Loyola University Chicago Loyola University Chicago

Loyola eCommons Loyola eCommons

Dissertations Theses and Dissertations

2017

Characterization of Multi-Albumin Pegylated Complexes Characterization of Multi-Albumin Pegylated Complexes

Synthesized Using "Click" Chemistry as Drug Delivery Systems Synthesized Using "Click" Chemistry as Drug Delivery Systems

Jonathan Alejandro Hill Loyola University Chicago

Follow this and additional works at: https://ecommons.luc.edu/luc_diss

Part of the Biochemistry Commons

Recommended Citation Recommended Citation Hill, Jonathan Alejandro, "Characterization of Multi-Albumin Pegylated Complexes Synthesized Using "Click" Chemistry as Drug Delivery Systems" (2017). Dissertations. 2813. https://ecommons.luc.edu/luc_diss/2813

This Dissertation is brought to you for free and open access by the Theses and Dissertations at Loyola eCommons. It has been accepted for inclusion in Dissertations by an authorized administrator of Loyola eCommons. For more information, please contact ecommons@luc.edu.

This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License. Copyright © 2017 Jonathan Alejandro Hill

LOYOLA UNIVERSITY CHICAGO

CHARACTERIZATION OF

MULTI-ALBUMIN PEGYLATED COMPLEXES

SYNTHESIZED USING “CLICK” CHEMISTRY

AS DRUG DELIVERY SYSTEMS

A DISSERTATION SUBMITTED TO

THE FACULTY OF THE GRADUATE SCHOOL

IN CANDIDACY FOR THE DEGREE OF

DOCTOR OF PHILOSOPHY

PROGRAM IN CHEMISTRY AND BIOCHEMISTRY

BY

JONATHAN A. HILL

CHICAGO, IL

AUGUST 2017

Copyright by Jonathan A. Hill, 2017 All rights reserved.

iii

ACKNOWLEDGEMENTS

To the following people, I express my undying thanks and gratitude for their guidance, patience,

support, and assistance throughout this process.

Thank you, Dr. Kenneth Olsen, for everything you have done to help me grow as a

student, as a researcher, and as a person.

To my committee, Drs. Miguel Ballicora, Dali Liu, and Yi Gao, thank you for your

advice and assistance along the way.

Without the help of Jennifer Hendrickson, Matthew Rhodes, Dzenita Huskic, Jacob

Parker, and Ross Carpino, none of this would have been possible.

To my wife, my parents, my siblings, and my friends, thank you for your constant

encouragement and support in everything I do.

To my wife, Alisa, for all of her love and support.

v

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ........................................................................................................... iii LIST OF TABLES ....................................................................................................................... viii LIST OF FIGURES ....................................................................................................................... ix LIST OF ABBREVIATIONS ........................................................................................................ xi CHAPTER I: INTRODUCTION .....................................................................................................1 Albumin-Based Nanoparticles ......................................................................................................1 nabTM Technology and Abraxane® ............................................................................................3 Albumin-Drug Conjugates .........................................................................................................5 Serum Albumin .............................................................................................................................6 Physiology .................................................................................................................................6 Clinical Applications .................................................................................................................7 Structural Basis for Ligand Binding ..........................................................................................8 Binding sites ...........................................................................................................................9 Fatty acid sites ....................................................................................................................10 Cys34 .................................................................................................................................11 The N-terminus and the Multi-metal binding site ..............................................................11 Drug Site 2 (DS2) ..............................................................................................................11 Drug Site 1 (DS1) ..............................................................................................................14 Albumin Metabolism .....................................................................................................................15 Human vs. Bovine Serum Albumin ............................................................................................17 Amino Acid Correlation ..........................................................................................................18 Structural Homology of DS1 and DS2 ....................................................................................19 Polyethylene Glycol (PEG) .........................................................................................................20 PEG-Protein Nanoparticles ......................................................................................................21 PEG Metabolism ......................................................................................................................23 Toxicity ....................................................................................................................................23 Advantages of PEGylation .......................................................................................................24 “Click” Chemistry .......................................................................................................................26 Azide-Alkyne Cycloaddition (AAC) .......................................................................................27 Cu(I)-catalyzed AAC (CuAAC) ...........................................................................................27 Strain-promoted AAC (SPAAC) ..........................................................................................28 Thiol-ene Reactions .................................................................................................................29 Summary .....................................................................................................................................30 CHAPTER II: STATEMENT OF RESEARCH ............................................................................32 CHAPTER III: MODIFICATION OF BOVINE SERUM ALBUMIN WITH MAL-PEG8 AS A DRUG DELIVERY SYSTEM................................................................................................35 Introduction .................................................................................................................................35

vi

Materials and Methods ................................................................................................................39 Materials ..................................................................................................................................39 Methods....................................................................................................................................39 Fatty acid removal .................................................................................................................40 PEGylation of FAF-BSA ......................................................................................................40 Circular dichroism (CD) .......................................................................................................40 SDS-PAGE ...........................................................................................................................41 Analytical ultracentrifugation (AUC) ...................................................................................41 Fluorimetric analyses ............................................................................................................42 Results and Discussion ...............................................................................................................43 BSAx-te-PEG8 Yield ................................................................................................................43 Effects on Secondary Structure ................................................................................................45 BSAx-te-PEG8 Speciation ........................................................................................................46 Ligand Binding Capabilities ....................................................................................................49 Conclusions .................................................................................................................................54 CHAPTER IV: CONJUGATION OF BOVINE SERUM ALBUMIN ONTO PEG8 VIA STRAIN-PROMOTED AZIDE-ALKYNE CYCLOADDITION AS A DRUG DELIVERY SYSTEM .....................................................................................................................................55 Introduction .................................................................................................................................55 Materials and Methods ................................................................................................................62 Materials ..................................................................................................................................62 Methods....................................................................................................................................62 Fatty acid removal .................................................................................................................63 Activation of FAF-BSA with DIBO .....................................................................................63 PEGylation of DIBO-activated FAF-BSA ............................................................................64 PEGylation optimization .......................................................................................................64 Fluorescent labeling ..............................................................................................................65 Circular dichroism (CD) .......................................................................................................65 SDS-PAGE ...........................................................................................................................66 Analytical ultracentrifugation (AUC) ...................................................................................66 Fluorimetric analyses ............................................................................................................67 Results and Discussion ...............................................................................................................68 Methodology Optimization ......................................................................................................68 Effects on Secondary Structure ................................................................................................70 Degree of Activation ................................................................................................................71 BSAx-R-123t-PEG8 Yields ......................................................................................................72 BSAx-R-123t-PEG8 Speciation ................................................................................................74 Ligand Binding Capabilities ....................................................................................................78 Conclusions .................................................................................................................................83 CHAPTER V: STUDIES ON THERMAL STABILITY AND THE EFFECTS OF PASTEURIZATION ON PEGYLATED MULTI-ALBUMIN COMPLEXES .........................84 Introduction .................................................................................................................................84 Materials and Methods ................................................................................................................88

vii

Materials ..................................................................................................................................88 Methods....................................................................................................................................88 Fatty acid removal .................................................................................................................89 PEGylation of FAF-BSA ......................................................................................................89 PEGylation of Cys34 .........................................................................................................89 PEGylation with SPAAC ...................................................................................................90 Purification of PEGylated BSA .........................................................................................90 Spectroscopic evaluation of aggregation ..............................................................................90 Circular dichroic evaluation of denaturation ........................................................................91 Pasteurization studies ............................................................................................................91 Circular dichroic evaluation of secondary structure ..........................................................91 Fluorimetric evaluation of naproxen binding ....................................................................92 Results and Discussion ...............................................................................................................93 Temperature of Aggregation (Tagg) ..........................................................................................93 Temperature of Denaturation (Tm) ...........................................................................................96 Effects of Pasteurization on Secondary Structure and Ligand Binding ...................................98 Conclusions ..................................................................................................................................102 CHAPTER VI: GENERAL CONCLUSIONS AND FUTURE RESEARCH ............................103 General Conclusions .................................................................................................................103 Future Research ........................................................................................................................105 REFERENCE LIST .....................................................................................................................108 VITA ............................................................................................................................................123

viii

LIST OF TABLES

Table 1. Contribution of the Major Classes of Serum Proteins to Osmotic Pressure. 7 Table 2. Example Ligands for Drug Site I and Drug Site II. 13 Table 3. Current Market-Available PEGylated Protein Products. 22 Table 4. Association Constants (Ka) and Number of Binding Sites (n) for BSAx-te-PEG8 for

Sulfamethoxazole and Naproxen. 50 Table 5. Nomenclature for the Conjugation of BSA onto PEG8 via Strain-Promoted Azide-

Alkyne Cycloaddition. 63 Table 6. Association Constants (Ka) and Number of Binding Sites (n) for BSAx-123t-PEG8

and BSAx-PEG4-123t-PEG8 for Sulfamethoxazole and Naproxen. 78 Table 7. Nomenclature for PEGylation of Bovine Serum Albumin via Thiol-Maleimide and

Strain-Promoted Azide-Alkyne Cycloaddition Reactions. 89 Table 8. Temperatures of Denaturation and Aggregation for PEGylated BSA Species. 94

ix

LIST OF FIGURES

Figure 1. Crystal structure of human serum albumin. 9 Figure 2. A selection of albumin binding sites. 10 Figure 3. Structure of Drug Site II in human serum albumin. 12 Figure 4. Structure of Drug Site I in human serum albumin. 14 Figure 5. Comparison of Drug Site II between human and bovine serum albumin. 19 Figure 6. Comparison of Drug Site I between human and bovine serum albumin. 20 Figure 7. Diagram of a PEGylated protein nanoparticle. 25 Figure 8. Cu(I)-catalyzed azide-alkyne cycloaddition. 27 Figure 9. Strain-promoted azide-alkyne cycloaddition. 29 Figure 10. Thiol-ene reaction. 30 Figure 11. Cys34, Drug Site I, and Drug Site II of bovine serum albumin. 37 Figure 12. Diagram of BSAx-te-PEG8. 38 Figure 13. Thiol-maleimide conjugation of bovine serum albumin onto PEG8. 39 Figure 14. Purification of thiol-maleimide PEGylated BSA via size exclusion

chromatography. 44 Figure 15. Circular dichroic spectra of FAF-BSA and BSAx-te-PEG8. 45 Figure 16. Electrophoretic separation of species of bovine serum albumin PEGylated via

thiol-maleimide click chemistry. 46 Figure 17. Analytical ultracentrifugation analysis of BSAx-te-PEG8. 48 Figure 18. Fluorimetric spectra and non-linear regression of BSAx-te-PEG8 in the presence

of naproxen. 51

x

Figure 19. Fluorimetric spectra and non-linear regression of BSAx-te-PEG8 in the presence of sulfamethoxazole. 53

Figure 20. Diagram of BSAx-R-123t-PEG8. 58 Figure 21. Conjugation of bovine serum albumin onto PEG8 via strain-promoted azide-alkyne

cycloaddition. 61 Figure 22. Optimization of the PEGylation procedure for DIBO-activated FAF-BSA. 69 Figure 23. Circular dichroic spectra of FAF-BSA and strain-promoted azide-alkyne

cycloaddition PEGylated products. 71 Figure 24. Purification of BSAx-R-123t-PEG8 via size exclusion chromatography. 73 Figure 25. Electrophoretic separation of species of bovine serum albumin PEGylated via

strain-promoted azide-alkyne click chemistry. 75 Figure 26. Analytical ultracentrifugation analysis of BSAx-123t-PEG8. 76 Figure 27. Analytical ultracentrifugation analysis of BSAx-PEG4-123t-PEG8. 77 Figure 28. Fluorimetric spectra and non-linear regression of BSAx-123t-PEG8 and

BSAx-PEG4-123t-PEG8 in the presence of naproxen. 80 Figure 29. Fluorimetric spectra and non-linear regression of BSAx-123t-PEG8 and

BSAx-PEG4-123t-PEG8 in the presence of sulfamethoxazole. 82 Figure 30. Change in absorbance as a function of temperature for the determination of Tagg. 95 Figure 31. Temperature-induced changes in circular dichroic spectra. 96 Figure 32. Change in helicity as a function of temperature for the determination of Tm. 97 Figure 33. Helicity after pasteurization in the absence and presence of naproxen. 99 Figure 34. Non-linear regression analyses of fluorimetric assays of PEGylated bovine serum

albumin before and after pasteurization. 101

xi

LIST OF ABBREVIATIONS

123t - 1,2,3-triazole AAC - Azide-alkyne cycloaddition AUC - Analytical ultracentrifugation βME - β-mercaptoethanol BSA - Bovine serum albumin BSAx-123t-PEG8 - Multiple BSA attached to PEG8 via 123t linkage BSAx-PEG4-123t-PEG8 - Multiple BSA attached to PEG8 via 123t linkage with PEG4 spacer BSAx-R-123t-PEG8 - Multiple BSA attached to PEG8 via 123t linkage with an unspecified spacer BSAx-R-PEG8 - Multiple BSA attached to PEG8 via an unspecified linkage BSAx-te-PEG8 - Multiple BSA attached to PEG8 via thioether linkage CD - Circular dichroism CrEL - Cremophor® EL CuAAC - Cu(I)-catalyzed azide-alkyne cycloaddition DDS - Drug delivery system DIBO-BSA - BSA activated by DIBO-NHS DIBO-NHS - Dibenzocyclooctyne-N-hydroxysuccinimidyl ester with no additional PEG spacer DIBO-PEG4-BSA - BSA activated by DIBO-PEG4-NHS DIBO-PEG4-NHS - Dibenzocyclooctyne-PEG4-N-hydroxysuccinimidyl ester

xii

DIBO-R-BSA - BSA activated by DIBO-R-NHS DIBO-R-NHS - Dibenzocyclooctyne-N-hydroxysuccinimidyl ester with an unspecified spacer DS1 - Drug Site I; Sudlow's Site I DS2 - Drug Site II; Sudlow's Site II FA5 - Fatty acid binding site 5 FAF-BSA - Fatty acid-free BSA FcRn - Neonatal Fc receptor Fluor-NHS - Fluorescein-5-EX N-hydroxysuccinimide ester FPLC - Fast protein liquid chromatography FrInt - Fractional intensity GdmCl - Guanidinium chloride gp18 - 18 kDa glycoprotein gp30 - 30 kDa glycoprotein gp60 - Albondin; 60 kDa glycoprotein HSA - Human serum albumin IgG - Immunoglobulin G Ka - Association constant Mal-PEG8 - Maleimido-functionalized PEG8 MOPS/NaCl - 50 mM 3-(N-Morpholino)propanesulfonic acid /0.15 M NaCl pH 7.2 buffer n - Number of binding sites N3-PEG8 - Azido-functionalized PEG8 nabTM - nanoparticle albumin-based

xiii

NAT - N-acetyl-L-tryptophan NHS - N-hydroxysuccinimidyl ester NP - Nanoparticle NPX - Naproxen NTS - N-terminus PC-DACTM - Preformed Conjugate-Drug Affinity Complex PEG - Polyethylene glycol PEG__(k) - __ (k)Da polyethylene glycol PEG8 - 8-armed polyetheylene glycol PTX - Paclitaxel SDS-PAGE - Sodium dodecyl sulfate polyacrylamide gel electrophoresis SEC - Size exclusion chromatography SMZ - Sulfamethoxazole SPAAC - Strain-promoted azide-alkyne cycloaddition Tagg - Temperature of aggregation te - Thioether Tm - Temperature of denaturation

1

CHAPTER I

INTRODUCTION

Since the inception of nanotechnology in the 1980s, nanoparticles (NP) have been explored as

more efficient and less toxic drug delivery systems (DDS). Many of these approaches have used

organic polymers as their core structures such as cyclodextrin nanosponges (Torne, Darandale,

Vavia, Trotta & Cavalli, 2013) and biodegradable polymers (Vila, Sánchez, Tobío, Calvo &

Alonso, 2002) while others synthesized homologs of naturally occurring molecules such as

liposomes (Batist et al., 2001) and self-assembling peptides (Boopathy & Davis, 2014). Another

group of NPs have used protein cores around which to build their DDSs. Perhaps the most

intriguing are albumin-based NPs.

Albumin provides an excellent core around which to design a DDS. It is biodegradable,

nontoxic, and easily customizable due to a large number of accessible functional groups on its

surface. More importantly, albumin is a natural mode of transport for both endogenous and

exogenous hydrophobic molecules (Elsadek & Kratz, 2012; Elzoghby, Samy & Elgindy, 2012;

Kratz, 2008; Peters, 1996).

Albumin-Based Nanoparticles

Several approaches have successfully produced nano-scale albumin-based DDSs.

Discussed below are just a few to demonstrate the breadth of methodologies developed.

Desolvation, developed by the Langer group, uses organic solvents such as ethanol and

acetone to dehydrate drug-loaded albumins. Sedimented protein is polymerized via a cross-

2

linking agent such as glutaraldehyde to provide stability before re-hydration (Langer et al., 2008;

Weber, Coester, Kreuter & Langer, 2000). Human serum albumin (HSA) NPs were prepared

using desolvation techniques and loaded with the chemotherapeutic doxorubicin. The NPs were

further modified by the addition of trastuzumab as a targeting mechanism. Targeted NPs showed

a cell viability of 20.1%, suggesting successful uptake and drug release (Anhorn, Wagner,

Kreuter, Langer & von Briesen, 2008).

Spray-drying is accomplished by passing a protein solution through a wire mesh at

elevated temperatures to generate a powder. Smooth, spherical bovine serum albumin (BSA)

NPs were successfully prepared at temperatures between 80-120oC in the presence of

Tween®-80. The morphology of the NPs was dependent on the surfactant content. Tween-80,

often used as an excipient in protein formulations, produced the most homogenous and spherical

NPs. NP size was most dependent on mesh size. The particle size increased as mesh size

increased. A 4.0 µm mesh generated NPs 733 nm in diameter while those from a 7.0 µm mesh

were 2609 nm (Lee, Heng, Ng, Chan & Tan, 2011).

Thermal gelation can be viewed as a two-phase process. Albumins are thermally

denatured and then non-covalently linked via electrostatic and hydrophobic/hydrophilic

interactions (Boye, Alli & Ismail, 1996). This process was successfully applied to the loading of

BSA-dextran and BSA-dextran-chitosan NPs with doxorubicin. Covalent BSA-dextran

conjugates were held at 80oC in the presence of chitosan. The resulting NPs consisted of chitosan

and dextran arms stretching out from an albumin core. Administration of doxorubicin-loaded

NPs to cancerous mice showed an increased survivability rate as compared to free doxorubicin

indicating an increased cell permeability for the conjugate (Qi, Yao, He, Yu & Huang, 2010).

Each of these systems has a distinct advantage in that they each provide a degree of fine control

3

over the final particle size via alteration of easily monitored parameters (pH, reaction time,

solvent content, reagent concentration, equipment used).

The drawbacks of each approach provide cause for concern. Reliance on organic solvents

and surfactants in the production of materials to be administered to living beings must be

minimized to avoid toxicity. Thermal denaturation and a highly-altered surface from extensive

cross-linking may elicit an immune response as well as potentially alter biological activity.

Necessarily high temperatures place limitations on potential drug candidates.

nabTM Technology and Abraxane®

Abraxane®, a paclitaxel (PTX) formulation, is the most successful albumin-based NPs to

date. Approved in 2005 by the US Food and Drug Administration for the treatment of metastatic

breast cancer, it is recognized as the first protein-based NP on the market. American Biosciences,

Inc., now Celgene Corporation, developed a method for producing nanoparticle albumin-based

(nabTM) carriers of which Abraxane was the first application (Desai, 2008).

The passage of an aqueous stream of HSA and hydrophobic drug through a high-pressure

jet forms the nab NP. Albumins adsorb to one another to which drugs preferentially bind. Upon

administration, the 130-nm shell dissociates (Elsadek & Kratz, 2012; Elzoghby et al., 2012;

Kratz, 2008).

Abraxane is the trade name for the complexation of PTX and nab albumin. This product

was developed as a response to the problems associated with Taxol®, the original formulation for

stand-alone PTX. In the absence of nab, PTX requires Cremophor® EL (CrEL), a poly-

ethoxylated castor oil, for solubilization. This necessitates infusion times of up to three hours and

pre-medication with corticosteroids and antihistamines to reduce the chance of hypersensitivity

reactions. Additionally, PTX delivery was hindered by micellular CrEL formation. Upon

4

administration of Taxol, CrEL aggregates, creating a hydrophobic environment into which PTX

readily diffuses. CrEL toxicity can be quite severe as it presents as neutropenia and sometimes

permanent neuropathy (Hawkins, Soon-Shiong & Desai, 2008).

The much simpler formulation of Abraxane, consisting of only HSA and PTX, allows for

shorter infusion times (30 minutes), no pre-medicating, a reduced risk of hypersensitivity

reactions, and no toxicity (Hawkins et al., 2008). Studies have reported 33% increase of

intracellular PTX concentrations with Abraxane (Gradishar et al., 2005). Dissociation of

Abraxane upon administration results in accumulation of albumins into tumor cells via both the

passive targeting phenomenon of the enhanced permeation and retention effect and active

targeting via albondin (gp60) endocytosis. It was determined that cancerous cells are able to

acquire up to 17% of a single dose of 111In-labeled albumin, indicating a degree of preference by

tumors for albumin (Stehle et al., 1999). By combining albumin’s two-fold tumor targeting with

the ability of Abraxane to house multiple PTX, an increased drug concentration was

unsurprising.

This same housing ability allows for Abraxane to be administered at a 50% higher dose

than Taxol. PTX introduced into the bloodstream via the Taxol is unbound and free to interact

with both cancerous and healthy tissues. nab technology provides a means of sequestering the

chemotherapeutic until it is inside its intended target. This sequestration results in decreased

toxicity, higher dosage rates, and a higher therapeutic window for Abraxane.

The wider application of nab technology is questionable. It has proven to be an effective

chemotherapy due to its high drug load, solubility, and improved administration. The gp60

targeting mechanism is not universal for all drugs and the dissolution of Abraxane would require

a targeting mechanism for each individual protein to be as effective for non-chemotherapeutics.

5

Additionally, dissolution of the nab NPs in the bloodstream results in the infusion of a large

quantity of individual albumin molecules which may greatly impact the osmotic pressure.

Albumin-Drug Conjugates

The techniques described result in polymerized albumins which provide a framework to

which the drug is non-covalently attached. Albumin-drug conjugates remain monomeric and

provide a means of solubilizing and increasing circulation time for bound ligands. Two

approaches have been employed to produce these conjugates. In the first, a drug is covalently

attached to albumin via a surface residue. In the second, the drug is covalently attached to a

ligand of albumin. Upon administration of this complex, the ligand binds to a high affinity site of

endogenous albumin, effectively non-covalently binding the drug.

The first albumin-drug conjugate to reach phase I/II clinical studies was a complex with

the antimitotic cancer drug methotrexate. Similar to PTX, methotrexate suffered from poor

solubility and severe toxicity associated with the drug itself and components of the formulation.

For the conjugate, methotrexate was covalently attached to HSA via lysine residues (Stehle,

Sinn, Wunder, Schrenk, Schütt et al., 1997). Clinical results for this complex were somewhat

promising in that three out of 17 patients showed a positive response to therapy. Other factors,

such as stomatitis, limited the dosage rates, and potentially limited the efficacy of treatment

(Kratz, 2008).

Long-lasting insulin (Levemir®) for the treatment of both type 1 and type 2 diabetes is a

prime example of the second type of albumin-drug conjugate. Novo Nordisk, Inc. developed a

method for producing recombinant human insulin in which the C-terminus had been shortened

by elimination of the B-chain terminal threonine. This excision allowed access to an adjacent

lysine which was covalently attached to myristic acid. Upon administration, the insulin-myristic

6

acid conjugate binds to albumin, resulting in an increase in circulatory half-life from four to six

minutes for native recombinant insulin to five to seven hours for the conjugate (Elsadek & Kratz,

2012; Keating, 2012).

Albumin-drug conjugates are attractive in their reasonably simple production protocols,

requiring no modifications of the protein outside of the attachment of the drug. While attachment

to the albumin provides an answer for solubilization, it does not necessarily protect the ligand

from the host. In the event of an immune response, the drug is susceptible to attack from

immunogenic agents which could result in early elimination, degradation of the therapeutic, or

dissection of the conjugate. If the drug were to be cleaved from the albumin before it reaches the

site of action, a potentially toxic molecule would be released into the bloodstream to interact

with healthy tissues.

Serum Albumin

Physiology

The robust nature of albumin, a highly-conserved protein found in the serum of all

mammals, has singled it out as a potential DDS component. It is a 66 kDa monomer produced by

the liver in quantities up to 12 g/day with a circulatory half-life of 20 days. Albumin can be found

in high quantities in skin, muscle, liver, and the gut as well as almost all bodily fluids (Lundblad,

2012; Peters, 1996). Its high production rate and long half-life make albumin the most abundant

protein in the serum.

Albumin is able to play a diverse set of roles in the body. Arguably its most important

function, albumin is the body’s main control on osmotic pressure (Table 1). Determined by the

number of particles in a system, not their size, osmotic pressure is the driving force between

7

tissues and blood. Despite albumin being many times smaller than fibrinogen, it is found at such

a high concentration that it accounts for 79% of the osmotic pressure (Guyton & Hall, 2006).

Table 1. Contribution of the Major Classes of Serum Proteins to Osmotic Pressure (Guyton & Hall, 2006).

