Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

RGD-containing ligands for targeting liposomal nanoparticles Cressman, Sonya 2007

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2007-317563.pdf [ 10.72MB ]
JSON: 831-1.0100426.json
JSON-LD: 831-1.0100426-ld.json
RDF/XML (Pretty): 831-1.0100426-rdf.xml
RDF/JSON: 831-1.0100426-rdf.json
Turtle: 831-1.0100426-turtle.txt
N-Triples: 831-1.0100426-rdf-ntriples.txt
Original Record: 831-1.0100426-source.json
Full Text

Full Text

RGD-CONTAINING LIGANDS FOR TARGETFNG LIPOSOMAL NANOPARTICLES  by  Sonya Cressman  B.Sc., The University o f Victoria, 2000  A THESIS S U B M I T T E D I N P A R T I A L F U L F I L L M E N T OF T H E R E Q U I R E M E N T S FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES (Biochemistry and Molecular Biology)  THE UNIVERSITY OF BRITISH C O L U M B I A Vancouver, Canada September 2007 © Sonya Cressman, 2007  11  ABSTRACT The use o f a targeting ligand to enhance the delivery o f liposomal nanoparticles (LNs) to specific cells in diseased tissue is an attractive strategy that has progressed very slowly since the idea originated over 25 years ago. The slow progress can be attributed, in part, to the difficulties involved in producing well-defined targeted L N systems. This work concerns the development o f peptide-based ligands for targeting L N s to cells that overexpress the a (33 integrin, which is considered a unique marker o f the tumor-associated v  endothelium. Initial work focused on using Fab' fragments o f monoclonal antibodies as targeting ligands and employing literature techniques for coupling these proteins to lipids that can then be inserted in preformed L N s . However, it was found that the coupling procedures resulted in low yields of poorly defined lipid-ligand conjugates that were difficult to quantitate when incorporated into L N . Attention was then given to the development o f a peptide-targeting agent that could be incorporated into an L N at the time that the L N is made. The RGD-containing cyclic peptide, c R G D f K , was employed as the targeting ligand. In order to characterize its binding to the 0 ^ 3 integrin, a fluorescently labeled analogue, c R G D f K - 4 8 8 ,  was -  synthesized and capillary electrophoresis was employed for analysis. This procedure proved advantageous for studying receptor ligand interactions, since it allowed for the binding to be characterized in solution without the need for covalent modification o f receptor or ligand. A 2:1 R G D ligand to integrin specific binding stoichiometry was revealed with the second binding event having a similar affinity as the first binding event.  Ill  The next phase o f these studies investigated the ability o f c R G D f K - 4 8 8 to bind to integrins on human umbilical vascular endothelial cells ( H U V E C ) and  subsequently  undergo endocytosis. This was compared to the binding and uptake properties o f a fluorescently labeled monoclonal antibody, L M 6 0 9 X , which specifically binds  a pY v  integrin. Using flow cytometry and fluorescence microscopy, it is shown that the R G D ligand  exhibited considerably  greater  uptake  following  incubation at  endocytosis  permitting temperatures (37°C) as compared to endocytosis inhibiting temperatures (4°C). A 7.4-fold increase in uptake o f the R G D - l i p i d was observed following a one-hour incubation with H U V E C at 37°C, as compared to 4°C. In contrast, only a 1.9 fold increase in cell-associated fluorescence was observed on incubation with L M 6 0 9 X at 37°C as compared to 4 ° C . It is suggested that this ability of R G D ligands to stimulate endocytosis may be o f utility for achieving enhanced intracellular delivery o f ligand-associated drugs i n anti-angiogenic applications. The cyclic peptide was then used to construct a fluorescently labeled RGD-spacerlipid construct o f defined molecular weight that could be incorporated into L N at the time of manufacture. It is shown that the resulting R G D - L N s bind to H U V E C with increasing avidity as the amount of RGD-spacer-lipid incorporated is increased. Further, these R G D L N s are straightforward to make and can load and retain anti-cancer drugs such as doxorubicin.  It is shown that R G D - L N s loaded with doxorubicin and incubated with  H U V E C selectively deliver the drug to the cytosol while non-targeted L N s are not internalized by the cell, suggesting potential utility as targeted drug delivery systems in vivo.  TABLE OF CONTENTS ABSTRACT.. T A B L E OF C O N T E N T S LIST OF T A B L E S LIST OF FIGURES ABBREVIATIONS ACKNOWLEDGEMENTS CO-AUTHORSHIP STATEMENT 1 Introduction to Targeted Liposomal Nanoparticles 1.1 Liposomal Nanoparticles 1.2 Liposomal Nanoparticles and Drug Delivery 1.3 L N s with Extended in vivo Circulation Lifetimes 1.4 Immunoliposomes (ILNs) 1.4.1 Targeting L N s to the Disease Site 1.5 Contemporary I L N s and I L N Production  1.5.1  ii iv vii viii ix xi xiii 1 1 2 4 5 5 8  ILNs in vivo  9  1.5.2 I L N s i n the Clinic 1.6 Small Molecule Targeting Ligands 1.6.1 Peptides with Disease Site Affinity 1.6.2 Arginine-Glycine-Aspartic A c i d ( R G D ) 1.7 RGD-Targeting Ligands and Affinity Analyses 1.8 Objectives, Evolution and Structure o f this Thesis 2 Preliminary Experiments 2.1 Introduction 2.2 Methods 2.2.1 Materials 2.2.2 Fab' Formation 2.2.3 M a l e i m i d e - P E G o o - D S P E Analysis 2.2.4 Fab'-PEG ooo-DSPE Conjugation 2.2.5 Fab'-PEG ooo-DSPE Purification 2.3 Results 2.3.1 Quality o f M a l - P E G - D S P E 2.3.2 F a b ' - M a l e i m i d e - P E G - D S P E Conjugation 2.3.3 Pure F a b ' - P E G - D S P E Preparations are Inherently Difficult to Achieve 2.4 Discussion 3 R G D - l i g a n d Synthesis and Binding to the Isolated a p 3 Integrin 3.1 Introduction : 3.1.1 Targeting Angiogenesis 3.1.2 The C E - F A Method for Measuring Binding to the a P 3 Integrins 20  2  2  v  v  3.2 Methods 3.2.1 Materials 3.2.2 Peptide Synthesis, Cyclization and Labeling 3.2.3 Capillary Electrophoresis ( C E ) Analysis 3.2.4 Capillary Electrophoresis Frontal Analysis Method 3.3 Results and Discussion  10 11 12 13 14 16 20 20 22 22 22 23 23 24 25 25 26 27 29 32 32 32 35 40 40 40 43 44 45  V  3.3.1 Peptide Synthesis, Labeling and Analysis 45 3.3.2 c R G D f K - 4 8 8 Peptide Purity 46 3.3.3 Binding to Isolated a fc Integrin as Measured by the C E - F A Method 47 4 Cellular Binding and Endocytosis o f RGD-Containing Targeting Ligands 51 4.1 Introduction 51 4.2 Methods 54 4.2.1 Materials 54 4.2.2 C e l l Culture 54 4.2.3 C e l l Binding 55 4.2.4 Cellular Uptake 55 4.2.5 Receptor Quantitation 56 4.2.6 Fluorescence Microscopy 57 4.3 Results.. 59 4.3.1 H U V E C Express 2.63 x 10 Integrins Per C e l l 59 4.3.2 c R G D f K - 4 8 8 Undergoes Extensive Endocytosis Following Binding to HUVECat37°C 61 4.3.3 Time Dependent Uptake o f c R G D f K - 4 8 8 by H U V E C . . . 63 4.3.4 a P 3 Expression Decreases U p o n R G D Binding 65 4.3.5 Endocytosis as Visualized by Fluorescence Microscopy...; 66 4.4 Discussion 69 5 Targeted Drug Delivery 72 5.1 Introduction ; 72 5.2 Methods 76 5.2.1 Materials and Reagents 78 5.2.2 Peptide Synthesis 78 5.2.3 c R G D f K - S U C C - D S P E (Compound I) Synthesis 79 5.2.4 c R G D f K - ( S U C C - 3 4 3 ) - D S P E (Compound II) Synthesis 80 5.2.5 c R ( P b f ) G D ( t - B u ) f K - P E G - C O O H and c R ( P b f ) G D ( t - B u ) f K - P E G - K ( M C A ) P E G - C O O H Synthesis (preceding compounds III and I V , respectively) 81 5.2.6 Synthesis o f c R ( P b f ) G D ( t - B u ) f K - P E G - D S P E (Compound III) and c R G D f K P E G - K ( M C A ) - P E G - D S P E (Compound I V ) 83 5.2.7 c R G D f K - P E G - K ( M C A ) - P E G - D S P E Formulation into L N s 84 5.2.8 Post-Insertion o f c R G D f K - P E G - K ( M C A ) - P E G - D S P E into Pre-Formed L N s 86 5.2.9 H P T S Loading 86 5.2.10 Doxorubicin Loading 87 5.2.11 Doxorubicin Leakage Assay 87 5.2.12 C e l l Culture and L N Binding Assays 88 5.2.13 Cellular H P T S Uptake 89 5.2.14 Doxorubicin Uptake into Cells as Determined by F l o w Cytometry 89 5.2.15 L N Uptake into Cells as Observed by Microscopy 90 5.3 Results and Discussion 91 5.3.1 Synthesis o f Targeting Lipids 91 5.3.2 R G D - L N Characterisation 92 5.3.3 Methods for Producing R G D - L N s 95 5.3.4 Drug Retention in RGD-Targeted L N s ; 96 v  5  v  3  2  2  2  2  2  2  2  2  2  2  VI  5.3.5 Cellular Uptake and Processing of R G D - L N s 5.3.6 In vitro Drug Delivery by R G D - L N s 5.3.7 Relation to Existing R G D - L N Data 6 Future W o r k 6.1 Pharmacokinetics of R G D - L N s 6.2 Anti-tumor Efficacy of RGD-Targeted Therapeutics 6.3 Choice o f Drug to be Delivered 6.4 Bimodal and Multifaceted Chemotherapy Regimes References APPENDIX  97 99 101 104 104 105 107 108 109 122  LIST OF TABLES T a b l e 1.1 S u m m a r y of in vitro binding c h a r a c t e r i s a t i o n s for R G D - c o n t a i n i n g targeting the a f3 integrin T a b l e 5.1 S u m m a r y of R G D - l i p i d s u s e d to study R G D - L N s in vivo T a b l e 5.2 S u m m a r y of R G D - l i p i d s s y n t h e s i z e d T a b l e 5.3 C h a r a c t e r i z a t i o n of R G D - L N s v  3  ligands 15 73 91 93  Vlll  LIST OF FIGURES F i g u r e 1.1 T y p e s of L N e m p l o y e d for drug delivery 2 F i g u r e 1.2 A n t i b o d y e n g i n e e r i n g 7 F i g u r e 2.1 C o m p a r i s o n of M a l - P E G - D S P E r e a g e n t s obtained f r o m different s u p p l i e r s . . . . 2 5 F i g u r e 2.2 A d d i t i o n of M a l - P E G - D S P E to F a b ' to form F a b ' - P E G - D S P E 26 F i g u r e 2.3 S i z e e x c l u s i o n purification of F a b ' - P E G - D S P E 27 F i g u r e 2.4 I o n - e x c h a n g e c h r o m a t o g r a p h y of F a b ' - P E G - D S P E c o n j u g a t e s 28 F i g u r e 2.5 T w o m e t h o d s of incorporating targeting ligands for I L N m a n u f a c t u r e 29 F i g u r e 3.1 T h e a n g i o g e n i c transition 32 F i g u r e 3.2 S c h e m a t i c of the C a p i l l a r y E l e c t r o p h o r e s i s Frontal A n a l y s i s M e t h o d 38 Figure 3.3 S y n t h e s i s of fluorescently l a b e l e d c R G D f K peptides 45 F i g u r e 3.4 E l e c t r o p h o r e g r a m d e m o n s t r a t i n g t h e purity of c R G D f K - 4 8 8 '.. 4 6 Figure 3.5 E l e c t r o p h o r e g r a m s p r o d u c e d for C E - F A e x p e r i m e n t s 48 F i g u r e 3.6 c R G D f K - 4 8 8 : a p integrin binding i s o t h e r m s g e n e r a t e d from the C E - F A method 49 F i g u r e 4.1 Binding a n d E n d o c y t o s i s of c R G D f K - 4 8 8 v i a the a p integrin 51 F i g u r e 4 . 2 Quantitation of the a p receptor o n cultured cells 59 F i g u r e 4 . 3 Binding of fluorescently l a b e l e d p e p t i d e s to H U V E C or M 2 1 L cells 61 F i g u r e 4 . 4 S p e c i f i c B i n d i n g a n d E n d o c y t o s i s of c R G D f K - 4 8 8 a n d L M 6 0 9 to H U V E C a n d M 2 1 L cells 62 Figure 4 . 5 C o m p e t i t i o n of c R G D f K - 4 8 8 : H U V E C binding with free c R G D f K 63 F i g u r e 4 . 6 T i m e - d e p e n d e n t uptake of c R G D f K - 4 8 8 a n d c R A D f K - 4 8 8 by H U V E C 65 F i g u r e 4 . 7 a p integrin s u r f a c e receptor e x p r e s s i o n after e x p o s u r e to c R G D f K 66 F i g u r e 4 . 8 B i n d i n g at 3 7 ° C a s s e e n by f l u o r e s c e n c e m i c r o s c o p y 67 F i g u r e 4 . 9 B i n d i n g at 4 ° C a s s e e n by f l u o r e s c e n c e m i c r o c o p y 68 F i g u r e 5.2 R G D - l i p i d s s y n t h e s i z e d 76 F i g u r e 5.3 S t e p s in the s y n t h e s i s of c R G D f K - P E G - K ( M C A ) - P E G - D S P E 77 F i g u r e 5.4 R G D - L N binding by H U V E C s a t 4 ° C a n d 3 7 ° C 94 F i g u r e 5.6 C o m p a r i s o n of formulation a n d post-insertion m e t h o d s for incorporating R G D lipids into L N s 95 F i g u r e 5.7 D o x o r u b i c i n is retained e q u a l l y well in R G D - L N s a n d non-targeted L N s :.. 9 6 F i g u r e 5.8 R G D - L N s a r e internalized a n d acidified by H U V E C 98 F i g u r e 5.9 U p t a k e of d o x o r u b i c i n p r e s e n t e d in t h e free form, in non-targeted L N o r R G D L N by cells e x p r e s s i n g v a r i o u s levels of the c t p integrin 100 v  3  v  v  v  3  3  3  2  v  3  2  IX  ABBREVIATIONS  LIF LM609X  Proton nuclear magnetic resonance nuclear overhauser effect Antibody binding capacity Aqueous acetonitrile Basic fibroblast growth factor Binding maximum Capillary electrophoresis Chloroform Cholesterol 4', 6-diamidino-2-phenylindole Dichloromethane Dynamic light scattering Dimethylformamide Deoxyribonucleic acid PEGylated Liposomal doxorubicin Distearylphosphatidyl choline Endothelial cell Extracellular matrix Ethylenediamine tetraacetic acid Enhanced permeability and retention Electrospray mass spectrometry Monomeric antibody fragment, variable region Dimeric antibody fragment, variable region Fetal bovine serum Antibody fragment, constant region 9-fluorenylmethyloxycarbonyl Fast protein liquid chromatography Antibody variable fragment H E P E S buffered saline 0 - B e n z o t r i a z o l e - N , N , N ' , N ' -tetramethyl-uronium-hexafluorophosphate 4-(2-hydroxyethyl)-1 -piperazineethanesulfonic acid N-hydroxybenzotriazole 8-hydroxypyrene-l, 3, 6-trisulfonic acid Human umbilical vein endothelial cells H a l f maximal inhibitory concentration Immunoglobin G Immunoliposomal nanoparticle Kilodalton Dissociation constant Observed rate constant Laser induced fluorescence Fluorescently labeled anti-a P3 antibody  LN  Liposomal nanoparticle  'HNMRNOE ABC aqACN bFGF Bmax  CE CHCb Choi DAPI DCM DLS DMF DNA Doxil DSPC EC ECM EDTA EPR ESI-MS Fab' Fab FBS Fc FMOC FPLC Fv HBS 2  FfBTU  HEPES HOBT HPTS HUVEC IC IgG ILN kD K 5 0  d  kobs  v  X  MCA ME MeOH MFI  Monoclonal antibody Matrix-assisted laser desorptive ionization time o f flight spectroscopy Maleimide polyethylene glycokooo distearoylphosphatidylethanolamine 7-methyl coumarin M o l a r Equivalents Methanol Mean fluorescence intensity  MLV  Multilamellar vesicle  MW NGR NME Pa Pbf  Molecular weight Asparagine-Glycine-Arginine N-mercaptoethylamine Pascal 2,2,4,6,7-pentamethyldihydrobenzofuran-5-sulfonyl side chain protecting group Phosphate buffered saline 10% (v/v) fetal bovine serum i n phosphate buffered saline Polyethylene glycol Pharmacokinetic Pounds per square inch Benzotriazol-1 -yl-oxytripyrrolidinophosphonium hexafluorophosphate Reticuloendothelial system Retention factor Arginine-Glycine-Aspartic A c i d Arginine-Glycine-Aspartic A c i d Liposomal Nanoparticles Rhodamine phosphatidyl ethanolamine Reversed phase-high pressure liquid chromatography Structure activity relationship Sodium dodecyl sulphate polyacrylamide gel electrophoresis Small interfering R N A Solid phase peptide synthesis H a l f time (s) 5 -carboxy-tetramethylrhodamine t-butyl side chain protecting group Triethylamine Trifluoroacetic acid Triisopropyl silane Thin layer chromatography Ultraviolet Vascular endothelial growth factor Benzyl side chain protecting group  mAb MALDI-TOF Mal-PEG-DSPE  PBS PBS/FBS PEG PK Psi PyBop RES Rf RGD RGD-LNs Rhod:PE RP-HPLC SAR SDS-PAGE siRNA SPPS tl/2  TAMRA t-Bu TEA TFA TIS TLC UV VEGF Z  XI  ACKNOWLEDGEMENTS I would like to thank my parents, Carol and T i m Cressman, who have inspired me to enjoy life and nature. From delivering those first lessons in basic reasoning to providing ongoing financial support, they have made this thesis possible. Secondly, I would like to thank Professor Cullis, a gifted leader, scientist, businessman and colleague. H i s mentorship has significantly changed my life and taught me the importance o f clear communication, focus and prioritization in research. Exceptional guidance is duly acknowledged from D r . Theresa A l l e n , a close mentor for this project. K i n d thanks are likewise given to my supervisory committee, D r . Roger Brownsey, and D r . D a v i d Chen for their insightful and intellectually stimulating feedback, which has greatly enriched the quality o f this thesis. I am forever indebted to my family and friends. First, a heartfelt thanks to those o f which have stood by me for many years; Desirae Rovere, Michael Guiterrez, Jean Cressman, Maxine Williams, Fred Newhouse and M a x w e l l Newhouse. Secondly, thanks to friends from U B C ; Barbara Lelj-Garolla D i Bard, Nicole Quenneville, Kaley Wilson, Lianne M c H a r d y , Chantal Levesque, Sue De Jong, Ian Dobson and D r . Scott Covey. Likewise, the members o f the Cullis lab have been supportive o f this project, and I thank each person for their friendship and collaboration. I extend my thanks to the U B C Biochemistry Department for providing a first class P h D program and the infrastructure for communication within the department and the scientific community. Likewise, thanks to several members o f the U B C Chemistry Department who.have facilitated this project, in particular; D r . D a v i d Chen and members  Xll  of the Chen lab, D r . John Sherman and members o f the Sherman lab, D r . Lawrence Mcintosh, D r . M a r k Okon and the Bioservices facility. I would like also to acknowledge my mentors from Gryphon Therapeutics, D r . Gerd Kochendoerfer, D r . Shiah-Yun Chen, and D r . Stephen Kent for sharing their expertise in peptide chemistry. A kind thank you is also extended to M r . D o n M o s l i n , my high school B i o l o g y teacher. A l o n g with the basics o f life sciences, he taught me the importance o f surrounding yourself with positive people. Finally,  financial  support from the National Science and Engineering Research  Council ( N S E R C ) , and Canadian Institute for Health Research ( C I H R ) in the form o f P h D studentships,  the American Peptide Society for travel grants to attend international  symposia and research operating grants provided by N S E R C , Pharmaceuticals Inc. is acknowledged.  C I H R and Tekmira  Xlll  CO-AUTHORSHIP STATEMENT ^ - N M R r b a s e d confirmation o f the cyclic structure o f c R G D f K was acquired by D r . M a r k Okon, who also aided in the interpretation o f the spectra. Chapters 3 and 4 describe work that is submitted for publication in the journal, Biochemistry. The cell-based binding characterization has been summarized in a manuscript written by myself and revised by Dr. Pieter Cullis. D r . D a v i d Chen, M s . Jane M a x w e l l and M s . Y i n Sun are listed as co-authors for their help with interpretation o f the cell-binding kinetics and characterization o f the c R G D f K - 4 8 8 peptide conjugates. Likewise, all C E experiments have been conducted i n collaboration with either D r . N i n g Fang or M s . Y i n Sun in D r . Chen's lab. The application o f capillary electrophoresis frontal analysis ( C E - F A ) to describe a 6 :integrin binding as presented in this thesis v  3  constitutes part o f a manuscript written by M s . Sun and revised by D r . Chen.  I am  responsible for initiating the project, experimental design, interpretation o f results and coauthorship o f this manuscript. A substantial part o f the manuscript and interpretation o f the data obtained by the C E - F A method is rationalized by new equations for determining multi-site binding parameters, which were developed by M s . Sun and D r . Chen.  These  equations have been included as an appendix (written entirely by M s . Sun and D r . Chen) to this work since they are applied to the results obtained in this thesis. The full manuscript has been submitted to the journal, Analytical Chemistry in July o f this year. Acquisition o f the M 2 1 and M 2 1 L melanoma cell lines was made possible through a material transfer agreement between Dr. D a v i d Cheresh and D r . Pieter Cullis, established in July o f 2006.  xiv Work done under my direction by Mr. Ian Dobson, a undergraduate student in Dr. Cullis' lab, contributed to the synthetic procedures that were developed in Chapter 5, and constitute a significant portion of the results compiled in a manuscript in preparation for submission, co-authored by Dr. Cullis, Mr. Dobson and myself.  1  1 Introduction to Targeted Liposomal Nanoparticles  1.1 Liposomal Nanoparticles The use o f liposomal nanoparticles as drug delivery agents has evolved from a line o f research originating over 40 years ago, based on the ability o f these unilamellar vesicles to entrap material in an aqueous compartment (Bangham et al., 1965). It was then known that most amphipathic membrane lipids form multilamellar vesicles ( M L V ) consisting of concentric bilayers when they are dispersed in aqueous media. M L V are relatively large (micron) sized structures, however they can be extruded through 100 nm pore size polycarbonate filters to produce unilamellar vesicles with a homogeneous size distribution (Olson et al., 1979; Szoka et al., 1980). Typically, the resulting liposomal nanoparticles (LNs) are 100 nm in diameter and each particle contains approximately 1.25 x 10 lipid 5  molecules (Winterhalter and Lasic, 1993). L N s have been widely used as models of biological membranes to study membrane permeability and transport across the bilayer (Madden et al, 1987). In addition to their utility as model membranes, drugs may be encapsulated within the particle's aqueous interior compartment. The ability o f L N s to deliver drugs preferentially to disease sites, such as solid tumors, can result in considerable improvements i n efficacy; therefore, L N s are widely studied for use in therapeutic applications (Figure 1.1).  2  Figure 1.1 Types of LN employed for drug delivery A ) L N s that encapsulate small molecule drugs B) " S t e a l t h ® " L N s have a hydrophilic surface coating (typically polyethylene glycol (PEG)) that enables them to survive in the circulatory system for extended periods o f time C) Targeted L N s are surface-modified with a targeting ligand to increase the accumulation o f the particle in target cells D ) L N s that encapsulate nucleic acid-based drugs such a plasmid, antisense oligonucleotides or s i R N A .  1.2 Liposomal Nanoparticles and Drug Delivery The first preparation o f a liposome with entrapped solute was characterized in 1965 by A . D . Bangham in Cambridge, U K (Bangham et al., 1965). The evolution of liposomes as drug delivery systems was subsequently accelerated in the 1980's by the development o f techniques to rapidly generate well defined nanoparticulate liposomal systems and to efficiently load them with drugs (Gabizon et al., 1985; Mayer et al., 1985; Szoka et al., 1980). The observation that long-circulating L N preferentially accumulate at sites o f disease, including sites o f infection, inflammation and tumors, due to the leaky nature o f the vasculature in these regions (Segal et al., 1975; Jain, 1987), gave a solid rationale for delivering drugs in L N systems. L N s have several features that have contributed to their success as a drug delivery system. Encapsulation within the aqueous cavity o f an L N can enhance the in vivo activity of drugs by protecting them from breakdown in the body and can reduce the toxic effects  3  of drugs, such as anti-cancer drugs, by reducing delivery to sensitive tissue (Forssen et al., 1981) . In addition, L N s are biocompatible, so that they may be used in vivo, and their physical properties can be readily manipulated. L N production and drug-loading techniques have been optimized and standardized such that L N s can be manufactured on a large scale. L N technology also offers flexibility such that the lipid composition may be varied to match the desired characteristics o f the drug that is being delivered, and can be used to optimize the halftime o f release o f the drug from the L N .  These features are summarized  by Maurer et al., 2001, and more recently by A l l e n and Martin, 2004, noting examples where L N encapsulation gave the associated drug long circulation lifetimes, enhanced accumulation at disease sites and increased efficacy for a variety o f drugs. The range o f material that may be encapsulated within the L N is diverse. L N s have been used for the encapsulation o f small molecule drugs including anti-cancer (Olson et al., 1982) and antifungal drugs (Mehta et al., 1984), and nucleic acid-based drugs such as plasmids for gene therapy (Ambegia et al., 2005; Choi et al., 2003), immunogenic D N A oligonucleotides ( M u i et a l , 2001) and s i R N A oligonucleotides (Zimmermann et al., 2006) (Figure 1.1). A review o f the current status o f modern drug delivery systems and their in vivo application has been published, noting six L N formulations that are clinically approved drugs and many others in advanced clinical trials (Allen and Cullis, 2004). The application o f L N s as carriers o f nucleic acids has required the development o f a sophisticated class o f nanoparticles for in vivo L N delivery. These particles must entrap high amounts o f nucleic acid (Jeffs et al., 2005), survive for prolonged amounts o f time i n the circulatory system (Fenske et al., 2002) and release their contents in the cytoplasm o f cells that internalize the particle (Kale and Torchilin, 2007; L i et al., 2005).  4  1.3 LNs with Extended in vivo Circulation Lifetimes  Despite the advantages o f encapsulating drugs within an L N , 25 years passed after the discovery o f liposomes, before the first L N drug formulation, Ambisome®, achieved regulatory approval for clinical use (Adler-Moore, 1994). This can be explained i n part by the complexity o f L N behavior when administered in vivo. A particularly important issue emerged from early in vivo experiments that showed that L N s were rapidly removed from the circulation by uptake o f the particle by fixed and free macrophages o f the reticuloendothelial system ( R E S ) that reside in the liver and spleen (Allen and Everest, 1983). A major breakthrough came from the observation that modification o f the L N surface could prevent the cells o f the R E S from recognizing and subsequently clearing the foreign particle. It was originally shown that the inclusion o f sialic acid (Allen and Chonn, 1987) or polyethylene glycol (PEG)-containing lipids, (Klibanov et al., 1990; Papahadjopoulos et al., 1991) in the L N formulation could significantly extend the circulation lifetime. The term "Stealth® liposome" describes L N s that can survive i n the blood stream for prolonged amounts o f time.  