Tams, and for you ever positive attitude towards the challenges of research.

Hanna: Så lycklig att jag hittade dig på andra sidan jordklotet, och att du och Andreas tog er an en förvirrad Fia-Lisa!

Jonas och Johanna: Ni har varit så otroligt värdefulla vänner och bollplank sedan vi tillsammans startade den här resan. Sara: även om du lämnade oss för smørrebrød och Carlsberg har det ändå alltid känts som att du varit här på något sätt.

Mina ”Swedish Fingers”, Ilse, Rebecca, Annelie and Erik: tack för alla ro-liga memes, diskussioner om böcker jag inte läst och allmänt vara en fantastisk supportgrupp

Till familjen jag fått: Mamma, Pappa, Gits, Lars, Marie-Louise, Nathali, Hampus, Erik, Niclas och Isabelle: Ni är det mest värdefulla jag har. Tack för att ni alla finns där när jag behöver det. Och Bellis: Tack för att jag har fått ta hand om låtsaskompisbebisen åt dig. Han har gett mig mycket glädje i svåra stunder, och någon att lägga skulden på när jag blir frustrerad över allt och inget.

Till Farfar som gav mig motiv till att fortsätta utbilda mig, Mormor som gav mig envisheten att fortsätta även när det är utmanande, Farmor som hjäl-per mig att behålla perspektiv på saker och ting, och Morfar som ständigt påminde mig om att härj inte har någon ålder. Ni har alla gett mig olika förut-sättningar för att jag skulle kunna ta mig hit.

Till familjen jag valt: Emil, Malin och Katrina, för vilka mitt hem är lika mycket mitt som ert: Tack för att ni är precis som ni är, amazeballs!

Kristian: Tack för att du är tålamodet jag inte har, lugnet i kaoset och trygg-heten i att allt kommer att ordna sig. Jesteś moim światem.

58

References

1 Brown TD, Whitehead KA, Mitragotri S. Materials for oral delivery of proteins and peptides. Nat Rev Mater 2020; 5: 127–148.

2 Tyagi P, Pechenov S, Subramony JA. Oral peptide delivery: Translational chal-lenges due to physiological effects. J Control Release 2018; 287: 167–176.

3 Homayun B, Ling X, Choi H-J. Challenges and Recent Progress in Oral Drug Delivery Systems for Biopharmaceuticals. Pharmaceutics 2019; 11. doi:10.3390/pharmaceutics11030129.

4 Fagerberg JH, Bergström CAS. Intestinal solubility and absorption of poorly water soluble compounds: predictions, challenges and solutions. Ther Deliv 2015; 6: 935–359.

5 Drucker DJ. Advances in oral peptide therapeutics. Nat Rev Drug Discov 2020; 19: 277–289.

6 Dokoumetzidis A, Macheras P. A century of dissolution research: From Noyes and Whitney to the Biopharmaceutics Classification System. Int J Pharm 2006; 321: 1–11.

7 Noyes AA, Whitney WR. THE RATE OF SOLUTION OF SOLID SUB-STANCES IN THEIR OWN SOLUTIONS. J Am Chem Soc 1897; 19: 930–934.

8 Sugano K, Kansy M, Artursson P, Avdeef A, Bendels S, Di L et al. Coexistence of passive and carrier-mediated processes in drug transport. Nat Rev Drug Dis-cov 2010; 9: 597–614.

9 Tannergren C, Petri N, Knutson L, Hedeland M, Bondesson U, Lennernäs H. Multiple transport mechanisms involved in the intestinal absorption and first-pass extraction of fexofenadine. Clin Pharmacol Ther 2003; 74: 423–436.

10 Poongavanam V, Doak BC, Kihlberg J. Opportunities and guidelines for dis-covery of orally absorbed drugs in beyond rule of 5 space. Curr Opin Chem Biol 2018; 44: 23–29.

11 Boyd BJ, Bergström CAS, Vinarov Z, Kuentz M, Brouwers J, Augustijns P et al. Successful oral delivery of poorly water-soluble drugs both depends on the intraluminal behavior of drugs and of appropriate advanced drug delivery sys-tems. Eur J Pharm Sci 2019; 137: 104967.

12 Rossi Sebastiano M, Doak BC, Backlund M, Poongavanam V, Over B, Er-mondi G et al. Impact of Dynamically Exposed Polarity on Permeability and Solubility of Chameleonic Drugs Beyond the Rule of 5. J Med Chem 2018; 61: 4189–4202.

13 FDA/CDER. Waiver of In Vivo Bioavailability and Bioequivalence Studies for Immediate-Release Solid Oral Dosage Forms Based on a Biopharmaceutics Classification System Guidance for Industry. https://www.fda.gov/me-dia/70963/download. 2017.