Protein Class

Average MW

(kDa)

Serum Concentration

(mM)

Osmotic Pressure (mm Hg)

% Total Osmotic Pressure

Albumin 66 0.68 21.8 79 Globulins 140 0.17 6 20 Fibrinogen 400 0.0075 0.2 1

This dependence on albumin as an osmotic regulator must be taken into account when

designing a DDS. A drastic increase in albumin concentration associated with a high dose of

albumin-bound drug would be accompanied by a drastic increase in osmotic pressure. This spike

would cause down-regulation of albumin gene-expression to compensate (Pietrangelo, Panduro,

Chowdhury & Shafritz, 1992). Decreased levels of native, unmodified albumin may lead to a

build-up of fatty acids, heavy metals, drugs, and their metabolites.

Clinical Applications

Albumin was first explored as a therapeutic agent during World War II. Albumin

solutions were administered as plasma expanders to wounded soldiers to minimize the effects of

blood loss on the battlefield. The first purification method involved multiple rounds of low-

temperature precipitations with ethanol to separate the protein components of blood. Albumin

was the fifth precipitate in the process and was originally deemed “Cohn Fraction V” (Cohn et

al., 1946; Peters, 1996). The peptide sequence of BSA was first published by Brown (1975)

while that of HSA was later independently reported by the Brown and Kostka labs (Behrens,

Spiekerman & Brown, 1975; Meloun, Morávek & Kostka, 1975). Since these initial works,

8

albumin has become a staple of hospitals and clinics. Unmodified albumin is still used as a

plasma expander to aid in post-operative recovery and for treatment of shock and burns. It is also

used to treat patients suffering from hypoalbuminemia, specifically those with acute liver failure.

In these cases, the liver is unable to produce the requisite amount of albumin and the body

requires supplementation to maintain normal functionality (Lundblad, 2012). It has also been

developed as a protein-based resin for sealing wounds by CryoLife, Inc. BioGlue® contains

compartments of aqueous glutaraldehyde and BSA solutions. These are mixed in the syringe tip

during application and allowed to bond in situ. The resin is a matrix of highly cross-linked

albumin strong enough to hold tissues together during the healing process and safe enough to be

used internally ("CryoLife: BioGlue Instructions for Use," 2017).

Structural Basis for Ligand Binding

Albumins contain approximately 585 amino acids arranged in a highly helical heart-

shaped tertiary structure (Figure 1). Three similar domains (I, II, and III) are each subdivided

into subdomains A and B. Each domain is composed of 10 helices with six in subdomain A and

four in subdomain B. Each subdomain is connected by a long loop. Conformational flexibility is

conferred by helical bending and subdomain orientation is controlled by the movement of the

loops. These types of movements open and close the various binding sites dotted throughout the

structure.

While albumin is extremely flexible, the core structure is maintained via 17 disulfide

bridges. Subdomain A is stapled together by four of these while two bridges stabilize subdomain

B. The sole exception is Subdomain IA, which contains only three cystine bridges (Bujacz,

2012).

9

Figure 1. Crystal structure of human serum albumin. (a) Space-filling model and (b) Ribbon structure highlighting the three domains and heart-shaped structure. Domains I (blue), II (yellow), and III (red) are held together by 17 disulfide bridges (gray). PDB ID: 1E7I (Bhattacharya, Grüne & Curry, 2000). All crystal structure images were prepared using UCSF Chimera (Pettersen et al., 2004) and rendered with POV-Ray ("Persistence of Vision Pty. Ltd.: Persistence of Vision (TM) Raytracer," 2004).

Binding sites. In addition to its role in controlling osmotic pressure, another crucial

function of albumin is its ability to bind a wide variety of ligands including fatty acids, heavy

metals, drugs, and drug by-products (Figure 2). Discussed here are just a few of the many

binding sites on albumin.

10

Figure 2. A selection of albumin binding sites. Fatty Acids are represented as yellow hydrophobic tails with red hydrophilic heads. PDB ID: 1E7I (Bhattacharya et al., 2000).

Fatty acid sites. Several sites have been identified for long- and medium-chain fatty acids

(Ashbrook, Spector & Fletcher, 1972; Ashbrook, Spector, Santos & Fletcher, 1975; Bhattacharya

et al., 2000). A preference for saturated, long-chain fatty acids was determined by examining

fatty acid content and association constants. Oleic (18:1, 33%), palmitic (16:0, 25%) and linoleic

acid (18:2, 20%) made up a majority of the bound fatty acid content of HSA. Five others

accounted for another 12.5% and ranged between C14-C20 and zero to four degrees of

unsaturation (A. Saifer & Goldman, 1961). Examining a series of fatty acids ranging from C6-

C18 and zero to two degrees of unsaturation confirmed this preference for saturated, long-chain

fatty acids. Affinity increased with carbon count and decreased as double bond content increased

(Spector, 1975). As in the case of Levemir, this binding site allows for tailoring of a drug

conjugate to take advantage of an innate function of albumin to form a DDS in vivo. The lack of

modifications to the protein allow it to maintain functionality and avoid degradation.

11

Cys34. A single free thiol found at residue 34 has been reported to participate in covalent

interactions with a variety of drugs and their metabolites, including bucillamine derivatives

(Narazaki, Hamada, Harada & Otagiri, 1996), nitrogen mustards (Noort, Hulst & Jansen, 2002),

and the reactive acetaminophen metabolite, N-acetyl-p-benzoquinoneimine (LeBlanc, Shiao,

Roy & Sleno, 2014). This residue is on the surface of the molecule, slightly buried in a shallow

cleft between two α-helices positioned well away from the two main drug binding sites. The

accessibility of Cys34 provides a great deal of control over conjugation of albumin to other

molecules. Given that it is the only sulfhydryl available, this allows for chemoselective, covalent

surface modification.

The N-terminus and the Multi-metal binding site. Many species of albumin, specifically

HSA and BSA, employ the X-Y-His motif at the N-terminus (NTS). This configuration allows

for high affinity binding of Cu(II) on the order of 1 pM (Rózga, Sokołowska, Protas & Bal,

2007). This strong association between the NTS and Cu(II) must be taken into account during

DDS development. Cu(I) is a catalyst for a set of biorthogonal reactions involving azides and

alkynes which will be discussed in more detail in other sections.

A secondary metal interaction site has been found. Deemed the Multi-metal binding site,

this area is known to interact with Cu(II), Ni(II), Zn(II), and Cd(II), however, none of these

species bind with high affinity. Typically, metal cations are not bound until the NTS has been

fully saturated (Bal, Christodoulou, Sadler & Tucker, 1998). Due to low affinities and little

impact on the NTS, this site has been largely ignored in favor of a better understanding of the

high-affinity NTS interaction with Cu(II).

Drug Site II (DS2). The two main drug binding sites of albumin were initially

characterized by Sudlow et al. (1975, 1976). One site has been deemed Sudlow Site II or Drug

12

Site II. It is a small, compact single chamber with a patch of polar residues at the entrance to the

cavity (Figure 3). Tyr411, Arg410 and Ser489 line the mouth of the site, serving as the only

polar residues in the entire space.

Figure 3. Structure of Drug Site II in human serum albumin. Polar residues (green) lining the mouth of the cavity are the only polar features of DS2. PDB ID: 2VDB (Lejon, Cramer & Nordberg, 2008).

A lack of hydrophilicity in the pocket is mirrored in its ligands which are small

hydrophobics with a terminal polar group (Table 2) (Yamasaki, Chuang, Maruyama & Otagiri,

2013). The hydrophilic head, typically a carboxylic acid, is used to anchor the molecule in the

pocket. This also leads to a high degree of stereospecificity for ligand orientation. By no means a

flexible site, DS2 has shown the ability to make slight adjustments to accommodate side-chain

phenyl groups (Ghuman et al., 2005; Yamasaki et al., 2013).

13

Tabl

e 2.

Exa

mpl

e Li

gand

s fo

r Dru

g Si

te I

and

Dru

g Si

te II

(Ghu

man

et a

l., 2

005;

Q. W

ang,

Zha

ng &

Ji,

2014

; Yam

asak

i et a

l., 2

013)

.

DS1

Lig

ands

D

S2 L

igan

ds

14

Drug Site I (DS1). Drug Site I or Sudlow Site I is a large, multi-chamber, mostly apolar

pocket with a few polar features that are key to the binding of ligands (Figure 4). A polar

environment is formed at the entrance to the cavity by Lys195, Lys199, Arg218, and Arg222.

Three residues are buried deep into the bottom of the site which help to orient molecules

(Tyr150, His242, Arg257). Ile264 splits the large chamber, giving a total of three subsites for

binding. Additionally, Trp214 resides near the mouth of the cavity. While this may participate in

hydrogen bonding with ligands and play a role sterically, it also serves as an innate fluorescent

probe for studying drug interactions (Ghuman et al., 2005; Yamasaki et al., 2013).

Figure 4. Structure of Drug Site I in human serum albumin. Polar residues (pink) at the entrance and back (orange) of the cavity are able to hydrogen bond with ligands. Ile264 (purple) divides the chamber into smaller subsites. Trp214 (red) provides an intrinsic probe for ligand binding. PDB ID: 2VDB (Lejon et al., 2008).

DS1 has been called a “large and flexible region.” It is able to bind a wide range of

ligands, often more than one at a time, and is not stereospecific (Kragh-Hansen, 1988).

Molecules that show a high affinity for DS1 are typically large, flat, and aromatic with at least

one terminal and one internal polar functional group (Table 2). Flat molecules are preferred for

15

DS1 due to Leu238 and Ala291 which are able to snugly fit the molecule into place. Terminal

polar moieties are needed to interact with the entryway residues while the internal polar group

interacts with the buried polar residues (Ghuman et al., 2005; Yamasaki et al., 2013).

While each of these sites exhibits preferential binding for certain types of molecules, they

are by no means exclusive for those molecules. Once the preferred site is filled, additional

molecules are often able to bind with lower affinities to other sites. In the case of naproxen,

crystal structures revealed three separate binding sites. DS2 was the preferred site as it was the

only site consistent between bovine, equine, and leporine serum albumins. Additional naproxen

were found in a fatty acid site and DS1 (Bujacz, Zielinski & Sekula, 2014).

Albumin Metabolism

Albumin has a half-life of 20 days, passing through circulation thousands of times before

being lysosomally degraded. The pathways used to evade early degradation, delivery of its

ligand payload, and ultimate catabolism provide useful knowledge for the informed design of

DDSs.

The Megalin/Cubilin complexes have been implicated as role players in saving albumin

from elimination through the kidneys. Cubilin, an extracellular protein, utilizes the

transmembrane protein megalin to shuttle albumin into the intracellular space. Mice missing one

or both of these proteins showed decreased albumin reabsorption, strongly suggesting a role in

albumin recycling (Christensen, Birn, Storm, Weyer & Nielsen, 2012).

Similar conclusions were drawn in the case of the neonatal Fc receptor (FcRn).

Immunoglobulin G (IgG) binds FcRn as a means to cross the placental barrier. Co-elution of the

IgG-FcRn complex unexpectedly coupled with albumin was the first indication of interaction.

Studies with FcRn deficient mice found a 40% decrease in albumin serum concentration, owing

16

to the elimination of albumin rather than the transmembrane movement provided by FcRn.

Under acidic conditions, such as those experienced throughout the lysosomal pathway, FcRn

binds with high affinity to both albumin and IgG and escorts them to a more neutral environment

(Kim et al., 2006; Larsen, Kuhlmann, Hvam & Howard, 2016; Merlot, Kalinowski &

Richardson, 2014). FcRn clearly plays a key role in the long half-life of albumin by way of

rescuing the protein from degradation.

Glycoproteins of 30 kDa (gp30) and 18 kDa (gp18) have been shown to play a role in

albumin catabolism. While native albumin appears able to escape elimination through the

kidneys via the Megalin/Cubilin complex and FcRn pathways, damaged albumins are tagged by

gp30 and gp18. The affinities of these proteins for altered albumins are 1000-fold stronger than

for native albumin. Alterations are often due to oxidative damage or non-enzymatic

glycosylation. Upon binding, the sequestered albumin is ushered to sites of degradation (Larsen

et al., 2016; Schnitzer, Sung, Horvat & Bravo, 1992). The albumin is lysosomally catabolized

and the amino acids re-enter the cycle of protein synthesis (Stehle, Sinn, Wunder, Schrenk,

Stewart et al., 1997).

The most extensively studied albumin receptor is gp60. This receptor is common in

continuous endothelia from various tissues. It selectively binds albumin and with the help of

caveolin-1, endocytoses the protein (Schnitzer, 1992). In healthy cells, albumin uses this

pathway to migrate between the bloodstream and tissue (Merlot et al., 2014).

This is not the case in cancerous cells. The fate of albumin and the mechanism of

chemotherapy release are not fully known. Studies focused on the cellular uptake of Abraxane

found that the endocytosis of albumin was accompanied by a spike in the concentration of

osteonectin. It is postulated that osteonectin binds to albumin, tagging it for degradation which

17

releases paclitaxel in the process (Desai, 2008; Hawkins et al., 2008; Tiruppathi, Song,

Bergenfeldt, Sass & Malik, 1997).

In addition to gp60 endocytosis, the enhanced permeation and retention effect must be

taken into account when considering cancerous cells. This is the passive trapping of

macromolecules in tumor tissues. The cells are growing rapidly and the lymphatic system is

unable to maintain sufficient drainage which limits the ability of molecules to exit the tumor

cells. Macromolecules, especially albumins, build up in the cells and are ultimately catabolized

for protein synthesis and energy.

Studies have explored the ability of modified albumins to be endocytosed into tumor

cells. One study covalently attached a low molecular weight protamine fragment to

chemotherapeutic-loaded BSA NPs to facilitate transfer across the blood-brain barrier (Lin et al.,

2016). Despite the modification, BSA was still successfully taken into tumor cells.

Both pathways have been exploited as ingress for chemotherapeutics into cancerous

tissues. The advantages are two-fold. The chance of toxicity due to free ligand is greatly reduced

when bound to albumin which also provides an innate targeting mechanism.

Human vs. Bovine Serum Albumin

To date, most clinical-use albumin formulations have been HSA-based. However, BSA

provides an alternative to address many concerns associated with a human blood-borne protein.

The main concern with using any human-derived blood product is the possibility of

disease. Human immunodeficiency virus, Hepatitis, Creutzfeldt-Jakob Disease, and West Nile

Virus are all potential pathogens that may be transferred with blood products. Despite the

stringent protocols employed to eliminate and avoid these blood-borne diseases, the stigma of a

formulation containing a human blood-based protein could decrease patient compliance.

18

Additionally, the supply and cost of HSA is a cause for concern. The two main avenues

for gathering HSA are donations and recombinant technology. Donation based materials are

often in limited supply and can cause a fluctuation in availability. Recombinant human albumin

has been developed with rice (He et al., 2011) and yeast (Recombumin® from Albumedix Ltd.)

(Bosse et al., 2005). These have shown comparable safety, tolerability, pharmacokinetics, and

pharmacodynamics to native HSA but are quite expensive.

As a by-product of the cattle industry, BSA provides a much more cost-effective and

readily available substitute for HSA. A BLAST® search shows 76% homology between HSA and

BSA, suggesting a high degree of structural similarity.

Amino Acid Correlation

Comparison of amino acid composition shows strong correlations for key residue types.

The solubility of albumin is tied to the number of external charged amino acids. Approximately

31% of the residues on both species are charged, confirming similar solubilities.

A single free thiol can be found in both albumins. Of the 35 cysteine residues, 34 are

paired to make disulfide bridges that stabilize the tertiary structure of the three domains. This

unpaired cysteine allows for more controlled surface modification at a single, specific site.

Functionalization via surface modifications is a key attractant to albumin as a DDS.

Lysine residues are often the sites targeted as their terminal amino groups are able to react with a

variety of functional groups, allowing flexibility in the type of conjugation chemistry used. Both

species contain 59 lysines, each of which is a potential location to anchor a targeting molecule,

covalent drug conjugate, cross-linker, or polyethylene glycol chain.

19

Structural Homology of DS1 and DS2

A closer look at the Sudlow sites reveals a strong contrast between species. The polar

residues of DS2 in HSA are present in BSA as residues Tyr410, Arg409, and Ser488 (Figure 5).

Conservation of the key residues implies similar binding capabilities and preferences for similar

molecules.

Figure 5. Comparison of Drug Site II between human and bovine serum albumin. (a) DS2 of HSA. PDB ID: 2VDB (Lejon et al., 2008). (b) DS2 of BSA. PDB ID: 4OR0 (Bujacz et al., 2014). Polar residues (green) are conserved between species.

As in HSA, DS1 of BSA contains a single tryptophan, Trp213. The polar features buried

in the cavity of DS1 are consistent between HSA and BSA (Tyr149, His241, and Arg256). Some

discrepancies are found at the entrance to DS1. The polar cluster of HSA is closely mimicked in

BSA (Arg194, Arg198, Arg217, and Lys221) (Figure 6). However, the substitution of arginines

for lysines may cause differences in binding capacities. The distal guanidinium group of arginine

allows for more hydrogen-bonding and stronger ligand interactions. The delocalized charge,

however, weakens the bonds with individual molecules (Li, Vorobyov & Allen, 2013).

20

Sterically, the guanidinium R-group is bulkier than the terminal amine of lysine which may

hinder entry and exit from DS1 in BSA. Despite these discrepancies, interchange of these two

proteins should be possible.

Figure 6. Comparison of Drug Site I between human and bovine serum albumin. (a) DS1 of HSA. PDB ID: 2VDB (Lejon et al., 2008). (b) DS1 of BSA. PDB ID:4OR0 (Bujacz et al., 2014). External (pink) and internal (orange) polar residues are well-conserved between species. Additionally, Trp (red) provides a native fluorophore in both forms of albumin.

Polyethylene Glycol (PEG)

Polyethylene glycol (PEG), a water-soluble polymer, is used in a wide variety of

industries for a wide variety of applications from control of the viscosity of printer ink to an anti-

foaming agent in food preparation to laxatives as pre-operative patient preparation. In laboratory

use, it has been employed as a protein precipitating agent as well as a means to facilitate cell

fusion (Zalipsky & Harris, 1997).

PEG is available in a wide range of sizes and functional groups. Entire companies have

been built around the production and customization of PEGs tailored for specific uses ("Creative

PEGworks: PEG Products," 2017; "JenKem Technology: PEG Products," 2017). PEGs from as

21

small as 400 Da (PEG400) to as large as 40,000 Da (PEG40k) have been examined as possible

surfactants and DDS components.

PEG-Protein Nanoparticles

PEG has played a role in the development of several successful protein-based NPs (Table

3) (Haag & Kratz, 2006; "US Food and Drug Administration: Drug Approval Packages," 2017;

Veronese, Mero & Pasut, 2009). The regulatory approval of the first PEGylated enzymes,

Adagen® and Oncaspar®, in the 1990s was the turning point for the change of perspective

towards PEG as a building block for future DDSs rather than a mere additive to formulations

(Duncan & Veronese, 2009). As a result, more research has been done on the metabolism,

toxicity, and benefits of PEGylation.

22

Tabl

e 3.

Cur

rent

Mar

ket-A

vaila

ble

PEG

ylat

ed P

rote

in P

rodu

cts (

Haa

g &

Kra

tz, 2

006;

"U

S Fo

od a

nd D

rug

Adm

inis

tratio

n: D

rug

App

rova

l Pac

kage

s," 2

017;

Ver

ones

e et

al.,

200

9).

Trad

e N

ame

PEG

MW

(c

ount

) Pr

otei

n In

dica

tion

Com

pany

A

ppro

val

Yea

r

Ada

gen®

5

kDa

(11-

17)

Ade

nosi

ne d

eam

inas

e Se

vere

com

bine

d im

mun

odef

icie

ncy

dise

ase

Sigm

a Ta

u Ph

arm

aceu

tical

s Inc

. 19

90

Onc

aspa

r®

5 kD

a (6

9-82

) A

spar

agin

ase

Leuk

emia

Si

gma

Tau

Phar

mac

eutic

als I

nc.

1994

PEG

-IN

TRO

N®

12

kD

a In

terf

eron

α2b

H

epat

itis C

Sc

herin

g-Pl

ough

C

orpo

ratio

n 20

01

Neu

last

a®

20 k

Da

Gra

nulo

cyte

-col

ony

stim

ulat

ing

fact

or

Neu

trope

nia

Am

gen

Inc.

20

02

PEG

ASY

S®

40 k

Da

Inte

rfer

on α

2a

Hep

atiti

s C

F. H

offm

ann-

La R

oche

Lt

d 20

02

Som

aver

t®

5 kD

a (4

-6)

G

row

th h

orm

one

anta

goni

st

Acr

omeg

aly

Phar

mac

ia &

Upj

ohn

Com

pany

20

03

Mac

ugen

®

20 k

Da

(2)

Ant

i-VEG

F ap

tam

er

Age

-rel

ated

mac

ular

de

gene

ratio

n V

alea

nt P

harm

aceu

tical

s In

tern

atio

nal I

nc.

2004

Mirc

era®

30

kD

a Er

ythr

opoi

etin

A

nem

ia a

ssoc

iate

d w

ith

chro

nic

kidn

ey d

isea

se

F. H

offm

ann-

La R

oche

Lt

d 20

07

Cim

zia®

40

kD

a (2

) A

nti-T

NF

Fab’

R

heum

atoi

d ar

thrit

is a

nd

Cro

hn’s

dis

ease

U

CB

, Inc

. 20

08

23

PEG Metabolism

Heavy PEGs (MW > 1000 Da) are minimally absorbed by the gastrointestinal tract. For

this reason, the oral administration of PEG conjugates is ineffective. The harsh environment of

the digestive system limits the types of conjugates that can be administered orally. PEG-protein

complexes risk damage due to enzymatic degradation before entering the bloodstream. For these

reasons, most PEGylated drugs are administered via injection or intravenously.

Increasing PEG molecular weight corresponds to an increased circulatory half-life. In a

study monitoring the distribution of 125I-labeled PEG between MWs of 6,000 Da (PEG6k) and

190,000 Da (PEG190k), urinary clearance decreased as molecular weight increased. PEG6k had

a half-life of 18 minutes following intravenous administration. PEG190k, nearly 30 times larger,

had a half-life of approximately 24 hours (Yamaoka, Tabata & Ikada, 1994). The size of the PEG

chain clearly has an effect on time retained by the body.

Evidence was found that PEG-protein conjugates were first broken down by degradation

of the protein followed by cleavage from the PEG backbone. Once separated, PEG is passed

through the kidneys while proteins are catabolized under normal pathways, producing non-toxic

metabolites (R. Webster et al., 2009). The safety of the remaining PEG must be examined to

determine the suitability of these conjugates as DDSs.

Toxicity

Toxicity has only been associated with extremely low molecular weight PEGs. Fatalities

have been reported in burn patients after treatment with an antimicrobial cream containing 95%

PEG300. It was found that in vivo, PEG400 and smaller were oxidized by alcohol dehydrogenase

into toxic diacid and hydroxyl acid metabolites in the blood, poisoning the patients. The rate of

oxidation decreased greatly with increased PEG size. The smallest PEG studied, ethylene glycol

24

(MW 62 Da), was degraded 32 times faster than the largest, octaethylene glycol (MW 370 Da)

(Herold, Keil & Bruns, 1989). Another study administered PEG400 and smaller to mice and rats.

Doses of 10 mL/kg or less proved fatal (Bartsch, Sponer, Dietmann & Fuchs, 1976). Intravenous

administration of PEG1400 and larger showed no toxicity or negative effects in rabbits and dogs

(Working, Newman, Johnson & Cornacoff, 1997).

Advantages of PEGylation

As Table 3 details, several proteins have been successfully attached to PEGs of various

sizes. The benefits of PEGylation are most strongly tied to the hydrophilicity of PEG itself. This

hydrophilicity, due to the oxygen content of the PEG backbone, means that PEG and its

conjugates are highly soluble and highly hydrated in aqueous environments. It is estimated that

PEG is able to adsorb three to five water molecules per ethoxylene subunit, thus greatly

increasing the hydrodynamic volume of the backbone. Increased hydrodynamic volume results in

a molecule that appears too large to pass through kidney filtration (Figure 7). Avoidance of

elimination via normal pathways leads to increased circulatory half-life (Plesner, Fee, Westh &

Nielsen, 2011; Veronese et al., 2009).

25

Figure 7. Diagram of a PEGylated protein nanoparticle. Covalent conjugation of PEG chains onto the surface of a protein greatly increases the hydrodynamic volume of the NP, conferring increased circulation lifetime and immunogenic masking.

Avoidance of degradation by immunogenic agents also increases the circulation half-lives

of these NPs. Chicken Immunoglobulin Y was administered to mice in native and PEGylated

forms. Protein conjugated with 3-13 PEG5k or PEG20k showed immune response levels drop to

as low as 1.6% detection (Gefen et al., 2013). PEGylation is able to mask bound protein from

immunogenic response as the antibodies are unable to sense the protein core due to steric

hindrance from the PEG backbones. This helps to protect the NP against damage and

degradation.

PEGylation confers a number of benefits to the attached protein core. However,

PEGylation may also be deleterious to the NP. The attachment of PEG is typically done through

covalent bonds with surface residues. Eliminating these charges and masking the protein surface

may lead to poor target and ligand recognition which, in turn, could greatly alter the biological

activity of the protein.