PEG-containing L N formulations have shown the greatest life-time  enhancement, and some optimized formulations, such as Doxil®, have reported circulation half-times o f up to 45 h i n humans (Gabizon et al., 1994). A long circulation half-life is a major advantage because the accumulation o f L N s at the site o f a tumor has been shown to increase with the amount o f time that the particles remain circulating i n the body (Huang et al., 1992). In 1995, a doxorubicin-containing L N (DOXIL®) became the first major anti-cancer L N to be approved.  It was initially approved for use against Aids-related Kaposi's  5  sarcoma (Northfelt et al., 1996), and subsequently for the treatment of refractory ovarian cancer in 1999 (Muggia and Hamilton, 2001).  1.4 Immunoliposomes (ILNs)  1.4.1 Targeting LNs to the Disease Site L N s containing anti-cancer drugs that accumulate at the tumor site without the aid o f a targeting molecule (defined as "passive" targeting) take advantage o f the enhanced permeability and retention ( E P R ) effect, caused by the leaky nature o f the neovasculature that supplies blood to the tumor (Maeda et al., 2000).  In addition, the  tumor  microenvironment does not have effective lymphatic drainage to remove particles from the tumor after they accumulate there. The E P R effect thus facilitates extravasation o f L N s from the circulation into tumor tissue and their subsequent retention there. Since the early days o f L N drug development, it has been hoped that L N s could be designed to deliver their contents to specific cells via a molecular targeting mechanism (Torchilin, 1985). In this concept, a targeting ligand on the surface o f an L N could aid in the binding of the LN-encapsulated drug specifically to the target cells at the disease site. It is hypothesized that the targeted L N should further improve the therapeutic index o f the drug over that o f non-targeted L N s . Commonly used anti-cancer drugs are often delivered below their effective dose due to their toxic side effects. Thus, i f a targeted L N can deliver drug to the tumour at concentrations greater than the effective dose, the anti-cancer effect should also increase.  6  Initiatives to covalently tether antibodies to liposomes began with work done i n the early 1980's by Heath and colleagues who demonstrated that the attachment o f anti-human erythrocyte antibodies to an L N surface could promote adherence o f the particle to human red blood cells in a mixed cell population (Heath et al., 1981, Leserman et al, 1981). Shortly  after,  researchers  began  to  investigate  antibody-targeted  liposomes,  or  immunoliposomal nanoparticles (ILNs) to deliver a variety o f drugs targeting diseases o f viral, oncologic and malarial origin (reviewed in Sapra and A l l e n , 2003). One o f the first in vivo studies with I L N s was conducted using an I g G antibody against the gp80 antigen present  on the avian myeloblastosis virus coat protein  (Dhananjaya and Antony, 1988). The resulting I L N s showed promising efficacy against the viral infection, with higher accumulation o f the particles i n target tissues, long circulation lifetimes and no immune response from the host.  The last finding was  surprising, since the IgG antibody used for targeting was produced in a rabbit while the experiment was done in a mouse. The authors do not acknowledge any response by the host's immune system to the foreign antibody. Recent in vivo studies report the opposite effect, that the I L N s are rapidly cleared. These findings have been attributed to recognition of foreign antibodies on the I L N ' s surface by the host's immune system, (Harding et al., 1997). Advancements, such as the somatic cell hybridization technique (Kohler and Milstein, 1975), and improvements i n antibody engineering (Padlan, 1991) have had a tremendous  impact on the progression o f antibody-based therapeutics.  Improved  formulation techniques have further enabled several antibody-based therapeutics to come to market with great success (Daugherty and Mrsny, 2006). The approval status and anti-  7  cancer effect o f the dozen or so approved antibody-based drugs has recently been reviewed (Imai and Takaoka, 2006). In 1997, a chimeric (part human and part mouse (Fig. 1.2A)) antibody, Rituxan®, used to treat B-cell lymphomas, became the first antibody to receive United States Federal Drug Advisory ( F D A ) approval. This was followed by Herceptin®, a humanized antibody (Figure 1.2B), approved in 1998 for the treatment o f Her-2/neu positive breast cancers.  Humanized antibody scaffolds are favored since their design  enables long circulation lifetimes of the antibody with a circulation half-time of 13 days (Stephens et al. 1995).  A current review o f F D A approved and antibody-based drug  candidates can be found i n Nayeem and Khan, 2006. Clinicians  are  now trying to  optimize the  combination o f antibodies  with  conventional anti-cancer drugs. In this regard it has been shown that the administration o f an LN-encapsulated drug in combination with Rituxan® improves the clinical outcome for patients with non-Hodgkin's lymphoma (Zaja et al. 2006).  i  iHuman sequences  Figure 1.2 Antibody engineering  A ) Chimeric monoclonal antibody (mAb). B ) Humanized m A b . C) Fab' fragments. Fab' fragments are typically made by enzymatic digestion o f the Fc region and subsequent reduction o f the disulphide between the antibody dimers D) Single chain variable fragment (scFv) antibodies, a recombinant antibody that contains the minimal protein sequences required for antigen binding.  8  The use o f antibody fragments for L N targeting such as Fab' (the antibodies variable region including the complementarity determining region (CDR)), or scFv (a recombinant protein made from the C D R ) is preferential to using whole antibodies for targeting since they are less prone to rapid clearance from the circulation (Maruyama et al., 1997). This is largely attributable to a reduction in the immune response to the species-specific sequences that exist i n the F c region o f the antibody when it is removed from the whole m A b (Gagne et al., 2002).  1.5 Contemporary ILNs and ILN Production Our understanding o f the desired in vivo characteristics o f I L N s has progressed together with the increased availability o f antibodies.  In the past decade,  important  concepts such as the need to target receptors that internalize the particle (Sapra and A l l e n , 2002) and new approaches to formulation and I L N production methods (Ishida et al., 1999) have also brought the technology closer to the clinic. A significant breakthrough in the advancement o f I L N production has been the use o f the "post-insertion" method (Ishida et al., 1999; Moreira et al., 2002). In this method, a targeting ligand, such as an antibody, is first coupled to a lipid anchor, and then coincubated with pre-formed L N s . The advantage that this approach holds over conventional I L N production methods is that the targeting construct can be made, analyzed and purified before it is incorporated into the L N formulation, resulting in a well-characterized system.  9  1.5.1 ILNs in vivo Unlike passive targeting, which relies on the E P R effect to enhance accumulation o f L N at disease sites due to the leaky nature o f the vasculature in these regions, active targeting implies the direct binding o f the L N to surface receptors on target cells followed by endocytosis o f the ligand that binds to these receptors. This can also result in internalization o f the L N into the target cells, thereby enhancing efficacy (Allen et al., 2002; Moreira et al., 2002). The  importance  o f I L N internalization was established through  a series o f  experiments using I L N s targeted with anti-CD 19 (internalizing) or anti-CD20 (noninternalizing) antibodies in a B-cell lymphoma mouse model (Sapra and A l l e n , 2002). Since both antibodies bind to antigens present on malignant B-cells, the effect o f targeting with an internalizing versus non-internalizing antibody could be demonstrated. This study showed that the internalized I L N could increase the life span of the mouse by 65.2% as opposed to 24.3% for non-internalized I L N . In another encouraging contribution by Park and colleagues, I L N s were made with scFv antibody fragments against the HER2/neu growth factor that is over-expressed on some breast cancer cells. The results reported cures of up to 50 % in tumor-bearing mice when the target cells expressed high antigen densities (Park et al., 2002). This work has resulted in many pre-clinical demonstrations of improved efficacy in animal models yet none have progressed to demonstrate efficacy i n humans.  This is  evidenced by a series o f recent publications that describe the gram-scale and G M P compliant production o f anti-HER2 scFv-PEG2ooo-distearoylphosphatidylethanolamine ( D S P E ) for post-insertion into L N s (Nellis et al., 2005a; Nellis et al., 2005b).  This is  10  perhaps the best characterized o f the I L N pre-clinical formulations, however, no clinical trial on this system has been initiated.  1.5.2 ILNs in the Clinic Despite encouraging pre-clinical  results,  only one  clinical  study  o f an I L N  doxorubicin formulation has been reported. It describes a Phase I clinical trial o f a targeted ILN,  made with Fab2 antibody fragments  (Matsumura et al., 2004).  directed against stomach cancer  antigens  The authors found that their I L N s showed the same in vivo  circulation kinetics as'non-targeted L N s and established a recommended dosing regime for phase II clinical trials. Further detail on this formulation is limited, as the supporting preclinical studies have not been reported. The major challenges that must be overcome before targeted L N s can be used i n the clinic include the accurate characterization o f these particles and the reproducible and costeffective manufacture o f these particles. With almost a hundred pre-clinical studies on targeted I L N s published, only one well-documented and convincing method for I L N production and characterization has been presented. This uses chemically defined s c F V PEG2000-DSPE lipids that are post-inserted into pre-formed L N s . Further work is needed to achieve well-characterized I L N systems that can be readily manufactured and where efficacy is well correlated with the composition, size and contents o f the I L N . A s shown in Chapter 2 of this thesis, techniques that involve the coupling o f targeting proteins to L N to produce I L N result in systems that are inherently difficult to characterize, manufacture and reproduce, and are also extremely expensive. For these reasons the bulk o f this thesis  11  concerns the development o f smaller molecule peptide targeted systems that allow more rigorous analysis and that can be incorporated into L N in a quantitative manner at the time of L N manufacture.  1.6 Small Molecule Targeting Ligands From many perspectives, the use o f smaller molecules with well-defined analytical characteristics is attractive for L N targeting. Although they have lower affinities for target molecules than antibodies, small molecules such as peptides and peptidomimetics are more straightforward to develop clinically than macromolecular drugs (Cho and Juliano, 1996; Livnah et al., 1996).  For example, small molecule ligands can be made to be  chemoselective during conjugation to lipid anchors by the use o f protecting groups that shield other sites until after the conjugation reaction is complete. Peptides may also be confined in space by conformational restrictions to mimic biological interactions between proteins and may also protect the peptide against enzymatic degradation in the body (Cho and Juliano, 1996; Livnah et al., 1996). These, along with several other advancements in the field of peptide science, have facilitated the development o f many new pharmaceuticals over the past 20 years. Over a hundred peptide-based drugs are now approved and generating substantial revenue.  Some hallmark examples o f successful peptide drugs on the market include  Fuzeon®, an anti-entry H I V inhibitor, Exubera®, an inhalable version o f insulin and Forteo®, an analogue of the human parathyroid hormone used for treatment of osteoporosis (reviewed in Watt, 2006).  12  1.6.1 Peptides with Disease Site Affinity  The search for peptides that might serve as affinity agents for targeting has been facilitated by work from Ruoslahti's group in the early 1990's using a method known as in vivo phage display. In these experiments, a library o f phage expressing different peptides on their surface were injected into a mouse and subsequently isolated from specific organs. The peptides that were associated with the recovered phage particles were then identified. Using this method, Ruoslahti and colleagues have identified peptides that w i l l specifically home to the brain (Pasqualini and Ruoslahti, 1996), heart (Zhang et al., 2005), prostate (Arap et al., 2002), and bladder (Lee et al., 2007; Pasqualini et al., 2000). Ruoslahti's in vivo phage display technology has also led to the identification o f peptides that bind to tumor vasculature (Ruoslahti, 2000). arganine  ( N G R ) and arginine-glycine-aspartic acid  The asparagine-glycine-  ( R G D ) tri-peptide motifs were  identified in several different phages that selectively bind to the tumor-associated endothelium, but not to the vasculature o f healthy tissues.  Both the N G R and the R G D  motifs have thus been pursued as tumor targeting agents. The N G R motif was later found to home to aminopeptidase N , a cell surface protein expressed on the vascular endothelium (Pasqualini et al., 2000). Peptides containing N G R have been shown to guide both high molecular weight proteins and L N to the site o f a tumor (Pastorino et al., 2006; Pastorino et al., 2004). The tumor targeting properties o f the R G D motif were consistent with targeting to the  civfc  integrin expressed during angiogenesis as established by Cheresh and Ruoslahti  13  (Brooks et al., 1994c; Cheresh, 1987; Cheresh et al., 1987). Subsequently, the use o f R G D as an anti-angiogenesis agent has been heavily pursued, resulting in drugs now i n late phase II clinical trials (Friess et al., 2006).  1.6.2 Arginine-Glycine-Aspartic Acid (RGD) Originally  identified  in  Ruoslahti's  group  as  a  component  of  fibronectin  (Pierschbacher and Ruoslahti, 1984), the R G D sequence is present in several other proteins of the extracellular matrix that interact with the cell through the integrin receptors.  This  w i l l be discussed in more detail i n Chapter 3 regarding the involvement o f RGD-containing proteins i n angiogenesis. Members o f Kessler's laboratory have produced detailed studies to relate structure to activity to optimize the pharmacophoric contacts between an ctypV selective R G D peptide and the a | B integrin (Dechantsreiter et al., 1999). The resulting v  3  compound has the cyclic amino acid sequence: Arginine-Glycine-Aspartic acid-n-methyld-Phenylalanine-Valine, (cRGD(nme)fV) and features an n-methylated  amide  bond  between the D and f residues. Merck has developed this compound as an anti-angiogenesis drug candidate with the trade name Cilengitide®. In phase I clinical results, Cilengitide® is effective without toxic side effects and has achieved complete and partial responses in patients with glioblastoma multiforme (Nabors et al., 2007). Cilengitide® is also currently employed i n several phase II clinical trials and has orphan drug status in Europe against pediatric glioma.  14  1.7 RGD-Targeting Ligands and Affinity Analyses The R G D sequence has also been pursued as an affinity agent for targeting to the tumor vasculature.  A range o f RGD-based constructs have been used in attempts to  increase the concentration o f drugs ( K i m and Lee, 2004), imaging agents (Chen et al., 2004b; L i et al., 2007a; W u et al., 2005), antibodies (Schraa et al., 2002) and L N s (using methods o f surface coupling) (Dubey et al., 2004; Janssen et al., 2003; Schiffelers et al., 2003) at the site of solid tumors.  While these studies have reported some encouraging  results, many aspects of the basic ability o f RGD-based targeting ligands to target to the a (33 integrin remain uncharacterized.  This is because the characterization o f most anti-  v  angiogenic RGD-ligands is done by an inhibition or competition assay, which reports a half maximal inhibitory concentration, or IC50 value. In this case, the IC50 value describes the half-maximal concentration o f R G D - d r u g that is required to compete with a naturally occurring ligand that binds to the a (3 integrin. A low IC50 value indicates that only a v  3  small amount o f the competitor is required in order to disrupt the integrimligand interaction. Many researchers use purified vitronectin as the integrin's natural ligand to be displaced by R G D (see table 1.1). When describing affinity, an IC50 value is generally not the best method, particularly for the describing binding to the dvfo integrins, which are known to bind to several different ligands (Brooks, 1996).  The IC50 value is therefore  strongly dependent on the particular ligand that is initially bound to the integrin.  A  summary o f some key in vitro binding characterizations o f RGD-ligands is given in Table 1.1.  15  Table 1.1 Summary of in vitro binding characterizations of RGD-containing ligands targeting the a v P integrin 3  RGD-construct Linear RGD-containing peptide  Publication Dechantsreiter et al, 1999  Cilengitide  Dechantsreiter et al, 1999  RGD-modified antibodies  Schraa et al, 2004  Activity assay Biotinylated vitronectin in competition with unlabelled RGD-ligands for immobilized a |3 binding. Biotinylated vitronectin in competition with unlabelled RGD-ligands for immobilized cc |3 binding. Radiolabeled RGD-antibody in competition with unlabelled RGD ligand for binding to a pV expressing cells Radiolabeled echistatin in competition with unlabelled RGD-ligands for binding to immobilized a (3 integrin Radiolabeled echistatin in competition with unlabelled RGD-ligands for immobilized a p binding Fluorescent RGD-LNs in competition with nonfluorescent RGD-LNs for binding to a pVexpressing cells v  Constants reported IC o=210nM 5  3  v  IC =0.58 nM 50  3  IC o=150 nM 5  v  PEGylated RGDimaging agents  Chen et al, 2004a  Octameric and tetrameric RGD-imaging agents  Li etal, 2007  RGD-LNs  Janssen et al, 2003  v  v  IC n=30.3 nM 5  3  IC o=l 0 nM for octamer vs. 35 nM for tetramer 5  3  v  No binding constants are reported however, non-specific LNs, are shown to be unable to compete with RGD-LNs for cell binding  F r o m the summary provided i n Table 1.1 it can be observed that different conditions for the competition assays are employed between research groups.  A s a result, the I C  5 0  values reported are highly subject to experimental conditions. The lack o f formal binding studies likely arises from the lack o f labeled R G D molecules to simplify the conduct o f the binding assays. In this thesis, this problem is directly addressed by the synthesis o f a well-characterized fluorescently-labeled R G D ligand.  This construct was used to measure equilibrium binding parameters, which  describe R G D - b i n d i n g to the a p V i n t e g r i n alone and to cc pVexpressing cells. v  v  Similar difficulties are encountered when measuring the binding o f targeted L N s to target cells. A solution to this problem was to incorporate a fluorescent label into the R G D -  16  lipid before L N formulation. This resulted in the ability to calculate K d values to describe binding for RGD-targeted L N systems  1.8 Objectives, Evolution and Structure of this Thesis The original aim of this thesis was to develop well-characterized I L N employing Fab' or scFv fragments o f Rituxan for targeting drug-containing I L N s to the C D 20 receptor on B cells to improve the treatment of non-Hodgkin's lymphoma.  Such an I L N  appeared to represent an interesting opportunity given the synergy observed clinically between Rituxan and non-targeted L N s containing doxorubicin (Zaja, 2006). Given the lack o f success over the last 25 years in moving a targeted L N into clinical testing, it was of particular interest was to develop an I L N that was sufficiently well-characterized that it could conceivably gain regulatory approval and move forward into the clinic i f the preclinical results looked promising. This requires a system where the molecular composition o f the targeting ligands is well-defined, where the targeting ligands can be produced in a cost-effective and reproducible manner, where the binding properties o f the targeting ligand for the receptor are well-understood, where the number o f targeting ligands per L N is well-defined and reproducible and where the I L N as a whole can be made in a straightforward and reproducible manner. Ideally, the targeting ligand employed would also stimulate endocytosis and intracellular delivery o f the I L N contents, which would be expected to result in improved potency. After considerable efforts over two years to develop an I L N targeted with Fab' derived' from the m A b , Rituxan, this approach was abandoned.  A s briefly summarized in  17  Chapter 2 o f this thesis, many difficulties were encountered. These ranged from the relatively trivial, for example the fact that starting reagents such as activated P E G - l i p i d s for coupling to proteins varied significantly in quality between different manufacturers, to the more fundamental, such as the fact that the stoichiometry o f attachment of P E G - l i p i d to Fab' could not be easily controlled and the isolation o f molecularly defined F a b ' - P E G - l i p i d targeting  ligands at reasonable  yield proved impossible. These issues led to  the  consideration o f alternative approaches. As  detailed in this Introduction, peptide-targeting agents such as RGD-based  targeting ligands have a number o f attractive features. They are small, and can be made to be chemoselective in the sense that only one function on the molecule is available for coupling to a P E G - l i p i d , for example, leading to a molecularly defined species. The small size also leads to the possibility that the ligand-PEG-lipid w i l l be sufficiently soluble in organic solvent so that it can be incorporated in a quantitative manner into L N at the time that the L N is made. In the case o f the RGD-containing ligands, the added attraction is the affinity for tumor vasculature and the resulting potential for the treatment o f a variety of solid tumors. It was therefore decided to develop an L N targeting ligand based on Cilengitide®, which is the most advanced o f the RGD-based anti-angiogenic agents. However, it was realized that many basic features relating to using this as a targeting ligand were not available. First, as indicated in the preceding Section, the interactions o f antiangiogenic RGD-containing ligands such as Cilengitide® with the 0 ^ 3 integrin in cells are largely characterized in terms o f IC50 values rather than binding parameters such as dissociation constants (Kd). This led to attempts to characterize the binding of Cilengitide® to the isolated  a (33 v  receptor, which required the synthesis o f a fluorescently labeled  18  analogue of Cilengitide and the identification o f a suitable technique for characterizing the ligand-receptor binding properties. I therefore synthesized c R G D f K - 4 8 8 , a fluorescently labeled Cilengitide® analogue and initiated the collaboration with Dr. David Chen to use capillary electrophoresis ( C E ) to measure, for the first time, the Kd values describing the binding o f this analogue with the isolated a p receptor. This work, described in Chapter 3, v  3  leads to the conclusion that c R G D f K - 4 8 8 binds to the isolated a (33 receptor v  with  micromolar affinity at a 2:1 stoichiometry. The second basic area that required characterization concerned the binding to the a P 3 receptor in intact cells and the influence o f R G D - l i g a n d binding on endocytosis. The v  literature on the effects o f the binding o f RGD-containing ligands on endocytosis is ambiguous, with one report suggesting that endocytosis o f RGD-peptides is independent o f binding to 0^3 integrin (Castel et al., 2001), whereas other reports suggest the opposite (Renigunta et al., 2006). I therefore  performed  the  studies  detailed in Chapter  characterizing the binding o f the c R G D f K - 4 8 8 ligand to the a | 3 v  3  4  integrin on human  umbilical vascular endothelial cells ( H U V E C ) , showing that c R G D f K - 4 8 8 binds to a B v  3  integrins in the membrane again with affinities in the micromolar range and also showing that such binding stimulated endocytosis of the ligand that was dependent on the presence of the a P 3 integrin on the cell surface. v  These studies set the stage for achieving one of the original aims of this thesis, namely, the construction o f a well-defined ligand-PEG-lipid that could be inserted in a quantitative manner into L N systems. I therefore synthesized c R G D f K - P E G 2 - K ( M C A ) PEG2-DSPE, an R G D - P E G - l i p i d that is fluorescently labeled in order to allow for straightforward estimates o f R G D incorporation into L N systems.  Further, our construct  19  contains a defined number o f ethylene glycol units in the P E G spacer moiety, resulting in a molecularly well-defined targeting ligand. R G D - P E G - l i p i d s previously constructed are not labeled and contain P E G with a range o f molecular weights. These studies are presented in Chapter 5, where it is shown that c R G D f K - P E G - K ( M C A ) - P E G - D S P E 2  2  can be inserted  into L N at the time the L N is made, that the resulting I L N systems target to and are endocytosed into H U V E C , and that such systems can be loaded with an anticancer drug such as doxorubicin and can deliver this drug to the interior o f these cells. A substantial amount o f time and effort has been spent on the syntheses and in vitro studies detailed above, such that this work is now ready for the next phase o f study that w i l l involve characterization o f the in vivo properties o f the L N systems constructed. In Chapter 6 some o f the issues that w i l l need to be addressed in such studies are discussed.  