59

14 EMA. Guideline on the investigation of bioequivalence (Rev.1). https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-in-

vestigation-bioequivalence-rev1_en.pdf. 2010.https://www.ema.europa.eu/en/documents/scientific-guideline/guideline-

investigation-bioequivalence-rev1_en.pdf (accessed 1 Apr2020). 15 Bergström CAS, Andersson SBE, Fagerberg JH, Ragnarsson G, Lindahl A. Is

the full potential of the biopharmaceutics classification system reached? Eur J Pharm Sci 2014; 57: 224–231.

16 Tsume Y, Mudie DM, Langguth P, Amidon GE, Amidon GL. The Biopharma-ceutics Classification System: Subclasses for in vivo predictive dissolution (IPD) methodology and IVIVC. Eur J Pharm Sci 2014; 57: 152–163.

17 Kawabata Y, Wada K, Nakatani M, Yamada S, Onoue S. Formulation design for poorly water-soluble drugs based on biopharmaceutics classification sys-tem: Basic approaches and practical applications. Int J Pharm 2011; 420: 1–10.

18 Augustijns P, Wuyts B, Hens B, Annaert P, Butler J, Brouwers J. A review of drug solubility in human intestinal fluids: Implications for the prediction of oral absorption. Eur J Pharm Sci 2014; 57: 322–332.

19 Zheng W, Jain A, Papoutsakis D, Dannenfelser R-M, Panicucci R, Garad S. Selection of oral bioavailability enhancing formulations during drug discovery. Drug Dev Ind Pharm 2012; 38: 235–247.

20 Williams HD, Trevaskis NL, Charman SA, Shanker RM, Charman WN, Pouton CW et al. Strategies to Address Low Drug Solubility in Discovery and Devel-opment. Pharmacol Rev 2013; 65: 315–499.

21 Feeney OM, Crum MF, McEvoy CL, Trevaskis NL, Williams HD, Pouton CW et al. 50years of oral lipid-based formulations: Provenance, progress and future perspectives. Adv Drug Deliv Rev 2016; 101: 167–194.

22 Schittny A, Huwyler J, Puchkov M. Mechanisms of increased bioavailability through amorphous solid dispersions: a review. Drug Deliv 2020; 27: 110–127.

23 Williams HD, Trevaskis NL, Charman SA, Shanker RM, Charman WN, Pouton CW et al. Strategies to Address Low Drug Solubility in Discovery and Devel-opment. Pharmacol Rev 2013; 65: 315–499.

24 Censi R, Di Martino P. Polymorph Impact on the Bioavailability and Stability of Poorly Soluble Drugs. Molecules 2015; 20: 18759–18776.

25 Grunenberg A, Henck J-O, Siesler HW. Theoretical derivation and practical application of energy/temperature diagrams as an instrument in preformulation studies of polymorphic drug substances. Int J Pharm 1996; 129: 147–158.

26 Buckley ST, Frank KJ, Fricker G, Brandl M. Biopharmaceutical classification of poorly soluble drugs with respect to “enabling formulations”. Eur J Pharm Sci 2013; 50: 8–16.

27 Ozaki S, Minamisono T, Yamashita T, Kato T, Kushida I. Supersaturation–Nu-cleation Behavior of Poorly Soluble Drugs and its Impact on the Oral Absorp-tion of Drugs in Thermodynamically High-Energy Forms. J Pharm Sci 2012; 101: 214–222.

28 Brouwers J, Brewster ME, Augustijns P. Supersaturating Drug Delivery Sys-tems: The Answer to Solubility-Limited Oral Bioavailability? J Pharm Sci 2009; 98: 2549–2572.

29 Augustijns P, Brewster ME. Supersaturating Drug Delivery Systems: Fast is Not Necessarily Good Enough. J Pharm Sci 2012; 101: 7–9.

60

30 Fong SYK, Bauer-Brandl A, Brandl M. Oral bioavailability enhancement through supersaturation: an update and meta-analysis. Expert Opin Drug Deliv 2017; 14: 403–426.

31 Hancock BC, Zografi G. Characteristics and Significance of the Amorphous State in Pharmaceutical Systems. J Pharm Sci 1997; 86: 1–12.

32 Baghel S, Cathcart H, O’Reilly NJ. Polymeric Amorphous Solid Dispersions: A Review of Amorphization, Crystallization, Stabilization, Solid-State Charac-terization, and Aqueous Solubilization of Biopharmaceutical Classification System Class II Drugs. J Pharm Sci 2016; 105: 2527–2544.

33 Bhugra C, Pikal MJ. Role of Thermodynamic, Molecular, and Kinetic Factors in Crystallization from the Amorphous State. J Pharm Sci 2008; 97: 1329–1349.

34 Kogermann K, Penkina A, Predbannikova K, Jeeger K, Veski P, Rantanen J et al. Dissolution testing of amorphous solid dispersions. Int J Pharm 2013; 444: 40–46.

35 Marsac PJ, Konno H, Rumondor ACF, Taylor LS. Recrystallization of Nifedi-pine and Felodipine from Amorphous Molecular Level Solid Dispersions Con-taining Poly(vinylpyrrolidone) and Sorbed Water. Pharm Res 2008; 25: 647–656.