The effects of PEGylation on albumin stability were monitored using PEG chains ranging

from PEG5k to PEG60k attached to BSA at Cys34 via the thiol-ene reaction. Circular dichroism

26

studies showed no loss of secondary structure after PEGylation. Upon application of heat, it was

found that attachment of PEG, slightly lowered the temperature of denaturation of 82.5oC for

unmodified BSA by 2-3oC. There was no direct correlation between the change of denaturation

temperature and PEG molecular weight. There does, however, appear to be a connection between

molecular weight and the temperature of aggregation. For PEGs between 10-40 kDa, aggregation

began at 81oC while unmodified BSA began to aggregate at 71oC (Plesner et al., 2011). It

appears that the stability of BSA is slightly compromised but the PEG backbone helps to stave

off aggregation.

It should be noted that while attaching many strands of small molecular weight PEG has

been successful, it has been proven to be more beneficial to use less PEGs with a total weight

equal to these small PEGs. The Williams group demonstrated that attaching two to five strands

of PEG30k to superoxide dismutase was more effective than 7-15 strands of PEG5k in

preserving biological activity, increasing lifetime and reducing immunogenicity (M. G. P. Saifer,

Somack & Williams, 1994).

“Click” Chemistry

“Click” chemistry denotes a series of reactions which are modular and give high yields.

Reaction conditions are simple and insensitive to water and oxygen. The stereospecificity of

these reactions is what makes them so attractive for bionconjugation. The final product is due

completely to the functional groups reacted. Groups attached to these reactive moieties may

enhance the reaction but may not change the final product (Kolb, Finn & Sharpless, 2001).

The basis for click chemistry begins with a modification of the Huisgen reaction between

azides and alkynes. This reaction was largely ignored because of the need for high temperatures

and pressures. The use of a Cu(I) catalyst as a way to run these reactions under less extreme

27

conditions paved the way for the rapid rise in popularity of click chemistry (Rostovtsev, Green,

Fokin & Sharpless, 2002; Tornøe, Christensen & Meldal, 2002). These reactions and their

applications have been extensively reviewed by Kolb et al. (2001) and Thirumurugan et al.

(2013).

Azide-Alkyne Cycloaddition (AAC)

Cu(I)-catalyzed AAC (CuAAC). The Cu(I)-mediated reaction between azides and

alkynes (Figure 8) is the prototypical click reaction. It is fast, high-yielding, and stereospecific,

giving only 1,4-disubstituted 1,2,3-triazoles. Reaction conditions do not require elevated

temperatures or pressures.

Figure 8. Cu(I)-catalyzed azide-alkyne cycloaddition. The reaction of a terminal azide with a terminal alkyne in the presence of Cu(I) with a compliment of ligand, reducing agent and suppressor reagent results in a single 1,4-disubstituted 1,2,3-triazole product. Kinetics may be enhanced by incorporating an electron-withdrawing group at the R3 position.

The employment of a copper catalyst is accompanied by a variety of complimentary

reagents. The accelerating ligand chelates the copper and helps maintain the proper coordination

environment. The catalyst is more cost-effective as a Cu(II) salt and reduce which is reduced in

situ by an agent such as ascorbate. However, this reduction often results in reactive oxygen

species which may modify amino acid side chains. To counteract this, aminoguanidine is

included (Hong, Presolski, Ma & Finn, 2009).

Despite the high solubility and stability of these reagents at physiological pH, this form

of AAC is not ideal for production of albumin-based DDSs. As previously discussed, the N-

28

terminus in many albumins is able to bind Cu(II) with an affinity of 1 pM. The bound Cu(II) will

be difficult to remove during work-up of the resulting NPs. All traces of copper must be removed

before introduction into a patient as additional, undesirable heavy metals are toxic to the host.

Cu(II) has been implicated in neurodegenerative diseases, including prion diseases and

Alzheimer’s disease (Rózga et al., 2007). While the copper catalyst and its compliment of

reagents are readily available and cost-effective, simpler AAC reactions have been developed.

Strain-promoted AAC (SPAAC). Copper-free click chemistry circumvents the need for

a Cu(I) catalyst by incorporating the alkyne into a strained ring such as a cyclooctyne or

cyclononyne (Figure 9a) (Agard, Prescher & Bertozzi, 2004). With a simple, unsubstituted

cyclooctyne, the reaction kinetics are considerably slower than those of CuAAC. Attachment of

electron withdrawing groups onto the ring adjacent to the alkyne has been used to increase

reaction rates. Difluoro- (Figure 9b) and Dibenzocyclooctynes (Figure 9c) returned rates back to

those of CuAAC without the need for a Cu(I) catalyst and have become the first choice for

SPAAC reactions (Ning, Guo, Wolfert & Boons, 2008; van Berkel et al., 2007).

29

Figure 9. Strain-promoted azide-alkyne cycloaddition. (a) By incorporating the alkyne into a strained ring system, the reaction with a terminal azide no longer requires a catalyst or the compliment of reagents needed for CuAAC. The same 1,4-disubstituted 1,2,3-triazole product is formed but with slower reaction kinetics. (b) Difluoro- and (c) Dibenzocyclooctynes have been shown to increase the kinetics to rates comparable to CuAAC (Ning et al., 2008; van Berkel et al., 2007).

Azides and alkynes are biorthogonal as both are virtually absent from nature. The sole

exceptions are azide-containing natural products found in toxic red algae on the Gulf Coast of

Florida. Additionally, both are virtually non-reactive with any naturally-occurring functional

group (Baskin & Bertozzi, 2009). This, coupled with the reduced toxicity due to the elimination

of the copper catalyst, make SPAAC an excellent candidate for conjugation chemistry with

albumin.

Thiol-ene Reactions

Thiol-ene reactions, while involving neither an azide nor an alkyne, are another class of

click chemistry (Figure 10). All of the criteria are met in that the reaction is simple and the

product formed is stereospecific and gives high yields (Lowe, 2010). This reaction is especially

applicable to proteins with exposed cysteine residues. The free thiol may be chemoselectively

30

reacted with an alkene-containing linker to attach drugs or targeting molecules (Chalker,

Bernardes, Lin & Davis, 2009).

Figure 10. Thiol-ene reaction. (a) A thiol from an exposed cysteine may be reacted with an alkene resulting in a thioether linkage. (b) A common form of the -ene used in these reactions is the maleimide functional group.

The most common functional group employed for protein modification at cysteines is the

maleimide. Incorporation of the electron-deficient moiety into the molecule to be attached to the

protein allows for fast and easy modification under physiological conditions. For example, a

maleimide-methotrexate prodrug was developed as a treatment for arthritis (Fiehn, Kratz, Sass,

Müller-Ladner & Neumann, 2008). Upon intravenous administration, this prodrug binds to

Cys34 of HSA in vivo. The conjugate is later enzymatically cleaved, releasing a bioactive

derivative of methotrexate. The maleimide-cysteine reaction is rapid and selective, greatly

increasing the efficacy of the methotrexate treatment.

Summary

BSA provides an intriguing option to build a nano-scale DDS. It is highly soluble, with a

plethora of binding pockets, and a large number of surface residues which can be used for

customization without the need for recombinant methodologies. Its structural similarities with

HSA suggests a similar degree of effectiveness can be achieved when incorporated into any of

the published NP methodologies.

31

A multitude of approaches have been attempted at formulating a successful albumin-

based DDS. Each was successful at solubilizing a hydrophobic pharmaceutical in the presence of

albumin but flaws in each must be addressed to move forward. Only one albumin-based NP has

successfully made it to market. Abraxane, while effective, still requires an infusion of a large

dose of albumin into the bloodstream, potentially affecting the osmotic pressure and

subsequently the production of native albumin.

Several PEGylated proteins have been approved, but none are DDSs. PEGylation has

proven to be advantageous as it confers increased hydrodynamic volume which leads to

increased circulation lifetimes and decreased immune response. However, PEGylation may

slightly destabilize the protein at elevated temperatures and steric hindrance due to the attached

PEGs may negatively affect biological activity and cellular recognition.

The biorthogonality of SPAAC and the selectivity of the thiol-ene reactions allow for

easy modification of surface residues of BSA. Due to the high affinity between the N-terminus of

albumins and Cu(II), traditional CuAAC cannot be used. A strained cyclooctyne further modified

by an electron-withdrawing group provides an even easier and simpler alternative.

A NP that can successfully harness the advantages of albumin, avoid the pitfalls

encountered in other production methods, and gain the advantages of PEGylation with simple

surface modifications would prove beneficial for the advancement of nanotechnology as DDSs.

32

CHAPTER II

STATEMENT OF RESEARCH

The aim of this research is to develop a multi-use albumin-based nanoparticle for drug delivery

in order to address the concerns associated with current methodologies and available products.

The most pressing concerns are the cost and risk of pathogens associated with human-sourced

blood proteins, the negative effect on osmotic pressure due to a large dose of albumins, the

potential immune response and loss of biological activity due to heavy modification of these

albumins, and the need for organic solvents for preparation of these products.

Most market available albumin formulations use human serum albumin. When sourced

from blood donations, supplies are limited and the risk of potential contamination with pathogens

such as human immunodeficiency virus and hepatitis must be considered. Recombinant albumin

provides a reliable but expensive way to avoid these human pathogens. An alternative approach

is to use bovine serum albumin (BSA). This species of albumin is cheaper and more readily

available than its human counterpart. The two are structurally similar. Specifically, key residues

of Drug Sites 1 and 2 are conserved, suggesting similar binding capabilities.

Covalent attachment of polyethylene glycol (PEG) chains to proteins has been shown to

greatly increase circulatory half-life and reduce immunogenic response. In this research, several

BSAs will be attached to an 8-arm polyethylene glycol backbone (BSAx-R-PEG8) using two

types of reactions, resulting in different “R” group linkers between protein and backbone. Use of

a branched PEG backbone minimizes the effect on osmotic pressure. Upon administration of a

33

single dose of drug, the total number of albumins introduced into the bloodstream will be greatly

reduced with BSAx-R-PEG8 as opposed to conventional albumin nanoparticles.

In the first part of this work, the lone, unpaired sulfhydryl of BSA at Cys34 will be

reacted with a maleimido-functionalized PEG8 (Mal-PEG8) resulting in a thioether (te) linkage.

This eliminates the need for organic solvents entirely, requires no additional modifications to the

protein and provides a chemoselective conjugation method. Yields, secondary structure,

speciation, and ligand binding capabilities will be determined for this form of PEGylated BSA

(BSAx-te-PEG8).

Secondly, BSAx-R-PEG8 will be prepared using strain-promoted azide-alkyne

cycloaddition. BSA, “activated” at a single superficial lysine with either a short- or long-chain

Dibenzocyclooctyne-N-hydroxysuccinimidyl ester (DIBO-NHS; DIBO-PEG4-NHS), will be

PEGylated via reaction with complimentary terminal azides (N3-PEG8). BSA and PEG8 will be

conjugated via a 1,2,3-triazole (123t). This approach requires a single modification of the

albumin and a minimal amount of organic solvent necessary to dissolve each activating agent.

PEGylated BSA activated with the short alkyne (BSAx-123t-PEG8) will be used to optimize the

PEGylation methodology. BSAx-123t-PEG8 and PEGylated BSA activated with the longer

alkyne (BSAx-PEG4-123t-PEG8) will be characterized in terms of secondary structure,

speciation, and ligand-binding capabilities.

The third area of research examines the effect of these methods of PEGylation on the

stability of BSA. Monitoring the temperatures of denaturation and aggregation of PEGylated

BSA has shown a reduction in thermal stability of the attached protein. Published works have

examined complexes containing a single albumin conjugated with multiple PEG backbones,

resulting in a higher degree of modification. These complexes showed a slight decrease in the

34

temperature of denaturation is accompanied by an increase in the temperature of aggregation.

BSAx-te-PEG8, BSAx-123t-PEG8, and BSAx-PEG4-123t-PEG8 complexes require a single point

of modification in order to attach multiple albumins to a single PEG backbone. Less

modification of albumin may have less effect on the thermal stability of the final product.

Traditional pasteurization of albumin requires N-acetyl-L-tryptophan and sodium

caprylate to stabilize the protein during the heating process. Both of these additives have shown a

preference for DS2, creating competition for potential ligands. In an effort to eliminate these

stabilizers, PEGylated BSA will be pasteurized in the absence of both reagents. The effects of

pasteurization on the helicity and ligand binding capabilities of PEGylation by both thiol-

maleimide and strain-promoted azide-alkyne cycloaddition reactions in conjunction with the pre-

loading of a DS2 ligand will be monitored.

Every drug delivery system must be non-toxic, non-immunogenic, cost-effective, easy to

produce, and applicable to a variety of ligands. By addressing the concerns associated with

previous approaches, these qualities can be achieved.

35

CHAPTER III

MODIFICATION OF BOVINE SERUM ALBUMIN WITH

MAL-PEG8 AS A DRUG DELIVERY SYSTEM

Introduction

Serum albumin has become an attractive option in the development of drug delivery systems

largely due to its role as a vehicle for hydrophobic molecules. Human serum albumin (HSA) has

been studied extensively for this purpose but the cost of production and risk of pathogenic

transfer are considerable. Bovine serum albumin (BSA) provides a more cost-effective and

readily available alternative while maintaining key structural features, including two drug

binding sites, DS1 and DS2.

Crystallographic data has shown that DS1 and DS2 are well-conserved between HSA and

BSA (Bujacz, Zielinski & Sekula, 2014; Lejon, Cramer & Nordberg, 2008). The polar

environment at the mouth of DS1 is the sole difference between binding sites of these species. In

HSA, the polar cluster consists of residues K195, K199, R218, and R222 while in BSA, the

entrance is coated in mostly arginines (R194, R198, R217, and K221). Arginine, while still

capable of hydrogen-bonding, has a more dispersed positive charge and a larger functional

group. Sterics as well as the delocalization of the positive charge weaken interactions with

individual polar molecules (Li, Vorobyov & Allen, 2013). Despite this difference, it is believed

that DS1 and DS2 on BSA will house similar molecules as HSA.

A single, free cysteine at position 34 is also conserved across species. In the body, Cys34

36

plays a key role as an antioxidant (Anraku, Chuang, Maruyama & Otagiri, 2013; Roche,

Rondeau, Singh, Tarnus & Bourdon, 2008) and is also capable of binding various ligands

(LeBlanc, Shiao, Roy & Sleno, 2014; Narazaki, Hamada, Harada & Otagiri, 1996; Noort, Hulst

& Jansen, 2002). In terms of drug delivery system development, the most attractive aspect of

Cys34 is that it provides an accessible, chemoselective site for surface modification.

Cysteine has been selectively modified using a variety of chemistries (Abbas, Xing &

Loh, 2014; Abegg et al., 2015; Mehtala, Kulczar, Lavan, Knipp & Wei, 2015) in order to

increase the circulatory half-lives of a variety of protein-based therapies (Léger et al., 2004;

Plesner, Fee, Westh & Nielsen, 2011; Stoddart et al., 2008). CJC-1134-PC is an application of

ConjuChem’s Preformed Conjugate-Drug Affinity Complex (PC-DACTM) technology for the

treatment of type 2 diabetes. This complex consists of maleimido-functionalized Exendin-4, a

homolog of glucagon-like peptide 1, covalently attached to recombinant human albumin via

interaction with Cys34. Exendin-4 has a short half-life, requiring twice daily injections to

maintain therapeutic levels. Conjugation onto albumin extends circulatory lifetime while

maintaining biological potency (Baggio, Huang, Cao & Drucker, 2008).

In the case of PC-DAC, the therapeutic agent is left exposed to the environment in order

to interact with cell surface targets. However, drug molecules need to be removed from the

environment. Direct conjugation of a drug to Cys34 provides no protection for the ligand,

subjecting the bound drug to possible degradation or alteration. The binding pockets of albumin

provide shelter for these molecules while keeping Cys34 free for surface modifications.

Cys34 is found in a shallow cleft at the surface of albumin in Domain IA (Figure 11).

DS2, found entirely in Domain IIIA, is likely unaffected by conjugation at a single residue in

Domain I. DS1, however, is found at the interface of Domains IB and IIA. Modification at Cys34

37

may cause conformational changes which may alter the binding characteristics of DS1.

Figure 11. Cys34, Drug Site I, and Drug Site II of bovine serum albumin. (a) Cys34 (brown) is located in Domain I (blue), adjacent to DS1 (pink) in Domain II (yellow). DS2 (green) in Domain III (red) is far removed from Cys34. (b) It is slightly buried in a shallow cleft on the surface of the protein. PDB ID: 4OR0 (Bujacz et al., 2014). All crystal structure images were prepared using UCSF Chimera (Pettersen et al., 2004) and rendered with POV-Ray ("Persistence of Vision Pty. Ltd.: Persistence of Vision (TM) Raytracer," 2004). A common modification for protein-based drug delivery systems is covalent attachment

of a polyethylene glycol (PEG) backbone. PEG-protein conjugates often consist of multiple PEG

chains covalently attached at several locations on the surface of a single protein core (Nischan &

Hackenberger, 2014; Veronese & Pasut, 2005). This approach results in a potential problem for

albumin-based drug delivery systems. Albumin is a key regulator in osmotic pressure.

Administration of a large dose of individually PEGylated albumins may lead to down-regulation

of albumin production (Pietrangelo, Panduro, Chowdhury & Shafritz, 1992). Down-regulation

would result in accumulation of materials normally transported by albumin, such as toxic

metabolites and heavy metals.

A multi-albumin complex would allow for a higher drug dose with introduction of fewer

38

albumins (Figure 12). Osmotic pressure is controlled by the number of particles, not their

individual sizes. In the case of serum proteins, with a MW of 400 kDa, fibrinogen is among the

largest but accounts for only 1% of total osmotic pressure due to low concentration (0.0075

mM). Albumin, however, is several times smaller (MW 66 kDa) but is found at a much higher

concentration (0.68 mM) and accounts for 79% of total osmotic pressure (Guyton & Hall, 2006).

For example, the same therapeutic dose of eight individual albumins pre-loaded with a drug

could be achieved with administration of a single eight-albumin particle, lessening the effects on

osmotic pressure of treatment.

Figure 12. Diagram of BSAx-te-PEG8. Thioether (te) linkages connect eight BSA molecules to a single PEG8 backbone via reaction between terminal maleimides and native Cys34 residues.

39

In this study, BSA will be attached to maleimido-functionalized PEG8 (Mal-PEG8) via

thiol-ene reaction (Figure 13). A thioether (te) linkage will connect BSA to the PEG8 backbone

via reaction between a terminal maleimide and native Cys34. The resulting BSAx-te-PEG8

(Figure 12) will be characterized in terms of yield, secondary structure, speciation, and ligand

binding.

Figure 13. Thiol-maleimide conjugation of bovine serum albumin onto PEG8. The free thiol of Cys34 reacts with Mal-PEG8 to give a thioether linkage between protein and PEG backbone.

Materials and Methods

Materials

Lyophilized BSA (Product Number A7906), sulfamethoxazole (SMZ; S7507), naproxen

(NPX; N8280), and all reagents not detailed were purchased from Sigma-Aldrich.

Methods

All experiments were carried out in 50 mM 4-Morpholinepropanesulfonic acid/0.15 M

NaCl pH 7.2 (MOPS/NaCl) buffer unless otherwise noted. Fast protein liquid chromatography

(FPLC) separations were performed on an ÄKTAprime plus chromatography system (Amersham

Biosciences; 11001313) with MOPS/NaCl mobile phase. Resulting fractions were concentrated

with 30 kDa MW cut-off centrifuge filters (Sartorius; VS2022). BSA concentrations were

calculated using absorbance values obtained on an Agilent 8453 UV-visible Spectroscopy

System at 280 nm with an extinction coefficient of 43.824 mM-1 cm-1 and a MW of 66,430 Da

40

(Hirayama, Akashi, Furuya & Fukuhara, 1990; "Thermo Scientific: Extinction Coefficients,"

2013).

Fatty acid removal. Fatty acid-free BSA (FAF-BSA) was prepared via an activated

charcoal method (Chen, 1967). In short, BSA was dissolved in MOPS/NaCl. The solution was

acidified to pH 3 and stirred on ice with 20-40 mesh particle size activated charcoal in order to

remove any bound fatty acids. Charcoal was pelleted at 10,000 rpm and the supernatant was

filtered through 0.2 µm cellulose filters to remove any remaining carbon. The de-fatted albumin

solution was returned to pH 7.2 before further purification. Dimers were removed by size

exclusion chromatography (SEC) on a pre-packed 120 mL 60 cm x 16 mm S300 FPLC column

(GE Healthcare; 17-1167-01).

PEGylation of FAF-BSA. The final product consisted of multiple BSA molecules

attached to a PEG8 backbone via a thioether (te) linkage formed by the reaction of terminal

maleimides of PEG8 and the sulfhydryl of Cys34 (BSAx-te-PEG8). 40 kDa Mal-PEG8 (Creative

PEGworks; PSB-864) was dissolved in MOPS/NaCl and added to 1 mM FAF-BSA at 8:1 FAF-

BSA:Mal-PEG8. The solution was allowed to incubate at room temperature for one week with

constant stirring. PEGylated FAF-BSA was purified via a two-phase SEC separation. The bulk

of unreacted FAF-BSA was removed by way of an S300 FPLC column, the remainder via a 320

mL 60 cm x 26 mm S500 column (GE Healthcare; 28-9356-07). Chromatograms from the S300

separation were processed in PrimeView 5.0 software to determine yields as the percentage of

total area count.

Circular dichroism (CD). The effects on secondary structure of the PEGylation

procedure were determined by CD. BSAx-te-PEG8 were assayed in 50 mM MOPS pH 7.2.

Studies were done on an Olis DSM 20 CD Spectrophotometer. Spectra were collected at 30oC

41

over the range of 190-260 nm with a step-size of 1 nm. Helicity was calculated with DichroWeb

(Whitmore & Wallace, 2008) and the analysis program CDSSTR (Sreerama & Woody, 2000).

Inclusion of a crystal structure for HSA (Wardell et al., 2002) allowed for use of reference set

SP175 (Janes, 2009; Lees, Miles, Wien & Wallace, 2006).

SDS-PAGE. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)

was performed as an initial characterization of BSAx-te-PEG8 species. The assay was performed

on a PhastSystem (Amersham Biosciences; 18-1018-23) with a 4-15% gradient PhastGel (GE

Healthcare; 17-0678-01), Coomassie R 350 stain (GE Healthcare; 17-0518-01), and high

molecular weight markers (GE Healthcare; 17-0615-01).

Analytical ultracentrifugation (AUC). AUC was carried out in a Beckman Coulter

ProteomeLab XL-A analytical ultracentrifuge using standard two sector cells with quartz

windows in four- or eight-hole rotors. All sedimentation velocity experiments were performed at

50,000 rpm with a preset scan number of 100 per cell. The reference sector was filled with 450

µL of MOPS/NaCl and the sample sector was filled with 430 µL of sample with an absorbance at

280 nm between 0.75 and 1.1. Path length was 1 cm. Buffer density was measured at 20.0oC in a

Mettler/Paar Calculating Density Meter DMA 55A and the buffer viscosity was measured in an

Anton Parr AMVn Automated Microviscometer. Data were analyzed using the program Sedfit

(Schuck, 2000; "Sedfit," 2017). Molecular weights and sedimentation coefficients (S20,w) were

determined using the continuous c(s) mode in the Sedfit program. Native monomeric BSA

yielded an S20,w value of 4.04 S which agreed with published values (Shulman, 1953; Taylor,

1952; Zhao et al., 2015). The c(s) plots were deconvoluted using non-linear least-squares fitting

algorithms available in the Origin (v. 9.0) software package (OriginLab Corporation,

Northampton, MA). Data were fit to multiple Gaussian components, increasing the number of

42

components until residual plots (data-fit) were random and the chi square value for the fit was

minimal. Values reported are percentages of the total area count of peaks corresponding to

PEGylated species. These area counts were calculated via the trapezoidal formula.

Fluorimetric analyses. Binding studies were performed on a Photon Technology

International fluorimeter using Felix software (version 1.1). SMZ, a DS1 ligand, and NPX, a

DS2 ligand, were studied to determine if PEGylation affected ligand binding capabilities. The

number of binding sites (n) and the association constant (Ka) were calculated for BSAx-te-PEG8.

In each case, the fluorophore monitored was Trp213, found at the entrance to DS1 in Domain

IIA.

For SMZ, 30 µM stock solutions of FAF-BSA and BSAx-te-PEG8 were prepared in 0.2

M Tris-HCl/0.1 M NaCl pH 7.4. SMZ was dissolved in methanol to a final stock concentration

of 30 mM. Working SMZ stocks were prepared at 3 mM and 1 mM by dilution with Tris buffer.

Monitoring emission between 300-380 nm upon excitation at 283 nm at 25oC, aliquots of

working SMZ stocks were titrated into a 3 µM BSA solution. 1 mM working stock was used up

to 20 µM SMZ. 3 mM working stock was used to raise the final concentration to 80 µM SMZ.

To determine n and Ka for NPX, 30 µM stock solutions of FAF-BSA and BSAx-te-PEG8

were prepared in 10 mM phosphate buffer pH 7.0. NPX was dissolved in methanol to a final

stock concentration of 45 mM. A working NPX stock was prepared at 0.15 mM by dilution with

phosphate buffer. Monitoring emission between 300-380 nm upon excitation at 295 nm at 25oC,

aliquots of working NPX stocks were titrated into a 0.725 µM BSA solution to a final

concentration of 3 µM NPX.

Blanks for each ligand at each concentration were necessary to account for any

fluorescence due to the ligand itself. SMZ had almost no signal under experimental conditions.

43

NPX, however, showed strong emission with maxima at approximately 350 nm. Particularly at

high concentrations, NPX signal obscured the BSA maxima which ranged from 344 to 350 nm.