20  2 Preliminary Experiments  2.1 Introduction The studies presented in this brief Chapter produced results that significantly changed the direction o f this thesis. These findings concern the difficulties encountered in constructing I L N s using "classical" procedures, which led to the abandonment o f attempts to attach either intact antibodies or Fab' fragments to L N s in order to construct I L N s . In particular, in a series o f studies using Fab' fragments o f anti-CD20 antibodies, (Rituxan®), it was found the non-chemoselective nature o f the coupling between Fab' fragments and maleimide-PEG-lipids makes it difficult  to obtain well-defined compounds.  Other  limitations also were encountered such as the poor quality o f the commercially available maleimide-PEG-lipids and a low yield o f synthesis.  A s a result, it was decided to  investigate targeting ligands, such as the R G D peptide, that can be incorporated into I L N in a straightforward and controllable manner. Rituxan® is an approved drug for the treatment o f B-cell lymphoma, and is a monoclonal antibody targeted to the C D 2 0 receptor on malignant B-cells (Maloney et al., 1997). It was initially reasoned that by using Rituxan® as a targeting ligand, beneficial effects arising from both antibody-mediated and LN-mediated anti-tumor efficacy would be observed. The approach taken to construct I L N s targeted by Rituxan® was to isolate the Fab' fragments from the intact antibody and couple these fragments to maleimide-PEG2oooD S P E in a quantitative manner.  The Fab' fragments were then used rather than intact  21  antibodies to avoid the potential rapid clearance from the circulation o f I L N s that contain the Fc portion of the antibody which is species specific (Harding et al., 1997).  22  2.2 Methods 2.2.1 Materials Unless otherwise stated, all reagents were purchased from Aldrich (Milwaukee, WI) and were o f reagent grade or higher.  A l l buffers were vacuum degassed, sparged with  helium and stored under argon until used. Sodium dodecyl sulphate polyacrylamide gel electrophoresis  (SDS-PAGE)  separation  was  performed  polyacrylamide gels and stained with Coomasie blue.  on  8%  non-denaturing  Pre-stained molecular weight  markers were purchased from Invitrogen (Carlsbad, C A ) .  2.2.2 Fab' Formation Fab' fragments  were made from commercially available Rituxan® ( B C Cancer  Agency, Vancouver, B C ) , a chimeric m A b (see Figure 1.2B).  The antibody was first  digested with immobilized pepsin (Pierce, Rockford, IL) for 8 h, followed by purification on Protein A affinity columns (Pierce, Rockford, I L ) .  The protein was then dialyzed  overnight, reduced with a solution of 100 m M Tris base, 65 m M N-mercaptoethylamine ( N M E ) p H 7.5 for 30 min and desalted on PD10 columns (Amersham Biosciences) equilibrated with "PEGylation buffer" (130 m M N a C l , 3 m M KC1, 10.4 m M N a P 0 , 5 2  4  m M E D T A , 1.7 m M KH2PO4 p H 7.0). The resulting product had an apparent molecular weight between 60 and 80 k D , as detected by S D S - P A G E .  23  2.2.3 Maleimide-PEG o-DSPE Analysis 200  The quality o f maleimide-PEG-lipids was assessed using a Bruker B i F l e x I V matrixassisted laser desorptive ionization time-of-flight mass spectrometer ( M A L D I - T O F ) with sinnapinic acid used as the matrix.  Maleimide-PEG-lipids from Nektar Polymers and  Avanti Polar Lipids were compared, resulting i n the finding that the Nektar product had significant quality issues and the Avanti product was chosen for use in further experiments.  2.2.4 Fab'-PEG oo-DSPE Conjugation 20  Conjugation  of Fab' fragments  (75  p M ) to maleimide-PEG2ooo-DSPE (herein  referred to as M a l - P E G - D S P E ) (Avanti Polar Lipids) was carried out over 8 h at room temperature under argon gas, using up to 10 molar equivalents o f M a l - P E G - D S P E .  The  p H was measured once all o f the reactants were dissolved and carefully adjusted with p H 8.0 PEGylation buffer. A series o f reactions were tested between p H 5-8, and the optimal p H to be employed for subsequent studies was found to be 7.0. The product, Fab'-PEG2oooD S P E typically formed after one hour and was detected by the appearance o f a band with greater retardation than the Fab' band by S D S - P A G E .  24  2.2.5 Fab'-PEG o-DSPE Purification 200  The reaction mixture was purified by size exclusion chromatography over 40 m l Sephadex C1-4B hydrated in PEGylation buffer.  The F a b ' - P E G o o - D S P E product eluted 20  between 7-10 m l and Fab' eluted after 10 m l . Alternatively, ion exchange chromatography was performed using a 1 m l S X Hitrap column on an A K T A fast protein liquid chromatography ( F P L C ) purification system ( G E Healthcare, U K ) .  The PEGylation  mixture was first dialyzed overnight against 20 m M H2PO4, p H 6.8 ( I E X buffer), then 1 mg of the total protein content protein was purified using a gradient of I E X buffer and I E X buffer containing 500 m M N a C l in order to elute the PEGylation mixture over 20 min.  25  2.3 Results  2.3.1 Quality of Mal-PEG-DSPE Initial attempts to conjugate Fab' fragments to M a l - P E G - D S P E were unsuccessful, bringing into question the quality o f the reagents used. The analysis shown i n Figure 2.1 resulted in the choice to use M a l - P E G - D S P E that was obtained from Avanti Polar Lipids for subsequent experiments.  A s shown in Figure 2.1 A , the product supplied by Nektar  produced a charge series by mass spectrometry that indicated a poor quality compound, whereas the Avanti product produced a spectrum consistent with a much purer compound. This can be seen in Figure 2 . I B where the charge series produced from repeat P E G units is devoid o f the contaminating peaks seen i n Figure 2.1 A .  Further analysis by ' H N M R  provided evidence that the maleimide functionality on the Avanti product was stable and accessible for conjugation.  Figure 2.1 Comparison of Mal-PEG-DSPE reagents obtained from different suppliers The quality of Mal-PEG-DSPE obtained from A) Nektar and B) Avanti Polar Lipids. Both lipids were analyzed immediately after receipt from the suppliers and dissolved in H 0 immediately before analysis by MALDI-TOF mass spectrometry. The two panels above show the spectra produced. The high degree of heterogeneity in spectrum A implies an impure product. 2  26  2.3.2 Fab'-Maleimide-PEG-DSPE Conjugation The overall yield (yield is defined as moles o f product produced as a percentage o f the moles of the limiting reactant) for the formation o f Fab' fragments from the intact antibody was 6 6 % . This yield decreases to less than 30% upon reaction with M a l - P E G D S P E using the optimized conjugation conditions detailed in the Methods section. Furthermore, a higher yield could not be obtained by increasing the molar equivalents o f M a l - P E G - D S P E in the reaction (see Figure 2 . 2 ) . This yield could also not be improved by adjusting the p H (between 5 . 5 to 8.0) or by increasing the concentration o f the Fab' fragments. The low yield is further decreased to less than 5 % overall upon purification by SEC. Molar Equivalents Mal-PEG-DSPE added to Fab  -  J  MW  10.5  1  2.5  5  10  I  Fab'  Mab Fab2  _ Fab'-PEG-DSPE Conjugates -Fab'  Figure 2.2 Addition of M a l - P E G - D S P E to Fab' to form F a b ' - P E G - D S P E  Analysis of the effect of adding increasing amounts of Mal-PEG-DSPE to Fab' fragments in an attempt drive the conjugation reaction to completion. Up to ten molar equivalents of Mal-PEG-DSPE were added to Fab' fragments, and the reaction mixture was separated by SDS-PAGE and stained with Coomassie blue reagent.  27  2.3.3 Pure Fab'-PEG-DSPE Preparations are Inherently Difficult to Achieve A t least two new bands could be observed upon conjugation o f Fab' fragments to M a l - P E G - D S P E . These were interpreted as mono- and di-PEG ooo-DSPE substituted Fab' 2  fragments. Some resolution between Fab' fragments and the higher M W conjugates could be obtained by size exclusion chromatography, but the overall yield significantly decreased since the products co-eluted with the unconjugated protein. The elution profile o f a typical size exclusion purification is shown in Figure 2.3. A.  6  8  12  10  e l u e n t  14  16  (ml)  Fraction eluted from size exclusion chromatography  B.  10  MW  11  12  13  . Fab'-PEG-DSPE conjugates -  Fab'  50kD  IIWW  'W ". r  :  . ...  F i g u r e 2.3 S i z e e x c l u s i o n p u r i f i c a t i o n o f F a b ' - P E G - D S P E  A ) Chromatogram from fractions eluting from a size exclusion column. F a b ' - P E G - D S P E conjugates and unreacted Fab' were detected by their relative absorbance at 280 nm. B ) SDS P A G E analysis o f fractions detected by S D S - P A G E .  28 Purification to homogeneity of Fab'-PEG-DSPE was also attempted using ion exchange chromatography.  This strategy was unsuccessful since the Fab'-PEG-DSPE  conjugates eluted before the salt gradient was applied (Figure 2.4). Attempts to adsorb the products to the column using different column matrices or eluting buffers were made without success.  Fab'-PEG-DSPE conjugates Fractions #1-5  Unreacted Fab' Fractions #16-18  Fractions eluted from IEX column i  B.  MW  P e  ^r Fabn  P  i  2  3  16  17  18  Figure 2.4 Ion-exchange chromatography of F a b ' - P E G - D S P E conjugates  A) Ion exchange chromatogram produced from the purification of Fab'-PEG-DSPE conjugates over an S X Hitrap column with absorbance at 214nm detected by an A K T A FPLC. The diagonal line indicates the start of a salt gradient used to elute the compounds from the column. B) Fraction analysis by SDS-PAGE stained with coomassie blue.  29  Discussion  2.4  Surface coupling, the classic method of introducing a targeting ligand into an L N , does not allow for characterization of the conjugated product before it is introduced into the L N formulation. This is because the maleimide-functionalized lipid is incorporated into the L N before the targeting ligand is subsequently conjugated to the L N surface (see Figure 2.5A).  Fab'-PEG-DSPE micelles  Figure 2.5 Two methods of incorporating targeting ligands for ILN manufacture  A) Surface coupling of antibody (or antibody fragments) to Mal-PEG-DSPE within the L N . The conjugate is formed at the surface of an L N . B) The post-insertion method (Ishida et al., 1999) where the conjugate is formed as a micelle and subsequently incorporated into the L N by co-incubation with pre-formed vesicles.  30 The post-insertion method, first introduced in 1999 (Ishida et al., 1999), provides the opportunity to analyze the conjugation product. In this method, the targeting ligand is first coupled to the functionalized lipid before it is transferred into an L N (see Figure 2.5B) and has many advantages as indicated in the Introduction. The observation that there is more than one site for conjugation on the Fab' molecule is not surprising, since other cysteine residues within the variable region of the Fab' fragment are known to exist. This illustrates the inherent difficulty in obtaining a welldefined I L N . This may potentially be overcome by the use of engineered antibody fragments that contain just one cysteine for conjugation. With this type of ligand, Nellis and colleagues have shown that performing the reaction at a relatively low pH (5.7) and limiting the time of the reaction helps to control non-specific conjugation to sites other than the engineered cysteine (Nellis et al., 2005a; Nellis et al., 2005b). In our experience the experimental variables could not be manipulated to control non-specific conjugation to Fab' fragments, indicating the absolute need for a molecule that has a specific site for conjugation.  The difficulty encountered here in using ion exchange chromatography to  resolve the Fab'-PEG-DSPE conjugates  has also been encountered by Nellis and  colleagues who found that by using a similar purification strategy, organic solvents, such as 28% aqueous isopropanol, could enable sufficient retention of SCFV-PEG2000-DSPE conjugates to an ion exchange chromatography column, which was not possible with aqueous eluting buffers. The poor conjugation yield and lack of a well-defined product reduce  the  attractiveness of the Fab'-PEG-DSPE targeting ligand. As an alternate approach to the maleimide functionality, some researchers prefer to use an amine reactive PEG-lipid which  31 may prove useful given that the resulting urethane bond is of high yield and chemical stability (Torchilin et al., 2001). However, the problem of non-specific conjugation sites remains. Other functionalities such as a hydrazine-functionalized PEG-DSPEs that are specific for carbohydrates on the Fc terminus of the antibody have been employed for conjugation as an alternative to maleimide-PEG-lipids.  Koning and colleagues found that when an  antibody was surface coupled to LNs via the hydrazine-PEG-lipid the resulting ILNs promote faster clearance than ILNs with the antibody randomly oriented on the surface by maleimide coupling (Koning et al., 2001). Thus, the true in vivo effect of rational I L N design is difficult to predict. In conclusion, the lack of chemoselectivity (i.e. having a defined reaction site) and the very poor overall yield presents difficulties for using proteins such as Fab' fragments to target clinically useful ILNs. A chemoselective ligand that has one site available for conjugation, such as the scFv or small synthetic ligands, seems essential, as does a reasonably efficient method of coupling the targeting ligand to a lipid anchor.  For the  purpose of this thesis, these findings have led to the abandonment of approaches relying on proteins as targeting ligands.  32  3 RGD-ligand Synthesis and Binding to the Isolated  a p3 v  Integrin  1  3.1  Introduction  3.1.1 Targeting Angiogenesis  Angiogenesis refers to the growth of new vasculature from pre-existing blood vessels.  Events such as tissue growth, repair and disease (e.g., tumor pathogenesis)  produce cytokines and other factors that disrupt the relatively quiescent state of vascular endothelial cells (EC) and cause them to transition to the angiogenic growth state (Boudreau et al., 1997). During this transition, the cell alters the array of integrins that are expressed on the cell surface (Stupack and Cheresh, 2002). During angiogenesis, the integrins function to sense changes in the extra cellular matrix (ECM) as the basal E C M is modified to form what is known as the provisional E C M . The proteins in the provisional E C M are ligands for the integrins that become expressed during angiogenesis (Brooks, 1996).  Figure 3.1 The angiogenic transition  A) Quiescent endothelial cells are supported by a basal E C M B) Angiogenic endothelial cells express different integrins that have ligands in the "provisional" E C M . The relative amounts of these integrins change between the different growth states. 1  A version of the chapter has been submitted for publication  33  The  a (33 v  integrins are one of the most interesting of the angiogenic integrins since  they are not expressed on quiescent endothelial cells (Brooks et al., 1994a). Furthermore, unlike other E C integrins, the ct 6 integrins do not bind to ligands in the basal E C M (Xu et v  3  al., 2001). The contact between ECs and proteins of the provisional E C M functions to mediate E C growth, migration and survival during angiogenesis (Brooks, 1996). Some of the proteins in the basal E C M contain ligands for the angiogenic integrins that become accessible upon proteolytic cleavage. A well-known example of a proteolytic modulator of the E C M is matrix metalloproteinase 2, which is known to be essential for angiogenic growth and migration of ECs in the case of tumor pathology (Giannelli et al., 1997). Two main pathways are known to stimulate angiogenesis, both of which are activated by different growth factors.  Vascular endothelial growth factor (VEGF) activates one  pathway and basic fibroblast growth factor (bFGF) activates the other pathway, which specifically promotes the expression of a 6 (Friedlander et al., 1995). bFGF is the most v  3  powerful mitogen used to stimulate blood vessel growth and has considerable utility in angiogenesis models (Doi et al., 2007; Presta et al., 2005). Cancer, tissue regeneration, and ocular degeneration are some of the compelling reasons why researchers have interest in modulating bFGF/a 6 v  3  signaling, potentially enabling control over neovascularization.  Thus, the interaction of the a 6 integrin with its natural ligands is an area of considerable v  3  research interest. During angiogenesis, ECs also secrete a 6 ligands known as matricellular proteins v  3  as part of the provisional E C M (Murphy-Ullrich, 2001).  The matricellular proteins  function to tune the E C response in a rapidly changing environment. Immobilized ligands,  34 such as those that are anchored into the E C M , promote adhesion whereas soluble or mobilized ligands can antagonize the 0^63 integrins and promote apoptosis (Brassard et al., 1999). This level of control is thought to attenuate the pro-angiogenic momentum once the desired vessel growth has occurred.  Secreted matricellular proteins are of particular  interest in anti-cancer strategies due to their ability to promote apoptosis of ECs (MurphyUllrich, 2001). Many  a 6 v  3  ligands within the provisional E C M interact with the integrin via an R G D  sequence, which is usually present on a flexible loop of the protein (see Ruoslahti, 2003 for a personal account of this finding). RGD-containing peptides that mimic the region of the ligand that interacts with the integrin can promote apoptosis by  a 6 v  3  antagonism (Buckley  et al., 1999). This finding has fueled the pursuit of the R G D sequence as an antiangiogenesis agent. Research from Kessler's group has led to the anti-angiogenesis compound, cRGDfN(Me)V, pursued now with the trade name Cilengitide®. This compound is a promising anti-angiogenesis agent with potency indicated by a low  IC50  value (0.58 nM)  (Dechantsreiter et al., 1999). Moreover, anti-cancer activity in humans has been observed and attributed to  a 6 v  3  antagonism of cells that constitute the tumor-associated vasculature  (Nabors et a l , 2007; Tucker, 2006). Despite extensive efforts devoted to finding integrin antagonists and the existence of high resolution structural data of the integrin, our understanding of the mechanisms that govern the switch from an inactive to high-affinity active state for integrins is far from complete (Puklin-Faucher et al., 2006).  35  3.1.2 The CE-FA Method for Measuring Binding to the a B Integrins v  3  Integrins are integral membrane proteins with binding sites on both sides of the membrane that participate actively in the formation of focal adhesion plaques (Petit and Thiery, 2000) and have a high range of intermolecular motion (Xiong et al., 2002). The equilibrium parameters characterizing R G D binding to these receptors remains poorly understood. In this chapter, an equilibrium assay to study the affinity of RGD-containing peptides for the  ct 63 V  integrins is developed.  The development of an equilibrium assay to study such a complex biological system requires a substantial effort. For example, well-characterized, high-purity ligands are essential to distinguish between non-specific effects of contaminating species. Likewise, efforts must be made to conserve the native conformation of the receptor so that the binding process reflects the in vivo situation. A variety of methods have been developed to characterize molecular interactions (Connors, 1987; Harding, 2001). Techniques applied to protein-ligand binding studies in the pharmaceutical and biomedical sciences include equilibrium dialysis, ultrafiltration, ultracentrifugation,  gel  filtration,  calorimetry,  microdialysis,  spectroscopy,  high  performance liquid chromatography (HPLC), surface plasmon resonance (SPR), and capillary electrophoresis (CE) (Bertucci and Domenici, 2002; Oravcova et al., 1996; Sebille et al., 1990). Due to the relatively short analysis time, low sample consumption ease of automation and high separation efficiency, C E has become a powerful technique for the characterization of protein-ligand interaction. Reproducible estimates of binding parameters for membrane proteins can be obtained with SPR techniques using an instrument such as Biacore®. One drawback to  36 SPR methods is that chemical modification of either protein or ligand is required for immobilization to a solid surface. The tethering reaction, usually achieved by amide or thiol conjugation, is non-chemoselective and the tethered species can exist as a number of different conformers.  In an effort to obtain further insight into the binding of R G D -  peptides to the integrin, we used solubilized ct 6 integrin and measured R G D binding v  3  using CE-FA. The following work presents a truly diffusion-controlled process for measuring the equilibrium binding of a therapeutically relevant membrane protein, ct B v  3  integrin, with a fluorescently labeled ligand, cyclic arginine-glycine-aspartic-acid-dphenylalanine-lysine (cRGDfK-488) or the control peptide, cyclic arginine-alanineaspartic-acid-d-phenylalanine-lysine (cRADfK-488). The RAD-containing peptide was chosen as a control, given that substitution of alanine for glycine has been shown to abolish the activity of RGD-peptides towards the a (3 integrin (Pfaff et al., 1994). v  3  Currently, several C E methods are available to study the equilibrium and kinetic binding properties (Busch et al., 1997c; Galbusera and Chen, 2003; Heegaard et al., 2002; Petrov et al., 2005; Rundlett and Armstrong, 1997; Rundlett and Armstrong, 2001b). These include normal affinity C E (ACE), Hummel-Dreyer method (HD), vacancy affinity C E (VACE), vacancy peak method (VP) and frontal analysis (CE-FA) (Busch et al., 1997a; Busch et al., 1997b; Kraak et al., 1992). The most suitable method can be chosen based on the  characteristics  of  the  binding  interaction,  namely,  the  speed  of  the  association/dissociation processes, the mobilities of the free species and the complex, each species' ability to absorb U V light or emit laser induced fluorescence, and the availability of the interacting species. In A C E and V A C E methods, measurable differences in electrophoretic mobilities between free ligand and the complex are required, and the  37 measurements are based on the changes in electrophoretic mobility of free protein or ligand due to complex formation (Busch et al., 1997b; Busch et al., 1997c; Rundlett and Armstrong, 2001a; Tanaka and Terabe, 2002). In H D and V P methods, the mobilities of free protein and the complex have to be similar, and the measurements are mainly based on the changes of the free ligand concentration. The method chosen for use in this study is C E - F A . In CE-FA, a relatively large amount (~ 100 nl) of pre-equilibrated mixture of protein and ligand is injected into the capillary filled with background electrolyte (BGE). The injection of a large volume of sample plug leads to the appearance of plateaus in the observed electropherogram. Partial separation of the binding species can be achieved based on the differences in their mobilities. The compound with a unique mobility in the capillary is separated from the mixture plateau to form another plateau. In the study of protein and ligand interaction, the mobility of the free ligand is usually different from the protein and the protein-ligand complex. The height of the plateau is directly related to the concentration of corresponding species therefore, the binding parameters can be determined by comparing the heights of the plateaus. A schematic of the CE-FA method has been provided in Figure 3.2.  38  electrophoretic mobility  Fluorescence Detection  Free lntegrin:cRGDfK-488 cRGDfK-488 complex  F r e e Integrin  Figure 3.2 Schematic of the Capillary Electrophoresis Frontal A n a l y s i s Method  An equilibrated mixture of free cRGDfK-488, cRGDfK-488 bound to a |33 integrin in a complex and free ctvfc integrin are injected into a capillary containing 20 m M Tris-HCl, pH 7.5, 2 m M MgCh and 0.2% Triton X-100. The three components are resolved by differences in their electrophoretic mobility. The amount of the free cRGDfK-488 or the control peptide, cRADfK-488, in complex with the a (3 integrin is detected by fluorescence. v  v  3  C E may also provide a unique advantage for the study of membrane proteins, which are inherently difficult to solubilize without disturbing the proteins tertiary structure. It is important for the field of drug discovery that these handling problems are overcome, since the most accessible targets for intravenous drug delivery reside at the surface of a cell. Different data analysis methods have been utilized for the estimation of binding parameters in C E (Tanaka and Terabe, 2002).  With the papers published to date, a  complex equilibrium model based on Scatchard analysis has been the most commonly used in CE-FA (Ostergaard and Heegaard, 2003). According to the assumptions used in this approach, binding parameters in cases where the binding stoichiometry is 1:1, and where non-cooperative binding occurs on multiple identical binding sites, can be obtained. However, there are many other cases relevant to biomolecules that do not fit the interaction model described (Bowser and Chen, 1998).  39  From this application of the C E - F A method, new information about the binding of RGD-containing ligands to the ct f3 integrin is provided. This behavior was characterized v  3  by a set of refined equations for determining the parameters of higher order binding. The full details of these equations are shown in Appendix 1 of this thesis.  