36 Laitinen R, Löbmann K, Strachan CJ, Grohganz H, Rades T. Emerging trends in the stabilization of amorphous drugs. Int J Pharm 2013; 453: 65–79.

37 He Y, Ho C. Amorphous Solid Dispersions: Utilization and Challenges in Drug Discovery and Development. J Pharm Sci 2015; 104: 3237–3258.

38 Kresge CT, Leonowicz ME, Roth WJ, Vartuli JC, Beck JS. Ordered mesopo-rous molecular sieves synthesized by a liquid-crystal template mechanism. Na-ture 1992; 359: 710–712.

39 Maleki A, Kettiger H, Schoubben A, Rosenholm JM, Ambrogi V, Hamidi M. Mesoporous silica materials: From physico-chemical properties to enhanced dissolution of poorly water-soluble drugs. J Control Release 2017; 262: 329–347.

40 Riikonen J, Xu W, Lehto V-P. Mesoporous systems for poorly soluble drugs – recent trends. Int J Pharm 2018; 536: 178–186.

41 Xu W, Riikonen J, Lehto V-P. Mesoporous systems for poorly soluble drugs. Int J Pharm 2013; 453: 181–197.

42 Valetti S, Wankar J, Ericson MB, Feiler A, Manet I. Mesoporous silica particles as a lipophilic drug vehicle investigated by fluorescence lifetime imaging. J Mater Chem B 2017; 5: 3201–3211.

43 Mäkilä E, Kivelä H, Shrestha N, Correia A, Kaasalainen M, Kukk E et al. In-fluence of Surface Chemistry on Ibuprofen Adsorption and Confinement in Mesoporous Silicon Microparticles. Langmuir 2016; 32: 13020–13029.

44 Bavnhøj CG, Knopp MM, Madsen CM, Löbmann K. The role interplay be-tween mesoporous silica pore volume and surface area and their effect on drug loading capacity. Int J Pharm 2019; 1: 100008.

45 Limnell T, Riikonen J, Salonen J, Kaukonen AM, Laitinen L, Hirvonen J et al. Surface chemistry and pore size affect carrier properties of mesoporous silicon microparticles. Int J Pharm 2007; 343: 141–147.

46 Forsgren J, Frykstrand S, Grandfield K, Mihranyan A, Strømme M. A Tem-plate-Free, Ultra-Adsorbing, High Surface Area Carbonate Nanostructure. PLOS ONE 2013; 8: e68486.

61

47 Frykstrand S, Forsgren J, Mihranyan A, Strømme M. On the pore forming mechanism of Upsalite, a micro- and mesoporous magnesium carbonate. Mi-croporous Mesoporous Mater 2014; 190: 99–104.

48 Cheung O, Zhang P, Frykstrand S, Zheng H, Yang T, Sommariva M et al. Nanostructure and pore size control of template-free synthesised mesoporous magnesium carbonate. RSC Adv 2016; 6: 74241–74249.

49 Zhang P, Forsgren J, Strømme M. Stabilisation of amorphous ibuprofen in Up-salite, a mesoporous magnesium carbonate, as an approach to increasing the aqueous solubility of poorly soluble drugs. Int J Pharm 2014; 472: 185–191.

50 Yang J, Alvebratt C, Zhang P, Zardán Gómez de la Torre T, Strømme M, Berg-ström CAS et al. Enhanced release of poorly water-soluble drugs from synergy between mesoporous magnesium carbonate and polymers. Int J Pharm 2017; 525: 183–190.

51 Zhang P, Zardán Gómez de la Torre T, Welch K, Bergström C, Strømme M. Supersaturation of poorly soluble drugs induced by mesoporous magnesium carbonate. Eur J Pharm Sci 2016; 93: 468–474.

52 Rezhdo O, Speciner L, Carrier R. Lipid-associated oral delivery: Mechanisms and analysis of oral absorption enhancement. J Control Release 2016; 240: 544–560.

53 Pouton CW, Porter CJH. Formulation of lipid-based delivery systems for oral administration: Materials, methods and strategies. Adv Drug Deliv Rev 2008; 60: 625–637.

54 Porter CJH, Trevaskis NL, Charman WN. Lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs. Nat Rev Drug Discov 2007; 6: 231–248.

55 Pouton CW. Lipid formulations for oral administration of drugs: non-emulsify-ing, self-emulsifying and ‘self-microemulsifying’ drug delivery systems. Eur J Pharm Sci 2000; 11: S93–S98.

56 Pouton CW. Formulation of poorly water-soluble drugs for oral administration: Physicochemical and physiological issues and the lipid formulation classifica-tion system. Eur J Pharm Sci 2006; 29: 278–287.

57 Carrière F. Impact of gastrointestinal lipolysis on oral lipid-based formulations and bioavailability of lipophilic drugs. Biochimie 2016; 125: 297–305.

58 Carlier H, Bernard A, Caselli C. Digestion and absorption of polyunsaturated fatty acids. Reprod Nutr Dev 1991; 31: 475–500.