Non-linear regression analyses were carried out using The Gnumeric Spreadsheet

(version 1.10.16) ("The Gnome Project: The Gnumeric Spreadsheet: Free, Fast, Accurate - pick

any three," 2011) and a modified Hill equation. In order to normalize samples, fractional

intensities (FrInt) were used in place of raw intensities.

FrInt = FrInt()*–FrInt(,- ∗ Ligand -

K5- + Ligand - +FrInt(,-

FrInt = (Int,89:,;)-5 − Int=89:,;)-5)(Int()*89:,;)-5 − Int=89:,;)-5)

K) =1K5

Results and Discussion

BSAx-te-PEG8 Yield

BSAx-te-PEG8 was purified via two size exclusion FPLC columns (Figure 14).

Integration of the S300 separation showed that the PEGylation procedure converted an average

of 36.1 ± 0.6% of BSA into BSAx-te-PEG8. In the initial purification, a small peak eluted at

approximately 40 mL suggests the presence of extremely high molecular weight material. This

same material was buried in the S500 separation, resulting in peak broadening for BSAx-te-PEG8

fractions.

44

(a)

(b)

Figu

re 1

4. P

urifi

catio

n of

thio

l-mal

eim

ide

PEG

ylat

ed b

ovin

e se

rum

alb

umin

via

siz

e ex

clus

ion

chro

mat

ogra

phy.

(a) I

nitia

l pur

ifica

tion

was

don

e on

an

S300

. (b)

Fin

al p

urifi

catio

n w

as d

one

an S

500

colu

mn.

FA

F-B

SA (-

--) a

nd B

SAx-

te-P

EG8 (

red)

.

020406080100

020

4060

8010

012

0

% Maximum Absorbance

Elut

ion

Volu

me

(mL)

020406080100

050

100

150

200

250

300

350

% Maximum Absorbance

Elut

ion

Volu

me

(mL)

45

Effects on Secondary Structure

Helicity was maintained after the PEGylation procedure. BSAx-te-PEG8 contained

minima at 210 nm and 220 nm which are characteristic of α-helices (Figure 15). The secondary

structure of FAF-BSA was 59.2 ± 2.2% α-helical. PEGylation resulted in a slight decrease in

helicity to 56.3 ± 1.3%. Both values are in agreement with published values which range from

56-66% helical content for BSA (Das et al., 2017; Moriyama et al., 2008). Given that Cys34 is

shallowly sheltered in a cleft between two helices, attachment of a large PEG chain may have

caused some loss of helicity.

Figure 15. Circular dichroic spectra of FAF-BSA and BSAx-te-PEG8. FAF-BSA (---) and BSAx-te-PEG8 (red) contain minima at 210 nm and 220 nm characteristic of α-helices.

-90

-80

-70

-60

-50

-40

-30

-20

-10

0

10

20

190 200 210 220 230 240 250 260

Ellip

ticity

(mill

ideg

rees

)

Wavelength (nm)

46

BSAx-te-PEG8 Speciation

Electrophoretic separation of PEGylated BSA via SDS-PAGE revealed the presence of

high molecular weight material (Figure 16). A faint band at 116 kDa corresponds to dimeric

BSA and BSA1-te-PEG8 (MW 106 kDa). A second faint band at 170 kDa corresponds to BSA2-

te-PEG8 (MW 172 kDa). A dark streak at MWs heavier than 220 kDa contains distinct banding,

most likely corresponding to BSA3-4-te-PEG8. Material retained in the stacking zone is made of

five or more PEGylated BSAs as the molecular weight cut-off for proteins in the separating zone

is 300 kDa. Smearing is most likely due to interaction of the PEG8 backbone with the

polyacrylamide gel.

Figure 16. Electrophoretic separation of species of bovine serum albumin PEGylated via thiol-maleimide click chemistry. Lane 1-2: BSAx-te-PEG8; Lane 3: FAF-BSA; Lane 4: High molecular weight marker. Several species of PEGylated BSA were found via AUC (Figure 17). 70% of the total

PEGylated area count was due to BSA2-6-te-PEG8. 11% corresponded to a single BSA attached

47

to a single PEG8 backbone. The remaining 19% consisted of very high molecular weight species.

Given that there are only eight arms on Mal-PEG8 and only one free cysteine on BSA, the largest

particle, BSA8-te-PEG8, would have a MW of 502 kDa. Peaks with a sedimentation coefficient

greater than 15 S have calculated molecular weights well above this mark. It is believed that the

extremely high molecular weight species are a combination of BSA7-8-te-PEG8 and PEGylated

species of a side-reaction of Mal-PEG8.

48

Peak 1 2 3 4 5 6 7 8 9 10 11 12 S20,w 2.12 4.24 6.06 7.58 8.79 10.6 12.7 15.8 18.2 19.7 21.8 24.2

MWcalc (kDa) 35 82.7 nr 194 259 343 473 704 nr nr 970 1160

MWtheor (kDa) 66 106 172 238 304 370 436 n/a

# BSA conjugated monomer 1 2 3 4 5 6 n/a

% Area Count n/a 11 9 14 17 15 15 19

Figure 17. Analytical ultracentrifugation analysis of BSAx-te-PEG8. % Area Count values are the percent of PEGylated species only. nr: not reported; n/a: not applicable. Maleimides react readily with sulfhydryls, however, Mal groups may also react with the

terminal amino group of lysine (Brewer & Riehm, 1967). At neutral pH, thiolation occurs 1000

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0 5 10 15 20 25 30

c(s)

S20,w

1

25

3

4

6

7

8

910

11 12

49

times faster than the lysine interaction. Given the high concentration of the reaction mixture, the

large number of superficial lysines, and the long incubation time, it is highly possible that

multiple lysines reacted with one or more PEG8 chains to create very high molecular weight

oligomers.

Results were similar to hemoglobin-based nanoparticles developed with the same Mal-

PEG8 (K. D. Webster et al., 2017). Hemoglobin, a 64 kDa tetrameric protein, contains two free

cysteines at position 93 of each β-subunit. Similar reaction conditions to those used in this study

resulted in a heterogeneous mixture of PEG8 complexed with predominantly three hemoglobins.

As many as six hemoglobin tetramers were found to be conjugated to a single PEG8 backbone.

BSA, similar in size to hemoglobin, behaved in a similar fashion during PEGylation. A similar

spread of conjugated protein was found, with a majority of PEGylated species falling between

two and six BSA molecules per PEG8 backbone.

Ligand Binding Capabilities

Fluorescent emission was monitored to determine the viability of DS1 and DS2 of PEG-

conjugated BSA. NPX was used to probe DS2. Published crystal structures of de-fatted bovine,

equine, and leporine serum albumins complexed with NPX indicate a preference for DS2

(Bujacz et al., 2014). Experimental values for NPX in the presence of FAF-BSA and BSAx-te-

PEG8 were consistent with published values (Table 4), suggesting that DS2 remains intact upon

PEGylation. While the crystal structure reported by Bujacz et al. show three NPX molecules

bound, DS2 has been the only high affinity site reported. Discrepancies in the n values are

possibly due to de-fatted BSA used in this work while fatty acid-containing albumin was used in

the published results.

50

Table 4. Association Constants (Ka) and Number of Binding Sites (n) for BSAx-te-PEG8 for Sulfamethoxazole and Naproxen.

SMZ NPX Sample Ka (x103 M-1) n Ka (x107 M-1) n

FAF-BSA 2.03 ± 0.18 0.9 ± 0.1 0.40 ± 0.13 0.4 ± 0.1 BSAx-te-PEG8 0.19 ± 0.05 0.8 ± 0.03 1.99 ± 0.17 0.9 ± 0.3

Lit. values1,2 3.55-15.6 0.9-1.0 0.12-3.70 0.6-3

1SMZ values were published by Naik et al. (2009; 2015), Rajendiran & Thulasidhasan (2015), and Wang et al. (2014). 2NPX values were reported by Banerjee et al. (2006), Bou-Abdallah et al. (2016), Fielding et al. (2005), and Honoré & Brodersen (1984). Bujacz et al. (2014) have reported a crystal structure (PDB ID: 4OR0) for BSA complexed with three NPX molecules. Binding of NPX to FAF-BSA and BSAx-te-PEG8 showed similar trends (Figure 18).

BSA in the absence of NPX had a maximum emission wavelength at 344 nm. With increasing

amounts of NPX, a red shift resulted in a maximum at 350 nm. This is consistent with binding to

a site adjacent to the fluorophore. In this case, the fluorophore studied was Trp213 found in

Domain IIA at the mouth of DS1. As ligand binds to DS2, slight conformational changes

exposed Trp213 to the solvent, causing a red shift.

51

(a)

[NPX] increasing

(b)

[NPX] increasing

(c)

Figure 18. Fluorimetric spectra and non-linear regression of BSAx-te-PEG8 in the presence of naproxen. (a) BSAx-te-PEG8 and (b) FAF-BSA spectra. BSA in the absence of NPX (---) had a maximum emission at 344 nm. Upon addition of NPX to a final concentration of 3 µM, maximum emission shifted to 350 nm and intensities increased. (c) Non-linear regression analyses for BSAx-te-PEG8 (red £) and FAF-BSA (black ¯). Experimental data is represented as symbols while calculated data points are lines of the same color.

4.0E+04

6.5E+04

9.0E+04

1.2E+05

320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

4.0E+04

6.5E+04

9.0E+04

1.2E+05

320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

0.0

0.2

0.4

0.6

0.8

1.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Frac

tiona

l Int

ensi

ty

[NPX] (µM)

52

DS1, however, appears to be greatly affected by the covalent modification of Cys34.

FAF-BSA values were similar to published values for both Ka and n for SMZ studies (Table 4).

However, BSAx-te-PEG8 demonstrated a greatly reduced affinity for SMZ. Spectra for both

FAF-BSA and BSAx-te-PEG8 followed similar trends (Figure 19). Intensities decreased as SMZ

concentration increased. The signal from Trp213 is quenched in the presence of the ligand

though it is not tightly bound in BSAx-te-PEG8. Modification of Cys34 in Domain IA, clearly

had an effect on binding to DS1 in Domains IB and IIA.

53

(a)

[SMZ]

increasing

(b)

[SMZ]

increasing

(c)

Figure 19. Fluorimetric spectra and non-linear regression of BSAx-te-PEG8 in the presence of sulfamethoxazole. (a) BSAx-te-PEG8 and (b) FAF-BSA spectra. BSA in the absence of SMZ (---) had a maximum emission at 346 nm. Intensities decreased with each addition of SMZ. (c) Non-linear regression analyses for BSAx-te-PEG8 (red £) and FAF-BSA (black ¯). Experimental data is represented as symbols while calculated data points are lines following the same color.

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

2.5E+06

300 310 320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

2.5E+06

300 310 320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

0.0

0.2

0.4

0.6

0.8

1.0

0 10 20 30 40 50 60 70 80

Frac

tiona

l Int

ensi

ty

[SMZ] (µM)

54

Conclusions

Mal-PEG8 was successfully used as a scaffolding to combine multiple native BSA

molecules into a single complex with drug binding capabilities. The reaction was simple to run,

requiring only a relatively high molar ratio of FAF-BSA:Mal-PEG8. A majority of PEGylated

species contained 2-6 BSA molecules though some species of molecular weights greater than

BSA8-te-PEG8 were found. While a larger spread of BSAx-te-PEG8 species were generated, the

complexes still showed the ability to effectively bind drugs to DS2, maintaining the key

functionality of these albumin-based nanoparticles.

55

CHAPTER IV

CONJUGATION OF BOVINE SERUM ALBUMIN ONTO PEG8 VIA

STRAIN-PROMOTED AZIDE-ALKYNE CYCLOADDITION

AS A DRUG DELIVERY SYSTEM

Introduction

The hydrophobicity of many drugs is a major hurdle for efficient delivery to target tissues. Drug

delivery systems (DDS) aim to increase the bioavailability of pharmaceuticals by increasing the

solubility and subsequently their efficacy. Increased solubility also reduces the risk of rapid

elimination and in vivo degradation.

The ideal DDS must be non-immunogenic, easily eliminated, and applicable to a range of

drugs. This versatility requires the ability to bind different types of ligands and allow for

targeting of specific cell types. It must protect the ligand from degradation while protecting the

host from toxicity by chaperoning bound drug to the site of action.

Incorporation of serum albumin into a DDS would address many of these points.

Albumin serves as a natural vehicle for insoluble materials by binding these molecules in a

variety of hydrophobic pockets found throughout its structure. Albumin-based DDSs have also

been shown to greatly increase the aqueous solubility of drugs (Khoder et al., 2016). By

sequestering a pharmaceutical, albumin provides protection against degradation, redox reactions,

and immunogenic agents. Additionally, the drug is no longer available to freely move through

the bloodstream and possibly damage healthy tissue, protecting the patient.

56

These hydrophobic cavities are able to non-covalently bind a wide variety of ligands,

making albumin a versatile drug carrier. Fatty acids (Bhattacharya, Grüne & Curry, 2000), heavy

metals (Bal, Christodoulou, Sadler & Tucker, 1998; Rózga, Sokołowska, Protas & Bal, 2007),

and metabolites such as bilirubin (Zunszain, Ghuman, McDonagh & Curry, 2008) are just a few

of the endogenous materials transported by albumin. The two main binding sites of extrinsic

hydrophobic molecules are Drug Site I (DS1) and Drug Site II (DS2). Both are mostly apolar

pockets with only a few hydrophilic residues. DS1, located at the interface between Domains IB

and IIA, is a large, multi-chambered cavity which is able to accommodate multiple ligands of

different sizes. DS2, found entirely in Domain IIIA, is less versatile and shows a higher degree

of stereospecificity due to a much more compact structure. These binding sites are reviewed in

greater detail by Ghuman et al. (2005) and Yamasaki et al. (2013).

Albumin also allows for specific cell targeting. Tumor cells have been shown to

accumulate albumins passively through the enhanced permeation and retention effect and

actively via albondin-mediated transcytosis (Desai, 2008; Hawkins, Soon-Shiong & Desai, 2008;

Tiruppathi, Song, Bergenfeldt, Sass & Malik, 1997). This provides a natural targeting ability for

chemotherapeutics and led to the approval of Abraxane®, an albumin-based DDS of paclitaxel

for treatment of metastatic breast cancer (Desai, 2008).

Additionally, albumins contain up to 59 lysine residues. The terminal amino groups of

these residues provide a potential site for surface modification with antibodies (Anhorn, Wagner,

Kreuter, Langer & von Briesen, 2008), vitamins (Leamon & Low, 1991), and other targeting

molecules.

One concern for using albumin as a drug-carrier is the effect on osmotic pressure upon

introduction of a large number of free, individual albumins. Approximately 79% of osmotic

57

pressure is credited to albumin (Guyton & Hall, 2006). For this reason, serum albumin levels are

tightly regulated by the body. A large infusion of free albumin would be accompanied by down-

regulation of albumin production (Pietrangelo, Panduro, Chowdhury & Shafritz, 1992). This

would lead to aggregation of fatty acids and other endogenous, insoluble biomaterials in blood

and tissue.

Conjugation of multiple bovine serum albumin (BSA) molecules onto a single eight-

armed polyethylene glycol backbone (BSAx-R-123t-PEG8) addresses this concern (Figure 20).

The linker between PEG8 and BSA in this diagram is a 1,2,3-triazole (123t) formed by strain-

promoted azide-alkyne cycloaddition (SPAAC) with a variable R group spacer between protein

and alkyne. Osmotic pressure is dependent on the number of molecules present, not their

individual masses. Delivering the same dose of albumin-drug via BSAx-R-123t-PEG8 would

have less effect on the osmotic pressure. For example, to deliver therapeutic levels of drug, PEG8

complexed with eight BSA molecules would require one-eighth of the dose of unmodified

albumin. Osmotic pressure would be affected by a single molecule as opposed to eight.

58

Figure 20. Diagram of BSAx-R-123t-PEG8. PEGylation run to completion would result in eight BSAs attached to a single PEG8 backbone. The linkers between PEG8 and BSA are a 1,2,3-triazole (123t) formed by SPAAC with a variable R group spacer between protein and alkyne.

PEG is a biologically inert polymer commercially available in a wide variety of sizes and

functional groups ("Creative PEGworks: PEG Products," 2017; "JenKem Technology: PEG

Products," 2017). Inclusion of a PEG backbone into a DDS would increase solubility of the

complex and provide additional protection of the attached BSA molecules from degradation and

in vivo modifications.

59

Human serum albumin (HSA) has been the main albumin studied for delivery of

synthetic hydrophobic molecules. HSA, being a human-sourced, blood-borne protein, presents an

increased risk of pathogen transfer. Recombinant human albumins have been developed but are

expensive to produce (Bosse et al., 2005; He et al., 2011). BSA provides a readily available and

inexpensive alternative to its human counterpart. Crystallographic data has shown that many

structural features, including DS1 and DS2, are well-conserved between HSA and BSA (Bujacz,

Zielinski & Sekula, 2014; Lejon, Cramer & Nordberg, 2008). The polar environment at the

mouth of DS1 is the sole difference between binding sites of these species. In HSA, the polar

cluster consists of residues K195, K199, R218, and R222 while in BSA, the entrance is coated in

mostly arginines (R194, R198, R217, and K221). While still capable of hydrogen-bonding,

arginine has a more dispersed positive charge and a larger functional group. Sterics as well as the

delocalization of the positive charge weaken interactions with individual polar molecules (Li,

Vorobyov & Allen, 2013). Despite this difference, it is believed that DS1 and DS2 on BSA will

house similar molecules as HSA.

While BSA is a natural product that shares a large number of similarities with

endogenous HSA, BSA is still a foreign molecule and is at risk of being treated as such upon

administration to humans. Studies have shown that oral and parenteral administration of

unmodified BSA to humans resulted in production of BSA antibodies. In some cases, no immune

response was invoked (Korenblat, Rothberg, Minden & Farr, 1968). In cross-species

administration of proteins, covalently attached PEG has been shown to effectively mask the

foreign molecules and greatly reduce immunogenic response (Gefen et al., 2013). A PEGylated

complex such as BSAx-R-123t-PEG8 (Figure 20) may hide the foreign BSA molecules from

immunogenic agents.

60

Studies have shown that PEGylated proteins are safely metabolized in two stages. First,

protein components are degraded. In the case of albumin, glycoproteins gp18 and gp30 bind to

albumin, tagging it for lysosomal catabolism (Larsen, Kuhlmann, Hvam & Howard, 2016;

Schnitzer, Sung, Horvat & Bravo, 1992; Stehle, Sinn, Wunder, Schrenk, Stewart et al., 1997).

The remaining PEG backbone is excreted through the kidneys (R. Webster et al., 2009). Safe

metabolism is key for any DDS. Once the drug has been delivered, deleterious effects of the

DDS or its metabolites are undesirable.

This BSAx-R-123t-PEG8 complex will be generated using “click” chemistry. The term

denotes reactions that are modular, able to give near quantitative yields, and are stereospecific.

These reactions are simple, easy to perform and insensitive to water and oxygen (Lowe, 2010).

SPAAC is a form of click chemistry in which a terminal azide is reacted with an alkyne

incorporated into a cyclooctyl or cyclononyl ring. To further strain the already reactive cyclic

structure, electron-withdrawing groups have been incorporated adjacent to the alkyne in order to

make it even more reactive with the complimentary azide. The most commonly utilized groups

are difluoro- and dibenzo- moieties (Ning, Guo, Wolfert & Boons, 2008; van Berkel et al.,

2007). SPAAC allows for a simple and straightforward production method which does not

require special equipment, high pressures, or elevated temperatures.

Additionally, SPAAC is a biorthogonal reaction which provides a stable, covalent linker

between protein and PEG. Neither alkyne nor azide are reactive with common functional groups

found in the body (Baskin & Bertozzi, 2009). This non-reactivity eliminates the possibility of

side-reactions between residual, unreacted SPAAC groups and native proteins upon

administration. The BSAx-R-123t-PEG8 complex will pass through a variety of environments as

61

it circulates through the body. The stable 1,2,3-triazole linker will reduce the chances of

dissolution of the PEG complex before delivery of ligand (Massarotti et al., 2014).

Figure 21. Conjugation of bovine serum albumin onto PEG8 via strain-promoted azide-alkyne cycloaddition. (a) BSA is first activated with DIBO-R-NHS at a single lysine. Activated BSA is attached to the PEG8 backbone via SPAAC. “R” denotes a variable spacer incorporated into the activating reagent. (b) DIBO-NHS and (c) DIBO-PEG4-NHS were used to activate BSA.

In this study, dibenzocyclooctyl-functionalized BSA molecules (DIBO-R-BSA) will be

conjugated onto an 8-arm PEG backbone functionalized with terminal azides (N3-PEG8) (Figure

21). BSA will be “activated” with one of two N-hydroxysuccinimidyl ester-functionalized DIBO

reagents (DIBO-R-NHS) via one of BSA’s 59 native lysine residues (Cline & Hanna, 1988).

A short activator, Dibenzocyclooctyne-N-hydroxysuccinimidyl ester (DIBO-NHS), was

used from the outset of this study. In an attempt to increase yields, the longer

62

Dibenzocyclooctyne-PEG4-N-hydroxysuccinimidyl ester (DIBO-PEG4-NHS) was added. It was

hypothesized that once a few BSA molecules attached to the backbone, additional proteins would

be unable to access the remaining PEG arms due to steric hindrance. The PEG4 spacer provided

an additional 14 Å between protein and alkyne. The reduced congestion at the attachment sites

between DIBO and N3 was expected to increase yields. Also, with easier access to the PEG

backbone, heterogeneity of the product was expected to decrease, resulting in fewer species of

BSAx-R-123t-PEG8.

A 1:1 activation ratio of DIBO:BSA will be used and the PEGylation methodology will

be optimized to give the highest yield of PEGylated protein. The degree of activation will also be

determined using Fluorescein-NHS (Fluor-NHS) to mimic the interaction of BSA and the

activating agents. The PEGylated complexes will then be characterized in terms of secondary

structure, speciation, and ligand-binding capabilities.

Materials and Methods

Materials

Lyophilized BSA (Product Number A7906), DIBO-NHS (761524), DIBO-PEG4-NHS

(764019), Fluorescein-5-EX N-hydroxysuccinimide ester (Fluor-NHS; F9551), sulfamethoxazole

(SMZ; S7507), naproxen (NPX; N8280), and all reagents not detailed were purchased from

Sigma-Aldrich.

Methods

All experiments were carried out in 50 mM 4-Morpholinepropanesulfonic acid/0.15 M

NaCl pH 7.2 (MOPS/NaCl) buffer unless otherwise noted. Fast protein liquid chromatography

(FPLC) separations were performed on an ÄKTAprime plus chromatography system (Amersham

Biosciences; 11001313) with MOPS/NaCl mobile phase. Resulting fractions were concentrated

63

with 30 kDa MW cut-off centrifuge filters (Sartorius; VS2022). Absorbance values were

obtained on an Agilent 8453 UV-visible Spectroscopy System. BSA concentrations were

calculated at 280 nm with an extinction coefficient of 43.824 mM-1 cm-1 and a MW of 66,430 Da

(Hirayama, Akashi, Furuya & Fukuhara, 1990; "Thermo Scientific: Extinction Coefficients,"

2013).

Fatty acid removal. Fatty acid-free BSA (FAF-BSA) was prepared via an activated

charcoal method (Chen, 1967). In short, BSA was dissolved in MOPS/NaCl. The solution was

acidified to pH 3 and stirred on ice with 20-40 mesh particle size activated charcoal in order to

remove any bound fatty acids. Charcoal was pelleted at 10,000 rpm and the supernatant was

filtered through 0.2 µm cellulose filters to remove any remaining carbon. The de-fatted albumin

solution was returned to pH 7.2 before further purification. Dimers were removed by size

exclusion chromatography (SEC) on a pre-packed 120 mL 60 cm x 16 mm S300 FPLC column

(GE Healthcare; 17-1167-01).

Activation of FAF-BSA with DIBO. Conjugation of BSA molecules onto a 40 kDa N3-

PEG8 backbone required two steps, activation and PEGylation (Figure 21). Table 5 details the

nomenclature used to discuss these reactions.

Table 5. Nomenclature for the Conjugation of BSA onto PEG8 via Strain-Promoted Azide-Alkyne Cycloaddition. Reaction with complimentary azide for all DIBO-functionalized BSA reactions results in a 1,2,3-triazole (123t).

Activator Activated BSA PEG Product DIBO-R-NHS

(generic) DIBO-R-BSA BSAx-R-123t-PEG8

DIBO-NHS (short) DIBO-BSA BSAx-123t-PEG8

DIBO-PEG4-NHS (long) DIBO-PEG4-BSA BSAx-PEG4-123t-PEG8

64

Activation involves the attachment of a DIBO functional group onto BSA via reaction

between an NHS-ester and the amino side-chain of lysine. DIBO-NHS and DIBO-PEG4-NHS

stocks were prepared in dry dimethyl sulfoxide and added to 1 mM FAF-BSA to give a final

ratio of 1:1 DIBO-R-NHS:FAF-BSA. Introduction of DIBO-R-NHS was made in small aliquots

with mixing between additions to avoid BSA denaturation due to the organic solvent. After two

hours at room temperature, excess activator was removed by triple-rinsing the sample through 30

kDa centrifuge filters ("Thermo Scientific: NHS-Azide and NHS-Phosphine Reagents," 2014).