40  3.2 Methods 3.2.1 Materials Unless specified, all reagents were obtained from Aldrich (Milwaukee, WI). Peptides were synthesized using 9-fluorenylmethyloxycarbonyl (FMOC) protected amino acids and resins obtained from E M D Biosciences (San Diego, CA) and dimethylformamide (DMF) as the main solvent. A l l solvents, of reagent grade or higher, were from Fisher Scientific (Nepean, ON, Canada). Reversed phase-high pressure liquid chromatography (RP-HPLC) was performed using gradients of aqueous acetonitrile (aqACN) containing 0.1% trifluoroacetic acid (TFA) on CI8 columns purchased from Grace Vydac (Hesperia, CA). Analytical gradients were applied using column no. 218TP5415, (dimensions: 150 mm x 4.6 mm) with a flow rate of 1 ml per min. A l l peptides were purified on a semi-preparative column: 218TP510 (250mm x 10mm) with a flow rate of 3 ml per min.  3.2.2 Peptide Synthesis, Cyclization and Labeling The linear sequences: D(t-Bu)fK(Z)R(Pbf)G (The RGD-containing peptide) or D(tBu)fK(Z)R(Pbf)A (An RAD-containing peptide) were synthesized by F M O C based solid phase peptide synthesis. Briefly, peptides were assembled on 0.5 mmoles of 2-chlorotrityl resin using 1.3 molar equivalents (to carboxylic acid) of the following activating agents: Obenzotriazole-N,N,N',N'-tetramethyl-uronium-hexafluoro-phosphate  (HBTU) (Advanced  Chemtech, Louisville, KT) and N-hydroxybenzotriazole (HOBT) (Advanced Chemtech,  41  Louisville, KT). Peptides were activated in situ with triethylamine (TEA) then coupled to the resin for at least one hour. F M O C removal steps were performed upon each amino acid addition by soaking the resin in a solution of 20% piperidine in D M F for five min, and repeated once. Between each step, the resin was washed with a constant flow of D M F for at least one min. Linear peptides were cleaved in 0.1% T F A in dichloromethane (DCM) with 2.5% triisopropyl silane (TIS) and H2O (v/v).  The crude peptide was analyzed by  RP- H L P C with a linear gradient of 30-90% aqACN over 30 min, (retention time = 16.3 min) and by electrospray mass spectrometry (ESI-MS). Lyophilized linear peptides were cyclized at 0.5 mM in D M F using benzotriazol-lyl-oxytripyrrolidinophosphonium hexafluorophosphate (PyBop) and HOBT to produce cR(Pbf)GD(t-Bu)fK(Z) or cR(Pbf)AD(t-Bu)fK(Z). Cyclization yields were typically over 95% and occurred within 30 min, as indicated by RP-HPLC, with a retention time of 19.5 min on the same gradient as the linear peptide and a difference of 18 in the observed M W by ESI-MS. ' H N M R NOE assignments for all a-amide bonds confirmed a head-to-tail cyclic compound. The lysine side chain was deprotected by hydrogenation with palladium over a carbon catalyst. The resulting compound (cR(Pbf)GD(t-Bu)fK or cR(Pbf)AD(t-Bu)fK) eluted at 11 min on the same RP-HPLC gradient as the fully protected peptide. Conjugation of the cyclized R G D peptide to the Alexa Fluor-488 pentafluorophenol ester (Invitrogen, Carlsbad, ON) was performed in DMF, with the pH adjusted to 7.2 with T E A . The products were purified by RP-HPLC with a gradient of 10-85% aqACN over 60 min. The desired compound eluted after 21 min for the Alexa Fluor-488 labeled peptides.  42 Deprotection of the arginine and aspartic acid side chains was performed using a solution of 95% aqueous T F A and the reaction was subsequently purified by RP-HPLC with a gradient between 0-65% aqACN. Both cRADfK-488 and cRGDfK-488 eluted at 19.5 min. The non-labeled peptide, cRGDfK (employed in Chapter 5), was precipitated into ice-cold diethyl ether, rinsed once, hydrated with distilled water and lyophilized. All peptides exhibited >95% purity, as determined by capillary electrophoresis (CE). Product identification was validated by ESI-MS as follows: cRGDfK-488 (observed M W 1120.6; calculated M W 1120.1); cRADfK-488 (observed M W 1134.2, calculated M W 1134.3); cRGDfK (observed M W 604.4, calculated M W 603.7). Peptides were lyophilized, weighed, and a standard curve for peptide quantification was established using the absorbance peak areas from chromatograms collected at 214 nm for peaks eluting between 12 and 13 min on a 0-65% A q A C N analytical gradient over 30 min for the Alexa Fluor-488 labeled peptides, 18 min for the TAMRA-labeled peptide and 8 min for the unlabelled peptide. Typically, a 0.5 m M scale synthesis would yield 450 mg of the dry linear peptide. The final yield of the unlabelled peptide was 20 dry mg (overall yield=l 1%) and the fluorescently labeled peptides was 8 mg (yield= 2%).  43  3.2.3 Capillary Electrophoresis (CE) Analysis C E experiments were carried out on a Beckman Coulter ProteomeLab PA800 (Beckman Coulter Inc., Fullerton, CA) instrument with a laser-induced fluorescence (LIF) detector (488 nm excitation and 520 nm emission). Uncoated fused-silica capillaries were used for analysis (70 cm total length, 60 cm length to detector, 50 um inner diameter, 360 um outer diameter, for CE-FA: 50 cm total length, 40 cm length to detector, 50 um inner diameter, 360 um outer diameter) (Polymicro Technologies, Phoenix, AZ). Prior to use, the capillary was rinsed with 1 M NaOH (30 min), M e O H (30 min), purified water (30 min), and finally the background electrolyte (30 min), and was conditioned overnight. At the beginning of each day, the capillary was rinsed with 1 M NaOH (10 min), followed by MeOH (10 min) and then with water (10 min). A voltage of +10 kV was then applied across the capillary and the fluorescence signals were recorded. The peptide was injected into a capillary filled with the background electrolyte (BGE) under a pressure of 1 psi for 5 s for the analytical runs or 0.5 psi for 30 to 99 s for the frontal analysis experiment and then a voltage of + 10 k V was applied when both inlet and outlet vials contained the B G E . The B G E for peptide analysis was phosphate buffered saline (PBS) and for frontal analysis, the B G E was an aqueous solution of 20 m M Tris-HCl, pH 7.5, 150 m M NaCl, 2 m M M g C l , 2  and 0.2% Triton X-100.  44  3.2.4 Capillary Electrophoresis Frontal Analysis Method Purified human integrin a 6 in 20 m M Tris-HCl, 150 m M NaCl, 2 m M M g C l , and v  3  2  0.2% Triton X-100, pH 7.5, was purchased from Chemicon International (Temecula, CA), catalog #CC1019. The initial concentration was 210 ug/ml, and for each experiment, 10 pi was employed. Purified peptides were re-dissolved in the aforementioned buffer at 36.03 m M , and then diluted to produce saturation isotherms using concentrations between 0-25 u M . The free ligand concentration ([Z]f ) was calculated by the peak height on an ree  electrophoregram produced from a sample containing the peptide alone. The amount of binding  ([IJbound)  was determined by subtracting the amount of the free ligand that  remained after reaching equilibrium with the receptor  ([PjtotaO-  The stoichiometry  constant, n, was further determined by the relationship between the amount of the bound ligand and total receptor concentration  ([Z]bound  / hP] total).  45  3.3 Results and Discussion  3.3.1 Peptide Synthesis, Labeling and Analysis The cyclic peptide containing arginine, glycine, aspartic acid, d-phenylalanine, lysine (cRGDfK) was synthesized as indicated in Methods, and was fluorescently labeled to produce cRGDfK-488 (see Figure 3.3). The negative control, cRADfK-488 was similarly synthesized using this scheme.  Figure 3.3 Synthesis of fluorescently labeled cRGDfK peptides  Solid phase peptide synthesis (SPPS) generated a linear peptide with protected lysine, aspartic acid and arginine side chains. Selective deprotection of the lysine side chain enabled site-specific conjugation to Alexa Fluor 488™ (in green), a) SPPS b) cleavage c) cyclization d) RP-HPLC e) selective deprotection f) RP-HPLC g) fluorophore conjugation h) RP-HPLC i) Removal of protecting groups j) RP-HPLC  46  3.3.2 cRGDfK-488 Peptide Purity Particular attention was paid to achieving the maximum purity of labeled peptides. The water-soluble compound, cRGDfK-488 was inherently difficult to separate from the unreacted fluorophore and additional purification steps were required to ensure good purity values before biological evaluation. Purity could not be assessed using RP-HPLC or T L C , since the fluorophore tends to co-elute with the peptide-conjugated fluorophore. In order to address this problem, C E proved to be a convenient method for analyzing the purity of the peptide-conjugated fluorochrome.  The cRGDfK-488 conjugate was  characterized by its electrophoretic mobility, which is determined by the charge to size ratio of the molecule. A n electrophoregram made by LIF detection best demonstrates the final purity of the cRGDfK-488 conjugate, since the unconjugated fluorophore is the principal contaminating species.  A n example of the high purity obtained for the  fluorescently labeled peptides is shown in Figure 3.4. It should be noted that an equivalent, but less intense, electropherogram is obtained by U V detection at 214nm.  35  AO  1  45 T i m e  SO  55  (min)  Figure 3.4 Electrophoregram demonstrating the purity of c R G D f K - 4 8 8  The above electrophoregram shows the high level of purity (i.e.>98%) that was obtained for the peptide ligands employed in this thesis. This level of purity was made possible by additional H P L C purification steps, which were found to be necessary in order to resolve the unconjugated fluorophore.  47  3.3.3 Binding to Isolated a B Method v  Integrin as Measured by the CE-FA  3  Particular attention was given to the design of a binding assay to measure the affinity of the RGD-ligands for the a f5 integrin. The assay should be non-perturbing to v  3  the overall protein structure and require relatively small amounts of the a |3 integrin, v  3  which is available in limited quantities. Use of the C E - F A method was first validated by a linear relationship that could be obtained for fluorescence signals and peptide concentrations (r >0.999). This calibration 2  curve was used to determine peptide concentrations from fluorescence signals. A  typical  sample data  point from  the  CE-FA  signal  and the resulting  electrophoregrams generated from C E - F A experiments is shown in Figure 3.5. It can be observed that as the amount of complex formed in the pre-equilibrated sample, the concentration of free peptide, [L]f, decreases, and the height of the free peptide plateau also decreases compared to the plateau obtained for the sample containing peptide only (dotdashed curve).  The electrophoregrams shown with lower and broader plateaus (solid  curve) in the Figure were obtained from the separation of the pre-equilibrated mixture of ct p3 integrin and cRGDfK-488. The C E - F A curve containing a broader plateau have v  similar plateaus on the right side, generated by the free peptide that is uncomplexed to the civfc  integrin after equilibrium is reached. The obtained electrophoregram also shows that  the complex migrates at nearly the same velocity as the free (unbound) a (33, but faster than v  the free peptide. The amount of cRGDfK-488 or cRADfK-488 that was bound to the receptor ([LJbound)  was calculated from these electrophoregrams as detailed in Methods. The  [LJbound  48 per total integrin  ([P]taO was o t  then used to generate the binding isotherms presented in  Figure 3.6.  Un  en  30 -r 25 -  i •  c  CD o CO CD  Flu  o  CD  20 -  c R G D f K - 4 8 8 alone, [L],c  [L]ix>und = [L]total - [L]lree  cRGDfK-488 + integrin mixture  = [L] o,a,(X-Y) 1  15 -  X  10 -  >  Y  5-  CO  a> CC  0 —i—  10  12  —i 14  —i  18  16  20  Time (min) Figure 3.5 Electrophoregrams p r o d u c e d for C E - F A experiments  The above electrophoregram was produced from injection of the pre-equilibrated mixture of a p integrin (0.295 pM) and peptide (2.40 pM). The difference in the height of the plateau corresponding to free cRGDfK-488 was used to calculate the amount of the peptide that was bound to the ci |3 integrin, and subsequently the equilibrium parameters v  3  v  3  The binding isotherms of cRGDfK-488 to the integrin a (3 v  3;  and the non-specific  binding of cRADfK-488 are shown in Figure 3.6A. Using the specific binding of cRGDfK488 (total binding minus non-specific, Figure 3.6B), a detailed analysis involving multi-site equilibrium binding revealed the presence of two binding sites for cRGDfK-488 on the integrin having affinities Kd,=3.18 x 10" M and Kd =2.84 x 10" M . Full details of this 6  7  2  calculation, which was performed by Dr. David Chen and Ms. Ying Sun, may be found in the Appendix.  49  0.0  5.0  10.0  15.0  20.0  25.0  0.0  5.0^6  1.0e-5  [Peptide] (MM)  1.5e-5  2.0e-5  2.5e-5  [Peptide] m  Figure 3.6 c R G D f K - 4 8 8 : a p* integrin binding isotherms generated from the C E - F A v  3  method  A . Closed circles represent total binding of cRGDfK-488. Open circles represent the nonspecific binding of cRADfK-488 to the same amount of 0^63 integrin. B. Specific binding of cRGDfK-488 obtained by subtracting non-specific from total binding. The curve represents a best-fit assuming two binding sites for the cRGDfK-488 on the integrin.  The C E - F A analysis of binding to peptides the isolated findings.  a |3 v  3  integrin resulted in three  First, the contribution of non-specific binding by cRADfK-488 was similar to  that seen for cellular a p3 (see Chapter 4) indicating that non-specific effects due to v  partitioning in the lipid bilayer are minimized for the cRGDfK-488 ligand. Second, it was found that a two-site binding model gave the best fit to the data, where the value of Kdi is slightly greater than K^2, suggesting that the second binding process is somewhat cooperative to the first binding. The two sites on the  a 63 v  molecules are saturated  gradually as ligand concentration in the sample mixture increases in free solution. It appears that a step-by-step, two-site binding process better describes the experimental data for the interaction of a (33 integrin and cRGDfK-488. v  50 The third finding revealed by C E - F A is that the binding of cRGDfK-488 to the a 6 v  3  integrin isolate occurs with a 1:2 integrin:RGD stoichiometry. This could be consistent with the proposal by Bednar and colleagues who proposed a two-step binding mechanism for RGD-peptidomimetics with the integrin that is present on activated platelets, GPIIb/IIIa (Bednar et al., 1997). Further, if the stoichiometry of RGD binding is of higher order, it may help to explain why multivalent RGD targeted imaging agents are more effective for binding to their cell target than their monovalent equivalents (Li et al., 2007b; Montet et al., 2006). This work demonstrates a new approach to deduce important binding characteristics of biomolecules employing one of the simplest C E methods, CE-FA. To understand the binding of membrane proteins with complex biological behavior, non-perturbing analytical methods such as CE-FA may prove to be increasingly useful.  51  4 Cellular Binding and Endocytosis of RGD-Containing Targeting Ligands 2  4.1 Introduction The ct B3 integrins have become important targets in the development of new v  anticancer strategies as they are expressed at high levels on the surface of many cancer cells such as gliomas (Gladson and Cheresh, 1991), melanomas (Albelda et al., 1990), ovarian carcinomas (Liapis et al., 1997) as well as tumor-associated endothelial cells (Brooks et al., 1994b). In addition, the a p 3 integrins are involved in angiogenesis (Brooks v  et al., 1994b), metastasis (Hieken et al., 1996) and resistance to radiotherapy (Albert et al., 2006; R. A . Smith and Giorgio, 2004). The a B3 integrins on tumor-associated endothelial v  cells are considered to be particularly important targets given that all tumors require a blood supply for survival (Folkman, 1990).  Figure 4.1 Binding and E n d o c y t o s i s of c R G D f K - 4 8 8 via the a P 3 integrin v  The findings in this Chapter suggest that the enhanced endocytosis of cRGDfK-488 is attributable to the presence of cellular a fo. Binding plus endocytosis is distinguished from binding alone by incubating the ligand and cells at endocytosis permissive (37°C) or non-permissive temperatures (4°C). v  A version of this chapter has been submitted for publication  52  The wide variety of natural ligands for the ctyfo integrins (e.g., vitronectin, fibronectin, osteopontin, denatured/proteolysed collagen and the foot and mouth disease virus) bind to the integrin by a conserved RGD tripeptide motif that is usually located on a flexible loop in the protein (Ruoslahti and Pierschbacher, 1987). RGD-containing peptides antagonize the a Ps integrin have anti-angiogenic activity both in vitro (Brassard et al., v  1999) and in vivo (Storgard et al., 1999), and are thought to mimic structural features at the binding site of the natural integrin ligands. Furthermore, improved drug delivery has been achieved when the targeting ligand is displayed in a multivalent form, a strategy also used by viruses and antibodies. When radionuclides are conjugated with dimeric (Chen et al., 2005) or (even better) tetrameric (Wu et al., 2005) RGD-containing peptides, localization to the site of a tumor is greatly enhanced. However, because of the complexity of their biology, the response of integrins to even monomeric R G D ligands remains poorly understood (Puklin-Faucher et al., 2006). This applies particularly to basic aspects such as endocytosis following the binding of RGD ligands to integrins, where some reports suggest that binding does not lead to endocytosis (Castel et al., 2001), whereas other reports assume that binding is accompanied by endocytosis (Balasubramanian and Kuppuswamy, 2003). In the following sections, the ability of the fluorescently labeled peptides described in Chapter 3 to bind to integrins on human umbilical vascular endothelial cells (HUVEC) and subsequently undergo endocytosis has been characterized and compared to the binding and uptake properties of a fluorescently labeled monoclonal antibody, LM609X. It is shown that the R G D ligand exhibited considerably greater uptake following incubation at endocytosis permitting temperatures (37°C) as compared to an endocytosis inhibiting  53 temperature (4°C). A 7.4-fold increase in uptake of the R G D peptide was observed following a one hour incubation with H U V E C at 37°C, as compared to 4°C. In contrast, only a 1.9 fold increase in cell-associated fluorescence was observed on incubation with LM609X at 37°C as compared to 4°C. Fluorescence microscopy studies provided further evidence supporting rapid endocytosis of the R G D peptide at 37°C as compared to LM609X. These results are discussed with regard to previous work indicating that R G D ligands enter cells by integrin-independent pathways. In addition, it is suggested that this ability of R G D ligands to stimulate endocytosis may be of considerable utility for achieving enhanced intracellular delivery of ligand-associated drugs in anti-angiogenic applications.  54  4.2 Methods  4.2.1 Materials Unless otherwise specified, all reagents were obtained from Aldrich (Milwaukee, WI). Peptides were synthesized and characterized as described in Section 3.2. A l l solvents were obtained from Fisher Scientific (Nepean, O N , Canada) and were of reagent grade or higher.  4.2.2 Cell Culture H U V E C seed cells, media and supplements were obtained from Cascade Biologies (Portland, OR). M21 and M21L melanoma cell lines (with and without ctyfo expression, respectively were obtained from Dr. D. Cheresh) and were maintained in a 5% C O 2 atmosphere Dubelco's Modified Eagle's Medium supplemented with 10% fetal bovine serum (FBS). Cells were grown to approximately 1.0 x 10 cells per 75 cm tissue culture 6  2  flask for H U V E C and 1.8 x 10 cells/flask for melanoma cell lines. At the time of harvest, 6  media was removed and cells were rinsed twice with 10 ml phosphate buffered saline (PBS) with C a  2+  and M g  2 +  (GIBCO product no. 14040). Adhered cells ( H U V E C and M21)  were harvested manually with a cell scraper to preserve the integrity of the integrins. Once unadhered, cell suspensions were washed, then centrifuged at 1100 rpm for 5 min and resuspended in 5% FBS in PBS (herein referred to as PBS/FBS) to a final volume that yielded 500,000 cells/ml.  55  4.2.3 Cell Binding Harvested cells were aliquoted in 200 pi volumes containing approximately 100,000 cells. Alexa Fluor-labeled peptides or antibodies, dissolved in PBS/FBS before addition to cells. The ligands were then added to cells and the final volume was adjusted to 500 u,l with PBS/FBS. Unbound ligands were removed by rinsing with 5 ml of FACS rinsing buffer (130 mM NaCl, 3 mM KC1, 10.4 m M Na P0 , 1.7 m M K H P 0 pH 7.4) followed by 2  4  2  4  centrifugation at 1100 rpm for 5 min and repeated twice. Unfixed cells were kept on ice and immediately analyzed by flow cytometry on a B D LSR II flow cytometer equipped with an air-cooled argon-ion laser. The instrument was calibrated weekly for fluorescence and light scattering, with 2 ^ M Calbrite beads. Light scattering and fluorescence channels were set to a logarithmic scale and 10,000 events were collected per sample. Data was analyzed with Flow Joe (Version 4.5.9, Stanford, CA). Electronic gates were established on unstained live cells based on their ability to exclude the uptake of propidium iodide. The mean fluorescent intensity (MFI) was plotted versus the concentration of peptide and fit to non-linear regression curves using Sigma Plot software from Jandel Scientific (Version 10.1, San Rafael, CA).  4.2.4 Cellular Uptake Harvested cells were aliquoted in 200 pi volumes containing approximately 100,000 cells. A n antibody saturation curve (Figure 4.2B) for receptor quantitation was made by adding 0-14 n M of the Alexa Fluor 488-labelled antibody that is specific for the ct P3 v  56 integrin, LM690X, (Chemicon International, Temecula, C A , USA) to the cells. The Alexa Fluor-labeled peptides, cPvADfK-488 and cRGDfK-488 were quantitated by a standard curve produced from the fluorescence emission at 520 nm following excitation at 495 nm. Once accurately prepared, solutions of the peptides were added to cells over a final concentration range of 0-5 p M . The final volume was adjusted to 500 pi with PBS/FBS. For kinetic binding experiments (Figures 4.4), the peptides were added at a final concentration of 13.5 p M . Each data point was prepared in triplicate. Ligands and cells were incubated at either 37°C or 4°C for one hour for equilibrium binding experiments and at either 0, 1,5, 15, 30, 60 or 90 min incubations for kinetic binding experiments. Unbound ligands were removed by rinsing with 5 ml of FACS rinsing buffer (130 m M NaCl, 3 m M KC1, 10.4 m M N a P 0 , 1.7 m M K H P 0 ) followed by centrifugation at 1100 rpm for 5 2  4  2  4  min, and repeated twice. Unfixed cells were kept on ice and immediately analyzed by flow cytometry on a B D L S R I I flow cytometer as described in Section 3.2.  4.2.5 Receptor Quantitation The amount of cell-bound antibody (and thus the number of a (33 integrins) was v  determined using a Quantum Simply Cellular kit for antigen quantitation (Bang Laboratories, Cat. No.815A, Fishers, IN). The kit contains microspheres with a series of known amounts of IgG antibody binding sites that were used to produce the X-axis of a standard curve. The Y-axis is a measurement of the signal that the labeled antibody, LM609X, emits when bound to the microspheres (see Figure 4.2A). LM609X was added to each series of microspheres in triplicate, and incubated at 4°C for one hour. The actual  57 amount of antibody used (2 ug) was higher than the manufacturer's recommended value to ensure saturation of the binding sites on the microspheres.  After binding, the excess  antibody was removed by washing with 5 ml of F A C S rinsing buffer, centrifuging at 1500 rpm, and repeated twice. The antibody was added to HUVECs, M21 and M21L over a concentration range between 0-14 n M . The number of a (B3 receptors could then be v  calculated from the maximum level of fluorescence by using the standard curve shown in Fig. 4.2A (see Results). To track the internalization of the receptor that resulted from RGD exposure (Figure 4.5), non-labeled cRGDfK was first added to each sample to achieve a final concentration of 7.0 p M . The samples were then incubated for 0, 15, 30, 60 or 90 min, rinsed and stained using 26 n M LM609X at 4°C for one hour, in the presence of 1.5 u M cRGDfK (the unlabelled peptide).  Samples were rinsed twice and analyzed by flow cytometry, and  statistical analyses were applied using GraphPad Instat Version 3.0. A p-value generated from a Tukey-Krammer multiple comparisons test that was greater than 0.05 was considered significant.  4.2.6 Fluorescence Microscopy H U V E C s were grown to confluence on two, eight-chambered, glass slides (BD, Franklin Lakes, NJ) to confluence.  Cells were rinsed three times with PBS/FBS and  ligands were added in PBS/FBS with 50 pg of rhodamine-conjugated dextran, M W 10,000 (Molecular Probes, Eugene, OR) as follows:  13.5 u M of cRGDfK-488 or cRADfK-488  (4ug) for a total volume of 200 pi, 1 ug of LM609X in 200 pi of PBS/FBS and for  58 comparison, a treatment control, 200 pi of PBS/FBS. Each slide was incubated at either 4°C or 37°C for 15 min and 1 h. Ligands were removed by washing each slide three times with PBS/FBS then fixed with 3.5% paraformaldehyde in PBS for 15 min. Immediately before imaging, chambers were removed, and the slides were prepared using Vectashield mounting media (Vector Laboratories, Burlingame, C A ) containing the blue nuclear stain, 4',6-diamidino-2-phenylindole (DAPI).  Cells were visualized on a Zeiss Axiovert 200  fluorescence microscope equipped with a Retiga 2000R camera. Images were captured under a bright field with a 0.3 s exposure time, and in the fluorescent field using three fluorescent filters with the following exposure times: red-3.1 s, green-8.1 s, and blue-0.005 s for each image.  Images were compiled using Openlab software (Version 5.0,  Improvision, Lexington, M A ) .  59  4.3  Results  4.3.1 HUVEC Express 2.63 x 10 Integrins Per Cell 5  a 63  Before characterizing the binding of cRGDfK to  v  integrin receptors on  HUVECs, it was important to establish that the H U V E C s employed express high levels of the  a p3 integrin receptors. v  The number of receptors was assayed by first incubating the  cells with saturating levels of the Alexa Fluor-488-conjugated antibody, LM609X, which is specific for a 6 integrin. The amount of antibody bound to these cells was quantitated v  3  with calibrated microspheres as detailed in Methods. The standard curve obtained for antibody fluorescence as a function of the antibody binding capacity of the microspheres is shown in Figure 4.2A.  ,  ,  0  2e<-5  —t  1  4e*5  6e*5  '  '  1  1  1  1  1  1  i  0  2  4  6  8  10  12  ABC/microsphere  r-  14  [LM609X] („M)  Figure 4.2 Quantitation of the a P3 receptor on cultured cells A) Standard curve for antibody-induced-fluorescence (MFI) of calibrated microspheres that have a known antibody binding capacity (ABC). MFI is a measure of the mean fluorescence intensity (arbitrary units) detected by a flow cytometer with n=3 ± SD. B) Regression curves for saturation of the 0^63 receptor on H U V E C (solid curve) M21 melanoma cells (medium dash) and M21L melanoma cells (dotted), with the antibody LM609X. The maximum number of binding sites ( B ) values were used to extrapolate v  max  60 the average number of receptors for each cell line as indicated in the text ± the standard error.  The standard curve in Figure 4.2A Could be fitted (r =0.998) by the linear relation: 2  y=11.89 + 0.002x  (1)  where y is the mean fluorescence intensity (MFI) and x is the total number of antibody binding sites available per microsphere. A plot of the cell-associated fluorescence as a function of antibody concentration for HUVECs, M21 and M21L cell lines is shown in Figure 4.2B. A good fit to this data (r >0.978) could be achieved using relation (2): 2  y = B *x/(K +x) max  d  (2)  Where y is the MFI per cell and x is the antibody concentration (nM), B  m a x  is the  maximum M F I per cell (reflecting saturation of all binding sites) and Kd is the antibodyintegrin dissociation constant. Using this method, H U V E C s were determined to have 2.63 ± 0.081 xlO a P 3 integrins per cell, in reasonable agreement with previous estimates of 5  v  4.22 ± 0.16 x 10 integrins per H U V E C (R. A . Smith and Giorgio, 2004). The measured 5  K value for LM609X binding to H U V E C was 14.4 nM. Similar assays for M21 and M21L d  cells revealed 56,700 ± 3500 and 1,400 ± 570 integrins per cell, respectively. H U V E C s were also saturated with cRGDfK-488 and cRADfK-488, demonstrating the total and non-specific binding respectively (Figure 4.3A). A large difference between the total and non-specific binding could be observed for H U V E C cells. Conversely, M21L  61  cells that do not express the  a 63 integrin, did not show an appreciable difference v  between  the total and non-specific binding curves (Figure 4.3B).  0  1  2  3  4  5  0  2  (Peptide) (jaM)  4  S  [Peptide] |iM  Figure 4.3 Binding of fluorescently labeled peptides to H U V E C or M21L cells  Binding isotherms for the fluorescently labeled R G D and R A D peptides to H U V E C or M21L cells following one hour incubations with the peptide ligands at 37°C. A) Saturation curves characterizing the binding of cRGDfK-488 (closed circles) and cRADfK-488 (open circles) to H U V E C (live cell population) as measured by mean fluorescence intensity following incubation at the peptide concentrations indicated. B) Saturation curves describing the binding of Alexa-Flour 488 labeled R G D (closed circles) and R A D (open circles) peptides to M21L cells that express low levels of the a (33 integrin. Measurements were done in triplicate with error bars representing the standard deviation. v  4.3.2 cRGDfK-488 Undergoes Binding to HUVEC at 37°C  Extensive  Endocytosis  Following  The cell-based saturation curves presented in the preceding Section reflect the total binding of cRGDfK-488 as assayed after one-hour incubations at 37 °C. The amount of bound ligand includes peptides that are bound to surface a (33 integrins and peptides that v  may be subsequently endocytosed. In order to investigate the influence of endocytosis on binding, the association of peptides with H U V E C following one hour incubations at 4°C (a  62 temperature which inhibits most cellular processes, including endocytosis) or 37°C (endocytosis permitting) was investigated. As shown in Figure 4.4A, a marked difference in the specific (i.e. total minus non-specific) binding of cRGDfK-488 at 4°C and 37°C was observed, corresponding to a 7.6 fold increase in the measured B  m a x  . In contrast, binding  of the antibody, LM609X, is only moderately elevated (1.9 fold) when endocytosis is permitted, as shown in Figure 4.4B. Similarly, M21L cells that express low levels of the a (33 integrin, exhibit a modest 2.8 fold increase in binding at the endocytosis permitting v  temperature (Figure 4.4C).  .  3  J  —  ,  ,  0  1  ,  2  *  6  [PeptideltiM  Figure 4.4 Specific Binding and E n d o c y t o s i s of c R G D f K - 4 8 8 and LM609 to H U V E C and M21L cells.  Saturation curves describing the binding of RGD peptide and mAb to a p 3 integrins on H U V E C following a one hour incubation at endocytosis enabling (37°C, closed circles) and endocytosis inhibiting (4°C, open circles) temperatures. A) Specific binding (total binding minus non-specific binding, as assayed employing the R A D peptide) of cRGDfK488 to H U V E C . B) Specific binding of the antibody, LM609X, to H U V E C C) Specific binding of cRGDfK-488 to M21L cells, which express a relatively low amount of the a f53 receptor. Error bars were calculated from the square root of the sum of squares of the error from non-specific binding and total binding. v  v  63 Information about the binding constants of cRGDfK-488 may be derived most appropriately from the 4°C curve shown in Figure 4.4A, given that this condition minimizes the effect of endocytosis of the ligand. Fitting the data to Equation (2) resulted in a K of 0.20 u M ± 0.04 (r >0.998). 2  d  The explicit involvement of the R G D sequence in H U V E C binding is further supported by the ability of unlabelled cRGDfK to compete against  cRGDfK-488 for  binding to H U V E C (Figure 4.5).  110  60 J  ,  ,  1  1  1  0  20  40  60  80  nM competitor (free cRGDfK)  Figure 4.5 Competition of c R G D f K - 4 8 8 : H U V E C binding with free c R G D f K  13.5 u M cRGDfK-488 was added to H U V E C at 37°C in the presence of increasing concentrations of the unlabelled peptide competitor, cRGDfK, incubated for one hour. The y-axis represents amount of cRGDfK-488 bound to H U V E C in the presence of the competitor. A measure of 100% bound is taken from the mean fluorescent intensity observed for cRGDfK-488 binding to HUVECs in the absence of any competitor.  64  4.3.3 Time Dependent Uptake of cRGDfK-488 by HUVEC In order to further characterize the increased binding at endocytosis permitting temperatures, a kinetic study of peptide binding to H U V E C s at either 4°C or 37°C was performed.  The time-dependent binding of cRGDfK-488 and the control peptide  (cRADfK-488) to H U V E C over 90 min is illustrated in Figure 4.6. It was found that a reasonable fit to the uptake curves shown in Figure 4.6 could be achieved using the first order equation:  y(t)=a(l-e-  kobst  )  (3)  where y is the mean fluorescent intensity of live cells detected on a flow cytometer, a is a constant and t is time, in seconds. At 4°C (Figure 4.6), where endocytosis is inhibited a good fit (r >0.969) could be achieved using Eq. 3 to describe binding of both cRGDfK-488 2  and cRADfK-488. The rate constant,  k bs, 0  associated with this binding was 0.0024 ± 4 . 1 5  x IO" [s" ] (half life, t, =288 s) for cRGDfK-488 and 0.009 ± 3.84 x 10" [s" ] (t =77 s) 4  1  4  1  /2  1/2  for cRADfK-488. Similarly, a good fit (r values >0.978) could be achieved for peptide 2  binding at 37°C for both cRGDfK-488 (k = 0.0007 ± 8.16 x 10" [s" ] ,t, =1004 s) and 5  obs  1  /2  cRADfK-488 (k = 0.0004 [s" ], (t, =1732 s). The greater uptake at 37°C associated with 1  obs  /2  slower rate constants could be consistent with endocytosis. In order to evaluate this more definitively, the fluorescence microscopy studies detailed in Section 4.3.5 were undertaken.  65  TimeCs)  Ttme{s)  Figure 4.6 Time-dependent uptake of cRGDfK-488 and c R A D f K - 4 8 8 by H U V E C  13.5 ixM cRGDfK-488 (closed circles) or cRADfK-488 (open circles) were added to H U V E C at A) endocytosis inhibiting (4°C) and B) endocytosis permitting (37°C) temperatures. The H U V E C s were incubated in the presence of 13.5 u M of labeled peptide. —k  The curves represent the best fit of the data to the equation y(t)=a(l-e  0  K  s  it  ).  4.3.4 a p 3 Expression Decreases Upon RGD Binding. v  In an attempt to show more directly that the a |33 integrin receptor is endocytosed v  following binding by the unlabelled peptide, cRGDfK,  was added at  saturating  concentrations to H U V E C cells for varying lengths of time at 37°C and the amount of integrin present after these incubations was quantitated by staining with LM609X (Figure 4.7).  Although not a significant difference (p>0.05), the amount of surface-exposed  integrin appeared to decrease by approximately 35% within 30 min, then subsequently increase, supporting the notion that peptide binding to integrin results in endocytosis and internalization of the receptor.  66  140 120 % 100 t 3 W  *.  80  °  60  c U  a  40 20  15  30  60  90  Time (min)  Figure  4.7 a 0 3 integrin surface receptor expression after e x p o s u r e to c R G D f K v  Expression of the ci |33 integrin after exposure to cRGDfK as assayed employing the LM609X mAb. H U V E C were incubated with 34 u M of (unlabelled) cRGDfK at 37°C for varying lengths of time and the level of a P 3 was then measured employing saturating amounts of LM609X at 4°C. Error bars represent the standard deviation of measurements performed in triplicate. v  v  4.3.5 Endocytosis as Visualized by Fluorescence Microscopy In order to directly visualize the ability of HUVECs to endocytose  cRGDfK-488  ligands, uptake was followed employing fluorescence microscopy. The appearance of cells following either 15 min or one hour incubations at 37°C are shown in Figure 4.8. As seen for the cells in suspension analyzed by flow-cytometry, the ligand cRGDfK-488  is  endocytosed. Co-localization of the green peptide label with the red endocytosis marker (rhodamine-conjugated dextran) commenced as early as 15 min and within one hour, the two labels co-localize within the cell.  Internalization of the LM609X antibody was  comparatively weak, consistent with the modest increase in B  m a x  with temperature observed  67  for LM609X binding to H U V E C (Figure 4.4B). The cRGDfK-488 peptide induced unique morphological changes (seen best in the bright field images) that are different that those seen for LM609X or cRADfK-488 treated cells. In particular, cells lost their adhesive properties, lifting off the surface. Few cells remained after cRGDfK-488 was removed by the required washing steps.  A) 15 minute incubation  B)l hour incubation  Figure 4.8 Binding at 37°C as s e e n by f l u o r e s c e n c e microscopy  Fluorescence microscopy of H U V E C following incubation with the LM609X mAb, cRGDfK-488 and cRADfK-488 ligands at 37°C. The Alexa Fluor 488 labeled ligands were added together with dextran rhodamine to visualize endocytosis. Fluorescence micrographs obtained A) following a 15 min incubation and B) following a 1 h incubation. The data indicates that cRGDfK-488 co-localizes with endocytosed vesicles while LM609X remains largely located at the surface of the cell. Non-specific binding by cRADfK-488 was not detected.  68  The same ligand incubations were carried out at 4°C with less overall binding detected for all peptide ligands and most cells remain adhered to the glass slide (Figure 4.9). Binding of the cRGDfK-488 ligand to cells at this temperature was barely detected. Similarly, as shown in Figure 4.4A,  binding at 4°C  is low compared to the level of could be  detected at 37°C.  A ) 15 m i n u t e i n c u b a t i o n  LM609X  PBS B)l hour incubation  L.M609X \  m :RCiDIK-48S  cRADfK-488  •  if' cRGDiK-488  m  eRAD(K-488  y  Figure 4.9 Binding at 4°C as seen by fluorescence microcopy Fluorescence images of H U V E C following incubation with the LM609X mAb, cRGDfK488 and cRADfK-488 ligands at 4°C. The Alexa Fluor 488 labeled ligands were added together with dextran rhodamine to visualize endocytosis. Fluorescence micrographs obtained A) following a 15 min incubation and B) following a 1 hour incubation.  69  4.4 Discussion The major finding of this work is that the cRGDfK-488 peptide binds to ct B v  3  integrins on H U V E C s and stimulates enhanced endocytosis when compared to the LM609X monoclonal antibody or the cRADfK-488 negative control. There are three interesting aspects of this observation. First, it is of interest to compare the ability of the RGD ligand employed here to stimulate endocytosis following binding to the a B integrin v  3  with the properties of other RGD ligands as reported by other investigators. The second area concerns the mechanism whereby the peptide ligand could stimulate greater endocytosis than other ligands such as the monoclonal antibody. The final point concerns the potential utility of the cRGDfK-488 peptide as a useful agent to target drugs in vivo. The relation between RGD peptide binding to a 6 integrins and subsequent uptake v  3  into cells has also been studied by other investigators. For melanoma cells, it has been reported that mAb binding to a B v  3  integrins results in receptor-mediated endocytosis,  whereas cRGDfK is taken up by an integrin-independent pathway (Castel et al., 2001). The conclusion that cRGDfK is taken up by an integrin independent process was indicated by the similar levels of uptake for H-cRGDfK and carboxyfluorescein-labeled cRGDfK for 3  both M21 and M21L cells at both 4°C and 37°C. These results are not in agreement with the results presented here, because little specific binding (and much reduced uptake) of the cRGDfK-488 peptide is observed for the M21L cell line that expresses low levels of receptors. Further, uptake of the cRADfK-488 control peptide, which binds nonspecifically to the H U V E C , is considerably reduced compared to the cRGDfK-488 ligand. It is possible that the high levels of non-specific uptake could have arisen from nonspecific binding effects. The M21L cells have only 40 times more receptors as compared  70 to M21L cells while H U V E C s express nearly 200 times more integrin receptors than M21L cells.  A second and related point is that the use of ligands with higher non-specific  background binding properties can lead to additional difficulties in detecting integrinspecific effects. In the study published by Castel and colleagues, the binding and uptake properties of R A D control ligands were not examined. Other studies demonstrating enhanced gene delivery of vectors targeted by R G D ligands support a specific receptor-mediated endocytotic process (Renigunta et al., 2006). Specific uptake of RGD-containing ligands is also supported by studies indicating that R G D internalization coincides with the activation of a kinase (S6K1) that is critical for cell growth. This effect is RGD-specific since the negative control (an RGE-containing peptide) is neither internalized nor active in S6K1 phosphorylation (Balasubramanian and Kuppuswamy, 2003). The next area of discussion concerns the mechanism whereby the cRGDfK-488 ligand could stimulate appreciably higher levels of internalization on binding to the a (3 v  3  integrin as compared to the integrin-specific mAb LM609X. A number of studies suggest that the interactions of RGD ligands with integrins have significantly different features than binding of mAb.  The studies presented in Chapter 3 indicate a 1:2 a 6 v  3  integrin:cRGDfK-488 peptide stoichiometry. A two-step process has been proposed for fluorescently labeled RGD-peptides that bind to the main integrin on activated platelets, GPIIb/IIIa.  In this study, a rearrangement process was proposed to follow the initial  binding event (Bednar et a l , 1997). The ratio of k ff/k 0  on  for the first step was observed to  be high compared to the second step, indicating that the subsequent rearrangement stabilizes a complex that is less prone to dissociation.  71 While it is not clear how these different binding characteristics of the RGD ligand to the a p 3 integrin, as compared to the mAb, lead to enhanced endocytosis for the R G D v  ligand, it is clear that the possibility exists for substantially different conformational and functional consequences for R G D binding as compared to mAb binding. It will be of interest to pursue the correlation between these differences and subsequent endocytotic events. The utility of RGD-containing peptides for the targeted delivery of associated therapeutics is clearly improved by an enhanced probability of endocytosis (Sapra and Allen, 2002). A related question concerns whether the specific type of R G D peptide influences the levels of endocytosis or whether all RGD ligands have the same endocytotic index. It will be of considerable interest to examine the endocytotic index of multivalent RGD targeting ligands, which demonstrate improved binding to the 0^63 integrins (Montet et al., 2006). In summary, the results presented here establish that the cRGDfK-488 stimulates enhanced endocytosis in H U V E C s as compared to the mAb LM609X, and that the uptake of the R G D ligand proceeds via an endocytotic process that involves specific interactions with the ct p3 integrin. It is anticipated that for RGD-targeted therapeutics the maximum v  therapeutic benefit will result from the use of targeting ligands that maximize intracellular delivery of associated cargo, pointing out a need to characterize targeting peptides according to their relative ability to stimulate endocytosis in target cells.  72  5 Targeted Drug Delivery  3  5.1 Introduction In order for LNs to reach cancer cells, they must extravasate from the blood stream into the tumor. The EPR effect facilitates L N retention at the site of a tumor and targeting ligands with affinity for the tumor-associated vasculature, such as cRGDfK, could help to increase the amount of the drug that is retained at the tumor site. Vascular targeting, as opposed to targeting tumor cells themselves, offers the additional advantage of being a relatively general therapy for the treatment of solid tumors that would not be expected to result in the generation of drug-resistant variants, given that all tumors require neovasculature in order to survive (Folkman, 1990). Studies on several published R G D - L N formulations have produced contradictory results regarding the effect that a targeting ligand has on the clearance of L N systems from the circulation. In some instances, RGD-LNs had the same circulation lifetimes as nontargeted LNs (Dubey et al., 2004; Xiong et al., 2005) however accelerated clearance has been reported for L N formulations with higher densities of the R G D ligand (Holig et al., 2004). A summary of recent in vivo publications describing the behavior of RGD-LNs is provided in Table 5.1. These studies differ in the synthesis and formulation of the R G D L N employed and the method of incorporating the conjugates into an L N membrane. These differences could account for the differences in their pharmacokinetic properties.  -5  A version of this chapter will be submitted for publication  73  Table 5.1 Summary of targeting lipids used to study R G D - L N s in vivo Authors  Targeting lipid  f  Method of LN incorporation  Major findings  Holig et al, 2002  Formulation  l mole % RGD-LNs are cleared more rapidly than 0.1 mole% RGD-LNs  Schifellers et al, 2003  Surface coupling  2.5 mole% RGD-LNs inhibit tumor growth no reported PK properties  Formulation  No difference in clearance between targeted and nontargeted LNs  Formulation  5 mole% RGD-LNs exhibit no difference in their clearance properties as compared to non-targeted LNs  it =*,  Xiong et al, 2005  Dubey et al, 2004  The synthetic route to forming targeted L N s can have an important impact on the ability of the resulting conjugate to be analyzed. For example, a commonly used technique for formulating a peptide into an L N is the maleimide-based, thioether conjugation method which can be done in an aqueous solvent (Schiffelers et al., 2003). RGD-based targeting lipids can be synthesized in an organic solvent, which can lead to a higher yield and a better characterized targeting ligand. Since an aqueous environment is required to maintain the L N ' s unilamellar structure, organic phase conjugation reactions must be conducted before L N formation and incorporated into the L N before the particle is formed (termed the "formulation" method) or using the post-insertion method (See Figure 5.1).  74  Figure 5.1 T e c h n i q u e s for incorporating an RGD-lipid into an LN  A) The formulation method where RGD-lipid incorporation precedes L N formation resulting in Type 1 RGD-LNs with the RGD ligand exposed at both surfaces of the bilayer B) The post-insertion method where RGD-lipid incorporation occurs after the L N formation, and the targeting ligand is located in the outer leaflet of the monolayer, resulting in Type 2 RGD-LNs.  Maleimide-based reactions typically tether the peptide by surface coupling, which occurs in an aqueous environment.  A major problem with such coupling is that the  reagents are highly susceptible to hydrolytic side reactions, and the resulting impurities cannot be subsequently removed from the L N formulation.  75 In the R G D - L N studies listed in Table 5.1 the purity of the peptide-lipid constructs used is not given. In general, the accurate modification of a peptide with a lipid, which is necessary in order to anchor the peptide in the L N bilayer, is a difficult process. This Chapter is devoted to achieving a well-characterized peptide-lipid conjugate that can be introduced into L N in a quantitative manner at the time that the L N is made. The first part of this work concerns the synthesis of an RGD-based targeting lipid that can be inserted into L N by the formulation method (Figure 5.1 A). The second part characterizes the overall properties of the targeted particle including parameters such as size, binding affinity, ability to load and retain drug as well as its ability to facilitate intracellular drug delivery.  76  5.2 Methods The following sections describe the synthetic strategies employed for the RGD-lipids shown in Figure 5.2.  q HC  J  — o - P= o  o o  I) c R G D f K - S U C C - D S P E Jj  O  O  NH  HN=t  II) R G D f K - G l y - ( S U C C - 3 4 3 ) - G l y - D S P E C  Q  3  O  O  "  NM  H  N  - . „  * A.  IM)cRGDfK-SUCC-(PEG) -DSPE  ® ^  2  O  o  IV) c R G D f K - S U C C - P E G - K ( M C A ) - P E G - D S P E 2  2  Figure 5.2 R G D - l i p i d s s y n t h e s i z e d  Compounds I-IV were synthesized for assessment as L N targeting agents. Compound IV was chosen for further studies, since the fluorescent labeling molecule enabled accurate quantitation and the presence of the P E G spacer reduced aggregation effects noted for compound 1.  77  Compounds I-III led to the development of compound IV, which was found to be the most appropriate for use in subsequent L N formulation, due to its good analytical and synthetic properties. The synthesis of this lead compound is shown in Figure 5.3.  Figure 5 . 3 Steps in the synthesis of c R G D f K - P E G - K ( M C A ) - P E G - D S P E 2  2  a) F M O C - P E G - O H , 3 h b) 20% piperidine in D M F c) FMOC-Lys(MCA)-OH, H B T U / H O B T D) 20% Piperidine in D M F e) F M O C - P E G - O H , PyBop/HOBT, 6 h f) cR(Pbf)GD(tBu)fK-succinic acid, PyBop/HOBT, 12 h. g) 0.1% T F A , 1% TIS, 3.9% H 0 , 95% D C M h) RP-HPLC i) DSPE, PyBOP/HOBT, CHC1 , 60°C j) Silica gel chromatography k) 95% TFA, 5% D C M , silica gel chromatography. 2  2  2  3  78  5.2.1 Materials and Reagents Unless otherwise specified, all chemicals and solvents were of reagent grade or higher and obtained from the same suppliers as described in Chapter 3. A l l lipids were purchased from Avanti Polar Lipids (Alabaster, A L ) . Reversed phase H P L C (RP-HPLC) was performed using gradients of aqueous acetonitrile containing 0.1% trifluoroacetic acid (TFA) (aqACN) applied to CI8 columns from Grace Vydac (Hesperia, CA). Analytical gradients were run using column no. 218TP5415 (150 mm X 4.6 mm) with a flow rate of 1 ml per min. Peptides were purified on a preparative scale column no. 218TP1022 (250mm X 22mm) with a flow rate of 10 ml per min.  5.2.2 Peptide Synthesis The protected cyclic peptide, cR(Pbf)GD(t-Bu)fK(Z), was synthesized as described in  Chapter  3.  Briefly,  the  linear  peptide  was  synthesized  by  fluorenylmethyoxycarbonyl (FMOC)-strategy solid phase peptide synthesis.  the  9-  Linear  peptides were then cyclized in D M F at a final peptide concentration of 0.5 mM, 1.3 molar equivalents  (to  peptide)  of  benzotriazol-l-yl-oxytripyrrolidinophosphonium  hexafluorophosphate (PyBop) and hydroxybenzotriazole (HOBT), and the pH was adjusted to 7.3 with T E A , with (Advanced Chemtech, Louisville, KT). The benzyl group on the lysine side chain was removed by hydrogenation using a palladium/carbon catalyst.  The amine-functionalized peptide, cR(Pbf)GD(t-Bu)K (100  m M in DMF), was modified with 10 molar equivalents of succinic anhydride to produce cR(Pbf)GD(t-Bu)fK(succ). The reaction mixture was purified with an RP-HPLC gradient  79 of 10-60% aqACN over one h. A 95% pure product eluting after 40 min was characterized both by ESI-MS (MW=1011.47 ±1, predicted=1012.18) and by analytical RP-HPLC using a 30-90% aqACN gradient, a product of >95% purity was obtained, with a retention time of 12 min.  5.2.3 cRGDfK-SUCC-DSPE (Compound I) Synthesis cRGDfK-SUCC-DSPE (i.e. Compound I in Figure 5.2) was the simplest design for an RGD-lipid of the four synthesized. The compound was first made by modifying the lipid, DSPE, with succinic anhydride (Avanti polar lipids Alabaster, A L ) , as follows: first, DSPE and 10 molar equivalents of succinic anhydride were co-dissolved in  CHCI3.  The  reaction was complete after one hour, and was monitored by disappearance of the free amine from DSPE (CHCl /MeOH (7.5/2.5) R f 0.6). The compound was precipitated into 3  H2O and lyophilized. The dry carboxylic acid-functionalized lipid was activated with 1.3 molar equivalents of PyBOP/HOBT for 1 min. Two molar equivalents of the peptide with a free amine, cR(Pbf)GD(t-bu)fK, was reacted for 5 h at room temperature. A UV-positive spot with an R f value of 0.84 T L C corresponded to product formation.  