59 Nieva-Echevarría B, Goicoechea E, Manzanos MJ, Guillén MD. Usefulness of 1H NMR in assessing the extent of lipid digestion. Food Chem 2015; 179: 182–190.

60 Tan A, Rao S, Prestidge CA. Transforming Lipid-Based Oral Drug Delivery Systems into Solid Dosage Forms: An Overview of Solid Carriers, Physico-chemical Properties, and Biopharmaceutical Performance. Pharm Res 2013; 30: 2993–3017.

61 Dening TJ, Rao S, Thomas N, Prestidge CA. Novel Nanostructured Solid Ma-terials for Modulating Oral Drug Delivery from Solid-State Lipid-Based Drug Delivery Systems. AAPS J 2016; 18: 23–40.

62 Mandić J, Zvonar Pobirk A, Vrečer F, Gašperlin M. Overview of solidification techniques for self-emulsifying drug delivery systems from industrial perspec-tive. Int J Pharm 2017; 533: 335–345.

62

63 Joyce P, Dening TJ, Meola TR, Schultz HB, Holm R, Thomas N et al. Solidi-fication to improve the biopharmaceutical performance of SEDDS: Opportuni-ties and challenges. Adv Drug Deliv Rev 2019; 142: 102–117.

64 Williams HD, Speybroeck MV, Augustijns P, Porter CJH. Lipid-Based Formu-lations Solidified Via Adsorption onto the Mesoporous Carrier Neusilin® US2: Effect of Drug Type and Formulation Composition on In Vitro Pharmaceutical Performance. J Pharm Sci 2014; 103: 1734–1746.

65 Gumaste SG, Dalrymple DM, Serajuddin ATM. Development of Solid SEDDS, V: Compaction and Drug Release Properties of Tablets Prepared by Adsorbing Lipid-Based Formulations onto Neusilin® US2. Pharm Res 2013; 30: 3186–3199.

66 Bolko Seljak K, Ilić IG, Gašperlin M, Zvonar Pobirk A. Self-microemulsifying tablets prepared by direct compression for improved resveratrol delivery. Int J Pharm 2018; 548: 263–275.

67 Aungst BJ. Optimizing Oral Bioavailability in Drug Discovery: An Overview of Design and Testing Strategies and Formulation Options. J Pharm Sci 2017; 106: 921–929.

68 Ayad MH. Rational formulation strategy from drug discovery profiling to hu-man proof of concept. Drug Delivery 2015; 22: 877–884.

69 Gad SC, Spainhour CB, Shoemake C, Pallman DRS, Stricker-Krongrad A, Downing PA et al. Tolerable Levels of Nonclinical Vehicles and Formulations Used in Studies by Multiple Routes in Multiple Species With Notes on Methods to Improve Utility: Int J Toxicol 2016. doi:10.1177/1091581815622442.

70 Persson LC, Porter CJH, Charman WN, Bergström CAS. Computational Pre-diction of Drug Solubility in Lipid Based Formulation Excipients. Pharm Res 2013; 30: 3225–3237.

71 Patel V, Lalani R, Bardoliwala D, Ghosh S, Misra A. Lipid-Based Oral Formu-lation Strategies for Lipophilic Drugs. AAPS PharmSciTech 2018; 19: 3609–3630.

72 Alskär LC, Porter CJH, Bergström CAS. Tools for Early Prediction of Drug Loading in Lipid-Based Formulations. Mol Pharm 2016; 13: 251–261.

73 Friesen DT, Shanker R, Crew M, Smithey DT, Curatolo WJ, Nightingale JAS. Hydroxypropyl Methylcellulose Acetate Succinate-Based Spray-Dried Disper-sions: An Overview. Mol Pharm 2008; 5: 1003–1019.

74 Vall M, Ferraz N, Cheung O, Strømme M, Zardán Gómez de la Torre T. Ex-ploring the Use of Amine Modified Mesoporous Magnesium Carbonate for the Delivery of Salicylic Acid in Topical Formulations: In Vitro Cytotoxicity and Drug Release Studies. Molecules 2019; 24: 1820.

75 Vall M, Zhang P, Gao A, Frykstrand S, Cheung O, Strømme M. Effects of amine modification of mesoporous magnesium carbonate on controlled drug release. Int J Pharm 2017; 524: 141–147.

76 Gumaste SG, Freire BOS, Serajuddin ATM. Development of solid SEDDS, VI: Effect of precoating of Neusilin® US2 with PVP on drug release from adsorbed self-emulsifying lipid-based formulations. Eur J Pharm Sci 2017; 110: 124–133.

77 Newman A, Knipp G, Zografi G. Assessing the performance of amorphous solid dispersions. J Pharm Sci 2012; 101: 1355–1377.

63

78 McCarthy CA, Faisal W, O’Shea JP, Murphy C, Ahern RJ, Ryan KB et al. In vitro dissolution models for the prediction of in vivo performance of an oral mesoporous silica formulation. J Control Release 2017; 250: 86–95.