PEGylation of DIBO-activated FAF-BSA. N3-PEG8 (Creative PEGworks; PSB-884)

was dissolved in MOPS/NaCl and added to 1 mM DIBO-activated FAF-BSA at 16:1 DIBO-R-

BSA:N3-PEG8. The solution was allowed to incubate at room temperature for one week with

constant stirring. PEGylated FAF-BSA was purified via a two-phase SEC separation. The bulk

of unreacted FAF-BSA was removed by way of an S300 FPLC column, the remainder via a 320

mL 60 cm x 26 mm S500 column (GE Healthcare; 28-9356-07). Chromatograms from the S300

separation were processed in PrimeView 5.0 software to determine yields as the percentage of

total area count.

PEGylation optimization. Optimization of the PEGylation methodology was determined

by monitoring the amount of BSAx-123t-PEG8 produced with varying PEGylation ratios,

temperatures, and incubation times. PEGylation ratios were monitored at 4, 8, 16, 32, and 128:1

DIBO-BSA:N3-PEG8. Reactions were carried out at 4oC, room temperature, and 32oC for one

week. The time trial was performed by determing the amount of product formed after two, five,

and seven days of incubation. Weekly assays were performed over the course of an additional

four weeks. The baseline conditions for all PEGylation reactions were 16:1 DIBO-BSA:N3-PEG8

65

at room temperature for one week with constant stirring. Samples were passed through an S500

column and the amount of product was determined as the percentage of total area count.

Fluorescent labeling. The degree of activation was determined using Fluor-NHS to

mimic the DIBO activators. Fluor-NHS has a MW of 590.56 g/mol which is similar to both DIBO-

NHS (MW 402.4) and DIBO-PEG4-NHS (MW 649.69). The activation procedure was followed,

substituting Fluor-NHS. Unreacted fluorophore was removed via triple rinse with 50 mM

Tricine/0.15 M NaCl pH 8.5 through 30 kDa centrifugation filters. The following calculations

used absorbance at 280 nm (λmax for BSA) and 495 nm (λmax for Flour) to determine

Fluorbound:FAF-BSA ("Thermo Scientific: NHS-Fluorescein," 2016).

MolarFluorBSA =

Abs/01 ∗ CAbs456 − CF ∗ Abs/01

CF = Abs456ofunreactedFluorAbs/01ofunreactedFluor

C =MW?@A Da ∗ E4566.E%

MWGHIJK ∗ 195

195 = E/016.E% of bound Fluor

MWFluor = 475.47 Da (MWFluor-NHS minus MWNHS which is lost upon reaction with lysine)

Circular dichroism (CD). The effects on secondary structure of the PEGylation

procedure were determined by CD. BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8 were assayed

in 50 mM MOPS pH 7.2 at 30oC on an Olis DSM 20 CD Spectrophotometer. Ellipticity was

monitored between 190 nm and 260 nm. Helicity was calculated with DichroWeb (Whitmore &

Wallace, 2008) and the analysis program CDSSTR (Sreerama & Woody, 2000). Inclusion of a

66

crystal structure for HSA (Wardell et al., 2002) allowed for use of reference set SP175 (Janes,

2009; Lees, Miles, Wien & Wallace, 2006).

SDS-PAGE. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)

was performed as an initial characterization of BSAx-R-123t-PEG8 species. The assay was

performed on a PhastSystem (Amersham Biosciences; 18-1018-23) with a 4-15% gradient

PhastGel (GE Healthcare; 17-0678-01), Coomassie R 350 stain (GE Healthcare; 17-0518-01),

and high molecular weight markers (GE Healthcare; 17-0615-01).

Analytical ultracentrifugation (AUC). AUC was carried out in a Beckman Coulter

ProteomeLab XL-A analytical ultracentrifuge using standard two sector cells with quartz

windows in four- or eight-hole rotors. All sedimentation velocity experiments were performed at

50,000 rpm with a preset scan number of 100 per cell. The reference sector was filled with 450

µL of MOPS/NaCl and the sample sector was filled with 430 µL of sample with an absorbance at

280 nm between 0.75 and 1.1. Path length was 1 cm. Buffer density was measured at 20.0oC in a

Mettler/Paar Calculating Density Meter DMA 55A and the buffer viscosity was measured in an

Anton Parr AMVn Automated Microviscometer. Data were analyzed using the program Sedfit

(Schuck, 2000; "Sedfit," 2017). Molecular weights and sedimentation coefficients (S20,w) were

determined using the continuous c(s) mode in the Sedfit program. Native monomeric BSA

yielded an S20,w value of 4.04 S which agreed with published values (Shulman, 1953; Taylor,

1952; Zhao et al., 2015). The c(s) plots were deconvoluted using non-linear least-squares fitting

algorithms available in the Origin (v. 9.0) software package (OriginLab Corporation,

Northampton, MA). Data were fit to multiple Gaussian components, increasing the number of

components until residual plots (data-fit) were random and the chi square value for the fit was

67

minimal. Values reported are percentages of the total area count of peaks corresponding to

PEGylated species. These area counts were calculated via the trapezoidal formula.

Fluorimetric analyses. Binding studies were performed on a Photon Technology

International fluorimeter using Felix software (version 1.1). SMZ, a DS1 ligand, and NPX, a

DS2 ligand, were studied to determine if PEGylation affected ligand binding capabilities. The

number of binding sites (n) and the association constant (Ka) were calculated for BSAx-123t-

PEG8 and BSAx-PEG4-123t-PEG8. In each case, the fluorophore monitored was Trp213, found at

the entrance to DS1 in Domain IIA.

For SMZ, 30 µM stock solutions of FAF-BSA, BSAx-123t-PEG8, and BSAx-PEG4-123t-

PEG8 were prepared in 0.2 M Tris-HCl/0.1 M NaCl pH 7.4. SMZ was dissolved in methanol to a

final stock concentration of 30 mM. Working SMZ stocks were prepared at 3 mM and 1 mM by

dilution with Tris buffer. Monitoring emission between 300-380 nm upon excitation at 283 nm at

25oC, aliquots of working SMZ stocks were titrated into a 3 µM BSA solution. 1 mM working

stock was used up to 20 µM SMZ. 3 mM working stock was used to raise the final concentration

to 80 µM SMZ.

To determine n and Ka for NPX, 30 µM stock solutions of FAF-BSA, BSAx-123t-PEG8,

and BSAx-PEG4-123t-PEG8 were prepared in 10 mM phosphate buffer pH 7.0. NPX was

dissolved in methanol to a final stock concentration of 45 mM. A working NPX stock was

prepared at 0.15 mM by dilution with phosphate buffer. Monitoring emission between 300-380

nm upon excitation at 295 nm at 25oC, aliquots of working NPX stocks were titrated into a 0.725

µM BSA solution to a final concentration of 3 µM NPX.

Blanks for each ligand at each concentration were necessary to account for any

fluorescence due to the ligand itself. SMZ had almost no signal under experimental conditions.

68

NPX, however, showed strong emission with maxima at approximately 350 nm. Particularly at

high concentrations, NPX signal obscured the BSA maxima which ranges from 344 to 350 nm.

Non-linear regression analyses were carried out using The Gnumeric Spreadsheet

(version 1.10.16) ("The Gnome Project: The Gnumeric Spreadsheet: Free, Fast, Accurate - pick

any three," 2011) and a modified Hill equation. In order to normalize samples, fractional

intensities (FrInt) were used in place of raw intensities.

FrInt = FrIntPQR–FrIntPTU ∗ Ligand U

KZU + Ligand U +FrIntPTU

FrInt = (IntT]^_T`QUZ − Int6]^_T`QUZ)

(IntPQR]^_T`QUZ − Int6]^_T`QUZ)

KQ =1KZ

Results and Discussion

Methodology Optimization

The PEGylation portion of the polymerization procedure was optimized by varying the

PEGylation ratios, temperature, and reaction time. FAF-BSA was activated at a 1:1 mole ratio

with DIBO-NHS, triple-rinsed with MOPS/NaCl, and concentrated to 1 mM with 30 kDa

centrifuge filters before PEGylation. Samples were evaluated for BSAx-123t-PEG8 content as the

percent of the total area count after separation on an S500 column.

69

(a)

PEGylation ratio

(DIBO-BSA:N3-PEG8) Temperature

(b)

Figure 22. Optimization of the PEGylation procedure for DIBO-activated FAF-BSA. (a) PEGylation ratio and temperature were varied to produce the highest yields. (b) A time trial of PEGylation at 1:1 activation with DIBO-NHS, 16:1 DIBO-BSA:N3-PEG8 at room temperature.

PEGylation ratio proved to be the most effective control over product formation (Figure

22). Varying PEGylation ratios from 4:1 DIBO-BSA:N3-PEG8 to 128:1 showed a maximum of

31% BSAx-123t-PEG8 product formed at 16:1. When this ratio was pushed to a higher excess of

DIBO, the lowest yield of 4% was obtained at 128:1. This could be due to dilution of the N3-

PEG8, resulting in less interaction between azide and alkyne. When in excess of N3-PEG8, a

similarly low 9% is obtained. At 16:1, the optimal balance between DIBO-BSA:N3-PEG8 ratio

and individual reactant concentrations was found.

9.0

16.6

30.9

18.8

4.2

28.330.9

27.1

0

5

10

15

20

25

30

35

40

45

50

4:1 8:1 16:1 32:1 128:1 4C RT 37C

% T

otal

Are

a C

ount

0

5

10

15

20

25

30

35

40

45

50

0 7 14 21 28 35

% T

otal

Are

a C

ount

Time (days)

70

Temperature proved to be the least effective control. Assays performed at room

temperature generated the highest yields. Slightly less product was formed at a lower

temperature (4oC) with similar results at an elevated temperature (37oC). The negligible

difference in yields suggests an indifference to temperature.

One week of reaction time was sufficient to generate a considerable amount of product.

Yields trended upwards over five weeks. While reacting for longer periods generated more

product, majority of the BSAx-123t-PEG8 was formed within the first seven days. Given an

already extensive reaction time, additional weeks of incubation were deemed unnecessary for

only slightly more product.

Effects on Secondary Structure

Helicity content for BSA has been reported between 56-66% (Das et al., 2017; Moriyama

et al., 2008). Characteristic minima at 210 nm and 220 nm were seen in all spectra of PEGylated

BSA (Figure 23). Unmodified FAF-BSA contained 59.2 ± 2.2% α-helix. Neither BSAx-123t-

PEG8 (63.4 ± 2.1%) nor BSAx-PEG4-123t-PEG8 (59.3 ± 3.5%) showed any loss in secondary

structure.

71

Figure 23. Circular dichroic spectra of FAF-BSA and strain-promoted azide-alkyne cycloaddition PEGylated products. FAF-BSA (---); BSAx-123t-PEG8 (blue); BSAx-PEG4-123t-PEG8 (orange). Degree of Activation

A 1:1 activator:FAF-BSA mole ratio was necessary to reduce the risk of over-activation

resulting in a single BSA making multiple connections to a single or multiple N3-PEG8

backbones. The degree of activation was quantified using Fluor-NHS that would mimic the

interactions between FAF-BSA and either activating agent. By monitoring the absorbance

maxima at 280 nm and 495 nm, an average of 0.75 ± 0.04 Fluor per BSA was calculated. By

controlling the mole ratio of activator:FAF-BSA, the odds of multiple attachments was greatly

decreased.

-80

-70

-60

-50

-40

-30

-20

-10

0

10

20

190 200 210 220 230 240 250 260

Ellip

ticity

(mill

ideg

rees

)

Wavelength (nm)

72

BSAx-R-123t-PEG8 Yields

While the risk of over-PEGylation was limited, so too was the yield of PEGylated

protein. By integrating chromatograms from the S300 separation (Figure 24a), yields were

calculated for both DIBO reagents. The hypothesis that an additional 14 Å would be sufficient to

greatly increase yields was proven incorrect. The shorter DIBO-NHS produced an average of

23.6% BSAx-123t-PEG8 (range: 19.5-31.0%). The longer DIBO-PEG4-NHS generated only

slightly more product (average: 25.3%, range: 19.1-36.5%). The additional length of the

extended alkyne activating agent did not result in substantially higher yields but does suggest

that further separation of the bulky BSA from the bulky backbone of PEG8 may enhance yields.

73

(a)

(b)

Figu

re 2

4. P

urifi

catio

n of

BSA

x-R

-123

t-PEG

8 via

siz

e ex

clus

ion

chro

mat

ogra

phy.

(a) I

nitia

l pur

ifica

tion

was

don

e on

an

S300

. (b

) Fin

al p

urifi

catio

n w

as d

one

an S

500

colu

mn.

FA

F-B

SA (-

--);

BSA

x-12

3t-P

EG8 (

blue

); B

SAx-

PEG

4-12

3t-P

EG8 (

oran

ge).

020406080100

020

4060

8010

012

0

% Maximum Absorbance

Elut

ion

Volu

me

(mL)

020406080100

050

100

150

200

250

300

350

% Maximum Absorbance

Elut

ion

Volu

me

(mL)

74

BSAx-R-123t-PEG8 Speciation

Purified BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8 were run through an SDS-PAGE

4-15% gradient gel. A band at approximately 116 kDa corresponds to a combination of dimeric

BSA and BSA1-R-123t-PEG8 conjugates (Figure 25). High molecular weight bands in the

separation zone and a heavy band retained in the stacking zone were seen for both BSAx-123t-

PEG8 and BSAx-PEG4-123t-PEG8. This suggests a minimum of three to four species of

PEGylated FAF-BSA were formed with either activating agent. The bands likely correspond to

PEG8 +2 FAF-BSA (172 kDa), +3 FAF-BSA (238 kDa), and +4 FAF-BSA (304 kDa). Smearing

in the separating zone was due to the interaction of the PEG backbone with the polyacrylamide

gel as well as variability of the specific lysine residue activated by the DIBO reagents.

Manufacturer specifications quote the upper limit of separation for these gels at 300 kDa for non-

PEGylated proteins. It is inferred that material retained in the stacking zone consists of a

minimum of five FAF-BSA molecules attached to the PEG8 backbone with a MW of

approximately 370 kDa.

75

Figure 25. Electrophoretic separation of species of bovine serum albumin PEGylated via strain-promoted azide-alkyne click chemistry. Lane 1: FAF-BSA; Lane 2-3: BSAx-123t-PEG8; Lane 4-5: BSAx-PEG4-123t-PEG8; Lane 6: High molecular weight marker.

The speciation of purified PEGylated BSA was examined by AUC. The majority of the

PEGylated species generated with both activators were complexes containing BSA2-4. These

species accounted for 55% of BSAx-123t-PEG8 (Figure 26) and 65% of BSAx-PEG4-123t-PEG8

(Figure 27). BSA5-7-R-123t-PEG8 conjugates were present in much lower concentrations, making

up 31% and 14% of short and long activator species, respectively. The shorter DIBO-BSA

resulted in 8% BSA1 species while the longer DIBO-PEG4-BSA generated 19%. The peaks

found at S20,w values greater than 15 S are thought to be a mixture of BSA8-R-123t-PEG8 species

and complexes formed from over-activated protein.

76

Peak 1 2 3 4 5 6 7 8 9 S20,w 3.23 4.65 6.26 7.47 9.09 10.3 12.1 14.1 16.8

MWcalc (kDa) 57.5 92.2 143 196 283 nr 396 493 609

MWtheor (kDa) 66 106 172 238 304 370 436 502 n/a

# BSA conjugated monomer 1 2 3 4 5 6 7 n/a

% Area Count n/a 8 17 18 20 11 11 9 6

Figure 26. Analytical ultracentrifugation analysis of BSAx-123t-PEG8. % Area Count values are the percent of PEGylated species only. nr: not reported; n/a: not applicable.

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.16

0 2 4 6 8 10 12 14 16 18 20

c(s)

S20,w

1

2

3 4

5

6

7

89

77

Peak 1 2 3 4 5 6 7 8 9 S20,w 3.03 4.65 6.06 7.27 8.89 10.5 12.1 13.9 15.8

MWcalc (kDa) 71.3 135 201 279 nr 481 nr nr nr

MWtheor (kDa) 66 106 172 238 304 370 436 502 n/a

# BSA conjugated monomer 1 2 3 4 5 6 7 n/a

% Area Count n/a 19 25 26 14 8 4 2 2

Figure 27. Analytical ultracentrifugation analysis of BSAx-PEG4-123t-PEG8. % Area Count values are the percent of PEGylated species only. nr: not reported; n/a: not applicable.

AUC results corresponded well with the SDS-PAGE analysis. Each purified sample

contained BSA1-R-123t-PEG8 with a much higher degree of BSA2-4-R-123t-PEG8. Higher

0.00

0.05

0.10

0.15

0.20

0.25

0 2 4 6 8 10 12 14 16 18 20

c(s)

S20,w

1

23

4

5

6

78

9

78

molecular weight material was present but a lower concentration. The longer DIBO-PEG4-NHS

generated only slightly more PEGylated species than DIBO-NHS based on SEC data. However,

the distribution of those species was smaller, generating a more homogenous final product.

Ligand Binding Capabilities

To compare the effects of PEGylation on the ability of conjugated BSA to bind ligands,

fluorescent emission intensity of Trp213 was monitored. The ability of PEGylated species to

bind ligands in DS2 was probed with NPX. Crystal structures of de-fatted bovine, equine, and

leporine serum albumins complexed with NPX have been reported (Bujacz et al., 2014). In each

species of serum albumin, NPX preferentially binds to DS2. A second molecule of NPX binds to

fatty acid binding site 6. In leporine serum albumin, a tertiary site for NPX is found in a

superficial cleft in subdomain IIIA while in BSA, the tertiary binding site is DS1. Experimental

values for NPX in the presence of FAF-BSA, BSAx-123t-PEG8, and BSAx-PEG4-123t-PEG8

were consistent with published values (Table 6). NPX was tightly bound by all species of BSA.

Table 6. Association Constants (Ka) and Number of Binding Sites (n) for BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8 for Sulfamethoxazole and Naproxen.

SMZ NPX Sample Ka (x103 M-1) n Ka (x107 M-1) n

FAF-BSA 2.03 ± 0.18 0.9 ± 0.1 0.40 ± 0.13 0.4 ± 0.1 BSAx-123t-PEG8 0.39 ± 0.29 0.8 ± 0.1 0.79 ± 0.61 2.4 ± 1.1 BSAx-PEG4-123t-PEG8 0.15 ± 0.06 0.8 ± 0.05 3.55 ± 1.6 0.9 ± 0.2

Lit. values1,2 3.55-15.6 0.9-1.0 0.12-3.70 0.6-3

1SMZ values were published by Naik et al. (2009; 2015), Rajendiran & Thulasidhasan (2015), and Wang et al. (2014). 2NPX values were reported by Banerjee et al. (2006), Bou-Abdallah et al. (2016), Fielding et al. (2005), and Honoré & Brodersen (1984). Bujacz et al. (2014) have reported a crystal structure (PDB ID: 4OR0) for BSA complexed with three NPX molecules.

79

Binding of NPX to FAF-BSA, BSAx-123t-PEG8, and BSAx-PEG4-123t-PEG8 showed

similar trends (Figure 28). BSA in the absence of NPX had a maximum emission wavelength at

344 nm. With increasing amounts of NPX, a red shift resulted in a maximum at 350 nm. This is

consistent with binding to a site adjacent to the fluorophore. In this case, the fluorophore studied

was Trp213 found in Domain IIA at the mouth of DS1. As ligand binds to DS2, slight

conformational changes exposed Trp213 to the solvent, causing a red shift.

80

(a)

[NPX] increasing

(b)

[NPX] increasing

(c)

Figure 28. Fluorimetric spectra and non-linear regression of BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8 in the presence of naproxen. Spectra for (a) BSAx-PEG4-123t-PEG8 and (b) FAF-BSA in the presence of increasing concentrations of NPX. BSAx-123t-PEG8 spectra followed similar trends as both FAF-BSA and BSAx-PEG4-123t-PEG8. (c) Non-linear regression analyses for FAF-BSA (black ¯), BSAx-123t-PEG8 (blue r), and BSAx-PEG4-123t-PEG8 (orange �). Experimental data is represented as symbols while calculated data points are lines.

0.0E+00

2.5E+04

5.0E+04

7.5E+04

1.0E+05

1.3E+05

320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

4.0E+04

6.5E+04

9.0E+04

1.2E+05

320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.0 0.5 1.0 1.5 2.0 2.5 3.0

Frac

tiona

l Int

ensi

ty

[NPX] (µM)

81

DS1, however, appears to be greatly affected by the covalent attachment of PEG8. FAF-

BSA values were similar to published values for both Ka and n for SMZ studies (Table 6).

However, both BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8 demonstrated a greatly reduced

affinity for SMZ. Spectra for all forms of BSA followed similar trends (Figure 29). Intensities

decreased as SMZ concentration increased. The signal from Trp213 was quenched in the

presence of the ligand though it was not tightly bound in either form of PEGylated BSA.

82

(a)

[SMZ]

increasing

(b)

[SMZ]

increasing

(c)

Figure 29. Fluorimetric spectra and non-linear regression of BSAx-R-123t-PEG8 in the presence of sulfamethoxazole. Spectra for (a) BSAx-123t-PEG8 and (b) FAF-BSA in the presence of increasing concentrations of SMZ. BSAx-PEG4-123t-PEG8 spectra followed similar trends as both FAF-BSA and BSAx-123t-PEG8. (c) Non-linear regression analyses for FAF-BSA (black ¯), BSAx-123t-PEG8 (blue r), and BSAx-PEG4-123t-PEG8 (orange �). Experimental data is represented as symbols while calculated data points are lines.

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

2.5E+06

300 310 320 330 340 350 360 370 380

Inte

nsity

Wavelength

0.0E+00

5.0E+05

1.0E+06

1.5E+06

2.0E+06

2.5E+06

300 310 320 330 340 350 360 370 380

Inte

nsity

Wavelength (nm)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 10 20 30 40 50 60 70 80

Frac

tiona

l Int

ensi

ty

[SMZ] (µM)

83

Similar results were seen for PEGylation at Cys34 with maleimido-PEG8. Upon

modification of this single residue in Domain IA, the affinity for SMZ was reduced to 0.19 x103

M-1. It was hypothesized that conformational changes induced by PEGylation at Cys34 altered

the structure of DS1 which is found in Domains IB and IIA. The conditions employed for

SPAAC PEGylation allow for selective modification of lysines, but do not preferentially alter

specific lysines on the surface of BSA. Given the results from the modification of Cys34, it is

possible that lysines in Domain I are preferentially altered by DIBO-R-NHS activators.

Conclusions

The conjugation of BSA onto a PEG8 backbone via SPAAC proved successful using both

forms of DIBO activator. The PEGylation process proved to be most dependent on the ratio of

DIBO-BSA:N3-PEG8 with an optimal reaction ratio of 16:1.

The shorter DIBO-NHS PEGylated an average of 23% FAF-BSA. The extended DIBO-

PEG4-NHS yielded only slightly more product at 25%. The additional 14 Å was of little

assistance in terms of yield. The longer activating agent did, however, produce a more

homogenous final product, generating a majority of BSA2-4-R-123t-PEG8 species.

In terms of ligand binding capabilities, neither activating agent greatly affected BSA’s

ability to bind NPX, a model DS2 ligand. DS1 was affected by PEGylation as evidenced by

greatly reduced association constants.

The PEGylation procedure proved to be successful regardless of the activating agent.

High molecular weight complexes of FAF-BSA were still capable of binding ligands to one of

the main drug binding sites. These complexes show promise as a possible drug delivery method.

84

CHAPTER V

STUDIES ON THERMAL STABILITY AND

THE EFFECTS OF PASTEURIZATION ON

PEGYLATED MULTI-ALBUMIN COMPLEXES

Introduction

PEGylation has become a common method of increasing circulatory lifetimes of protein-based

therapies. Due to increased hydrodynamic volume conferred by the attachment of a polyethylene

glycol (PEG) backbone, the PEGylated complexes appear several times larger than molecular

weight would suggest (Plesner, Fee, Westh & Nielsen, 2011; Veronese, Mero & Pasut, 2009).

Increased size allows the particle to escape elimination and remain in circulation for longer

periods of time.

In order to increase solubility and provide protection for hydrophobic drugs, serum

albumins have been incorporated into drug carriers (Anhorn, Wagner, Kreuter, Langer & von

Briesen, 2008; Desai, 2008). Bovine serum albumin (BSA) contains a multitude of binding sites

which are tailored for endogenous molecules including fatty acids and heavy metals

(Bhattacharya, Grüne & Curry, 2000; Rózga, Sokołowska, Protas & Bal, 2007). Fatty acid

binding pockets, have been found to bind drugs but are typically secondary or tertiary options.

The primary hydrophobic cavities for drug binding were initially characterized by Sudlow et al.

(1975, 1976). Deemed Drug Sites I and II (DS1 and DS2), these sites are able to house a variety

of ligands (Ghuman et al., 2005; Yamasaki, Chuang, Maruyama & Otagiri, 2013).

85

DS1 is found at the interface of Domains IB and IIA while DS2 resides in Domain IIIA. Access

to these sites is controlled via flexing of the local tertiary structure.

BSA contains three highly helical domains (I, II, and III), each divided into two

subdomains (A and B) which are connected by long loops. Conformational flexibility of each

domain is conferred by helical bending and subdomain orientation is controlled by movement of

the loops. These types of movements open and close the various binding sites dotted throughout

the structure.

While the exterior of albumin is somewhat flexible, the interior of the protein is rigidly

maintained via 17 disulfide bridges. Subdomain A is stapled together by four of these while two

bridges stabilize subdomain B. An exception is Subdomain IA, which contains only three bridges

(Bujacz, 2012). A single free cysteine at position 34 is found on the surface of Domain I. This

superficial residue allows for selective, covalent attachment of a PEG chain.