The reaction  mixture was purified over silica gel and concentrated. The side chain protecting groups from the resulting cR(Pbf)GD(t-bu)fK-succ-DSPE conjugate were removed by methods similar to those used in Chapter 3 to yield the deprotected peptide-lipid conjugate, cRGDfK-succ-DSPE. The final product was precipitated into H2O and lyophilized (yield 14%).  80  5.2.4 cRGDfK-(SUCC-343) -DSPE (Compound II) Synthesis 3  The synthesis of cRGDfK-(SUCC-343) -DSPE (i.e. compound II in Figure 5.2) was 3  attempted in order to make a relatively inexpensive and analytically defined spacer molecule, which functions to distance the RGD-peptide from the lipid head group, could be used to improve the analytical properties of compound I. The strategy for the SUCC-343 spacer is similar to the solid phase synthesis (SPPS) procedure used to obtain defined length polyamides (Kochendoerfer et al., 2003). For the SUCC-343 spacer, 1 mmol W A N G (NovaBiochem, San Diego, C A ) resin was loaded with FMOC-Gly-OH following activation with 5 molar equivalents of N , N ' diisopropylcarbodiimide, and 0.1 molar equivalents of dimethylaminopyrimidine (DMAP) was added to this solution and coupled to the resin for 2 h. F M O C removal and amine detection was conducted as described previously. Sixteen molar equivalents (ME) to amine of succinic anhydride, T E A and HOBT were dissolved in D M F and coupled to the resin under N and allowed to react for 30 min. Sixteen M E of carbodiimadazole (CDI) (Sigma2  Aldrich, St. Louis, MO) HOBT and T E A were used to activate the resin for 30 mins. Following activation, 16 M E of 4,9 diaoxa-1, 12-dodecane, diamine (referred to as 343) was coupled for 1 hr. The succinic anhydride/343 couplings were repeated twice to yield (343-SUCC) -Gly-OH. 3  A final FMOC-Gly coupling was performed using 8 M E of  FMOC-Gly-OH, HOBT, H B T U and T E A and coupled to the resin for 1 h. The W A N G resin was cleaved 3x in 10 ml of 25% T F A in D C M for 5 min each. After each cleavage, the resin was rinsed with 10 ml MeOH and 10 ml D C M . The final 30 ml of each cleavage fraction was collected in a small round bottom flask and concentrated to approximately 0.5 ml. The crude cleavage mixture was precipitated into ice-cold diethyl ether. Fifteen ml of  81 diethyl ether ( E M D Biosciences, San Diego, C A ) was mixed into the 0.5 ml sample, forming a precipitate. The resulting precipitate was characterized by H P L C and T L C . Subsequent purification by silica gel chromatography, using 100 g of silica powder (SiliCycle, Quebec City, Quebec) was found to be necessary to produce one homogenous UV-positive band in the pooled fraction as detected by T L C (CHCl /MeOH 7.5/2.5 3  Rf=0.74) and one main peak as by H P L C (gradient 10-85% AqACN/30 min, retention time=13.4 min). The product was hydrated with distilled water and lyophilized to yield solid FMOC-Gly-(343-succ)3-Gly-OH an N-terminally protected spacer molecule. The side chain protected peptide was coupled to the spacer by activating the spacer's Cterminal glycine with 1.3 molar equivalents of PyBOP/HOBT and coupling the aminecontaining peptide for about 5 h in D M F . The resulting construct was purified by preparative H P L C using a gradient of 10-65% AqACN/60 min eluting after 20 min. The product produced one peak that eluted after 18.8 min. The F M O C was removed and the purification was repeated, producing a compound with an analytical retention time of 15.5 min. Coupling to SUCC-DSPE was performed as above, and detected by iodine staining. After the final deprotection, (TLC:CHCl /MeOH (9/1) R f 0.65) the overall yield was less 3  than 1%.  5.2.5 cR(Pbf)GD(t-Bu)fK-PEG -COOH and cR(Pbf)GD(t-Bu)fK-PEG K(MCA)-PEG -COOH Synthesis (preceding compounds III and IV, respectively) 2  2  2  It was rationalized that the use of a short PEG-based spacer might likewise improve upon the poor analytical properties of compound I, and the poor synthesis yield observed  82  for compound II. The peptide-spacer constructs preceding the formation of compounds III and IV were made using a commercially available PEG and a solid phase peptide-spacer coupling strategy (see Figure 5.3). Compound IV evolved from compound III as it became apparent  that a fluorescently labeled molecule was required for achieving well  characterized R G D - L N systems. The spacer-peptide conjugates were assembled in a similar manner for compounds III and IV, as follows: 0 . 1 1 2 mmole 2-chlorotrityl resin was loaded with 2.5 molar equivalents of 0-(N-Fmoc-3-aminopropyl)-0'-(N-diglycoyl-3-aminopropyl)-diethyleneglycol (FmocN H - P E G 2 - C O O H ) over the course of 3 h, under nitrogen gas with gentle stirring. F M O C was removed by soaking the resin in a solution of 2 0 % piperidine in DMF for 5 min, and repeated once. Ninhydrin assays were used to monitor the success of SPPS reactions. For compound IV, which contains a fluorescent label, 2 . 5 molar equivalents of Fmoc-Lys(MCA)-OH were coupled to the free amine using 1.2 molar equivalents to carboxylic acid of HOBT, H B T U and T E A for one h then subsequently, 5 molar equivalents of the second F m o c - N H - P E G 2 - C O O H unit was incorporated over 3 h using 1.2 molar equivalents to carboxylic acid of: PyBOP, HOBT and TEA.  In order to monitor the success of each  coupling reaction, trial cleavages (performed in a solution of 0 . 1 % T F A in D C M , 2 . 5 % H 2 O , 2 . 5 % TIS (v/v)) proved to be more reliable than ninhydrin assays from this point on, presumably due to the shielding effect of PEG.  The product (after step d of Figure 5.3)  could be detected with a retention time of 21 min on an analytical RP-HPLC gradient from 0-85%o  aqACN.  Successful F M O C removal resulted in a shift in retention to  using the same gradient.  14.5  min  1.2 molar equivalents (to amine) of the cyclic peptide,  cR(Pbf)GD(t-Bu)fK(succ) was coupled overnight to the spacers from compound III and IV  83 under nitrogen gas using 1.2 molar equivalents of PyBop and HOBT and T E A to acid. The products  cR(Pbf)GD(t-Bu)fK-PEG -COOH  and  2  cR(Pbf)GD(t-Bu)fK-PEG -K(MCA)2  P E G - C O O H was purified directly by preparative RP-HPLC, (0-85% aqACN) eluting after 2  36 min (compound III) or 40 min (compound IV). The dry yield of the peptide-spacer construct was 94 mg for compound III and 80 mg for  compound  IV. Both  products  were  clearly resolved by  T L C (chloroform  (CHCl )/methanol (MeOH), 8.8/2.2), Rf=0.18, and were purified to >95% purity as 3  indicated by RP-HPLC (0-85% aqACN).  5.2.6 Synthesis of cR(Pbf)GD(t-Bu)fK-PEG -DSPE (Compound III) and cRGDfK-PEG -K(MCA)-PEG -DSPE (Compound IV) 2  2  2  The peptide spacer-constructs were conjugated with the lipid anchor, DSPE, in a similar manner to form compound III and compound IV. The purified and lyophilized peptide-PEG constructs were dissolved in 2 ml D C M , and activated with 5 molar equivalents of N-hydroxysuccinimde and diisopropylcarbodiimide.  The active ester  formed within 30 min and 3.5 molar equivalents to acid of distearylphosphatidyl ethanolamine (DSPE) was dissolved in 1 ml CHCI3 at 60°C. The active ester was added to the lipid and 1.2 molar equivalents (to ester) of T E A were added. The reaction occurred within one h at 60°C and the resulting product was purified on lOOg of silica gel, mobile phase MeOH/ CHCI3. The product,  C  R(PbF)GD(t-Bu)fK-PEG -K(MCA)-PEG -DSPE, 2  2  eluted between 30-50% methanol in CHC1 and was identified by T L C (CHCl /MeOH 3  8.8/2.2 R f 0.74)  3  84 The compounds were concentrated and redissolved in 95% aqueous T F A for deprotection of Arg and Asp side chains. The final deprotected products formed within 2 h.  T F A was removed by evaporation and precipitated into ice-cold diethyl ether,  centrifuged, rinsed with diethyl ether, and centrifuged again.  A yield of 40.8 mg of  cRGDfK-SUCC-PEG (MCA)-PEG -DSPE (compound IV) and 22 mg of cRGDfK-SUCC2  2  PEG -DSPE (compound III) resulted. ESI-MS values for compound IV were M W 2  2415.5±1, (predicted M W 2414.89), and for compound III, M W 1751.6±1 (predicted M W 1752.16).  Purity  was  confirmed  by  T L C detected  with  iodine  and U V  (MeOH/EtOH/H O/TEA (6/3/1/1) R f 0.20 for III and 0.25 for IV). Excitation (325 nm) and 2  emission (Max 389 nm) spectra were collected for the construct between 280 and 500 nm using a Varian Cary Eclipse Fluorometer. Since compound IV had the best analytical properties, namely that it could be detected by H P L C and M S methods, and had the highest yield, it was pursued as the lead compound.  5.2.7 cRGDfK-PEG -K(MCA)-PEG -DSPE Formulation into LNs 2  2  Empty RGD-LNs with different RGD content were made by the formulation method shown in Figure 5.1 A. Variations in the RGD content were made by increasing the mole % of cRGDfK-PEG -K(MCA)-PEG -DSPE (Compound IV, herein referred to as the 2  2  RGD-lipid) in the lipid formulation from which the L N were formed. In order to detect the L N by a fluorescent signal that could be related back to the concentration of the L N particles, a lipid fluorescently labeled with rhodamine was  85 included in the formulation.  This label could be used to detect the L N concentration  whereas the M C A label is used for detection of the RGD-lipid. The series of RGD-LNs containing different amounts of RGD were formulated as follows:  0.5  mole%  RGD-LN  was  made  by  incorporating:  54  mole  %  distearylphosphatidylethanolamine (DSPC) : 45 mole % Cholesterol (Choi) : 0.5 mole% Rhodamine-labeled phosphatidylethanolamine (Rhod:PE) : and 0.5 mole % RGD-lipid (i.e., 54:45:0.5:0.5 mole:mole), 1 mole% R G D - L N (53.5:45:0.5: mole:mole), 2.5 mole% (52:45:0.5:2.5 mole:mole), and 5 mole%: (49.5: 45:0.5:5 mole:mole). Non-targeted LNs were formulated as (54.5:45:0.5:0 mole:mole).  Dry lipids were dissolved in C H C I 3 at  60°C. The solvent was removed under a stream of nitrogen gas to form a thin film, and residual solvent was removed by lyophilisation. For empty vesicles, the thin film was hydrated in 100 m M HEPES buffer containing 144 m M NaCl, pH 7.5 (HBS) warmed to 60°C.  The resulting multilamellar vesicles were extruded through two polycarbonate  filters with a 100 nm pore size ten times employing an Extruder (Northern Lipids, Vancouver, Canada) to form unilamellar LNs. The amount of surface-exposed RGD-lipid per liposome was estimated assuming that one L N contains approximately 148,641 lipids (based on the average diameter of the particle as 116 nm, and the average surface area of a lipid head group as 0.5 nm ). For this diameter, the proportion of those lipids residing in 2  the outer leaflet is approximately 57% of the total lipid composition.  86  5.2.8 Post-Insertion of cRGDfK-PEG -K(MCA)-PEG -DSPE into PreFormed LNs 2  2  LNs containing 1% R G D were also made by post-insertion of the RGD-lipid into preformed (55:44:0:0 mole:mole) LNs. The procedure, depicted in Figure 5.IB, involved incubating LNs with micelles composed of the RGD lipid at 65 °C for 2 h. Incorporation of the RGD-lipid into the L N was measured by emission at 389 nm, following excitation at 325 nm in HBS containing 1% Tween 20, and compared against a standard curve made with known amounts of the RGD-lipid. The mean diameter of the LNs was characterized by dynamic light scattering (DLS) using a Nicomp 370 particle sizer for size distribution analysis.  LNs were further  characterized by a phosphate assay using the method of Fiske and Subbarow, and by fluorescence emission at 590 nm, upon excitation at 530 nm.  5.2.9 HPTS Loading In order to track intracellular delivery of L N contents, the pH sensitive probe, 8hydroxypyrene-1, 3, 6-trisulfonic acid (HPTS) (Invitrogen, Carlsbad, C A ) was loaded into Non-targeted  L N and 0.5% R G D - L N  (i.e. HPTS-LN (55:45:0:0  mole:mole) and  H P T S : R G D - L N (54.5:45:0:0.5 mole:mole)), were formed from lipid dispersed in 10 ml of HBS containing 40 m M HPTS, extruded and dialyzed three times for 12 h against 4 L HBS. The resulting HPTS:LNs were characterized by D L S , phosphorus content and HPTS fluorescence emission at 450 nm following excitation at 395 nm. The contribution from the M C A group of the RGD-lipid was found to be negligible at this wavelength.  87  5.2.10 Doxorubicin Loading Doxorubicin was loaded into preformed non targeted L N s and 0.5 % RGD-LNs using the ammonium sulphate method of drug loading (Haran et al., 1993). Briefly, L N s were hydrated in 10 ml of 300 m M ammonium sulfate, extruded and dialyzed twice over 12 h against 150 m M NaCl. Phosphate assays were conducted to determine the amount of doxorubicin required to achieve a drug-to-lipid ratio of 0.2 (mole:mole). Doxorubicin was dissolved in 1 ml of 150 m M NaCl, and the appropriate volume was delivered to each solution of LNs. Doxorubicin and LNs were co-incubated at 55°C for 1 h to load the drug, then the free drug was removed by size exclusion chromatography, with 40 ml Sephadex G50 equilibrated in 150 m M NaCl.  Doxorubicin-encapsulated LNs (Dox:LNs) eluted  within the first 10 ml fraction and free doxorubicin eluted over the following 15 ml fraction. The resulting particles were sized by D L S , and assayed for phosphate content. The amount of encapsulated material was determined by RP-HPLC, using a standard curve made from the integrated peak area of doxorubicin's 280 nm U V absorbance.  Free  doxorubicin elutes on a 0-85% aqACN gradient at 15.8 min.  5.2.11 Doxorubicin Leakage Assay The amount of doxorubicin that was retained in the L N s over time was measured by a doxorubicin leakage assay.  Doxorubicin-loaded 0.5% R G D - L N and non-targeted L N  (i.e. Dox:RGD-LN and Dox:LN), corresponding to a concentration of 0.25 m M free doxorubicin, were incubated with 30% (v/v) FBS, total volume 1.5 ml, at 37°C over 24 h. During this time, 100 pi was removed in duplicate at 0, 30 min, 1, 2, 5 and 24 h time  88 points. The sample was applied to a spin column of 750 pi Sephadex G50, and centrifuged at 100 rpm for 1 min. One ml of 10% Triton X-100 (v/v) in 150 m M NaCl was added to the eluent in order to release the LN-entrapped doxorubicin.  Doxorubicin content was  subsequently assayed for fluorimetry, using the fluorescent emission of doxorubicin at 590 nm when compared to a standard curve made from free doxorubicin in 10% Triton X-100, 150 m M NaCl.  5.2.12 Cell Culture and LN Binding Assays H U V E C , M21, and M21L cells were maintained and harvested for binding studies as described in Chapter 4 with minor modifications. Rhodamine-labeled LNs varying in their R G D content (i.e., containing between 0-5 mole% of the targeting lipid) were added to cells to construct binding isotherms (Figure 5.4). LNs were adjusted to have equivalent phosphate content by dilution with HBS, then added to cells at concentrations between 0 and 4 m M (total lipid content). The final volume was adjusted to 500 u.1 with PBS/FBS. Cells were incubated with LNs for 1 h at either 4°C or 37°C. After incubation, unbound LNs were removed with 5 ml of FBS/PBS, followed by centrifugation at 1100 rpm for 5 min, repeated twice. Unfixed cells were kept on ice and immediately analyzed by flow cytometry using the same instrument protocol described in Section 4.2.3.  The mean  fluorescent intensity was plotted versus the concentration of L N . Statistical analyses were conducted using GraphPad Instat, Version 3.0. A Tukey-Kramer multiple comparisons test was used, with n values of 3, representing triplicate data points. P values greater than 0.05 were considered insignificant.  89  5.2.13 Cellular HPTS Uptake HPTS-loaded, 0.5% RGD-targeted and non-targeted LNs were added to cells and incubated at 37°C for 0, 5, 15, 30 and 90 min time points. The cells were rinsed and assayed for cell-associated HPTS fluorescence by flow cytometry. At neutral pH, HPTS fluoresces green (excitation at 488 nm and emission at 519 nm). Acidification of the vesicles (pH -6.5-7) was monitored by detecting blue fluorescent emission at 455 nm following excitation at 405 nm. The contribution of non-targeted HPTS-containing LNs to the observed cellular fluorescence was minimal and this value was subtracted from the fluorescence signals produced by RGD-LNs.  5.2.14 Doxorubicin Cytometry The amount  Uptake into Cells  of doxorubicin entrapped  as Determined by Flow  inside the Dox:LNs was  measured  quantitatively by H P L C peak area integration and compared to a standard curve as described for the loading of doxorubicin (Section 5.2.10). Eighty u M of doxorubicin (either encapsulated within the L N or as the free drug) was delivered to each aliquot of M21, M21L or H U V E C cells. The mixture was incubated for 90 min at 37°C then assayed for cell-associated doxorubicin fluorescence by flow cytometry.  90  5.2.15 LN Uptake into Cells as Observed by Microscopy H U V E C cells were grown to confluence on four chambered glass slides (BD, Franklin Lakes, NJ).  Cells were rinsed three times with PBS/FBS. 0.265 m M of  doxorubicin was added for a final volume of 125 pi. For HPTS uptake, 1.36 m M (total lipid concentration) was added with a final volume of 200 pi. LNs or free drug were added to the cells, the mixture was incubated at 37°C for 30 min, rinsed 2X with PBS/FBS, the cells were then fixed with 3.5% paraformaldehyde in PBS for 15 min, and rinsed once again with PBS/FBS. Immediately before imaging, chambers were removed and slides were prepared using Vectashield mounting media (Vector Laboratories, Burlingame, USA) containing the blue nuclear stain, 4',6-diamidino-2-phenylindole (DAPI).  Images were  captured using a Zeiss Axiovert 200 fluorescence microscope equipped with a Retiga 200R camera.  Those acquired under a bright field used a 0.3 s exposure time and in the  fluorescent field were acquired using three fluorescent filters with the following exposure times: red 3.5 s, green 5.5 s, DAPI 0.5 s.  91  5.3 Results and Discussion 5.3.1 Synthesis of Targeting Lipids Of the four RGD-containing targeting ligands synthesized (see Figure 5.3), compound IV had the highest yield and best analytical characteristics. The M C A fluorescent-labeling molecule was also found to be useful during purification steps, since the ability to detect the compound facilitated chromatographic fractionation.  Further,  detection by the M C A fluorophore in compound IV was found to be essential once the targeting lipid was incorporated into an L N . A summary of the properties of the synthesized compounds which lead to the selection of compound IV as the lead RGD-lipid can be found in Table 5.2, which relates their production, characterization and L N formulation properties.  Compound  Synthesis  Length of  Purity  Yield  Conclusion  spacer(A)^  I  Facile, cost effective  12.0  >85% (TLC)  ~10mg (14%)  Poor characterization Induces LN aggregation  II  Difficult  97.5  <50% (TLC)  <1 %  Difficult to analyze Low yield  III  Facile, cost effective  > 85%  27mg (22%)  >95%  40.8mg (42%)  IV  Facile  43.0  75.7  Excellent synthesis Difficult to analyze Lead candidate for LN formulation  •(•total b o n d length f r o m phosphate to e-amine o f c R G D f K * c a l c u l a t e d f r o m peptide+spacer to the l i p i d c o u p l i n g  RGD-LNs made with compound I did not allow incorporation of RGD-lipid at levels higher than one percent of the total lipid composition. This is because the resulting LNs  92 produced a non-uniform size distribution with evidence of aggregation. LNs made with less than one mole percent of the targeting lipid could be made without aggregation, however definitive measures of the content in the resulting L N could not be achieved because the RGD-lipid compound was not labeled. Analytical difficulties were also encountered for compounds II and III, namely that, upon lipid conjugation, they became increasingly difficult to identify by standard methods such as H P L C or T L C . T L C analyses of compounds I-III were based on iodine staining only. The use of a fluorescent label as shown in compound IV enabled accurate analytical characterization. Ionization by ESI-MS was ineffective for compounds I and II. The use of a PEG-based spacer used in compounds III and IV greatly improved ESI-MS ionization, enabling a defined mass to confirm the identity of both compounds. It is also important to note that the PEG-lipid conjugates used in this study differ from commonly used P E G polymers, which are an average molecular weight of P E G repeats.  The compounds  synthesized here employ homogenous spacers of a defined molecular weight. Table 5.2 also lists the approximate length of the spacer based on the bond distances between the peptide and lipid head group. This estimate does not consider the bending and rotation of the spacer in aqueous solutions, and could be better estimated using an SPR-based technique (Jeppesen et al., 2001) to describe the extension of PEG-spacers in water. Using this method, the length of typical PEG2000 spacers was found to be approximately 200A.  5.3.2 RGD-LN Characterisation The RGD-lipid chosen for further studies had good formulation characteristics. When combined with other lipids such as DSPC, cholesterol, and a fluorescently labeled-  93 PE (Rhodamine-PE), LNs exhibiting a regular Gaussian size distribution with an average diameter of approximately 100 nm could be formed using up to 5 mole % of the RGD-lipid without obvious signs of aggregation. As summarized in Table 5.3, the L N particle size increased slightly, with a moderate increase in the standard deviation, when 2.5 or 5 mole % of the RGD-lipid were included in the formulation. The apparent binding constants (expressed as K d values) shown in Table 5.3 were calculated from the specific binding at 4°C (data presented in Figure 5.4). It is important to note that the K d value observed accounts for a sum of several possible cellular processes and has been simplified to describe the binding of RGD-LNs to live H U V E C cells.  Table 5.3 Characterization of R G D - L N s H U V E C Binding I N formulation L N formulation  Number of R G D molecules/LN ¥  Size (nm)  Non-targeted  0  102.4 ± 2 7 . 3  0.5% R G D  423  102.2 ± 32.0 nm  K d (nM) *  f  n/a 2.59 ±  0.58  1% R G D  846  94.2 ± 20.7 nm  1.98  ±0.35  2.5% R G D  2113  128.9 ± 68.0 nm  0.38  ±0.07  5 % RGD  4226  117.4 ± 6 6 . 0 nm  0.59 ±  0.02  ¥ this value was estimated using the assumption that 100% of the RGD-lipid was incorporated into the L N , and that 57% of that value reside on the outer leaflet of the L N lipid bilayer * error is expressed as ± the standard error of the estimate from non linear regression t the K d value is expressed as the concentration of R G D - L N required to achieve half saturation of H U V E C cells and is derived from the data presented in Figure 5.4  94  T 0  ———r— 1  i 2  [LN](mM total lipid)  i 3  .(00 J  , 0  _ _  , 1  r  i 3  2  |LN](mM total lipid)  Figure 5.4 RGD-LN binding by HUVECs at 4°C and 3 7 ° C A) Saturation curves describing the binding of L N to H U V E C following a one hour incubation at 4°C. The LNs contain between 0-5 mole percent RGD-lipid. B) The same binding experiment conducted at 37°C.  The apparent K d values measured from the data presented in Figure 5.4 decrease as the proportion of RGD-lipid is increased to 2.5 mole%. The decrease in observed K d values as the amount of RGD per L N is increased suggests that the improved binding of the L N can be attributed to the increased R G D valency of the L N . These findings are consistent with results published by Montet and colleagues that reported improved binding for multivalent presentations of the RGD-ligand compared to monovalent (Montet et al., 2006). The minor increase in K d when the RGD-lipid is incorporated at the 5 mole% level could arise if the RGD-lipid becomes somewhat masked when incorporated at excessive levels. For example, lateral segregation of the RGD-lipid into dimers or larger aggregates may occur. It can also be observed from the data in Figure 5.4A that the B  m a x  values  observed for the 5 mole% RGD-LNs is higher than the other formulations by 1.7 fold, which could be due to some limited aggregation of the RGD-LNs at these high contents of RGD-lipid.  95  5.3.3  Methods for Producing RGD-LNs  The post-insertion and formulation methods were compared with regard to their ability to incorporate RGD-lipid into LNs (refer to Figure 5.1 for a schematic of these methods). In order to quantitate the amount of the RGD-lipid that was incorporated into the L N by either method, the fluorescent emission and excitation spectra were obtained for the RGD-lipid (Figure 5.6A). The amount of the targeting ligand within Type 1 and Type 2 RGD-LNs was subsequently measured by fluorometry conducted in the presence of 1% Tween 20 in order to disrupt the L N structure. A linear increase in the fluorescence signal corresponded to increasing amounts of RGD-LNs, therefore, a standard curve was established from known amounts of the RGD-lipid in the presence of non-targeted LNs. The amount of RGD-lipid present in L N s made by the two different techniques was calculated as shown in Figure 5.6B.  J4S  }70  395  Wavctengtn (nm)  t « RGD + LNs  LNt  Formulation Postmsmion method method  Figure 5.6 Comparison of formulation and post-insertion methods for incorporating RGD-lipids into LNs  A) Excitation (red) and emission (blue) spectra obtained for RGD-LNs with the RGD-lipid embedded in the bilayer. The spectra for non-targeted LNs are shown by dashed lines. B) Fluorescent measurements of the relative amounts of the RGD-lipid in RGD-LNs made by either the formulation method or the post-insertion method. The starting concentration of RGD-lipid (1 mole%) plus non-targeted LNs represents the maximum possible RGD-lipid incorporation that was used before the RGD-lipid was incorporated into an L N .  96 When compared to the starting material (i.e. RGD-lipid + LN) it was found that the amount of RGD-lipid incorporated into RGD-LNs made via the post-insertion method was slightly reduced (p<0.05).  