79 Williams HD, Sassene P, Kleberg K, Bakala-N’Goma J-C, Calderone M, Jan-nin V et al. Toward the Establishment of Standardized In Vitro Tests for Lipid-Based Formulations, Part 1: Method Parameterization and Comparison of In Vitro Digestion Profiles Across a Range of Representative Formulations. J Pharm Sci 2012; 101: 3360–3380.

80 Knopp MM, Chourak N, Khan F, Wendelboe J, Langguth P, Rades T et al. Effect of polymer type and drug dose on the in vitro and in vivo behavior of amorphous solid dispersions. Eur J Pharm Biopharm 2016; 105: 106–114.

81 Hong S, Shen S, Tan DCT, Ng WK, Liu X, Chia LSO et al. High drug load, stable, manufacturable and bioavailable fenofibrate formulations in mesopo-rous silica: a comparison of spray drying versus solvent impregnation methods. Drug Deliv 2016; 23: 316–327.

82 Mathers A, Hassouna F, Malinová L, Merna J, Růžička K, Fulem M. Impact of Hot-Melt Extrusion Processing Conditions on Physicochemical Properties of Amorphous Solid Dispersions Containing Thermally Labile Acrylic Copoly-mer. J Pharm Sci 2020; 109: 1008–1019.

83 Berben P, Bauer-Brandl A, Brandl M, Faller B, Flaten GE, Jacobsen A-C et al. Drug permeability profiling using cell-free permeation tools: Overview and ap-plications. Eur J Pharm Sci 2018; 119: 219–233.

84 Berben P, Brouwers J, Augustijns P. Assessment of Passive Intestinal Permea-bility Using an Artificial Membrane Insert System. J Pharm Sci 2018; 107: 250–256.

85 Wuyts B, Riethorst D, Brouwers J, Tack J, Annaert P, Augustijns P. Evaluation of fasted state human intestinal fluid as apical solvent system in the Caco-2 absorption model and comparison with FaSSIF. Eur J Pharm Sci 2015; 67: 126–135.

86 Bittermann K, Goss K-U. Predicting apparent passive permeability of Caco-2 and MDCK cell-monolayers: A mechanistic model. PLOS ONE 2017; 12: e0190319.

87 Flaten GE, Dhanikula AB, Luthman K, Brandl M. Drug permeability across a phospholipid vesicle based barrier: A novel approach for studying passive dif-fusion. Eur J Pharm Sci 2006; 27: 80–90.

88 di Cagno M, Bibi HA, Bauer-Brandl A. New biomimetic barrier PermeapadTM for efficient investigation of passive permeability of drugs. Eur J Pharm Sci 2015; 73: 29–34.

89 Hubatsch I, Ragnarsson EGE, Artursson P. Determination of drug permeability and prediction of drug absorption in Caco-2 monolayers. Nat Protoc 2007; 2: 2111–2119.

90 Frank KJ, Westedt U, Rosenblatt KM, Hölig P, Rosenberg J, Mägerlein M et al. What Is the Mechanism Behind Increased Permeation Rate of a Poorly Sol-uble Drug from Aqueous Dispersions of an Amorphous Solid Dispersion? J Pharm Sci 2014; 103: 1779–1786.

91 Keemink J, Bergström CAS. Caco-2 Cell Conditions Enabling Studies of Drug Absorption from Digestible Lipid-Based Formulations. Pharm Res 2018; 35: 74.

64

92 Ueda K, Higashi K, Limwikrant W, Sekine S, Horie T, Yamamoto K et al. Mechanistic Differences in Permeation Behavior of Supersaturated and Solu-bilized Solutions of Carbamazepine Revealed by Nuclear Magnetic Resonance Measurements. Mol Pharm 2012; 9: 3023–3033.

93 Sadhukha T, Layek B, Prabha S. Incorporation of lipolysis in monolayer per-meability studies of lipid-based oral drug delivery systems. Drug Deliv Transl Res 2018; 8: 375–386.

94 Dahan A, Miller JM. The Solubility–Permeability Interplay and Its Implications in Formulation Design and Development for Poorly Soluble Drugs. AAPS J 2012; 14: 244–251.

95 Porat D, Dahan A. Active intestinal drug absorption and the solubility-permea-bility interplay. Int J Pharm 2018; 537: 84–93.

96 Buckley ST, Fischer SM, Fricker G, Brandl M. In vitro models to evaluate the permeability of poorly soluble drug entities: Challenges and perspectives. Eur J Pharm Sci 2012; 45: 235–250.

97 Stewart AM, Grass ME, Brodeur TJ, Goodwin AK, Morgen MM, Friesen DT et al. Impact of Drug-Rich Colloids of Itraconazole and HPMCAS on Mem-brane Flux in Vitro and Oral Bioavailability in Rats. Mol Pharm 2017; 14: 2437–2449.

98 Pestieau A, Evrard B. In vitro biphasic dissolution tests and their suitability for establishing in vitro-in vivo correlations: A historical review. Eur J Pharm Sci 2017; 102: 203–219.