Modifications of cysteines are often made via thiol-maleimide reactions, resulting in a

thioether linkage (Anhorn et al., 2008; Manjula et al., 2003; K. D. Webster et al., 2017). This

class of reaction has become attractive due to typically simple reaction conditions as it can

readily proceed in an aqueous environment and requires neither catalyst nor high temperatures.

In addition to this free cysteine, 59 lysines are scattered throughout BSA. Terminal amino

groups allow for modification with a number of different functional groups. One of the more

common reactions is via an N-hydroxysuccinimide (NHS) ester (Cline & Hanna, 1988). The

reactivity between NHS and lysine provides a convenient way to anchor another functional

group to the protein.

The abundance and easy access to these lysines allows for attachment of a foreign moiety

without the need for recombinant techniques. Strain-promoted azide-alkyne cycloaddition

86

(SPAAC), a class of “click” chemistry, results a highly stable 1,2,3-triazole linker. Additionally,

both azide and alkyne are biorthogonal in that they are absent from nature and non-reactive with

naturally-occurring functional groups (Agard, Prescher & Bertozzi, 2004; Baskin & Bertozzi,

2009). By using the NHS-lysine reaction, a strained alkyne may be covalently attached to BSA.

PEG functionalized with the complimentary azido moiety may then be reacted with the alkyne-

activated BSA to link the two molecules.

PEGylation, while providing many benefits to the modified protein, have also been

shown to affect thermal stability, which may be described by the temperatures of denaturation

(Tm) and aggregation (Tagg). Tm is the temperature at which a protein is 50% unfolded (K. W.

Olsen, 1994). Tagg describes the ability of both native and non-native forms of a protein to form

oligomerize (Plesner et al., 2011; Wetzel et al., 1980). Both are important parameters to consider

during the development of protein-based drug delivery systems as they factor into shelf-life and

terminal sterilization.

In the case of BSA, typical Tm is approximately 82oC as determined by differential

scanning calorimetry (Plesner et al., 2011). Attachment of a single linear 40 kDa PEG chain at

Cys34, lowered Tm to 80oC. Tagg, on the other hand, increased upon PEGylation. Unmodified

BSA aggregated at 71oC while the 40 kDa PEG-BSA species had a Tagg of 82oC.

Steam sterilization, or autoclaving, is the most common approach used in pharmaceutical

production. Materials are heated to 121oC at 1.5 bar for a minimum of 15 minutes. These

conditions are likely to denature protein components. For this reason, protein-based formulations

are sterilized aseptically, an approach where each component is sterilized individually before

preparation of the final product (Baez & Assaf-Anid, 2008; Yaman, 2001).

87

Albumin-based components are pasteurized at 60oC for ten hours. In order to maintain

helicity, the stabilizing reagents N-acetyl-L-tryptophan (NAT) and sodium caprylate are included

at mole ratios of 5.5:1 stabilizer:albumin (Anraku et al., 2004; Christiansen & Skotland, 2010;

Farcet, Kindermann, Modrof & Kreil, 2012; Gellis et al., 1948).

High concentrations of these stabilizers have been implicated in the reduced efficacy of

albumin solutions administered to patients with impaired liver function. For these patients,

albumin, which is produced in the liver, is typically found at extremely low concentrations.

Exogenous albumin, which has been pasteurized with NAT and caprylate, is often administered

intravenously to compensate. Both reagents have been shown to have an affinity for DS2, a site

also shared by metabolites commonly formed at high levels in patients with liver disease (Kawai

et al., 2017).

In healthy individuals, NAT and caprylate bind to albumin. Albumin production is

unimpaired and routine transport of NAT, caprylate, and other metabolites is unaffected. In

impaired individuals, elevated levels of these toxins are often found due to direct competition

with stabilizers for a single high affinity binding site (Stange et al., 2011).

These additives also pose a potential problem for production of albumin-based drug

delivery systems. Their preference for DS2 results in competition between NAT, caprylate, and

the drug molecule being administered. For naproxen (NPX), a known DS2 ligand, affinity was

40 times lower in pharmaceutical-grade albumin solutions (i.e. pasteurized with NAT and

caprylate) than in research-grade albumin (i.e. un-pasteurized) (H. Olsen, Andersen, Nordbo,

Kongsgaard & Bormer, 2004).

In this study, the effects of PEGylation on the thermal stability of BSA were examined.

Three PEGylated BSA complexes were prepared. In the first case, native Cys34 of BSA was

88

used as a site of direct attachment of an eight-armed PEG backbone with terminal maleimides

(Mal-PEG8) resulting in a thioether linker. For the two others, BSA were attached to PEG8 via

SPAAC click chemistry. BSA was “activated” with either a short- or long-chain

dibenzocyclooctyne-functionalized N-hydroxysuccinimidyl ester (DIBO-NHS; DIBO-PEG4-

NHS). Activated BSA were attached to an azido-functionalized PEG8 backbone (N3-PEG8)

resulting in a 1,4-disubstituted, 1,2,3-triazole. Changes in the absorbance at 300 nm as a function

of temperature were used to calculate Tagg while Tm was determined for each PEG species by

circular dichroism (CD). PEGylated BSA, pre-loaded with NPX, were pasteurized in the absence

of NAT and caprylate to determine if a combination of PEGylation and bound ligand would

allow for omission of these stabilizing compounds.

Materials and Methods

Materials

Lyophilized BSA (Product Number A7906), DIBO-NHS (761524), DIBO-PEG4-NHS

(764019), NPX (N8280), and all reagents not detailed were purchased from Sigma-Aldrich.

Methods

All experiments were carried out in 50 mM 4-Morpholinepropanesulfonic acid/0.15 M

NaCl pH 7.2 (MOPS/NaCl) buffer unless otherwise noted. Fast protein liquid chromatography

(FPLC) separations were performed on an ÄKTAprime plus chromatography system (Amersham

Biosciences; 11001313) with MOPS/NaCl mobile phase. Resulting fractions were concentrated

with 30 kDa molecular weight cut-off centrifuge filters (Sartorius; VS2022). Absorbance values

were obtained on an Agilent 8453 UV-visible Spectroscopy System at 280 nm. Concentrations of

BSA were calculated with an extinction coefficient of 43.824 mM-1 cm-1 and a molecular weight

89

of 66,430 Da (Hirayama, Akashi, Furuya & Fukuhara, 1990; "Thermo Scientific: Extinction

Coefficients," 2013).

Fatty acid removal. Fatty acid-free BSA (FAF-BSA) was prepared via an activated

charcoal method (Chen, 1967). In short, BSA was dissolved in MOPS/NaCl. The solution was

acidified to pH 3 and stirred on ice with 20-40 mesh particle size activated charcoal in order to

remove any bound fatty acids. Charcoal was pelleted at 10,000 rpm and the supernatant was

filtered through 0.2 µm cellulose filters to remove any remaining carbon. The de-fatted albumin

solution was returned to pH 7.2 before further purification. Dimers were removed by size

exclusion chromatography (SEC) on a pre-packed 120 mL 60 cm x 16 mm S300 FPLC column

(GE Healthcare; 17-1167-01).

PEGylation of FAF-BSA. Table 7 details the nomenclature used to discuss these

reactions. Cys34, requiring no further activation, was attached to PEG8 with a thioether (te)

linkage formed by reaction with Mal-PEG8. For SPAAC reactions, the complimentary azide for

all DIBO-functionalized BSA reactions was N3-PEG8, resulting in a 1,2,3-triazole (123t).

Table 7. Nomenclature for PEGylation of Bovine Serum Albumin via Thiol-Maleimide and Strain-Promoted Azide-Alkyne Cycloaddition Reactions.

Activator Activated BSA PEG8 PEG Product Thiol-Maleimide

n/a n/a Mal-PEG8 BSAx-te-PEG8 Strain-Promoted Azide-Alkyne Cycloaddition

DIBO-R-NHS (generic) DIBO-R-BSA

N3-PEG8

BSAx-R-123t-PEG8

DIBO-NHS (short) DIBO-BSA BSAx-123t-PEG8

DIBO-PEG4-NHS (long) DIBO-PEG4-BSA BSAx-PEG4-123t-PEG8

90

PEGylation of Cys34. 40 kDa Mal-PEG8 (Creative PEGworks; PSB-864) was dissolved

in MOPS/NaCl and added to 1 mM FAF-BSA at 8:1 FAF-BSA:Mal-PEG8. The solution was

allowed to incubate at room temperature for one week with constant stirring.

PEGylation with SPAAC. Conjugation of BSA molecules onto a 40 kDa N3-PEG8

backbone required two steps, activation and PEGylation. Activation involves the attachment of a

DIBO functional group onto BSA via reaction between an NHS-ester and the amino side-chain

of lysine. DIBO-NHS and DIBO-PEG4-NHS stocks were prepared in dry dimethyl sulfoxide and

added to 1 mM FAF-BSA to give a final ratio of 1:1 DIBO-R-NHS:FAF-BSA. Introduction of

DIBO-R-NHS was made in small aliquots with mixing between additions to avoid BSA

denaturation due to the organic solvent. After two hours at room temperature, excess activator

was removed by triple-rinsing the sample through 30 kDa centrifuge filters ("Thermo Scientific:

NHS-Azide and NHS-Phosphine Reagents," 2014).

N3-PEG8 (Creative PEGworks; PSB-884) dissolved in MOPS/NaCl was added to 1 mM

DIBO-activated FAF-BSA at 16:1 DIBO-R-BSA:N3-PEG8. The solution was allowed to

incubate at room temperature for one week with constant stirring.

Purification of PEGylated BSA. PEGylated FAF-BSA was purified via a two-phase SEC

separation. The bulk of unreacted FAF-BSA was removed by way of an S300 FPLC column, the

remainder via a 320 mL 60 cm x 26 mm S500 column (GE Healthcare; 28-9356-07).

Spectroscopic evaluation of aggregation. Tagg was determined by monitoring

absorbance at 300 nm as samples were heated from 30 oC to 100oC over two hours. Samples

were initially assayed in MOPS/NaCl. Due to a lack of aggregation of the PEGylated samples,

the buffer was further modified with 1.5 M Guanidinium chloride (GdmCl) and 0.4% (v/v) 2-

Mercaptoethanol (βME). The maximum temperature was held for an additional 20 minutes to

91

ensure the maximum absorbance had been reached. Tagg was calculated following the two-state

method detailed by Olsen (1994).

Additionally, the absorbance at 300 nm was monitored for samples prepared in

MOPS/NaCl/0.4% βME/1.5 M GdmCl at 30oC for 2 hours 20 minutes. This ensured that

unfolding was not due solely to the high concentrations of denaturing reagents.

Circular dichroic evaluation of denaturation. The effect of PEGylation on the thermal

stability of each of these complexes was determined by CD. Samples were diluted to 0.15 mg/mL

in 50 mM MOPS pH 7.2 and heated from 30oC to 90oC in an Olis DSM 20 CD

Spectrophotometer. Spectra were collected at 30, 40, 50, 60, 63, 66, 70, 80, and 90oC over the

range of 190-260 nm with a 1 nm step-size. Helicity was calculated with DichroWeb (Whitmore

& Wallace, 2008) and the analysis program CDSSTR (Sreerama & Woody, 2000). Inclusion of a

crystal structure for human serum albumin (Wardell et al., 2002) allowed for use of reference set

SP175 (Janes, 2009; Lees, Miles, Wien & Wallace, 2006). Helicity at each temperature was

compared to that of 30oC to calculate the % Remaining Helicity. A plot of % Remaining Helicity

as a function of temperature was used to calculate Tm.

Pasteurization studies. To determine the necessity for the stabilizing agents NAT and

caprylate in pasteurization, PEGylated BSA samples were diluted to 0.15 mg/mL in 50 mM MOPS

pH 7.2 and incubated at 60oC for ten hours in the presence and absence of NPX. A 150 mM NPX

stock prepared in methanol was further diluted to a working stock concentration of 0.5 mM with

MOPS buffer. Samples were loaded at a mole ratio of 3:1 BSA:NPX prior to incubation. Control

samples were held at 4oC for the duration of all experiments. All samples were brought to room

temperature before further analysis.

92

Circular dichroic evaluation of secondary structure. CD spectra of pasteurized samples

were collected at 30oC under the conditions described above. For neat solutions of NPX, no

signal was observed for NPX under these conditions.

Fluorimetric evaluation of naproxen binding. Binding studies were performed on a

Photon Technology International fluorimeter using Felix software (version 1.1). NPX was

studied to determine if pasteurization in the absence of stabilizing agents affected ligand binding

in DS2. The number of binding sites (n) and the association constant (Ka) were calculated.

30 kDa centrifuge filters were used to concentrate pasteurized BSA and exchange MOPS

buffer for 10 mM phosphate buffer pH 7.0. 30 µM BSA stock solutions were prepared in

phosphate buffer. NPX was dissolved in methanol to a final stock concentration of 45 mM. A

working NPX stock was prepared at 0.15 mM by dilution with phosphate buffer. Monitoring

emission between 300-380 nm upon excitation at 295 nm at 25oC, aliquots of working NPX

stocks were titrated into a 0.725 µM BSA solution to a final concentration of 3 µM NPX.

NPX blanks at each concentration were necessary to account for any fluorescence due to

the ligand itself. NPX showed strong emission with maxima at approximately 350 nm.

Particularly at high concentrations, NPX signal obscured the BSA maxima which ranges from

344 to 350 nm.

Non-linear regression analyses were carried out using The Gnumeric Spreadsheet (version

1.10.16) ("The Gnome Project: The Gnumeric Spreadsheet: Free, Fast, Accurate - pick any

three," 2011) and a modified Hill equation. In order to normalize samples, fractional intensities

(FrInt) were used in place of raw intensities.

93

FrInt = FrInt()*–FrInt(,- ∗ Ligand -

K5- + Ligand - +FrInt(,-

FrInt = (Int,89:,;)-5 − Int=89:,;)-5)(Int()*89:,;)-5 − Int=89:,;)-5)

K) =1K5

Results and Discussion

Temperature of Aggregation (Tagg)

Initial experiments were performed in MOPS/NaCl (Figure 30a). However, neither FAF-

BSA nor any of the PEGylated species peaked under these conditions. In order to further weaken

the secondary structure of BSA, 0.4% (v/v) βME was added to MOPS/NaCl to reduce the 17

disulfide bridges of BSA. FAF-BSA and BSAx-123t-PEG8 denatured and aggregated under these

conditions but BSAx-te-PEG8 and BSAx-PEG4-123t-PEG8 still did not peak (Figure 30b).

MOPS/NaCl/0.4% (v/v) βME was further modified with 1.5 M GdmCl. With a high

concentration of a strong chaotrope and reducing agent, all species of BSA peaked (Figure 30c).

The decrease in signal after maximum absorbance was due to aggregates precipitating and

settling. Samples incubated at 30oC for the full run-time showed no change in absorbance at 300

nm, indicating that denaturation was not due solely to the presence of high concentrations of

GdmCl and βME.

FAF-BSA had the lowest Tagg of 72.5oC (Table 8). BSAx-123t-PEG8 (91.8oC) and BSAx-

PEG4-123t-PEG8 (92.9oC) had significantly higher Tagg values with similar peak shapes. BSAx-

te-PEG8 reached the 50% mark at a higher temperature (77.5oC) than FAF-BSA but the signal

began increasing at a lower temperature.

94

Table 8. Temperatures of Denaturation and Aggregation for PEGylated BSA Species.

Sample Tagg (oC) Tm (oC) ΔTm (oC) FAF-BSA 72.5 ± 5.3 81.2 ± 4.1 0 BSAx-te-PEG8 77.5 ± 5.3 77.7 ± 0.9 -3.5 BSAx-123t-PEG8 91.8 ± 2.0 69.4 ± 3.6 -11.8 BSAx-PEG4-123t-PEG8 92.9 ± 2.6 76.0 ± 3.2 -5.2

The broadening of the BSAx-te-PEG8 peak was attributed to the dissolution of the

thioether linker in the presence of high concentrations of excess thiol in the form of βME.

Studies have explored the dissociation of antibody-drug conjugates due to this phenomenon

(Alley et al., 2008; Shen et al., 2012). Following intravenous administration to mice, the

thioether linkage reverted back to thiol-containing antibody and maleimido-functionalized drug,

which covalently bound to Cys34 of native serum albumin. To stabilize the thioether,

bromination of the maleimide moiety and hydrolysis of the succinimide to succinic acid have

been explored and provide intriguing options (Fontaine, Reid, Robinson, Ashley & Santi, 2015;

Smith et al., 2015).

The shape of BSAx-te-PEG8 in Figure 30c is due to the gradual dissolution of the

thioether and subsequent denaturation and aggregation of the released BSA. The triazole of

BSAx-R-123t-PEG8 species provides a thermally stable linker and results in a sharper peak.

95

(a) (b)

(c)

Figure 30. Change in absorbance as a function of temperature for the determination of Tagg. (a) Aggregation was minimal in MOPS/NaCl. (b) Addition of βME to the buffer resulted in aggregation of some species. (c) All samples aggregated at measurable temperatures in 50 mM MOPS/0.15 M NaCl/0.4% βME/1.5 M GdmCl pH 7.2. FAF-BSA (black); BSAx-te-PEG8 (red); BSAx-123t-PEG8 (blue); BSAx-PEG4-123t-PEG8 (orange). Dashed lines correspond to 50% aggregation.

0.0

0.5

1.0

1.5

2.0

2.5

40 50 60 70 80 90 100

Abs

orba

nce

(AU

)

Temperature (oC)

0.0

0.5

1.0

1.5

2.0

2.5

40 50 60 70 80 90 100

Abs

orba

nce

(AU

)

Temperature (oC)

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

40 50 60 70 80 90 100

Frac

tiona

l Abs

orba

nce

Temperature (oC)

96

Temperature of Denaturation (Tm)

CD spectra for all species of BSA showed characteristic minima at 210 nm and 220 nm at

30oC, indicating highly α-helical protein (Figure 31). The helicity for each species ranged

between 59-63% which is in agreement with published values of 56-66% (Das et al., 2017;

Moriyama et al., 2008). As temperature increased, these valleys became less defined and helicity

decreased.

Figure 31. Temperature-induced changes in circular dichroic spectra. Definition of valleys at 210 and 220 nm of FAF-BSA was lost as temperature increased. BSAx-te-PEG8, BSAx-123t-PEG8, and BSAx-PEG4-123t-PEG8 followed this same trend.

-85

-70

-55

-40

-25

-10

5

20

190 200 210 220 230 240 250 260

Ellip

ticity

(mill

ideg

rees

)

Wavelength (nm)

30oC

90oC

97

Tm was calculated as the temperature at which 50% of the original helical content was

lost. For FAF-BSA and all three types of PEGylated complexes, helicity slowly decreased up to

60oC where the values dropped rapidly to minima at 90oC (Figure 32).

(a) (b)

(c) (d)

Figure 32. Change in helicity as a function of temperature for the determination of Tm. (a) FAF-BSA; (b) BSAx-te-PEG8; (c) BSAx-123t-PEG8; (d) BSAx-PEG4-123t-PEG8. n/a: not applicable.

Unmodified FAF-BSA had the highest Tm at 81.2oC (Table 8). This is in agreement with

the published value of 82.5oC (Plesner et al., 2011).

The effect seen with PEGylated species on Tm is not due to the type or size of PEG

backbone used as evidenced by BSAx-te-PEG8. For modifications at Cys34, neither a straight-

chain nor branched PEG caused a significant change in denaturation temperature. Attachment of

0

20

40

60

80

100

30 40 50 60 70 80 90

% R

emai

ning

Hel

icity

Temperature (oC)

0

20

40

60

80

100

30 40 50 60 70 80 90

% R

emai

ning

Hel

icity

Temperature (oC)

0

20

40

60

80

100

30 40 50 60 70 80 90

% R

emai

ning

Hel

icity

Temperature (oC)

0

20

40

60

80

100

30 40 50 60 70 80 90

% R

emai

ning

Hel

icity

Temperature (oC)

98

a branched 40 kDa PEG8 backbone resulted in a decrease in Tm by 3.5oC, which is similar to the

published 2oC decrease with a straight chain 40 kDa Mal-PEG. As reported by Plesner et al.

(2011), the decrease in Tm of PEGylated BSA was approximately 2oC for all species of linear 5-

60 kDa PEG, regardless of the molecular weight of the PEG chain.

The drastic change in lysine-modified PEGylated BSA may be related to the proximity of

DIBO to the surface of the protein. Modification of a lysine with the extended DIBO-PEG4-NHS

activator resulted in a decrease in Tm by 5.2oC while the largest difference of 11.8oC was seen

with the shorter DIBO-NHS activator. DIBO, being a bulky, hydrophobic group, is capable of

altering the surface chemistry of BSA. This is likely the case for BSAx-123t-PEG8, where the

activating agent lacks a spacer. In BSAx-PEG4-123t-PEG8, the added space between DIBO and

the surface of BSA lessened the effect of hydrophobic interactions. Addition of a nonpolar patch

onto the exterior of BSA greatly affected solvent interaction and may have promoted

denaturation.

Effects of Pasteurization on Secondary Structure and Ligand Binding

The effects of pasteurization without stabilizing reagents was monitored by CD and NPX

binding studies. In the absence of NPX, up to 18% of total helicity was lost (Figure 33). BSAx-

PEG4-123t-PEG8 lost only 7%, further strengthening the argument that increasing the distance

between protein and PEG8 may further stabilize the complex.

Helicity was maintained after addition of NPX at 4oC. Upon pasteurization, all samples

consistently lost 7-10% of the original helicity. NPX has been found to bind with an association

constant (Ka) on the order of 107 M-1. Addition of a tightly bound ligand proved to stabilize all of

these species of BSA. This is most significant for BSAx-te-PEG8 as helicity was restored to

levels comparable to BSAx-R-123t-PEG8 species.

99

60oC 4oC 60oC w/o NPX w/ NPX w/ NPX

% Helicity (% Remaining)

FAF-BSA (black)

BSAx-te-PEG8 (red)

BSAx-123t-PEG8 (blue)

BSAx-PEG4-123t-PEG8 (orange)

4oC w/o NPX 57.1 +/- 4.0 56.3 ± 2.5 55.5 ± 2.1 56.5 ± 1.7

60oC w/o NPX 46.8 +/- 2.6 48.0 ± 3.6 49.7 ± 3.1 52.5 ± 2.1

(81.8) (85.2) (89.5) (92.9) 4oC w/ NPX

59.4 +/- 0.9 56.7 ± 2.1 59.5 ± 0.7 57.3 ± 0.6 (104.0) (100.6) (107.2) (101.5)

60oC w/ NPX 51.3 +/- 4.3 52.7 ± 1.4 49.7 ± 2.5 50.5 ± 0.7

(89.7) (93.5) (89.5) (89.4) Figure 33. Helicity after pasteurization in the absence and presence of naproxen. Samples held at 4oC in the absence of NPX were treated as controls and set to 100% helicity (---). All other values were calculated against these controls.

82

104

9085

101

9389

107

8993

101

89

0

10

20

30

40

50

60

70

80

90

100

% R

emai

ning

Hel

icity

100

A significant amount of helicity was lost upon pasteurization of BSA species without

ligand bound. Ligand binding capabilities of these samples was determined using NPX to probe

the high affinity site DS2.

Prior to pasteurization, all samples were able to tightly bind NPX with Ka values in the

range of 107 M-1 which correspond well with published values (Figure 34). After pasteurization,

these values decreased by several fold for all species of BSA except BSAx-123t-PEG8. This

suggests major structural changes that greatly affect DS2.

101

(a

) (b

)

Bef

ore

past

euriz

atio

n A

fter p

aste

uriz

atio

n

Sam

ple

Ka (

x107 M

-1)

n K

a (x1

07 M-1

) Fo

ld d

ecre

ase

n

FAF-

BSA

0.

40 ±

0.1

3 0.

4 ±

0.1

0.09

± 0

.05

4.4

1.5

± 0.

3 B

SAx-

te-P

EG8

1.99

± 0

.17

0.9

± 0.

3 0.

15 ±

0.0

0 13

1.

0 ±

0.2

BSA

x-12

3t-P

EG8

0.79

± 0

.61

2.4

± 1.

1 0.

57 ±

0.2

6 1.

4 0.

5 ±

0.0

BSA

x-PE

G4-

123t

-PEG

8 3.

55 ±

1.6

0.

9 ±

0.2

0.03

± 0

.00

118

0.2

± 0.

3 Li

t. va

lues

1 0.

12-3

.70

0.6-

3 n/

a n/

a n/

a Fi

gure

34.

Non

-line

ar re

gres

sion

ana

lyse

s of f

luor

imet

ric a

ssay

s of P

EGyl

ated

bov

ine

seru

m a

lbum

in b

efor

e an

d af

ter p

aste

uriz

atio

n.

FAF-

BSA

(bla

ck ¯

); B

SAx-

te-P

EG8 (

red £

); B

SAx-

123t

-PEG

8 (bl

ue r

); B

SAx-

PEG

4-12

3t-P

EG8 (

oran

ge �

). 1 N

PX v

alue

s wer

e re

porte

d by

Ban

erje

e et

al.

(200

6), B

ou-A

bdal

lah

et a

l. (2

016)

, Fie

ldin

g et

al.