No significant decrease in the amount of RGD-lipid was  detected for RGD-LNs made by the formulation method (p>0.05). Because the formulation method was more straightforward and led to effectively complete insertion it was the method of choice in subsequent experiments.  5.3.4  Drug Retention in RGD-Targeted LNs  In some cases, it has been shown that when targeting ligands are incorporated into LNs, the permeability of the L N membrane to the drug may also increase (Ishida et al., 1999; Nallamothu et al., 2006). For this reason, the release of a commonly used anticancer drug, doxorubicin, from L N containing RGD-lipid was investigated.  Figure 5.7 Doxorubicin is retained equally well in RGD-LNs and non-targeted LNs A) Emission spectrum (following excitation at 490 nm) of 1.5 u,g of doxorubicin dissolved in water (dashed line) or 10% Triton X - 1 0 0 (solid line). B) Release of doxorubicin from RGD-LNs made via the formulation method (solid line) and non-targeted LNs (dashed line) over time. Error is given as ± the standard deviation of three data points from the average doxorubicin emission (590 nm) determined in the presence of 10% Triton X-100.  97 Doxorubicin was loaded into L N and R G D - L N (containing 1 mole% RGD-lipid) as indicated in Methods and the L N were incubated ion aqueous buffer over 24 h. At specified times aliquots were withdrawn, free drug removed by the spin column procedure indicated in Methods and the amount of doxorubicin remaining in the L N , upon release with Triton X-100, was assayed (see Methods).  Little difference could be observed between the  release of doxorubicin from RGD-LNs or non-targeted LNs (Figure 5.7B) indicating that the RGD-lipid did not significantly (p>0.05) enhance the permeability of the L N membrane to doxorubicin over a 24 hour period.  5.3.5 Cellular Uptake and Processing of RGD-LNs. The internalization and subsequent acidification of an R G D - L N within the cell, was tracked with the pH-sensitive, membrane impermeable probe HPTS.  When HPTS is  acidified, the emission spectrum shifts to shorter wavelengths. HPTS was loaded into LNs as indicated in Methods and the LNs were added to HUVECs.  From the fluorescent  microscope images, green fluorescence in punctate cytoplasmic vesicles could be observed within 10 min when delivered by RGD-LNs, but not by non-targeted LNs (Figure 5.8A and B). The signal detected after a 40 min incubation (Figure 5.8D) is more diffuse, suggesting that the fluorophore escapes the endosome at some time between 10 and 40 min following association of the R G D - L N with the H U V E C .  98  A  B  C  D  E  2-  Time (min)  ViJIIH  Figure 5.8 R G D - L N s are internalized and acidified by H U V E C  Panels A - D show fluorescent microscopy images demonstrating LN-associated HPTS uptake by H U V E C , while panel E reflects data obtained by flow cytometry which shows that the L N experiences an increasingly acidic environment as reflected by a progressive blue shift in the HPTS fluorescence. LNs loaded with HPTS were co-incubated with H U V E C cells at 37°C. A) Non-targeted LNs, 10 min incubation B) 1% RGD-targeted LNs, 10 min incubation. C) and D) results from a 40 min incubation, with non targeted and RGD-LNs, respectively. E) Ratio of blue fluorescence to green fluorescence for cellassociated HPTS-containing RGD-LNs over time as measured by flow cytometry (see text). Error represents the standard deviation of a series of 3 independent experiments. The significance level compared to the one minute time point is marked as follows: ** indicating p<0.01 and *** indicating p< 0.001.  Figure 5.8E indicates the acidification as a function of incubation time using flow cytometric detection of neutral endocytotic environments indicated by green fluorescent cells (excitation at 495 nm, emission at 519 nm) or an acidic environment indicated by blue fluorescent cells (excitation at 405 nm, emission at 455 nm). These results suggest that internalization of the particle commences as early as 5 min, and a significant increase in the relative blue fluorescence is observed after 60 min. From this data it is suggested that, upon binding, the contents within the particle are processed in a manner that coincides with endosomal maturation.  99  5.3.6 In vitro Drug Delivery by RGD-LNs In section 5.3.3 it was shown that, once entrapped within an L N particle, more than 50% of doxorubicin is retained following incubation at 37°C for 24 h. In this section, it is shown that RGD-LNs can deliver doxorubicin to the cytosol of HUVECs.  Fluorescent  microscope images comparing the delivery of doxorubicin by LNs versus the free drug are presented in Figure 5.9A-D.  As expected for an in vitro cellular system with a  hydrophobic drug that has affinity for the cell nucleus, the free drug rapidly permeates the plasma membrane and stains nuclear material, thus co-localizing with the nuclear stain, DAPI (Figure 5.9A). compartments.  Conversely, 1% RGD-LNs deliver doxorubicin to cytoplasmic  Doxorubicin delivered by non-targeted LNs show minimal background  staining with uptake levels similar to the saline control.  100  B  A  4  m  * *  c  D  30 u M  • Free Doxorubicin 450  .p.*  T  RGD Liposomes  • 0.1%  2  HUVEC  M21  M21L  Figure 5.9 Uptake of doxorubicin presented in the free form, in non-targeted L N or R G D - L N by cells expressing various levels of the a B integrin v  3  A-D) Doxorubicin uptake by H U V E C cells as monitored by fluorescence microscopy following a 30 min incubation at 37°C with the same amount of doxorubicin (0.265 mM) delivered to cells either as A ) Free drug, B) 1% R G D - L N , C) Non-targeted L N or D) Saline background control. E) Doxorubicin uptake (as measured by flow cytometry) by three different a pVexpressing cell lines following addition of 0.08 m M of doxorubicin in either free, R G D - L N or non-targeted L N form to cells and incubation for one hour at 37°C prior to analysis. v  101 The ability of RGD-LNs to deliver doxorubicin to a pVexpressing cells is shown in v  the flow cytometry experiments summarized in Figure 5.9E. It was shown in Chapter 4 that H U V E C s express approximately 263,000 civfo integrins per cell, whereas M21 and M21L melanoma cells express approximately 56,700 and 1,400 integrins per cell, respectively. RGD-LNs deliver increased amounts of doxorubicin to H U V E C s compared to non-targeted L N (p<0.05), and appear to deliver modestly higher amounts to M21 cells over the non-targeted LNs, however this increase is not statistically significant (p>0.05). In contrast, M21L cells that express a relatively low amount of the receptor do not show enhanced uptake of doxorubicin-containing R G D - L N .  5.3.7 Relation to Existing RGD-LN Data In this section, the results presented in the previous sections are summarized and related to RGD-LNs previously used in vivo. The discussion focuses on the four in vivo studies on RGD-LNs summarized in Table 5.1. In the present work, it was found that a high synthetic yield and good analytical properties enabled the compound, cRGDfK-PEG2K(MCA)-PEG2-DSPE, to be quantitatively formulated into LNs at the time of manufacture and subsequently used to examine the effects of RGD-targeted LNs in vitro. Synthesis of RGD-lipid constructs without the fluorescent label suffered the effects of low overall yield, poor analytical characteristics and inaccurate formulation. Since none of the studies listed in Table 5.1 use a fluorescent label or other means of accurately tracing the incorporation of RGD-lipids into LNs, it is likely that the actual amount of R G D in the formulation is unknown. This uncertainty could explain the different pharmacokinetic effects that were  102 observed in these studies, namely that some are rapidly cleared from the body and some are not. Furthermore, since quantitative analytical data has not been provided in any of these studies, it is possible that some unreacted or impure material may have been included in their L N formulations. In such case, the mole percent of the RGD-lipid would change, which could potentially affect the clearance of the particle from the circulation. In the work published by Schifellers et al, a Cilengitide® analogue modified with a thioacetyl group for incorporation into pre-formed LNs via the surface coupling method was used.  The authors assume that all of the peptide is conjugated to a membrane-  embedded maleimide-PEG-lipid, but do not address how much RGD-PEG-lipid is actually present in the membrane. Furthermore, although the resulting thioether bond that joins the maleimide-PEG-lipid to the peptide is not stable, and reverses in aqueous environments (Dunphy and Linder, 1998), which may be an issue for production and storage of the drug i f it were to progress to the clinic. Like many other targeted L N formulations, RGD-LNs produced by both Xiong and Dubey et al. show equivalent circulation lifetimes to non-targeted LNs, which is different from the studies published by Holig et al who claim that LNs containing 0.1 mole % of the RGD-lipid are not rapidly cleared from the circulation, however, LNs with 1 mole% of the RGD-lipid are rapidly cleared. Without knowing the true mole percentage of the targeting lipid in the L N at the time of i.v. administration, it is difficult to discuss the effect that the targeted L N has in vivo. The results presented here show that incorporation of the RGD-lipids synthesized here in an L N formulation do not result in the leakage of entrapped doxorubicin. Doxorubicin is an amphiphilic drug that is well-retained in many different L N formulations  103 (Gabizon et al., 2006), therefore it is perhaps not surprising that a good release profile was achieved using targeted systems. The experiments with L N containing HPTS demonstrate that ct |33 integrinv  expressing cells rapidly internalize RGD-LNs. Using intravital microscopy, Janssen and colleagues have shown that the RGD-targeted imaging agents associate with tumor vasculature within 30 min (Janssen et al., 2003). If LNs can be engineered to bind their target rapidly and become rapidly internalized, without toxic or immunogenic effects, extended circulation lifetimes may not be as important for RGD-LNs as for non-targeted LNs. In a related vein, since the a |33 integrin target is exposed to the vasculature, it is a v  highly accessible target for drugs that are administered via the circulatory system. This differs from targets within the tumor tissue, which require extravasation from the bloodstream to enable interactions between L N and the targeted cell.  104  6  Future Work  6.1 Pharmacokinetics of RGD-LNs In the previous chapter, the in vitro properties of RGD-LNs were investigated. It is expected that these results will lead to the use of these RGD-LNs in animal tumor models. Given that the use of an RGD-targeting entity can promote L N clearance in vivo, the pharmacokinetics of these compounds will be of significant concern (Holig et al., 2004). Despite the publication of over 50 in vivo efficacy studies of RGD-targeted therapeutics (being either protein, nanoparticulate or small molecule conjugates), little is known about the P K effects of the various RGD motifs in vivo (Temming et al., 2005). Perhaps the best P K characterization of RGD-targeted therapeutics has been produced from the use of tumor imaging agents (reviewed in Haubner et al., 2003). In this application, the uptake of RGDconjugates by the liver or the rapid elimination of the compound by renal filtration will likely affect the quality of the image produced. Efforts to minimize hepatic uptake and maximize the tumor to background signal have included modification of the RGD-peptideimaging agent with a hydrophilic moiety such as a sugar (Haubner et al., 2001) or P E G chain (Chen et al., 2004c). Nonetheless, the P K behavior of RGD-PEG-imaging agents is 125  not entirely predictable. For example, it has been shown that an  I-PEG-RGD conjugate  is taken up by the liver and cleared more rapidly than the non-pegylated version (Chen et al., 2004c) yet a contradictory effect is observed for  64  C u - R G D conjugates in which  pegylation does not significantly affect their hepatic uptake but improves upon their tumor retention (Chen et al., 2004a).  105 The P K behavior of RGD-LNs is different than imaging agents mainly because of the large difference in the size of the L N particles, which limits renal clearance of the L N systems. However, large particles can also lead to enhanced accumulation by the RES due to opsonization of the particle and subsequent uptake by macrophages. Furthermore, the P K characteristics can be complicated by the fact that cells of the RES express the 0^63 integrin albeit in relatively low amounts (Singh et al., 2001). Thus, it will be important to establish a distinction between non-specific clearance effects and RGD:integrin binding to non-targeted cells. A potential in vivo experiment to address this concern might include a receptor quantitation assay such as the one presented in Chapter 4. In this case, the number of a 6 v  3  integrins on cells of the RES that affect the biodistribution of the R G D - L N  particles within the body. Flow cytometric analysis of both targeted and non-targeted cells after L N administration may prove useful to aid this understanding.  6.2 Anti-tumor Efficacy of RGD-Targeted Therapeutics A l l of the published studies describing the ability of the RGD-motif to target therapeutics to the disease sites report some degree of efficacy leading to the continued interest in the R G D motif for vascular targeting (Temming et al., 2005). The new generation RGD-containing constructs will likely employ multivalent display of the R G D sequence since it is well demonstrated that increased efficacy has been attributed to increasing the valency of the RGD-moiety per agent targeted (Li et al., 2007b; Montet et al., 2006; Wu et al., 2005).  106 It may also be expected that the treatment of tumors that are resistant to conventional chemotherapy regimes, will benefit from RGD-mediated delivery, since RGD-doxorubicin conjugates can bestow efficacy against the doxorubicin-insensitive C26 colon carcinoma cell line (Schiffelers et al., 2003). The distinction between anti-tumor efficacy due to direct destruction of tumor cells and efficacy through an anti-angiogenesis mechanism will be an important issue in future efficacy studies. Good evidence has been produced both by Pastorino et al, that shows a synergistic effect of targeting (Pastorino et al., 2006).  The finding by Janssen and  colleagues that RGD-LNs cluster in the tumor blood vessels while RAD-LNs and nontargeted LNs are extravasated and dispersed in tumor tissue further suggests that this kind of distinction can be made (Janssen et al., 2003). In Chapter 5, it was established that the RGD-LNs that exhibited the highest level of binding contained 2.5 mole% or greater of the RGD-lipid.  Whether these systems will  exhibit the best therapeutic properties remains to be determined. As indicated above, it is possible that the presence of high levels of RGD targeting ligands will result in more rapid clearance by the RES, which may effectively reduce the amount of L N that can access the tumor vasculature. Future studies must establish the optimum levels of RGD targeting ligands that are compatible with long circulation lifetimes, enhanced delivery to tumor sites and maximum association with tumor vasculature. In order to be therapeutically advantageous, such systems must clearly exhibit improved efficacy over non-targeted L N formulations that are not bound to and internalized by the target cells.  107  6.3 Choice of Drug to be Delivered The drug that is encapsulated within the RGD-targeted L N is also an important variable that will have to be optimized in future studies. As presented in Chapter 5, the incorporation of RGD-lipids in to L N formulations did not significantly change the drug release properties of doxorubicin, which is generally considered not to be a cell-cycle specific drug (Gardner, 2000). Cell-cycle specific drugs such as vinorelbine have been shown to be synergistic with anti-angiogenic agents when administered as a free drug and may be better suited in an L N formulation targeted to tumor neovasculature (Han et al., 2005). In addition, the release rates of cell-cycle specific drugs from within the L N have been optimized for one tumor model and correlated with improved efficacy (Johnston et al., 2006). Finally, a second variable to be optimized in targeted L N s is the drug-to-lipid ratio, which can dramatically affect the potency of the L N system. This may be important i f the density of integrins on target tissue is relatively low, where targeted L N that contains high levels of highly potent drug would be most logically employed. For example, vincristine is a highly potent drug as compared to doxorubicin, as the dose at which maximum efficacy is observed (with reasonable toxicity levels) in mice is approximately 20 mg/kg for doxorubicin as compared to only 2 mg/kg for vincristine. Thus vincristine is approximately ten times more potent than doxorubicin, and it has been calculated that, for L N loaded with vincristine at a drug-to-lipid ratio (wt/wt) of 1, only 5 L N would be required to kill a target cell (Johnston et al., 2006).  108 Optimization of defined RGD-LNs such as those described in this thesis will likely be achieved by determination of the most appropriate formulation of the drug in vivo as well as in vitro.  6.4 Bimodal and Multifaceted Chemotherapy Regimes RGD-LNs may be considered a bimodal chemotherapy regime given that the R G D targeting ligand is anti-angiogenic and that LNs carry a cytotoxic drug. Using bimodal strategies that interfere with different cellular processes may prove to be particularly beneficial in the treatment of drug resistant tumors (Pastorino et al., 2006). If drug-loaded RGD-LNs were engineered  to contain an imaging agent, a  chemotherapy regime could include the simultaneous visualization and treatment of a tumor. This scenario is likely, since RGD-based imaging agents have progressed rapidly towards the clinic (Beer et al., 2006). The findings that these agents are non-toxic and can deliver low-dose positron emission to the site of a tumor for example, have resulted in the production of PET images superior to those acquired by non-targeted systems. The design of L N systems that contain radio-labeled or fluorescently labeled agents is relatively straightforward and offers the possibility of directly visualizing the biodistribution of RGD-targeted L N in relation to target tissue. Overall, it is anticipated that further research and optimization of the nano-design of targeted drug delivery vehicles will lead to advanced chemotherapy regimes with improved therapeutic outcomes for cancer patients.  109  References Adler-Moore, J., 1994. AmBisome targeting to fungal infections. Bone Marrow Transplant, 14 Suppl 5: S3-7. Albelda, S.M. et al., 1990. Integrin distribution in malignant melanoma: association of the beta 3 subunit with tumor progression. Cancer Res, 50(20): 6757-64. Albert, J.M. et al., 2006. Integrin alpha v beta 3 antagonist Cilengitide enhances efficacy of radiotherapy in endothelial cell and non-small-cell lung cancer models. Int J Radiat Oncol Biol Phys, 65(5): 1536-43. Allen, T . M . and Chonn, A . , 1987. Large unilamellar liposomes with low uptake into the reticuloendothelial system. FEBS Lett, 223(1): 42-6. Allen, T . M . and Cullis, P.R., 2004. Drug delivery systems: entering the mainstream. Science, 303(5665): 1818-22. Allen, T . M . and Everest, J.M., 1983. Effect of liposome size and drug release properties on pharmacokinetics of encapsulated drug in rats. J Pharmacol Exp Ther, 226(2): 53944. Allen, T . M . and Martin, F.J., 2004. Advantages of liposomal delivery systems for anthracyclines. Semin Oncol, 31(6 Suppl 13): 5-15. Allen, T.M., Sapra, P. and Moase, E., 2002. Use of the post-insertion method for the formation of ligand-coupled liposomes. Cell Mol Biol Lett, 7(3): 889-94. Ambegia, E. et al., 2005. Stabilized plasmid-lipid particles containing PEG-diacylglycerols exhibit extended circulation lifetimes and tumor selective gene expression. Biochim Biophys Acta, 1669(2): 155-63. Arap, W. et al., 2002. Targeting the prostate for destruction through a vascular address. Proc Natl Acad Sci U S A , 99(3): 1527-31. Balasubramanian, S. and Kuppuswamy, D., 2003. RGD-containing peptides activate S6K1 through beta3 integrin in adult cardiac muscle cells. J Biol Chem, 278(43): 4221424. Bangham, A . D . , Standish, M . M . and Watkins, J.C., 1965. Diffusion of univalent ions across the lamellae of swollen phospholipids. J Mol Biol, 13(1): 238-52. Bednar, B. et al., 1997. Flow Cytometric Measurement of Kinetic and Equilibrium Binding Parameters of Arginine-Glycine-Aspartic Acid Ligands in Binding to Glycoprotein Ilb/IIIa on Platelets. Cytometry, 28: 58-65.  110 Beer, A.J. et a l , 2006. Positron emission tomography using [18F]Galacto-RGD identifies the level of integrin alpha(v)beta3 expression in man. Clin Cancer Res, 12(13): 3942-9. Bertucci, C. and Domenici, E., 2002. Reversible and covalent binding of drugs to human serum albumin: Methodological approaches and physiological relevance. Current Medicinal Chemistry, 9(15): 1463-1481. Boudreau, N . , Andrews, C , Srebrow, A . , Ravanpay, A . and Cheresh, D.A., 1997. Induction of the angiogenic phenotype by Hox D3. J Cell Biol, 139(1): 257-64. Bowser, M.T. and Chen, D.D.Y., 1998. Higher order equilibria and their effect on analyte migration behavior in capillary electrophoresis. Analytical Chemistry, 70(15): 3261-3270. Brassard, D.L. et al., 1999. Integrin alpha(v)beta(3)-mediated activation of apoptosis. Exp Cell Res, 251(1): 33-45. Brooks, P.C., 1996. Role of integrins in angiogenesis. Eur J Cancer, 32A(14): 2423-9. Brooks, P.C., Clark, R . A . and Cheresh, D.A., 1994a. Requirement of vascular integrin alpha v beta 3 for angiogenesis. Science, 264(5158): 569-71. Brooks, P.C. et al., 1994c. Integrin alpha v beta 3 antagonists promote tumor regression by inducing apoptosis of angiogenic bloodvessels. Cell, 79(7): 1157-64. Buckley, C D . et al., 1999. RGD peptides induce apoptosis by direct caspase-3 activation. Nature, 397(6719): 534-9. Busch, M.H.A., Boelens, H.F.M., Kraak, J.C. and Poppe, H . , 1997a. Vacancy affinity capillary electrophoresis, a new method for measuring association constants. Journal of Chromatography A , 775(1-2): 313-326. Busch, M.H.A., Carels, L . B . , Boelens, H.F.M., Kraak, J.C. and Poppe, H . , 1997b. Comparison of five methods for the study of drug-protein binding in affinity capillary electrophoresis. Journal of Chromatography A , 777(2): 311-328. Busch, M.H.A., Kraak, J.C. and Poppe, H . , 1997c. Principles and limitations of methods available for the determination of binding constants with affinity capillary electrophoresis. Journal of Chromatography A , 777(2): 329-353. Castel, S. et al., 2001. RGD peptides and monoclonal antibodies, antagonists of alpha(v)integrin, enter the cells by independent endocytic pathways. Lab Invest, 81(12): 1615-26. Chen, X . et a l , 2004a. Pegylated Arg-Gly-Asp peptide: 64Cu labeling and PET imaging of brain tumor alphavbeta3-integrin expression. J Nucl Med, 45(10): 1776-83.  Ill  Chen, X . et al., 2004b. MicroPET imaging of breast cancer alphav-integrin expression with 64Cu-labeled dimeric RGD peptides. Mol Imaging Biol, 6(5): 350-9. Chen, X . , Park, R., Shahinian, A . H . , Bading, J.R. and Conti, P.S., 2004c. Pharmacokinetics and tumor retention of 1251-labeled R G D peptide are improved by PEGylation. Nucl Med Biol, 31(1): 11-9. Chen, X . , Plasencia, C , Hou, Y . and Neamati, N . , 2005. Synthesis and biological evaluation of dimeric R G D peptide-paclitaxel conjugate as a model for integrintargeted drug delivery. J Med Chem, 48(4): 1098-106. Cheresh, D.A., 1987. Human endothelial cells synthesize and express an Arg-Gly-Aspdirected adhesion receptor involved in attachment to fibrinogen and von Willebrand factor. Proc Natl Acad Sci U S A , 84(18): 6471-5. Cheresh, D . A . et al., 1987. A n Arg-Gly-Asp-directed receptor on the surface of human melanoma cells exists in an divalent cation-dependent functional complex with the disialoganglioside GD2. J Cell Biol, 105(3): 1163-73. Cho, M.J. and Juliano, R., 1996. Macromolecular versus small-molecule therapeutics: drug discovery, development and clinical considerations. Trends Biotechnol, 14(5): 1538. Choi, J.S., MacKay, J.A. and Szoka, F.C., Jr., 2003. Low-pH-sensitive PEG-stabilized plasmid-lipid nanoparticles: preparation and characterization. Bioconjug Chem, 14(2): 420-9. Colton, I.J., Carbeck, J.D., Rao, J. and Whitesides, G . M . , 1998. Affinity capillary electrophoresis: A physical-organic tool for studying interactions in biomolecular recognition. Electrophoresis, 19(3): 367-382. Connors, K . A . , 1987. Binding Constant. The Measurement of Molecular Complex Stability. John Wiley & Sons, New York. Daugherty, A . L . and Mrsny, R.J., 2006. Formulation and delivery issues for monoclonal antibody therapeutics. Adv Drug Deliv Rev, 58(5-6): 686-706. Dechantsreiter, M . A . et al., 1999. N-Methylated cyclic RGD peptides as highly active and selective alpha(V)beta(3) integrin antagonists. J Med Chem, 42(16): 3033-40. Dhananjaya, K . V . and Antony, A., 1988. Pharmacokinetics and chemotherapeutic efficacy of adriamycin encapsulated in immunoliposomes against avian myeloblastosis virus infection. J Virol Methods, 19(2): 121-9.  