99 Li J, Tsinman K, Tsinman O, Wigman L. Using pH Gradient Dissolution with In-Situ Flux Measurement to Evaluate Bioavailability and DDI for Formulated Poorly Soluble Drug Products. AAPS PharmSciTech 2018; 19: 2898–2907.

100 Borbás E, Kádár S, Tsinman K, Tsinman O, Csicsák D, Takács-Novák K et al. Prediction of Bioequivalence and Food Effect Using Flux- and Solubility-Based Methods. Mol Pharm 2019; 16: 4121–4130.

101 Tsinman K, Tsinman O, Lingamaneni R, Zhu S, Riebesehl B, Grandeury A et al. Ranking Itraconazole Formulations Based on the Flux through Artificial Lipophilic Membrane. Pharm Res 2018; 35: 161.

102 Sironi D, Christensen M, Rosenberg J, Bauer-Brandl A, Brandl M. Evaluation of a dynamic dissolution/permeation model: Mutual influence of dissolution and barrier-flux under non-steady state conditions. Int J Pharm 2017; 522: 50–57.

103 Xu H, Shi Y, Vela S, Marroum P, Gao P. Developing Quantitative In Vitro–In Vivo Correlation for Fenofibrate Immediate-Release Formulations With the Biphasic Dissolution-Partition Test Method. J Pharm Sci 2018; 107: 476–487.

104 Jankovic S, O’Dwyer PJ, Box KJ, Imanidis G, Reppas C, Kuentz M. Biphasic drug release testing coupled with diffusing wave spectroscopy for mechanistic understanding of solid dispersion performance. Eur J Pharm Sci 2019; 137: 105001.

105 O’Dwyer PJ, Imanidis G, Box KJ, Reppas C. On the Usefulness of Two Small-Scale In Vitro Setups in the Evaluation of Luminal Precipitation of Lipophilic Weak Bases in Early Formulation Development. Pharmaceutics 2020; 12. doi:10.3390/pharmaceutics12030272.

106 O’Dwyer PJ, Litou C, Box KJ, Dressman JB, Kostewicz ES, Kuentz M et al. In vitro methods to assess drug precipitation in the fasted small intestine - a PEARRL review. J Pharm Pharmacol 2019; 71: 536–556.

65

107 Sironi D, Rosenberg J, Bauer-Brandl A, Brandl M. Dynamic dissolution-/per-meation-testing of nano- and microparticle formulations of fenofibrate. Eur J Pharm Sci 2017; 96: 20–27.

108 Ruponen M, Visti M, Ojarinta R, Laitinen R. Permeability of glibenclamide through a PAMPA membrane: The effect of co-amorphization. Eur J Pharm Biopharm 2018; 129: 247–256.

109 Korjamo T, Heikkinen AT, Waltari P, Mönkkönen J. The Asymmetry of the Unstirred Water Layer in Permeability Experiments. Pharm Res 2008; 25: 1714.

110 Kuentz M. Analytical technologies for real-time drug dissolution and precipi-tation testing on a small scale. J Pharm Pharmacol 2015; 67: 143–159.

111 Nir I, Lu X. In Situ UV Fiber Optics for Dissolution Testing – What, Why, and Where We Are After 30 Years. Dissolut Techn 2018; 25: 70–77.

112 Berthelsen R, Klitgaard M, Rades T, Müllertz A. In vitro digestion models to evaluate lipid based drug delivery systems; present status and current trends. Adv Drug Deliv Rev 2019; 142: 35–49.

113 Beig A, Miller JM, Lindley D, Carr RA, Zocharski P, Agbaria R et al. Head-To-Head Comparison of Different Solubility-Enabling Formulations of Etopo-side and Their Consequent Solubility–Permeability Interplay. J Pharm Sci 2015; 104: 2941–2947.

114 Alskär LC, Keemink J, Johannesson J, Porter CJH, Bergström CAS. Impact of Drug Physicochemical Properties on Lipolysis-Triggered Drug Supersaturation and Precipitation from Lipid-Based Formulations. Mol Pharm 2018; 15: 4733–4744.

115 Eerdenbrugh BV, Baird JA, Taylor LS. Crystallization tendency of active phar-maceutical ingredients following rapid solvent evaporation—classification and comparison with crystallization tendency from undercooled melts. J Pharm Sci 2010; 99: 3826–3838.

116 Alhalaweh A, Alzghoul A, Mahlin D, Bergström CAS. Physical stability of drugs after storage above and below the glass transition temperature: Relation-ship to glass-forming ability. Int J Pharm 2015; 495: 312–317.

117 Fagerberg JH, Tsinman O, Sun N, Tsinman K, Avdeef A, Bergström CAS. Dis-solution Rate and Apparent Solubility of Poorly Soluble Drugs in Biorelevant Dissolution Media. Mol Pharm 2010; 7: 1419–1430.