(200

5), a

nd H

onor

é &

Bro

ders

en (1

984)

. Buj

acz

et a

l. (2

014)

ha

ve re

porte

d a

crys

tal s

truct

ure

(PD

B ID

: 4O

R0)

for B

SA c

ompl

exed

with

thre

e N

PX m

olec

ules

. n/a

: not

app

licab

le.

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Fractional Intensity

[NPX

] (µM

)

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Fractional Intensity

[NPX

] (µM

)

102

Conclusions

Attachment of a PEG chain to BSA clearly affected the thermal stability of the protein.

PEGylation, while lowering Tm, increased Tagg. Use of a short activating agent resulted in

alteration of the surface of BSA due to interaction with the hydrophobic DIBO group, resulting

in a marked decrease in Tm. Added space between the protein surface and the DIBO group

lessened the effect of the nonpolar DIBO.

Mal-PEG8 attached at the lone free thiol of BSA showed an increase in Tagg but peak

broadening may be due to the dissolution of the thioether linker. Both BSAx-123t-PEG8 and

BSAx-PEG4-123t-PEG8 effectively blocked aggregation, as seen by considerably higher Tagg.

PEGylation and pre-loading of complexes with a high affinity ligand are insufficient

measures for maintenance of helicity during pasteurization. Sterilization of PEG species in the

absence of ligand resulted in significant loss of helicity along with reduction of ligand binding

capabilities. Pre-loading PEGylated BSA species mitigated these losses but still resulted in some

loss of helicity. A combination of effects is needed to stabilize the complex. Binding NPX prior

to pasteurization helps to prevent denaturation while PEGylation if more effective at preventing

aggregation.

103

CHAPTER VI

GENERAL CONCLUSIONS AND FUTURE RESEARCH

General Conclusions

Three different species of PEGylated BSA have been developed using thiol-maleimide and

strain-promoted azide-alkyne cycloaddition “click” chemistries. All three products were mixtures

of predominantly BSA2-4-R-PEG8 with some amount of unmodified BSA and higher molecular

weight species. BSAx-PEG4-123t-PEG8 was the most homogeneous sample, containing the

fewest higher molecular weight species while BSAx-te-PEG8 was the least homogeneous,

containing a large number of higher molecular weight species. The wider distribution of BSAx-

te-PEG8 was unexpected due to the presence of a single PEGylation site at Cys34. Formation of

higher molecular weight species was attributed to the side-reaction of Mal-PEG8 with native

lysines.

All species of BSAx-R-PEG8 demonstrated similar ligand binding capabilities. DS2 was

unaffected as association constants for naproxen were within the range of reported values for

unmodified BSA. Interactions between DS1 and sulfamethoxazole, however, were greatly

weakened by PEGylation. While the number of binding sites for each remained in agreement

with the literature, association constants were several times smaller for PEGylated BSA than

unmodified FAF-BSA.

The difference in effects of PEGylation on these binding sites was likely due to the

placement of the PEG backbone. In the case of BSAx-te-PEG8, the main site of modification was

104

Cys34 in Domain IA. DS1 is located at the interface of Domains IB and IIA. Conformational

changes due to PEGylation in the same domain altered the cavity of DS1 and greatly reduced

binding capabilities. DS2, on the other hand, is located in Domain IIIA. Modifications made at

Cys34 appear to have little to no effect on ligand binding to DS2.

For BSAx-123t-PEG8 and BSAx-PEG4-123t-PEG8, the exact site of modification is

unknown as both activating agents were attached by an NHS-lysine reaction. BSA contains 59

lysines. 20 appear in Domains I and III while 19 are found in Domain II. Given the high number

and even distribution of potential sites of modification, a preference for a given domain is not

readily apparent. However, similar drug binding profiles between BSAx-te-PEG8 and BSAx-R-

123t-PEG8 of this study suggest a preference for modification of Domain I.

Temperature of unfolding was most affected by activation with the short DIBO-NHS.

Both BSAx-te-PEG8 and BSAx-PEG4-123t-PEG8 showed similar denaturation curves to

unmodified FAF-BSA. BSAx-123t-PEG8 was the least thermally stable of the complexes. This

was attributed to the proximity of the hydrophobic DIBO group to the surface of BSA altering

solvent interaction. Moving the DIBO further from the protein surface, as in BSAx-PEG4-123t-

PEG8, or complete elimination of DIBO (BSAx-te-PEG8), restored Tm to similar values for that

of unmodified FAF-BSA.

The reverse was found for the temperature of aggregation. BSAx-R-123t-PEG8 species

aggregated at much higher temperatures than either FAF-BSA or BSAx-te-PEG8. While Cys34

modified FAF-BSA had a higher Tagg, the signal began increasing at a much lower temperature

than FAF-BSA, likely due to the dissolution of the thioether linker. This instability was

potentially due to the high concentration of excess thiol from 2-Mercaptoethanol in the

denaturing buffer.

105

PEGylation and pre-loading of a high affinity DS2 ligand proved insufficient to allow for

omission of stabilizing agents used in the pasteurization process. A significant decrease in

helicity was seen for all PEGylated BSA species following pasteurization in the absence of

naproxen. When sterilized with naproxen bound, a moderate loss of helicity was seen for all

species of BSA. A combination of PEGylation and pre-loading the complex with a tightly bound

ligand reduced the denaturing effects of pasteurization but did not completely eliminate them.

Future Research

The future of this project may be aimed at answering four questions: 1) Can a

homogeneous BSAx-R-PEG8 product be generated with higher yields? 2) Can a fatty acid which

does not compete for DS2 be substituted for caprylate? 3) Can these PEGylated species

effectively release drugs? and 4) Does PEGylation affect cellular uptake?

The formation of three different forms of PEGylated BSA has been verified but the

production process may still be improved. Yields and homogeneity of the products are

potentially tied to the distance between BSA and the PEG8 backbone. An additional 14 Å

between BSA and PEG8 was gained by incorporation of a PEG4 linker into the activating agent.

This modification produced slightly higher yields and a more homogeneous final product.

Extension of the PEG arms is the easiest way to increase this distance. The largest available

eight-armed PEG is 40 kDa, each branch having an average molecular weight of 5 kDa which

corresponds to 113 subunits. By extending these branches, the congestion around the PEG

backbone will be relieved. Less steric hindrance would allow for BSA molecules to have easier

access to the terminal functional groups. The amount of PEGylated product generated would

increase while the number of different species formed would decrease.

106

Caprylate, one of the stabilizing agents used in pasteurization, is a saturated eight-carbon

fatty acid (8:0) with an affinity for DS2 (Kawai et al., 2017). Oleate (18:1), palmitate (16:0), and

linoleate (18:2) account for the majority of bound fatty acids to human serum albumin (A. Saifer

& Goldman, 1961). One of these compounds may prove to be an alternative stabilizing agent.

BSA contains seven fatty acid binding sites (Bhattacharya, Grüne & Curry, 2000;

Rizzuti, Bartucci, Sportelli & Guzzi, 2015; Simard et al., 2005). Fatty acid site 5 (FA5) resides in

Domain III and contains hydrophilic features (Y400 and K524) at the entrance to the cavity

which interact with the polar head group of bound fatty acids. This cavity has been shown to

bind long chain fatty acids with the highest affinity of the seven sites (Rizzuti et al., 2015;

Simard et al., 2005).

DS2 of BSA is also located in Domain III with polar residues at Y410, R409, and S488.

Use of a single long chain fatty acid which preferentially binds to FA5 may provide a substitute

for caprylate in the pasteurization process. This would reduce competition for DS2, however, the

proximity of FA5 to DS2 may alter binding profiles. By binding to a site away from DS2,

thermal stability may be enhanced while maintaining ligand binding capabilities.

Also of interest is the ability to release bound drugs. Reported results show that these

PEGylated BSA species are able to bind ligands, if only in DS2. If addition of a DS2 ligand is

required before sterilization, the storage stability of the drug-BSA complex must be evaluated to

avoid dissociation of the complex before administration. Diffusion-based experiments

monitoring the effects of pH, temperature, time, etc. will provide information as to the shelf-life

of these formulated products.

Gp60 has been identified as a cell surface receptor for albumin and plays a key role in the

efficacy of the non-PEGylated albumin-based therapy Abraxane® (Desai, 2008; Schnitzer, 1992).

107

Albumin-based nanoparticles prepared via conjugation of BSA with either polystyrene (Wang,

Tiruppathi, Cho, Minshall & Malik, 2011) or PEGylated liposomes (Thöle, Nobmann, Huwyler,

Bartmann & Fricker, 2002) were successfully endocytosed into cells in vitro. This suggests that

the degree of PEGylation in BSAx-te-PEG8, BSAx-123t-PEG8, and BSAx-PEG4-123t-PEG8 will

not necessarily inhibit gp60-mediated transcytosis. In order to evaluate the viability as drug

delivery systems, the ability of cells to endocytose these PEGylated complexes must be assessed.

These PEGylated species utilize more cost-effective components, require no specialized

equipment, and modify the albumin core at a bare minimum of sites. By maintaining albumin

close to its native form while conferring the benefits of PEGylation, potential alternatives to

current drug delivery systems and protein-based therapies have been developed.

108

REFERENCE LIST

Abbas, A., Xing, B. & Loh, T.-P. (2014). Allenamides as Orthogonal Handles for Selective Modification of Cysteine in Peptides and Proteins. Angewandte Chemie International Edition, 53, 7491-7494. doi:10.1002/anie.201403121

Abegg, D., Frei, R., Cerato, L., Prasad Hari, D., Wang, C., Waser, J. & Adibekian, A.

(2015). Proteome-Wide Profiling of Targets of Cysteine reactive Small Molecules by Using Ethynyl Benziodoxolone Reagents. Angewandte Chemie International Edition, 54, 10852-10857. doi:10.1002/anie.201505641

Agard, N. J., Prescher, J. A. & Bertozzi, C. R. (2004). A Strain-Promoted [3+2] Azide-

Alkyne Cycloaddition for Covalent Modification of Biomolecules in Living Systems. Journal of the American Chemical Society, 126, 15046-15047. doi:10.1021/ja044996f

Alley, S. C., Benjamin, D. R., Jeffrey, S. C., Okeley, N. M., Meyer, D. L., Sanderson, R.

J. & Senter, P. D. (2008). Contribution of Linker Stability to the Activities of Anticancer Immunoconjugates. Bioconjugate Chemistry, 19, 759-765. doi:10.1021/bc7004329

Anhorn, M. G., Wagner, S., Kreuter, J., Langer, K. & von Briesen, H. (2008). Specific

Targeting of HER2 Overexpressing Breast Cancer Cells with Doxorubicin-Loaded Trastuzumab-Modified Human Serum Albumin Nanoparticles. Bioconjugate Chemistry, 19, 2321-2331. doi:10.1021/bc8002452

Anraku, M., Chuang, V. T. G., Maruyama, T. & Otagiri, M. (2013). Redox properties of

serum albumin. Biochimica et Biophysica Acta, 1830, 5465-5472. doi:10.1016/j.bbagen.2013.04.036

Anraku, M., Tsurusaki, Y., Watanabe, H., Maruyama, T., Kragh-Hansen, U. & Otagiri,

M. (2004). Stabilizing mechanisms in commercial albumin preparations: octanoate and N-acetyl-L-tryptophanate protect human serum albumin against heat and oxidative stress. Biochimica et Biophysica Acta, 1702, 9-17. doi:10.1016/j.bbapap.2004.07.002

Ashbrook, J. D., Spector, A. A. & Fletcher, J. E. (1972). Medium Chain Fatty Acid

Binding to Human Plasma Albumin. The Journal of Biological Chemistry, 247(21), 7038-7042. Ashbrook, J. D., Spector, A. A., Santos, E. C. & Fletcher, J. E. (1975). Long Chain Fatty

Acid Binding to Human Plasma Albumin. The Journal of Biological Chemistry, 250(6), 2333-2338.

109

Baez, H. A. & Assaf-Anid, N. M. (2008). Novel and Conventional Approaches to Sterilization. Chemical Engineering, 115(8), 42-45

Baggio, L. L., Huang, Q., Cao, X. & Drucker, D. J. (2008). An Albumin-Exendin-4

Conjugate Engages Central and Peripheral Circuits Regulating Murine Energy and Glucose Homeostasis. Gastroenterology, 134, 1137-1147. doi:10.1053/j.gastro.2008.01.017

Bal, W., Christodoulou, J., Sadler, P. J. & Tucker, A. (1998). Multi-metal binding site of

serum albumin. Journal of Inorganic Biochemistry, 70, 33-39. Banerjee, T., Singh, S. K. & Kishore, N. (2006). Binding of Naproxen and Amitriptyline

to Bovine Serum Albumin: Biophysical Aspects. The Journal of Physical Chemistry B, 110, 24147-24156. doi:10.1021/jp062734p

Bartsch, W., Sponer, G., Dietmann, K. & Fuchs, G. (1976). Acute Toxicity of Various

Solvents in the Mouse and Rat. LD50 of ethanol, diethylacetamide, dimethylformamide, dimethylsulfoxide, glycerine, N-methylpyrrolidone, polyethylene glycol 400, 1,2-propanediol and Tween 20. Drug Research, 26(8), 1581-1583.

Baskin, J. M. & Bertozzi, C. R. (2009). Copper-free Click Chemistry. In J. Lahann (Ed.),

Click Chemistry for Biotechnology and Materials Science (pp. 29-51). Chichester, UK: John Wiley and Sons, Ltd.

Batist, G., Ramakrishnan, G., Rao, C. S., Chandrasekharan, A., Gutheil, J., Guthrie, T., . .

. Lee, L. W. (2001). Reduced Cardiotoxicity and Preserved Antitumor Efficacy of Liposome-Encapsulated Doxorubicin and Cyclophosphamide Compared With Conventional Doxorubicin and Cyclophosphamide in a Randomized, Multicenter Trial of Metastatic Breast Cancer. Journal of Clinical Oncology, 19(5), 1444-1454. doi:10.1200/JCO.2001.19.5.1444

Behrens, P. Q., Spiekerman, A. M. & Brown, J. R. (1975). Structure of human serum

albumin. Federation proceedings, 34, 591. Bhattacharya, A. A., Grüne, T. & Curry, S. (2000). Crystallographic Analysis Reveals

Common Modes of Binding of Medium and Long-chain Fatty Acids to Human Serum Albumin. Journal of Molecular Biology, 303, 721-732. doi:10.1006/jmbi.2000.4158

Boopathy, A. V. & Davis, M. E. (2014). Self-Assembling Peptide-Based Delivery of

Therapeutics for Myocardial Infarction. In K. K. Jain (Ed.), Drug Delivery System (Second ed., pp. 159-164). New York, NY: Humana Press.

Bosse, D., Praus, M., Kiessling, P., Nyman, L., Andresen, C., Waters, J. & Schindel, F.

(2005). Phase I Comparability of Recombinant Human Albumin and Human Serum Albumin. The Journal of Clinical Pharmacology, 45, 57-67. doi:10.1177/0091270004269646

110

Bou-Abdallah, F., Sprague, S. E., Smith, B. M. & Giffune, T. R. (2016). Binding thermodynamics of Diclofenac and Naproxen with human and bovine serum albumins: A calorimetric and spectroscopic study. The Journal of Chemical Thermodynamics, 103, 299-309. doi:10.1016/j.jct.2016.08.020

Boye, J. I., Alli, I. & Ismail, A. A. (1996). Interactions Involved in the Gelation of

Bovine Serum Albumin. Journal of Agricultural and Food Chemistry, 44, 996-1004. doi:10.1021/jf950529t

Brewer, C. F. & Riehm, J. P. (1967). Evidence for possible nonspecific reactions between

N-ethylmaleimide and proteins. Analytical biochemistry, 18(2), 248-255. Brown, J. R. (1975). Structure of bovine serum albumin. Federation Proceedings,

Federation of American Societies for Experimental Biology, 34(1), 591. Bujacz, A. (2012). Structures of bovine, equine and leporine serum albumin. Acta

Crystallographica Section D: Biological Crystallography, D68, 1278-1289. doi:10.1107/S0907444912027047

Bujacz, A., Zielinski, K. & Sekula, B. (2014). Structural studies of bovine, equine, and

leporine serum albumin complexes with naproxen. Proteins, 82, 2199-2208. doi:10.1002/prot.24583

Chalker, J. M., Bernardes, G. J. L., Lin, Y. A. & Davis, B. G. (2009). Chemical

Modification of Proteins at Cysteine: Opportunities in Chemistry and Biology. Chemistry-An Asian Journal, 4, 630-640. doi:10.1002/asia.200800427

Chen, R. F. (1967). Removal of Fatty Acids from Serum Albumin by Charcoal

Treatment. The Journal of Biological Chemistry, 242(2), 173-181. Christensen, E. I., Birn, H., Storm, T., Weyer, K. & Nielsen, R. (2012). Endocytic

Receptors in the Renal Proximal Tubule. Physiology, 27, 223-236. doi:10.1152/physiol.00022.2012

Christiansen, C. & Skotland, T. (2010). Changes of protein solutions during storage: a

study of albumin pharmaceutical preparations. Biotechnology and Applied Biochemistry, 55, 121-130. doi:10.1042/BA20090239

Cline, G. W. & Hanna, S. B. (1988). Kinetics and Mechanisms of the Aminolysis of N-

Hydroxysuccinimide Esters in Aqueous Buffers. Journal of Organic Chemistry, 53, 3583-3586.

111

Cohn, E. J., Strong, L. E., Hughes, W. L., Jr., Mulford, D. J., Ashworth, J. N., Melin, M. & Taylor, H. L. (1946). Preparation and Properties of Serum and Plasma Proteins. IV. A System for the Separation into Fractions of the Protein and Lipoprotein Components of Biological Tissues and Fluids. Journal of the American Chemical Society, 68(3), 459-475. doi:10.1021/ja01207a034

Creative PEGworks: PEG Products. (2017). Retrieved from

http://www.creativepegworks.com/product.php?cid=8 CryoLife: BioGlue Instructions for Use. (2017). Retrieved from

http://www.cryolife.com/wp-content/uploads/stories/assets/docs/BG_Surgical_Adhesive_Syringe_IFU_dom.pdf

Das, S., Islam, M. M., Jana, G. C., Patra, A., Jha, P. K. & Hossain, M. (2017). Molecular

binding of toxic phenothiazinium derivatives, azures to bovine serum albumin: A comparative spectroscopic, calorimetric, and in silico study. Journal of Molecular Recognition, e2609. doi:10.1002/jmr.2609

Desai, N. (2008). Nab Technology: A Drug Delivery Platform Utilising Endothelial gp60

Receptor-based Transport and Tumour-derived SPARC for Targeting. Drug Delivery Report(Winter 2007/2008), 37-41.

Duncan, R. & Veronese, F. M. (2009). PEGylated protein conjugates: A new class of

therapeutics for the 21st century. In F. M. Veronese (Ed.), Milestones in Drug Therapy: PEGylated Protein Drugs: Basic Science and Clinical Applications: Basic Science and Clinical Applications (pp. 1-9). Basel, CH: Birkhauser.

Elsadek, B. & Kratz, F. (2012). Impact of albumin on drug delivery - New applications

on the horizon. Journal of Controlled Release, 157, 4-28. doi:10.1016/j.jconrel.2011.09.069 Elzoghby, A. O., Samy, W. M. & Elgindy, N. A. (2012). Albumin-based nanoparticles as

potential controlled release drug delivery systems. Journal of Controlled Release, 157, 168-182. doi:10.1016/j.jconrel.2011.07.031

Farcet, M. R., Kindermann, J., Modrof, J. & Kreil, T. R. (2012). Inactivation of hepatitis

A variants during heat treatment (pasteurization) of human serum albumin. Transfusion, 52, 181-187. doi:10.1111/j.1537-2995.2011.03251.x

Fiehn, C., Kratz, F., Sass, G., Müller-Ladner, U. & Neumann, E. (2008). Targeted drug

delivery by in vivo coupling to endogenous albumin: an albumin-binding prodrug of methotrexate (MTX) is better than MTX in the treatment of murine collagen-induced arthritis. Annals of the Rheumatic Diseases, 67, 1188-1191. doi:10.1136/ard.2007.086843

112

Fielding, L., Rutherford, S. & Fletcher, D. (2005). Determination of protein-ligand binding affinity by NMR: observations from serum albumin model systems. Magnetic Resonance in Chemistry, 43, 463-470. doi:10.1002/mrc.1574

Fontaine, S. D., Reid, R., Robinson, L., Ashley, G. W. & Santi, D. V. (2015). Long-term

stabilization of maleimide-thiol conjugates. Bioconjugate Chemistry, 26, 145-152. doi:10.1021/bc5005262

Gefen, T., Vaya, J., Khatib, S., Harkevich, N., Artoul, F., Heller, E. D., . . . Aizenshtein,

E. (2013). The impact of PEGylation on protein immunogenicity. International Immunopharmacology, 15, 254-259. doi:10.1016/j.intimp.2012.12.012

Gellis, S. S., Neefe, J. R., Stokes, J., Jr., Strong, L. E., Janeway, C. A. & Scatchard, G.

(1948). Chemical, Clinical, and Immunological Studies on the Products of Human Plasma Fractionation. XXXVI. Inactivation of the Virus of Homologous Serum Hepatitis in Solutions of Normal Human Serum Albumin by Means of Heat. The Journal of Clinical Investigation, 27(2), 239-244.

Ghuman, J., Zunszain, P. A., Petitpas, I., Bhattacharya, A. A., Otagiri, M. & Curry, S.

(2005). Structural Basis of the Drug-binding Specificity of Human Serum Albumin. Journal of Molecular Biology, 353, 38-52. doi:10.1016/j.jmb.2005.07.075

Gradishar, W. J., Tjulandin, S., Davidson, N., Shaw, H., Desai, N., Bhar, P., . . .

O'Shaughnessy, J. (2005). Phase III Trial of Nanoparticle Albumin-Bound Paclitaxel Compared With Polyethylated Castor Oil-Based Paclitaxel in Women With Breast Cancer. Journal of Clinical Oncology, 23(31), 7794-7803. doi:10.1200/jco.2005.04.937

Guyton, A. C. & Hall, J. E. (2006). The Microcirculation and the Lymphatic System:

Capillary Fluid Exchange, Interstitial Fluid, and Lymph Flow. In Textbook of medical physiology (Vol. 11, pp. 181-194).

Haag, R. & Kratz, F. (2006). Polymer Therapeutics: Concepts and Applications.

Angewandte Chemie International Edition, 45, 1198-1215. doi:10.1002/anie.200502113 Hawkins, M. J., Soon-Shiong, P. & Desai, N. (2008). Protein nanoparticles as drug

carriers in clinical medicine. Advanced Drug Delivery Reviews, 60, 876-885. doi:10.1016/j.addr.2007.08.044

He, Y., Ning, T., Xie, T., Qiu, Q., Zhang, L., Sun, Y., . . . Yang, D. (2011). Large-scale

production of functional human serum albumin from transgenic rice seeds. Proceedings of the National Academy of Sciences, 108(47), 19078-19083. doi:10.1073/pnas.1109736108

Herold, D. A., Keil, K. & Bruns, D. E. (1989). Oxidation of Polyethylene Glycols by

Alcohol Dehydrogenase. Biochemical Pharmacology, 38(1), 73-76. doi:10.1016/0006-2952(89)90151-2

113

Hirayama, K., Akashi, S., Furuya, M. & Fukuhara, K.-i. (1990). Rapid Confirmation and Revision of the Primary Structure of Bovine Serum Albumin by ESIMS and Frit-FAB LC/MS. Biochemical and Biophysical Research Communications, 173(2), 639-646. doi:10.1016/S0006-291X(05)80083-X

Hong, V., Presolski, S. I., Ma, C. & Finn, M. G. (2009). Analysis and Optimization of

Copper-Catalyzed Azide-Alkyne Cycloaddition for Bioconjugation. Angewandte Chemie International Edition, 48, 9879-9883. doi:10.1002/anie.200905087

Honoré, B. & Brodersen, R. (1984). Albumin Binding of Anti-Inflammatory Drugs.

Utility of a Site-Oriented versus a Stoichiometric Analysis. Molecular Pharmacology, 25, 137-150.

Janes, R. W. (2009). Reference Datasets for Protein Circular Dichroism and Synchrotron

Radiation Circular Dichroism Spectroscopic Analyses. In B. A. Wallace & R. W. Janes (Eds.), Modern Techniques for Circular Dichroism and Synchrotron Radiation Circular Dichroism Spectroscopy (pp. 183-201). Amsterdam, Netherlands: IOS Press BV.

JenKem Technology: PEG Products. (2017). Retrieved from

http://www.jenkemusa.com/products Kawai, A., Chuang, V. T. G., Kouno, Y., Yamasaki, K., Miyamoto, S., Anraku, M. &

Otagiri, M. (in press). (2017). Crystallographic analysis of the ternary complex of octanoate and N-acetyl-L-methionine with human serum albumin reveals the mode of their stabilizing interactions. Biochimica et Biophysica Acta. doi:10.1016/j.bbapap.2017.04.004

Keating, G. M. (2012). Insulin Detemir. Drugs, 72(17), 2255-2287.

doi:10.2165/11470200-000000000-00000 Khoder, M., Abdelkader, H., ElShaer, A., Karam, A., Najlah, M. & Alany, R. G. (2016).