112 Doi, K . et al., 2007. Enhanced angiogenesis by gelatin hydrogels incorporating basic fibroblast growth factor in rabbit model of hind limb ischemia. Heart Vessels, 22(2): 104-8. Dubey, P.K., Mishra, V . , Jain, S., Mahor, S. and Vyas, S.P., 2004. Liposomes modified with cyclic R G D peptide for tumor targeting. J Drug Target, 12(5): 257-64. Dunphy, J.T. and Linder, M . E . , 1998. Signalling functions of protein palmitoylation. Biochim Biophys Acta, 1436(1-2): 245-61. Fanali, S., 1997. Controlling enantioselectivity in chiral capillary electrophoresis with inclusion-complexation. Journal of Chromatography A , 792(1-2): 227-267. Fenske, D.B., MacLachlan, I. and Cullis, P.R., 2002. Stabilized plasmid-lipid particles: a systemic gene therapy vector. Methods Enzymol, 346: 36-71. Folkman, J., 1990. What is the evidence that tumors are angiogenesis dependent? J Natl Cancer Inst, 82(1): 4-6. Forssen, E.A and Tokes, ZA., 1981 Use of anionic liposomes for the reduction of chronic doxorubicin-induced cardiotoxicity. Proc Natl Acad Sci U S A 78(3): 1873-7. Friedlander, M . et al., 1995. Definition of two angiogenic pathways by distinct alpha v integrins. Science, 270(5241): 1500-2. Friess, H . et al., 2006. A randomized multi-center phase II trial of the angiogenesis inhibitor Cilengitide (EMD 121974) and gemcitabine compared with gemcitabine alone in advanced unresectable pancreatic cancer. B M C Cancer, 6: 285. Gabizon, A . et al., 1994. Prolonged circulation time and enhanced accumulation in malignant exudates of doxorubicin encapsulated in polyethylene-glycol coated liposomes. Cancer Res, 54(4): 987-92. Gabizon, A . , Goren, D., Fuks, Z., Meshorer, A . and Barenholz, Y . , 1985. Superior therapeutic activity of liposome-associated adriamycin in a murine metastatic tumour model. Br J Cancer, 51(5): 681-9. Gabizon, A . A . , Shmeeda, H . and Zalipsky, S., 2006. Pros and cons of the liposome platform in cancer drug targeting. J Liposome Res, 16(3): 175-83. Gagne, J.F., Desormeaux, A . , Perron, S., Tremblay, M.J. and Bergeron, M.G., 2002. Targeted delivery of indinavir to HIV-1 primary reservoirs with immunoliposomes. Biochim Biophys Acta, 1558(2): 198-210. Galbusera, C. and Chen, D.D.Y., 2003. Molecular interaction in capillary electrophoresis. Current Opinion in Biotechnology, 14(1): 126-130.  113  Gardner, S.N., 2000. A mechanistic, predictive model of dose-response curves for cell cycle phase-specific and -nonspecific drugs. Cancer Res, 60(5): 1417-25. Giannelli, G., Falk-Marzillier, J., Schiraldi, O., Stetler-Stevenson, W.G. and Quaranta, V . , 1997. Induction of cell migration by matrix metalloprotease-2 cleavage of laminin5. Science, 277(5323): 225-8. Gladson, C.L. and Cheresh, D.A., 1991. Glioblastoma expression of vitronectin and the alpha v beta 3 integrin. Adhesion mechanism for transformed glial cells. J Clin Invest, 88(6): 1924-32. Han, G.Z., Liu, Z.J., Shimoi, K . and Zhu, B.T., 2005. Synergism between the anticancer actions of 2-methoxyestradiol and microtubule-disrupting agents in human breast cancer. Cancer Res, 65(2): 387-93. Haran, G., Cohen, R., Bar, L . K . and Barenholz, Y . , 1993. Transmembrane ammonium sulfate gradients in liposomes produce efficient and stable entrapment of amphipathic weak bases. Biochim Biophys Acta, 1151(2): 201-15. Harding, J.A., Engbers, C M . , Newman, M.S., Goldstein, N.I. and Zalipsky, S., 1997. Immunogenicity and pharmacokinetic attributes of poly(ethylene glycol)-grafted immunoliposomes. Biochim Biophys Acta, 1327(2): 181-92. Harding, S.E., Chowdhry, B. Z., 2001. Protein-Ligand Interactions: Hydrodynamics and Colorimetry. oxford University Press, oxford. Haubner, R. et al., 2001. Glycosylated RGD-containing peptides: tracer for tumor targeting and angiogenesis imaging with improved biokinetics. J Nucl Med, 42(2): 326-36. Haubner, R.H., Wester, H.J., Weber, W.A. and Schwaiger, M . , 2003. Radiotracer-based strategies to image angiogenesis. Q J N u c l Med, 47(3): 189-99. Heath, T.D., Macher, B . A . and Papahadjopoulos, D., 1981. Covalent attachment of immunoglobulins to liposomes via glycosphingolipids. Biochim Biophys Acta, 640(1): 66-81. Heegaard, N.H.H., Nissen, M . H . and Chen, D.D.Y., 2002. Applications of on-line weak affinity interactions in free solution capillary electrophoresis. Electrophoresis, 23(6): 815-822. Hieken, T.J. et al., 1996. Beta3 integrin expression in melanoma predicts subsequent metastasis. J Surg Res, 63(1): 169-73. Holig, P. et al., 2004. Novel R G D lipopeptides for the targeting of liposomes to integrinexpressing endothelial and melanoma cells. Protein Eng Des Sel, 17(5): 433-41.  114  Hood, J.D. et al., 2002. Tumor regression by targeted gene delivery to the neovasculature. Science, 296(5577): 2404-7. Huang, S.K. et al., 1992. Pharmacokinetics and therapeutics of sterically stabilized liposomes in mice bearing C-26 colon carcinoma. Cancer Res, 52(24): 6774-81. Imai, K . and Takaoka, A . , 2006. Comparing antibody and small-molecule therapies for cancer. Nat Rev Cancer, 6(9): 714-27. Ishida, T., Iden, D.L. and Allen, T . M . , 1999. A combinatorial approach to producing sterically stabilized (Stealth) immunoliposomal drugs. FEBS Lett, 460(1): 129-33. Jain, R.K., 1987. Transport of molecules across tumor vasculature. Cancer Metastasis Rev, 6(4): 559-93. Janssen, A.P. et al., 2003. Peptide-targeted PEG-liposomes in anti-angiogenic therapy. Int J Pharm, 254(1): 55-8. Jeffs, L . B . et al., 2005. A scalable, extrusion-free method for efficient liposomal encapsulation of plasmid D N A . Pharm Res, 22(3): 362-72. Jeppesen, C. et al., 2001. Impact of polymer tether length on multiple ligand-receptor bond formation. Science, 293(5529): 465-8. Johnston, M.J. et al., 2006. Therapeutically optimized rates of drug release can be achieved by varying the drug-to-lipid ratio in liposomal vincristine formulations. Biochim Biophys Acta, 1758(1): 55-64. Kale, A . A . and Torchilin, V.P., 2007. Design, synthesis, and characterization of pHsensitive PEG-PE conjugates for stimuli-sensitive pharmaceutical nanocarriers: the effect of substitutes at the hydrazone linkage on the ph stability of PEG-PE conjugates. Bioconjug Chem, 18(2): 363-70. Kim, H . et al., 2006. The pharmacokinetics of rituximab following an intravitreal injection. Exp Eye Res, 82(5): 760-6. Kim, J.W. and Lee, H.S., 2004. Tumor targeting by doxorubicin-RGD-4C peptide conjugate in an orthotopic mouse hepatoma model. Int J Mol Med, 14(4): 529-35. Klibanov, A . L . , Maruyama, K., Torchilin, V.P. and Huang, L . , 1990. Amphipathic polyethyleneglycols effectively prolong the circulation time of liposomes. FEBS Lett, 268(1): 235-7. Kochendoerfer, G.G. et a l , 2003. Design and chemical synthesis of a homogeneous polymer-modified erythropoiesis protein. Science, 299(5608): 884-7.  115  Kohler, G. and Milstein, C , 1975. Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, 256(5517): 495-7. Koning, G.A. et al., 2001. Pharmacokinetics of differently designed immunoliposome formulations in rats with or without hepatic colon cancer metastases. Pharm Res, 18(9): 1291-8. Kraak, J . C , Busch, S. and Poppe, H . , 1992. Study of Protein Drug-Binding Using Capillary Zone Electrophoresis. Journal of Chromatography, 608(1-2): 257-264. Lee, S.M. et al., 2007. Targeting bladder tumor cells in vivo and in the urine with a peptide identified by phage display. Mol Cancer Res, 5(1): 11-9. Leserman, L . D., Machy, P., Barbet, J., 1981. Cell-specific drug transfer from liposomes bearing monoclonal antibodies. Nature 293(5829):226-8 L i , W., Huang, Z., MacKay, J.A., Grube, S. and Szoka, F . C , Jr., 2005. Low-pH-sensitive poly(ethylene glycol) (PEG)-stabilized plasmid nanolipoparticles: effects of P E G chain length, lipid composition and assembly conditions on gene delivery. J Gene Med, 7(1): 67-79. L i , Z.B. et al., 2007. 64Cu-Labeled Tetrameric and Octameric R G D Peptides for SmallAnimal PET of Tumor {alpha}v{beta}3 Integrin Expression. J Nucl Med, 48(7): 1162-71. Liapis, H . , Adler, L . M . , Wick, M.R. and Rader, J.S., 1997. Expression of alpha(v)beta3 integrin is less frequent in ovarian epithelial tumors of low malignant potential in contrast to ovarian carcinomas. Hum Pathol, 28(4): 443-9. Livnah, O. et al., 1996. Functional mimicry of a protein hormone by a peptide agonist: the EPO receptor complex at 2.8 A . Science, 273(5274): 464-71. Maeda, H . , Wu, J., Sawa, T., Matsumura, Y . and Hori, K . , 2000. Tumor vascular permeability and the EPR effect in macromolecular therapeutics: a review. J Control Release, 65(1-2): 271-84. Maloney D. G., Grillo-Lopez A . J., White C. A., Bodkin D., Schilder R. J, Neidhart J. A . , JanakiramanN., Foon K. A., Liles T. M . , Dallaire B. K., Wey K., Royston I., Davis T., Levy R. (1997) IDEC-C2B8 (Rituximab) anti-CD20 monoclonal antibody therapy in patients with relapsed low-grade non-Hodgkin's lymphoma. Blood 90(6);2188-95. Martino, R. et al., 2002. Cyclophosphamide, pegylated liposomal doxorubicin (Caelyx), vincristine and prednisone (CCOP) in elderly patients with diffuse large B-cell lymphoma: results from a prospective phase II study. Haematologica, 87(8): 822-7.  116  Maruyama, K . , Takahashi, N . , Tagawa, T., Nagaike, K . and Iwatsuru, M . , 1997. Immunoliposomes bearing polyethyleneglycol-coupled Fab' fragment show prolonged circulation time and high extravasation into targeted solid tumors in vivo. FEBS Lett, 413(1): 177-80. Matsumura, Y. et al., 2004. Phase I and pharmacokinetic study of MCC-465, a doxorubicin (DXR) encapsulated in PEG immunoliposome, in patients with metastatic stomach cancer. Ann Oncol, 15(3): 517-25. Maurer, N . , Fenske, D.B. and Cullis, P.R., 2001. Developments in liposomal drug delivery systems. Expert Opin Biol Ther, 1(6): 923-47. Mayer, L.D., Bally, M . B . , Hope, M.J. and Cullis, P.R., 1985. Uptake of antineoplastic agents into large unilamellar vesicles in response to a membrane potential. Biochim Biophys Acta, 816(2): 294-302. Mayer, L.D., Hope, M.J. and Cullis, P.R., 1986. Vesicles of variable sizes produced by a rapid extrusion procedure. Biochim Biophys Acta, 858(1): 161-8. Mehta, R., Lopez-Berestein, G., Hopfer, R., Mills, K . and Juliano, R.L., 1984. Liposomal amphotericin B is toxic to fungal cells but not to mammalian cells. Biochim Biophys Acta, 770(2): 230-4. Montet, X . , Funovics, M . , Montet-Abou, K . , Weissleder, R. and Josephson, L . , 2006. Multivalent effects of R G D peptides obtained by nanoparticle display. J Med Chem, 49(20): 6087-93. Moreira, J.N., Ishida, T., Gaspar, R. and Allen, T . M . , 2002. Use of the post-insertion technique to insert peptide ligands into pre-formed stealth liposomes with retention of binding activity and cytotoxicity. Pharm Res, 19(3): 265-9. Muggia, F. and Hamilton, A . , 2001. Phase III data on Caelyx in ovarian cancer. Eur J Cancer, 37 Suppl 9: SI5-8. Mui, B., Raney, S.G., Semple, S.C. and Hope, M.J., 2001. Immune stimulation by a CpGcontaining oligodeoxynucleotide is enhanced when encapsulated and delivered in lipid particles. J Pharmacol Exp Ther, 298(3): 1185-92. Murphy-Ullrich, J.E., 2001. The de-adhesive activity of matricellular proteins: is intermediate cell adhesion an adaptive state? J Clin Invest, 107(7): 785-90. Nabors, L . B . et al., 2007. Phase I and correlative biology study of cilengitide in patients with recurrent malignant glioma. J Clin Oncol, 25(13): 1651-7.  117 Nallamothu, R. et al., 2006. A targeted liposome delivery system for combretastatin A4: formulation optimization through drug loading and in vitro release studies. P D A J Pharm Sci Technol, 60(3): 144-55. Nayeem, M . S . and Khan, R.H., 2006. Recombinant antibodies in cancer therapy. Curr Protein Pept Sci, 7(2): 165-70. Nellis, D.F. et al., 2005a. Preclinical manufacture of an anti-HER2 scFv-PEG-DSPE, liposome-inserting conjugate. 1. Gram-scale production and purification. Biotechnol Prog, 21(1): 205-20. Nellis, D.F. et al., 2005b. Preclinical manufacture of anti-HER2 liposome-inserting, scFvPEG-lipid conjugate. 2. Conjugate micelle identity, purity, stability, and potency analysis. Biotechnol Prog, 21(1): 221-32. Northfelt, D.W. et al., 1996. Doxorubicin encapsulated in liposomes containing surfacebound polyethylene glycol: pharmacokinetics, tumor localization, and safety in patients with AIDS-related Kaposi's sarcoma. J Clin Pharmacol, 36(1): 55-63. Olson, F., Hunt, C.A., Szoka, F.C., Vail, W.J. and Papahadjopoulos, D., 1979. Preparation of liposomes of defined size distribution by extrusion through polycarbonate membranes. Biochim Biophys Acta, 557(1): 9-23. Olson, F., Mayhew, E., Maslow, D., Rustum, Y . and Szoka, F., 1982. Characterization, toxicity and therapeutic efficacy of adriamycin encapsulated in liposomes. Eur J Cancer Clin Oncol, 18(2): 167-76. Oravcova, J., Bohs, B . and Lindner, W., 1996. Drug-protein binding studies - New trends in analytical and experimental methodology. Journal of Chromatography B Biomedical Applications, 677(1): 1-28. Ostergaard, J. and Heegaard, N.H.H., 2003. Capillary electrophoresis frontal analysis: Principles and applications for the study of drug-plasma protein binding. Electrophoresis, 24(17): 2903-2913. Padlan, E.A., 1991. A possible procedure for reducing the immunogenicity of antibody variable domains while preserving their ligand-binding properties. M o l Immunol, 28(4-5): 489-98. Papahadjopoulos, D. et al., 1991. Sterically stabilized liposomes: improvements in pharmacokinetics and antitumor therapeutic efficacy. Proc Natl Acad Sci U S A , 88(24): 11460-4. Park, J.W. et al., 2002. Anti-HER2 immunoliposomes: enhanced efficacy attributable to targeted delivery. Clin Cancer Res, 8(4): 1172-81.  118 Pasqualini, R. et al., 2000. Aminopeptidase N is a receptor for tumor-homing peptides and a target for inhibiting angiogenesis. Cancer Res, 60(3): 722-7. Pasqualini, R. and Ruoslahti, E., 1996. Organ targeting in vivo using phage display peptide libraries. Nature, 380(6572): 364-6. Pastorino, F. et al., 2006. Targeting liposomal chemotherapy via both tumor cell-specific and tumor vasculature-specific ligands potentiates therapeutic efficacy. Cancer Res, 66(20): 10073-82. Pastorino, F. et al., 2004. Targeted delivery of oncogene-selective antisense oligonucleotides in neuroectodermal tumors: therapeutic implications. Ann N Y Acad Sci, 1028: 90-103. Petit, V . and Thiery, J.P., 2000. Focal adhesions: structure and dynamics. Biol Cell, 92(7): 477-94. Petrov, A . , Okhonin, V . , Berezovski, M . and Krylov, S.N., 2005. Kinetic capillary electrophoresis (KCE): A conceptual platform for kinetic homogeneous affinity methods. Journal of the American Chemical Society, 127(48): 17104-17110. Pierschbacher, M.D. and Ruoslahti, E., 1984. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature, 309(5963): 30-3. Presta, M . et al., 2005. Fibroblast growth factor/fibroblast growth factor receptor system in angiogenesis. Cytokine Growth Factor Rev, 16(2): 159-78. Puklin-Faucher, E., Gao, M . , Schulten, K . and Vogel, V . , 2006. How the headpiece hinge angle is opened: New insights into the dynamics of integrin activation. J Cell Biol, 175(2): 349-60. R. A . Smith and Giorgio, T.D., 2004. Quantitation and Kinetics of CD51 Surface Receptor Expression: Implications for Targeted Delivery. Annals of Biomedical Engineering, 32(5): 635-644. Renigunta, A . et al., 2006. D N A transfer into human lung cells is improved with Tat-RGD peptide by caveoli-mediated endocytosis. Bioconjug Chem, 17(2): 327-34. Rundlett, K . L . and Armstrong, D.W., 2001a. Methods for the determination of binding constants by capillary electrophoresis. Electrophoresis, 22(7): 1419-1427. Ruoslahti, E., 2000. Targeting tumor vasculature with homing peptides from phage display. Semin Cancer Biol, 10(6): 435-42. Ruoslahti, E., 2003. The RGD story: a personal account. Matrix Biol, 22(6): 459-65.  119 Ruoslahti, E. and Pierschbacher, M.D., 1987. New perspectives in cell adhesion: RGD and integrins. Science, 238(4826): 491-7. Safra, T. et al., 2000. Pegylated liposomal doxorubicin (doxil): reduced clinical cardiotoxicity in patients reaching or exceeding cumulative doses of 500 mg/m2. Ann Oncol, 11(8): 1029-33. Sapra, P. and Allen, T.M., 2002. Internalizing antibodies are necessary for improved therapeutic efficacy of antibody-targeted liposomal drugs. Cancer Res, 62(24): 7190-4. Sapra, P. and Allen, T.M., 2003. Ligand-targeted liposomal anticancer drugs. Prog Lipid Res, 42(5): 439-62. Scatchard, G., Ann, N . Y., 1949. Acad. Sci., 51(660). Schiffelers, R . M . et al., 2003. Anti-tumor efficacy of tumor vasculature-targeted liposomal doxorubicin. J Control Release, 91(1-2): 115-22. Schraa, A . J . et al., 2002. Targeting of RGD-modified proteins to tumor vasculature: a pharmacokinetic and cellular distribution study. Int J Cancer, 102(5): 469-75. Sebille, B., Zini, R., Madjar, C.V., Thuaud, N . and Tillement, J.P., 1990. Separation Procedures Used to Reveal and Follow Drug Protein-Binding. Journal of Chromatography-Biomedical Applications, 531: 51-77. Segal, A.W., Gregoriadis, G., and Black, C D . , 1975. Liposomes as vehicles for the local release of drugs. Clin Sci Mol Med. 49(2):99-106. Singh, B., Rawlings, N . and Kaur, A., 2001. Expression of integrin alphavbeta3 in pig, dog and cattle. Histol Histopathol, 16(4): 1037-46. Stephens, S., Emtage, S., Vetterlein, 0 „ Chaplin, L., Bebbington, C , Nesbitt, A., Sopwith, M . , Athwal, D., Novak, C , Bodmer, M . , 1995. Comprehensive pharmacokinetics of a humanized antibody and analysis of residual anti-idiotypic responses. Immunology 85(4):668-74. Storgard, C M . et al., 1999. Decreased angiogenesis and arthritic disease in rabbits treated with an alphavbeta3 antagonist. J Clin Invest, 103(1): 47-54. Stupack, D.G. and Cheresh, D.A., 2002. E C M remodeling regulates angiogenesis: endothelial integrins look for new ligands. Sci STKE, 2002(119): PE7. Szoka, F. et al., 1980. Preparation of unilamellar liposomes of intermediate size (0.1-0.2 mumol) by a combination of reverse phase evaporation and extrusion through polycarbonate membranes. Biochim Biophys Acta, 601(3): 559-71.  120  Tanaka, Y . and Terabe, S., 2002. Estimation of binding constants by capillary electrophoresis. Journal of Chromatography B-Analytical Technologies in the Biomedical and Life Sciences, 768(1): 81-92. Temming, K., Schiffelers, R . M . , Molema, G. and Kok, R.J., 2005. RGD-based strategies for selective delivery of therapeutics and imaging agents to the tumour vasculature. Drug Resist Updat, 8(6): 381-402. Torchilin, V.P., 1985. Liposomes as targetable drug carriers. Crit Rev Ther Drug Carrier Syst, 2(1): 65-115. Torchilin, V.P. et al., 2001. p-Nitrophenylcarbonyl-PEG-PE-liposomes: fast and simple attachment of specific ligands, including monoclonal antibodies, to distal ends of PEG chains via p-nitrophenylcarbonyl groups. Biochim Biophys Acta, 1511(2): 397-411. Tucker, G . C , 2006. Integrins: molecular targets in cancer therapy. Curr Oncol Rep, 8(2): 96-103. Watt, P., 2006 Screening for drugs from the natural repetoire of biodiverse protein folds. Nature Biotechnology 24(2).T 77-183 Winterhalter, M . , and Lasic, D. D., 1993. Liposomes stability and formulation: experimental parameters and theories on the size distribution. Chem Phys Lipids. 64(l-3):35-43 Wu, Y . et al., 2005. microPET imaging of glioma integrin {alpha}v{beta}3 expression using (64)Cu-labeled tetrameric RGD peptide. J Nucl Med, 46(10): 1707-18. Xiong, J.P. et al., 2002. Crystal structure of the extracellular segment of integrin alpha Vbeta3 in complex with an Arg-Gly-Asp ligand. Science, 296(5565): 151-5. Xiong, X . B . et al., 2005. Intracellular delivery of doxorubicin with RGD-modified sterically stabilized liposomes for an improved antitumor efficacy: in vitro and in vivo. J Pharm Sci, 94(8): 1782-93. X u , J. et al., 2001. Proteolytic exposure of a cryptic site within collagen type IV is required for angiogenesis and tumor growth in vivo. J Cell Biol, 154(5): 1069-79. Zhang, L., Hoffman, J.A. and Ruoslahti, E., 2005. Molecular profiling of heart endothelial cells. Circulation, 112(11): 1601-11. Zimmermann, T.S. et al., 2006. RNAi-mediated gene silencing in non-human primates. Nature, 441(7089): 111-4.  121 F., Tomadini, V . , Zaccaria, A . , Lenoci, M . , Battista, M . , Molinari, A . , L., Fabbri, A . , Battista, R., Cabras, M . , G., Gallamini, A., Fanini, R., 2006. CHOP-rituximab with pegylatated liposomal doxorubicin for the treatment of elderly patients with diffuse large B-cell lypmphoma. Leukemia Lymphoma 47(10): 2174-2180.  122 APPENDIX  Theory for multi-site equilibrium calculations Currently the most often used approach to determining the binding characteristics for interactions with higher order stoichiometry is to evaluate the changes of a physical parameter of the system with a Scatchard Plot:(Connors, 1987) For a Scatchard Plot to provide meaningful information, certain conditions have to be satisfied. The ligand should have only one active site, and the multiple binding sites on the protein should be identical, distinguishable, and independent. However, in practice, these conditions may not be easily met. Specific binding events often differ from non-specific bindings, and the binding of the first ligand often affects the second binding, favorably of adversely, depending on the function of the protein. To properly address these properties, the following equations are derived and used to process the data used in this work. As the binding of the species (protein-ligand) occurs with a 1:1 stoichiometry, the equilibrium is given by: (1)  [PL] [P][L]  (2)  where [P], [L] and [PL] are the concentrations of the protein, the ligand and the proteinligand complex, respectively. The binding constant, Kb, is given by Eq 2. The fraction of the proteins bound, f^p, is an important factor that can be used to determine the affinity between the protein and the ligand:  fb,P  ~  [A  [PL]  tn  [P]/+[PL]  (3)  123 The subscripts b, f, and t denote bound, free, and total concentrations of corresponding species in the solution, respectively. In C E - F A , the injected sample plug contains preequilibrated proteins and ligands. The total concentration of protein present in the sample mixture, [P] , is kept constant, and the total ligand concentration, [L] , is varied in each C E t  t  process. As species start to migrate through the capillary under a high voltage, the free ligands are partially separated from the proteins and the protein-ligand complexes. As shown in Figure 1A, the free ligand concentration can be calculated by using a common calibration curve, obtained by injecting samples containing only the ligand. The calculated [L]f can be used to estimate the binding parameters. In CE-FA, [P]b is always hard to measure because the mobilities of free protein and protein-ligand complex are often very close. Therefore, the average number of ligand molecules bound per protein molecule, /, is introduced as  [ZV[.P]r  base on the relation of [P]b and [L]b in different interaction  stoichiometrics. For a 1:1 binding, I is defined as _[L] _  J  b  [PI  With a plot of  [PL] [P\ +[PL] f  [£V[-P]r  _  K [L] b  f  l + K [L] h  f  vs. [L]f, the binding constant, Kb, can be determined by a non-  linear curve fit.  Higher Order Binding Processes Specific binding of small molecules to macromolecules, such as enzymes and other proteins, poly-nucleic acids, and synthetic polymers, is an important area that often requires consideration of multiple binding sites (Scatchard, 1949). Scatchard plot (I vs. [L]/) have been used to process data obtained in several C E methods, such as C E - F A , H D , and V P (Colton et al., 1998; Fanali, 1997; Rundlett and Armstrong, 2001a).  124 Because the assumptions used in the Scatchard method are often not valid for biological molecules, a more general model should be developed. For a 1:2 interaction, Eq 1 and Eq 2 can be used to describe the first step binding, with Kbi, replacing Kb as the binding constant. The following steps can be expressed as the following: (5)  "  '  [PL][L]  (6)  f  and (7)  *" ~ [PL„_{\[L]  f  The overall binding constant, K, is generally calculated as the product of binding constants of each step:  In this study, we will focus on the specific interaction with a small number of binding sites (i.e., n = 2 and/or 3), which is common for binding of macromolecules in real biological systems. For a protein-ligand binding with 1:2 or 1:3 stoichiometry, inserting [PL], [PL2] and [PLs] into Eq 9, the / value for 1:2 binding is obtained from: J  [L\ [PI  [PL] 2[PL ] +  [P] +[PL] f  K [L] +2K K [L]  2  bx  + [PL ] 2  \  +  and the / value for 1:3 binding is obtained from:  f  hl  2  h2  f  K \L] K K [L] h  f+  hX  b2  2  f  125 j _ [L\ _ [PI  [PI] + 2[PZ ] 3[PZ ] 2  +  3  [ / ] + [ ^ ] [PZ ] + [PZ ] >  /  _ K [L], H  \ +  +  2  3  2K K [Lf  +  hl  b2  + 3K K K [Lf  f  bl  b2  b3  (  H  )  f  K [L] +K K [Lf +K K K [Lf H  f  bl  b2  f  bl  b2  b3  f  For the case of multiple types of binding sites, the overall / can be defined as: /,=|/,  (12)  Most commonly, i is 1 or 2 as reported in the literature. If non-specific binding also exists, / is the sum of I cific and I nonspecific, and Eq 4, 10 and 11 still describe describe the specific spe  binding process, i f the non-specific binding can be account for. Thus, the appropriate isotherm describing higher order equilibrium for each type of sites is generalized:  03)  I—^H  Most of the parameters in Eq 13 are defined earlier, except for N, which is the maximum number of binding sites available on each protein molecule. The individual binding constant for each step, Kf,„, can be determined by plotting  vs. the free ligand  concentration, [L]/, and the number of binding sites can be determined by fitting the experimental results with the nth order equation, such as Eq 10, 11 and 13. It should be noted that the unit of the binding constant obtained from Scatchard Plots is always in M " , 1  forcing all multiple bindings to a pseudo 1:1 stoichiometry. The overall binding constant obtained by Eq 4, 10 and 11 have units of M " , M " , or M " , depending on the overall 1  2  3  stoichiometry. Strictly speaking the binding constants are unitless. The units are used in this work only to follow the conventions practiced in most current literature.  126 There are also cases where the concentration of ligand present in the B G E is much greater than that of protein in B G E  ([L], »  [P]i),  in which case the binding sites on the  protein molecules are saturated, and the concentration of the unsaturated species [PL2],...,  ([PL],  are negligible. Eq 13 can be simplified in these situations as the  [PL„.j])  following: nKJL]" 1 =  — l— K [L]"  (14)  x  l +  n  f  Due to the similarities of the C E techniques, these equations can be also applied to Hummel-Dreyer (HD) and vacancy peak (VP) methods for the determination of binding parameters  in either  cooperative  interactions (Busch et al., 1997b).  or  non-cooperative  multiple-site protein-ligand  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items