118 Alskär LC, Parrow A, Keemink J, Johansson P, Abrahamsson B, Bergström CAS. Effect of lipids on absorption of carvedilol in dogs: Is coadministration of lipids as efficient as a lipid-based formulation? J Control Release 2019; 304: 90–100.

119 Keemink J, Mårtensson E, Bergström CAS. Lipolysis-Permeation Setup for Simultaneous Study of Digestion and Absorption in Vitro. Mol Pharm 2019; 16: 921–930.

120 Dening TJ, Thomas N, Rao S, van Looveren C, Cuyckens F, Holm R et al. Montmorillonite and Laponite Clay Materials for the Solidification of Lipid-Based Formulations for the Basic Drug Blonanserin: In Vitro and in Vivo In-vestigations. Mol Pharm 2018; 15: 4148–4160.

121 Dening TJ, Rao S, Thomas N, Prestidge CA. Silica encapsulated lipid-based drug delivery systems for reducing the fed/fasted variations of ziprasidone in vitro. Eur J Pharm Biopharm 2016; 101: 33–42.

66

122 Edueng K, Mahlin D, Larsson P, Bergström CAS. Mechanism-based selection of stabilization strategy for amorphous formulations: Insights into crystalliza-tion pathways. J Control Release 2017; 256: 193–202.

123 Dening TJ, Joyce P, Webber JL, Beattie DA, Prestidge CA. Inorganic surface chemistry and nanostructure controls lipolytic product speciation and partition-ing during the digestion of inorganic-lipid hybrid particles. J Colloid Interface Sci 2018; 532: 666–679.

124 Åhlén M, Cheung O, Strømme M. Amorphous Mesoporous Magnesium Car-bonate as a Functional Support for UV-Blocking Semiconductor Nanoparticles for Cosmetic Applications. ACS Omega 2019; 4: 4429–4436.

125 Carlton RA. Polarized Light Microscopy. In: Pharmaceutical Microscopy. Springer: New York, NY, 2011.

126 Blaabjerg LI, Grohganz H, Lindenberg E, Löbmann K, Müllertz A, Rades T. The Influence of Polymers on the Supersaturation Potential of Poor and Good Glass Formers. Pharmaceutics 2018; 10: 164.

127 Avdeef A, Tsinman O. Miniaturized Rotating Disk Intrinsic Dissolution Rate Measurement: Effects of Buffer Capacity in Comparisons to Traditional Wood’s Apparatus. Pharm Res 2008; 25: 2613–2627.

128 Andersson SBE, Alvebratt C, Bevernage J, Bonneau D, da Costa Mathews C, Dattani R et al. Interlaboratory Validation of Small-Scale Solubility and Disso-lution Measurements of Poorly Water-Soluble Drugs. J Pharm Sci 2016; 105: 2864–2872.

129 Tsinman K, Avdeef A, Tsinman O, Voloboy D. Powder Dissolution Method for Estimating Rotating Disk Intrinsic Dissolution Rates of Low Solubility Drugs. Pharm Res 2009; 26: 2093–2100.

130 Joyce P, Tan A, Whitby CP, Prestidge CA. The Role of Porous Nanostructure in Controlling Lipase-Mediated Digestion of Lipid Loaded into Silica Particles. Langmuir 2014; 30: 2779–2788.

131 Joyce P, J. Barnes T, J. Boyd B, A. Prestidge C. Porous nanostructure controls kinetics, disposition and self-assembly structure of lipid digestion products. RSC Adv 2016; 6: 78385–78395.

132 Nieva-Echevarría B, Goicoechea E, Manzanos MJ, Guillén MD. A method based on 1H NMR spectral data useful to evaluate the hydrolysis level in com-plex lipid mixtures. Food Res Int 2014; 66: 379–387.

133 Eliasen JN, Berthelsen R, Slot AL, Müllertz A. Evaluating side‐by‐side dif-fusion models for studying drug supersaturation in an absorptive environment: a case example of fenofibrate and felodipine. J Pharm Pharmacol 2020; 72: 371–384.

134 Borbás E, Sinkó B, Tsinman O, Tsinman K, Kiserdei É, Démuth B et al. Inves-tigation and Mathematical Description of the Real Driving Force of Passive Transport of Drug Molecules from Supersaturated Solutions. Mol Pharm 2016; 13: 3816–3826.

135 Ortiz C, Ferreira ML, Barbosa O, Santos JCS dos, Rodrigues RC, Berenguer-Murcia Á et al. Novozym 435: the “perfect” lipase immobilized biocatalyst? Catal Sci Technol 2019; 9: 2380–2420.

136 Phan S, Salentinig S, Hawley A, Boyd BJ. Immobilised Lipase for In Vitro Lipolysis Experiments. J Pharm Sci 2015; 104: 1311–1318.

67

137 Mosgaard MD, Sassene P, Mu H, Rades T, Müllertz A. Development of a high-throughput in vitro intestinal lipolysis model for rapid screening of lipid-based drug delivery systems. Eur J Pharm Biopharm 2015; 94: 493–500.