Efficient approach to enhance drug solubility by particle engineering of bovine serum albumin. International Journal of Pharmaceutics, 515, 740-748. doi:10.1016/j.ijpharm.2016.11.019

Kim, J., Bronson, C. L., Hayton, W. L., Radmacher, M. D., Roopenian, D. C., Robinson,

J. M. & Anderson, C. L. (2006). Albumin turnover: FcRn-mediated recycling saves as much albumin from degradation as the liver produces. American Journal of Physiology: Gastrointestinal and Liver Physiology, 290, G352-G360. doi:10.1152/ajpgi.00286.2005

Kolb, H. C., Finn, M. G. & Sharpless, K. B. (2001). Click Chemistry: Diverse Chemical

Function from a Few Good Reactions. Angewandte Chemie International Edition, 40, 2004-2021. doi:10.1002/1521-3773(20010601)40:11<2004::AID-ANIE2004>3.0.CO;2-5

Korenblat, P. E., Rothberg, R. M., Minden, P. & Farr, R. S. (1968). Immune responses of

human adults after oral and parenteral exposure to bovine serum albumin. Journal of allergy, 41(4), 226-235.

114

Kragh-Hansen, U. (1988). Evidence for a Large and Flexible Region of Human Serum Albumin Possessing High Affinity Binding Sites for Salicylate, Warfarin, and Other Ligands. Molecular Pharmacology, 34, 160-171.

Kratz, F. (2008). Albumin as a drug carrier: Design of prodrugs, drug conjugates and

nanoparticles. Journal of Controlled Release, 132, 171-183. doi:10.1016/j.jconrel.2008.05.010 Langer, K., Anhorn, M. G., Steinhauser, I., Dreis, S., Celebi, D., Schrickel, N., . . . Vogel,

V. (2008). Human serum albumin (HSA) nanoparticles: Reproducibility of preparation process and kinetics of enzymatic degradation. International Journal of Pharmaceutics, 347, 109-117. doi:10.1016/j.ijpharm.2007.06.028

Larsen, M. T., Kuhlmann, M., Hvam, M. L. & Howard, K. A. (2016). Albumin-based

drug delivery: harnessing nature to cure disease. Molecular and Cellular Therapies, 4(3). doi:10.1186/s40591-016-0048-8

Leamon, C. P. & Low, P. S. (1991). Delivery of macromolecules into living cells: A

method that exploits folate receptor endocytosis. Proceedings of the National Academy of Sciences of the United States of America, 88, 5572-5576.

LeBlanc, A., Shiao, T. C., Roy, R. & Sleno, L. (2014). Absolute Quantitation of NAPQI-

Modified Rat Serum Albumin by LC-MS/MS: Monitoring Acetaminophen Covalent Binding in Vivo. Chemical Research in Toxicology, 27, 1632-1639. doi:10.1021/tx500284g

Lee, S. H., Heng, D., Ng, W. K., Chan, H.-K. & Tan, R. B. H. (2011). Nano spray drying:

A novel method for preparing protein nanoparticles for protein therapy. International Journal of Pharmaceutics, 403, 192-200. doi:10.1016/j.ijpharm.2010.10.012

Lees, J. G., Miles, A. J., Wien, F. & Wallace, B. A. (2006). A reference database for

circular dichroism spectroscopy covering fold and secondary structure space. Bioinformatics, 22(16), 1955-1962. doi:10.1093/bioinformatics/btl327

Léger, R., Benquet, C., Huang, X., Quraishi, O., van Wyk, P. & Bridon, D. (2004).

Kringle 5 peptide–albumin conjugates with anti-migratory activity. Bioorganic & Medicinal Chemistry Letters, 14, 841-845. doi:10.1016/j.bmcl.2003.12.025

Lejon, S., Cramer, J. F. & Nordberg, P. (2008). Structural basis for the binding of

naproxen to human serum albumin in the presence of fatty acids and the GA module. Acta Crystallographica Section F, Structural Biology and Crystallization Communications, F64, 64-69. doi:10.1107/S174430910706770X

Li, L., Vorobyov, I. & Allen, T. W. (2013). The Different Interactions of Lysine and

Arginine Side Chains with Lipid Membranes. The Journal of Physical Chemistry B, 117, 11906-11920. doi:10.1021/jp405418y

115

Lin, T., Zhao, P., Jiang, Y., Tang, Y., Jin, H., Pan, Z., . . . Huang, Y. (2016). Blood–Brain-Barrier-Penetrating Albumin Nanoparticles for Biomimetic Drug Delivery via Albumin-Binding Protein Pathways for Antiglioma Therapy. American Chemical Society Nano, 10, 9999-10012. doi:10.1021/acsnano.6b04268

Lowe, A. B. (2010). Thiol-ene “click” reactions and recent applications in polymer and

materials synthesis. Polymer Chemistry, 1, 17-36. doi:10.1039/B9PY00216B Lundblad, R. L. (2012). Albumin. In Biotechnology of Plasma Proteins (pp. 83-182).

Boca Raton, FL: CRC Press. Manjula, B. N., Tsai, A., Upadhya, R., Perumalsamy, K., Smith, P. K., Malavalli, A., . . .

Acharya, A. S. (2003). Site-Specific PEGylation of Hemoglobin at Cys-93(β):  Correlation between the Colligative Properties of the PEGylated Protein and the Length of the Conjugated PEG Chain. Bioconjugate Chemistry, 14, 464-472. doi:10.1021/bc0200733

Massarotti, A., Aprile, S., Mercalli, V., Del Grosso, E., Grosa, G., Sorba, G. & Tron, G.

C. (2014). Are 1,4- and 1,5-Disubstituted 1,2,3-Triazoles Good Pharmacophoric Groups? ChemMedChem, 9, 2497-2508. doi:10.1002/cmdc.201402233

Mehtala, J. G., Kulczar, C., Lavan, M., Knipp, G. & Wei, A. (2015). Cys34-PEGylated

Human Serum Albumin for Drug Binding and Delivery. Bioconjugate Chemistry, 26, 941-949. doi:10.1021/acs.bioconjchem.5b00143

Meloun, B., Morávek, L. & Kostka, V. (1975). Complete Amino Acid Sequence of

Human Serum Albumin. Federation of European Biochemical Societies Letters, 58(1), 134-137. doi:10.1016/0014-5793(75)80242-0

Merlot, A. M., Kalinowski, D. S. & Richardson, D. R. (2014). Unraveling the mysteries

of serum albumin-more than just a serum protein. frontiers in Physiology, 5(299). doi:10.3389/fphys.2014.00299

Moriyama, Y., Watanabe, E., Kobayashi, K., Harano, H., Inui, E. & Takeda, K. (2008).

Secondary Structural Change of Bovine Serum Albumin in Thermal Denaturation up to 130 oC and Protective Effect of Sodium Dodecyl Sulfate on the Change. Journal of Physical Chemistry B, 112, 16585-16589. doi:10.1021/jp8067624

Naik, P. N., Chimatadar, S. A. & Nandibewoor, S. T. (2009). Study on the interaction

between antibacterial drug and bovine serum albumin: A spectroscopic approach. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 73, 841-845. doi:10.1016/j.saa.2009.04.018

116

Naik, P. N., Nandibewoor, S. T. & Chimatadar, S. A. (2015). Non-covalent binding analysis of sulfamethoxazole to human serum albumin: Fluorescence spectroscopy, UV–vis, FT-IR, voltammetric and molecular modeling. Journal of Pharmaceutical Analysis, 5(3), 143-152. doi:http://dx.doi.org/10.1016/j.jpha.2015.01.003

Narazaki, R., Hamada, M., Harada, K. & Otagiri, M. (1996). Covalent Binding Between

Bucillamine Derivatives and Human Serum Albumin. Pharmaceutical Research, 13(9), 1317-1321.

Ning, X., Guo, J., Wolfert, M. A. & Boons, G.-J. (2008). Visualizing Metabolically

Labeled Glycoconjugates of Living Cells by Copper-Free and Fast Huisgen Cycloadditions. Angewandte Chemie International Edition, 47, 2253-2255. doi:10.1002/anie.200705456

Nischan, N. & Hackenberger, C. P. R. (2014). Site-specific PEGylation of Proteins:

Recent Developments. The Journal of Organic Chemistry, 79, 10727-10733. doi:10.1021/jo502136n

Noort, D., Hulst, A. G. & Jansen, R. (2002). Covalent binding of nitrogen mustards to the

cysteine-34 residue in human serum albumin. Archives of Toxicology, 76, 83-88. doi:10.1007/s00204-001-0318-2

Olsen, H., Andersen, A., Nordbø, A., Kongsgaard, U. E. & Børmer, O. P. (2004).

Pharmaceutical-grade albumin: impaired drug-binding capacity in vitro. BMC Clinical Pharmacology, 4. doi:10.1186/1472-6904-4-4

Olsen, K. W. (1994). Thermal Denaturation Procedures for Hemoglobin. In J. Everse, K.

D. Vandegriff & R. M. Winslow (Eds.), Hemoglobins. Part B, Biochemical and Analytical Methods (First ed., Vol. 231, pp. 514-524). San Diego: Academic Press, Inc.

Persistence of Vision Pty. Ltd.: Persistence of Vision (TM) Raytracer. (2004). Version

3.6. Retrieved from http://www.povray.org/ Peters, T., Jr. (1996). All about Albumin: Biochemistry, Genetics, and Medical

Applications. San Diego, CA: Academic Press, Inc. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt, D. M., Meng, E.

C. & Ferrin, T. E. (2004). UCSF Chimera-A Visualization System for Exploratory Research and Analysis. Journal of Computational Chemistry, 25(13), 1605-1612. doi:10.1002/jcc.20084

Pietrangelo, A., Panduro, A., Chowdhury, J. R. & Shafritz, D. A. (1992). Albumin Gene

Expression is Down-regulated by Albumin or Macromolecule Infusion in the Rat. The Journal of Clinical Investigation, 89, 1755-1760. doi:10.1172/JCI 115778

117

Plesner, B., Fee, C. J., Westh, P. & Nielsen, A. D. (2011). Effects of PEG size on structure, function and stability of PEGylated BSA. European Journal of Pharmaceutics and Biopharmaceutics, 79, 399-405. doi:10.1016/j.ejpb.2011.05.003

Qi, J., Yao, P., He, F., Yu, C. & Huang, C. (2010). Nanoparticles with dextran/chitosan

shell and BSA/chitosan core-Doxorubicin loading and delivery. International Journal of Pharmaceutics, 393, 176-184. doi:10.1016/j.ijpharm.2010.03.063

Rajendiran, N. & Thulasidhasan, J. (2015). Interaction of sulfanilamide and

sulfamethoxazole with bovine serum albumin and adenine: Spectroscopic and molecular docking investigations. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 144, 183-191. doi:10.1016/j.saa.2015.01.127

Rizzuti, B., Bartucci, R., Sportelli, L. & Guzzi, R. (2015). Fatty acid binding into the

highest affinity site of human serum albumin observed in molecular dynamics simulation. Archives of Biochemistry and Biophysics, 579, 18-25. doi:10.1016/j.abb.2015.05.018

Roche, M., Rondeau, P., Singh, N. R., Tarnus, E. & Bourdon, E. (2008). The antioxidant

properties of serum albumin. Federation of European Biochemical Societies Letters, 582, 1783-1787. doi:10.1016/j.febslet.2008.04.057

Rostovtsev, V. V., Green, L. G., Fokin, V. V. & Sharpless, K. B. (2002). A Stepwise

Huisgen Cycloaddition Process: Copper(I)-Catalyzed Regioselective "Ligation" of Azides and Terminal Alkynes. Angewandte Chemie International Edition, 41(14), 2596-2599. doi:10.1002/1521-3773(20020715)41:14<2596::AID-ANIE2596>3.0.CO;2-4

Rózga, M., Sokołowska, M., Protas, A. M. & Bal, W. (2007). Human serum albumin

coordinates Cu(II) at its N-terminal binding site with 1 pM affinity. Journal of Biological Inorganic Chemistry, 12, 913-918. doi:10.1007/s00775-007-0244-8

Saifer, A. & Goldman, L. (1961). The free fatty acids bound to human serum albumin.

Journal of Lipid Research, 2(3), 268-270. Saifer, M. G. P., Somack, R. & Williams, L. D. (1994). Plasma Clearance and

Immunologic Properties of Long-Acting Superoxide Dismutase Prepared Using 35,000 to 120,000 Dalton Poly-ethylene Glycol. Advances in Experimental Medicine and Biology, 366, 377-387.

Schnitzer, J. E. (1992). gp60 is an albumin-binding glycoprotein expressed by continuous

endothelium involved in albumin transcytosis. American Journal of Physiology, 262, H246-H254.

118

Schnitzer, J. E., Sung, A., Horvat, R. & Bravo, J. (1992). Preferential Interaction of Albumin-binding Proteins, gp30 and gp18, with Conformationally Modified Albumins. Presence in Many Cells and Tissues with a Possible Role in Catabolism. The Journal of Biological Chemistry, 267(34), 24544-24553.

Schuck, P. (2000). Size-Distribution Analysis of Macromolecules by Sedimentation

Velocity Ultracentrifugation and Lamm Equation Modeling. Biophysical Journal, 78, 1606-1619. doi:10.1016/S0006-3495(00)76713-0

Sedfit. (2017). Retrieved from http://www.analyticalultracentrifugation.com/default.htm Shen, B.-Q., Xu, K., Liu, L., Raab, H., Bhakta, S., Kenrick, M., . . . Junutula, J. R.

(2012). Conjugation site modulates the in vivo stability and therapeutic activity of antibody-drug conjugates. Nature Biotechnology, 30(2), 184-189. doi:10.1038/nbt.2108

Shulman, S. (1953). The Determination of Sedimentation Constants with the Oil-Turbine

and Spinco Ultracentrifuges. Archives of biochemistry and biophysics, 44(1), 230-240. doi:10.1016/0003-9861(53)90028-9

Simard, J. R., Zunszain, P. A., Ha, C. E., Yang, J. S., Bhagavan, N. V., Petitpas, I., . . .

Hamilton, J. A. (2005). Locating high-affinity fatty acid-binding sites on albumin by x-ray crystallography and NMR spectroscopy. Proceedings of the National Academy of Sciences of the United States of America, 102(50), 17958-17963. doi:10.1073/pnas.0506440102

Smith, M. E. B., Caspersen, M. B., Robinson, E., Morais, M., Maruani, A., Nunes, J. P.

M., . . . Chudasama, V. (2015). A platform for efficient, thiol-stable conjugation to albumin's native single accessible cysteine. Organic & Biomolecular Chemistry, 13, 7946-7949. doi:10.1039/C5OB01205H

Spector, A. A. (1975). Fatty acid binding to plasma albumin. Journal of Lipid Research,

16, 165-179. Sreerama, N. & Woody, R. W. (2000). Estimation of Protein Secondary Structure from

Circular Dichroism Spectra: Comparison of CONTIN, SELCON, and CDSSTR Methods with an Expanded Reference Set. Analytical Biochemistry, 287, 252-260. doi:10.1006/abio.2000.4880

Stange, J., Stiffel, M., Goetze, A., Strube, S., Gruenert, J., Klammt, S., . . . Reisinger, E.

(2011). Industrial Stabilizers Caprylate and N-Acetyltryptophanate Reduce the Efficacy of Albumin in Liver Patients. Liver Transplantation, 17, 705-709. doi:10.1002/lt.22237

Stehle, G., Sinn, H., Wunder, A., Schrenk, H. H., Schütt, S., Maier-Borst, W. & Heene,

D. L. (1997). The loading rate determines tumor targeting properties of methotrexate-albumin conjugates in rats. Anti-Cancer Drugs, 8, 677-685.

119

Stehle, G., Sinn, H., Wunder, A., Schrenk, H. H., Stewart, J. C. M., Hartung, G., . . . Heene, D. L. (1997). Plasma protein (albumin) catabolism by the tumor itself-implications for tumor metabolism and the genesis of cachexia. Critical Reviews in Oncology/Hematology, 26(2), 77-100.

Stehle, G., Wunder, A., Schrenk, H. H., Hartung, G., Heene, D. L. & Sinn, H. (1999).

Albumin-based drug carriers: comparison between serum albumins of different species on pharmacokinetics and tumor uptake of the conjugate. Anti-Cancer Drugs, 10, 785-790.

Stoddart, C. A., Nault, G., Galkina, S. A., Thibaudeau, K., Bakis, P., Bousquet-Gagnon,

N., . . . Quraishi, O. (2008). Albumin-conjugated C34 Peptide HIV-1 Fusion Inhibitor: Equipotent to C34 and T-20 in vitro with sustained activity in SCID-hu Thy/Liv mice. The Journal of Biological Chemistry, 283(49), 34045-34052. doi:10.1074/jbc.M805536200

Sudlow, G., Birkett, D. J. & Wade, D. N. (1975). The Characterization of Two Specific

Drug Binding Sites on Human Serum Albumin. Molecular Pharmacology, 11, 824-832. Sudlow, G., Birkett, D. J. & Wade, D. N. (1976). Further Characterization of Specific

Drug Binding Sites on Human Serum Albumin. Molecular Pharmacology, 12, 1052-1061. Taylor, J. F. (1952). The Determination of Sedimentation Constant with the Spinco

Ultracentrifuge. Archives of biochemistry and biophysics, 36(2), 357-364. doi:10.1016/0003-9861(52)90421-9

The Gnome Project: The Gnumeric Spreadsheet: Free, Fast, Accurate - pick any three.

(2011). 1.10.16. Retrieved from http://www.gnumeric.org/ Thermo Scientific: Extinction Coefficients. (2013). Retrieved from

https://tools.thermofisher.com/content/sfs/brochures/TR0006-Extinction-coefficients.pdf Thermo Scientific: NHS-Azide and NHS-Phosphine Reagents. (2014). Retrieved from

https://tools.thermofisher.com/content/sfs/manuals/MAN0011695_NHSAzide_NHSPhosphine_Reag_UG.pdf

Thermo Scientific: NHS-Fluorescein. (2016). Retrieved from

https://tools.thermofisher.com/content/sfs/manuals/MAN0011647_NHSFluorescein_UG.pdf Thirumurugan, P., Matosiuk, D. & Jozwiak, K. (2013). Click Chemistry for Drug

Development and Diverse Chemical-Biology Applications. Chemical Reviews, 113, 4905-4979. doi:10.1021/cr200409f

Thöle, M., Nobmann, S., Huwyler, J., Bartmann, A. & Fricker, G. (2002). Uptake of

Cationized Albumin Coupled Liposomes by Cultured Porcine Brain Microvessel Endothelial Cells and Intact Brain Capillaries. Journal of Drug Targeting, 10(4), 337-344. doi:10.1080/10611860290031840

120

Tiruppathi, C., Song, W., Bergenfeldt, M., Sass, P. & Malik, A. B. (1997). Gp60 Activation Mediates Albumin Transcytosis in Endothelial Cells by Tyrosine Kinase-dependent Pathway. The Journal of Biological Chemistry, 272(41), 25968-25975. doi:10.1074/jbc.272.41.25968

Torne, S., Darandale, S., Vavia, P., Trotta, F. & Cavalli, R. (2013). Cyclodextrin-based

nanosponges: effective nanocarrier for Tamoxifen delivery. Pharmaceutical Development and Technology, 18(3), 619-625. doi:10.3109/10837450.2011.649855

Tornøe, C. W., Christensen, C. & Meldal, M. (2002). Peptidotriazoles on Solid Phase:

[1,2,3]-Triazoles by Regiospecific Copper(I)-Catalyzed 1,3-Dipolar Cycloadditions of Terminal Alkynes to Azides. Journal of Organic Chemistry, 67, 3057-3064. doi:10.1021/jo011148j

US Food and Drug Administration: Drug Approval Packages. (2017). Retrieved from

http://www.fda.gov/Drugs/InformationOnDrugs/ApprovedDrugs/default.htm van Berkel, S. S., Dirks, A. J., Debets, M. F., van Delft, F. L., Cornelissen, J. J. L. M.,

Nolte, R. J. M. & Rutjes, F. P. J. T. (2007). Metal-Free Triazole Formation as a Tool for Bioconjugation. ChemBioChem, 8, 1504-1508. doi:10.1002/cbic.200700278

Veronese, F. M. & Pasut, G. (2005). PEGylation, successful approach to drug delivery.

Drug Discovery Today, 10(21), 1451-1458. doi:10.1016/S1359-6446(05)03575-0 Veronese, F. M., Mero, A. & Pasut, G. (2009). Protein PEGylation, basic science and

biological applications. In F. M. Veronese (Ed.), Milestones in Drug Therapy: PEGylated Protein Drugs: Basic Science and Clinical Applications: Basic Science and Clinical Applications (pp. 11-31). Basel, CH: Birkhauser.

Vila, A., Sánchez, A., Tobío, M., Calvo, P. & Alonso, M. J. (2002). Design of

biodegradable particles for protein deilvery. Journal of Controlled Release, 78, 15-24. doi:10.1016/S0168-3659(01)00486-2

Wang, Q., Zhang, S.-R. & Ji, X. (2014). Investigation of interaction of antibacterial drug

sulfamethoxazole with human serum albumin by molecular modeling and multi-spectroscopic method. Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy, 124, 84-90. doi:10.1016/j.saa.2013.12.100

Wang, Z., Tiruppathi, C., Cho, J., Minshall, R. D. & Malik, A. B. (2011). Delivery of

Nanoparticle-complexed Drugs across the Vascular Endothelial Barrier via Caveolae. International Union of Biochemistry and Molecular Biology Life, 63(8), 659-667. doi:10.1002/iub.485

Wardell, M., Wang, Z., Ho, J. X., Robert, J., Ruker, F., Ruble, J. & Carter, D. C. (2002).

The Atomic Structure of Human Methemalbumin at 1.9 Å. Biochemical and Biophysical Research Communications, 291(4), 813-819. doi:10.1006/bbrc.2002.6540

121

Weber, C., Coester, C., Kreuter, J. & Langer, K. (2000). Desolvation process and surface characterisation of protein nanoparticles. International Journal of Pharmaceuticals, 194, 91-102.

Webster, K. D., Dahhan, D., Otto, A. M., Frosti, C. L., Dean, W. L., Chaires, J. B. &

Olsen, K. W. (2017). “Inside-Out” PEGylation of Bovine β-Cross-Linked Hemoglobin. Artificial Organs, 45, 351-358. doi:10.1111/aor.12928

Webster, R., Elliott, V., Park, B. K., Walker, D., Hankin, M. & Taupin, P. (2009). PEG

and PEG conjugates toxicity: towards an understanding of the toxicity of PEG and its relevance to PEGylated biologicals. In F. M. Veronese (Ed.), Milestones in Drug Therapy: PEGylated Protein Drugs: Basic Science and Clinical Applications: Basic Science and Clinical Applications (pp. 127-146). Basel, CH: Birkhauser.

Wetzel, R., Becker, M., Behlke, J., Billwitz, H., Böhm, S., Ebert, B., . . . Lassmann, G.

(1980). Temperature Behaviour of Human Serum Albumin. European Journal of Biochemistry, 104, 469-478. doi:10.1111/j.1432-1033.1980.tb04449.x

Whitmore, L. & Wallace, B. A. (2008). Protein Secondary Structure Analyses from

Circular Dichroism Spectroscopy: Methods and Reference Databases. Biopolymers, 89(5), 392-400. doi:10.1002/bip.20853

Working, P. K., Newman, M. S., Johnson, J. & Cornacoff, J. B. (1997). Safety of

Poly(ethylene glycol) and Poly(ethylene glycol) Derivatives. In J. M. Harris & S. Zalipsky (Eds.), ACS Symposium Series, 680: Poly(ethylene glycol): Chemistry and Biological Applications (pp. 45-57). Washington, DC: American Chemical Society.

Yaman, A. (2001). Alternative methods of terminal sterilization for biologically active

macromolecules. Current Opinion in Drug Discovery & Development, 4(6), 760-763. Yamaoka, T., Tabata, Y. & Ikada, Y. (1994). Distribution and Tissue Uptake of

Poly(ethylene glycol) with Different Molecular Weights after Intravenous Administration to Mice. Journal of Pharmaceutical Sciences, 83(4), 601-606. doi:10.1002/jps.2600830432

Yamasaki, K., Chuang, V. T. G., Maruyama, T. & Otagiri, M. (2013). Albumin-drug

interaction and its clinical implication. Biochimica et Biophysica Acta, 1830, 5435-5443. doi:10.1016/j.bbagen.2013.05.005

Zalipsky, S. & Harris, J. M. (1997). Introduction to Chemistry and Biological

Applications of Poly(ethylene glycol). In J. M. Harris & S. Zalipsky (Eds.), ACS Symposium Series, 680: Poly(ethylene glycol): Chemistry and Biological Applications (pp. 1-13). Washington, DC: American Chemical Society.

122

Zhao, H., Ghirlando, R., Alfonso, C., Arisaka, F., Attali, I., Bain, D. L., . . . Schuck, P. (2015). A Multilaboratory Comparison of Calibration Accuracy and the Performance of External References in Analytical Ultracentrifugation. Public Library of Science one, 10(5), e0126420. doi:10.1371/journal.pone.0126420

Zunszain, P. A., Ghuman, J., McDonagh, A. F. & Curry, S. (2008). Crystallographic

Analysis of Human Serum Albumin Complexed With 4Z,15E-bilirubin-IXα. Journal of Molecular Biology, 381, 394-406. doi:10.1016/j.jmb.2008.06.016

123

VITA

Jonathan Alejandro Hill was born and raised in Lumberton, North Carolina. He earned a

Bachelor of Science in Chemistry in 2005 from the University of North Carolina at Greensboro.

After graduation, he accepted a position as an Analytical Chemist at Syngenta Crop Protection,

LLC in Greensboro, NC. In 2011, his academic career was resumed by pursuing his doctorate at

Loyola University Chicago. This work is the culmination of those efforts.