138 Lea T. Epithelial Cell Models; General Introduction. In: Verhoeckx K, Cotter P, López-Expósito I, Kleiveland C, Lea T, Mackie A et al. (eds). The Impact of Food Bioactives on Health: in vitro and ex vivo models. Springer International Publishing: Cham, 2015, pp 95–102.

139 Zhang P, Zardán Gómez de la Torre T, Forsgren J, Bergström CAS, Strømme M. Diffusion-Controlled Drug Release From the Mesoporous Magnesium Car-bonate Upsalite®. J Pharm Sci 2016; 105: 657–663.

140 Ozaki S, Kushida I, Yamashita T, Hasebe T, Shirai O, Kano K. Inhibition of Crystal Nucleation and Growth by Water-Soluble Polymers and its Impact on the Supersaturation Profiles of Amorphous Drugs. J Pharm Sci 2013; 102: 2273–2281.

141 Rath N, Haq H, Balendiran G. Fenofibric acid. Acta Crystallogr C 2005; 61: o81-4.

142 Van Speybroeck M, Williams HD, Nguyen T-H, Anby MU, Porter CJH, Au-gustijns P. Incomplete Desorption of Liquid Excipients Reduces the in Vitro and in Vivo Performance of Self-Emulsifying Drug Delivery Systems Solidi-fied by Adsorption onto an Inorganic Mesoporous Carrier. Mol Pharm 2012; 9: 2750–2760.

143 Dening TJ, Joyce P, Rao S, Thomas N, Prestidge CA. Nanostructured Mont-morillonite Clay for Controlling the Lipase-Mediated Digestion of Medium Chain Triglycerides. ACS Appl Mater Interfaces 2016; 8: 32732–32742.

144 Almoselhy RIM, Allam MH, El-Kalyoubi MH, El-Sharkawy AA. 1H NMR spectral analysis as a new aspect to evaluate the stability of some edible oils. Ann Agric Sci 2014; 59: 201–206.

145 Alonso-Salces RM, Holland MV, Guillou C. 1H-NMR fingerprinting to evalu-ate the stability of olive oil. Food Control 2011; 22: 2041–2046.

146 Warren DB, Benameur H, Porter CJH, Pouton CW. Using polymeric precipita-tion inhibitors to improve the absorption of poorly water-soluble drugs: A mechanistic basis for utility. J Drug Target 2010; 18: 704–731.

147 Raina SA, Zhang GGZ, Alonzo DE, Wu J, Zhu D, Catron ND et al. Enhance-ments and Limits in Drug Membrane Transport Using Supersaturated Solutions of Poorly Water Soluble Drugs. J Pharm Sci 2014; 103: 2736–2748.

148 McConnell EL, Basit AW, Murdan S. Measurements of rat and mouse gastro-intestinal pH, fluid and lymphoid tissue, and implications for in-vivo experi-ments. J Pharm Pharmacol 2008; 60: 63–70.

149 Sjögren E, Abrahamsson B, Augustijns P, Becker D, Bolger MB, Brewster M et al. In vivo methods for drug absorption – Comparative physiologies, model selection, correlations with in vitro methods (IVIVC), and applications for for-mulation/API/excipient characterization including food effects. Eur J Pharm Sci 2014; 57: 99–151.

150 Maurer JM, Schellekens RCA, Rieke HM van, Wanke C, Iordanov V, Stellaard F et al. Gastrointestinal pH and Transit Time Profiling in Healthy Volunteers Using the IntelliCap System Confirms Ileo-Colonic Release of ColoPulse Tab-lets. PLOS ONE 2015; 10: e0129076.

68

151 Koziolek M, Grimm M, Becker D, Iordanov V, Zou H, Shimizu J et al. Inves-tigation of pH and Temperature Profiles in the GI Tract of Fasted Human Sub-jects Using the Intellicap® System. J Pharm Sci 2015; 104: 2855–2863.

152 Zhang D, Lee Y-C, Shabani Z, Frankenfeld Lamm C, Zhu W, Li Y et al. Pro-cessing Impact on Performance of Solid Dispersions. Pharmaceutics 2018; 10. doi:10.3390/pharmaceutics10030142.

153 Riethorst D, Mols R, Duchateau G, Tack J, Brouwers J, Augustijns P. Charac-terization of Human Duodenal Fluids in Fasted and Fed State Conditions. J Pharm Sci 2016; 105: 673–681.

154 Sironi D, Rosenberg J, Bauer-Brandl A, Brandl M. PermeaLoopTM, a novel in vitro tool for small-scale drug-dissolution/permeation studies. J Pharm Biomed 2018; 156: 247–251.

155 Mudie DM, Shi Y, Ping H, Gao P, Amidon GL, Amidon GE. Mechanistic anal-ysis of solute transport in an in vitro physiological two-phase dissolution appa-ratus. Biopharm Drug Dispos 2012; 33: 378–402.

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 291

Editor: The Dean of the Faculty of Pharmacy

A doctoral dissertation from the Faculty of Pharmacy, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofPharmacy. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Pharmacy”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-425586

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2021