Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development and characterization of Paclitaxel loaded polymeric films based on polysaccharides-graft-poly… Shi, Ruiwen 2003

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

Item Metadata

Download

Media
831-ubc_2003-792587.pdf [ 11.76MB ]
Metadata
JSON: 831-1.0091213.json
JSON-LD: 831-1.0091213-ld.json
RDF/XML (Pretty): 831-1.0091213-rdf.xml
RDF/JSON: 831-1.0091213-rdf.json
Turtle: 831-1.0091213-turtle.txt
N-Triples: 831-1.0091213-rdf-ntriples.txt
Original Record: 831-1.0091213-source.json
Full Text
831-1.0091213-fulltext.txt
Citation
831-1.0091213.ris

Full Text

DEVELOPMENT AND CHARACTERIZATION OF PACLITAXEL LOADED POLYMERIC FILMS BASED ON POLYSACCHARIDES-GRAFT-POLY(e-CAPROLACTONE) FOR THE PREVENTION OF SURGICAL ADHESIONS by RUIWEN SHI B. Eng., Tianjin University, China, 1987 M . Eng., Tianjin University, China, 1990 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES Faculty of Pharmaceutical Sciences Division of Pharmaceutics and Biopharmaceutics We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA March 2003 © Ruiwen Shi, 2003 i n p r e s e n t i n g this thes is in partial fu l f i lment o f the r e q u i r e m e n t s fo r an a d v a n c e d d e g r e e at the Univers i ty of Brit ish C o l u m b i a , I agree that t h e Library shall m a k e it f reely avai lable f o r re fe rence a n d s tudy . I fur ther agree that p e r m i s s i o n f o r e x t e n s i v e c o p y i n g of this thesis fo r scho lar ly p u r p o s e s may b e g ranted by the h e a d o f m y d e p a r t m e n t o r b y his o r h e r representat ives . It is u n d e r s t o o d that c o p y i n g o r p u b l i c a t i o n of this thesis for f inancial gain shal l n o t b e a l l o w e d w i t h o u t m y w r i t t e n p e r m i s s i o n . D e p a r t m e n t T h e Univers i ty o f Brit ish C o l u m b i a V a n c o u v e r , C a n a d a D a t e D E - 6 (2/88) ABSTRACT Novel amphiphilic graft copolymers with hydrophilic hydroxypropylcellulose (HPC) and dextrans (Dx) as main chains and hydrophobic poly(s-caprolactone) (PCL) as side chains were synthesized and characterized. Paclitaxel loaded polymeric films based on these graft copolymers were developed and characterized for the prevention of surgical adhesions. Hydroxypropylcellulose-graft-poly(s-caprolactone) (HPC-g-PCL) and dextran-graft-poly(s-caprolactone) (Dx-g-PCL) were synthesized via a "graft from" process, in which the PCL side chains were formed by a ring-opening polymerization of s-caprolactone (CL) initiated by the hydroxyl groups on the main chains of HPC or dextrans. HPC-g-PCL with a molar substitution of C L (MSCL ) in the range of 8.6 to 10.1 was synthesized by bulk polymerization without using any catalyst. Dx-g-PCLs were synthesized by solution polymerization using dimethyl sulfoxide as a solvent and stannous 2-ethylhexanoate as a catalyst. Dextrans with two different molecular weights, 70,000 (Dx70) and 500,000 (Dx500), were used as the main chains and the M S C L was determined to be in the range of 1.2 to 1.3. The films of the copolymers with or without paclitaxel were cast from tetrahydrofuran (THF) solutions. Both paclitaxel and the graft copolymers could be readily dissolved in THF and homogenous film formulations with paclitaxel loading up to 10% (w/w) were obtained. Due to the higher M S C L of HPC-g-PCL, PCL-rich micro-crystalline regions formed in the HPC-g-PCL films. Films of HPC-g-PCL were more hydrophobic than the Dx-g-PCL films and showed less water uptake and swelling. Controlled release of paclitaxel from the copolymer films was achieved. A n increase in i i initial paclitaxel loading resulted in an increase in the release rate for all the copolymer films. At 1% paclitaxel loading, the release rate from the Dx-g-PCL films was significantly higher than from the HPC-g-PCL films due to increased swelling and therefore greater diffusion rate of paclitaxel in the Dx-g-PCL matrices. At loadings of 5% and 10%, crystallization of paclitaxel occurred in the Dx-g-PCL matrices during incubation in the release media. The onset of this crystallization event led to a reduction in paclitaxel release rate from the films. At 10% loading, release rates from the Dx-g-PCL and HPC-g-PCL films were similar. No significant degradation was detected after the HPC-g-PCL films were incubated in aqueous media at 37 °C for 3 months and this was considered unsuitable for a surgical adhesion formulation. Dx-g-PCL films with or without paclitaxel were evaluated in rat models of surgical adhesions. The films were biocompatible and showed good handling characteristics when being applied to a surgical site. In a rat cecal abrasion model, Dx500-g-PCL films with paclitaxel loadings of 0.1% and 0.5% significantly reduced adhesion formation. Dx500-g-PCL films with no paclitaxel showed a barrier effect and reduced the incidence of adhesion formation. iii TABLE OF CONTENTS A B S T R A C T — i i T A B L E OF CONTENTS iv LIST OF TABLES —- xiii LIST OF FIGURES xv LIST OF SCHEMES xxiii LIST OF ABBREVIATIONS xxiv A C K N O W L E D G E M E N T xxix CHAPTER 1. PROJECT OVERVIEW AND BACKGROUND 1.1 PROJECT OVERVIEW 1 1.2 SURGICAL ADHESIONS 2 1.2.1 The problem and incidence 2 1.2.2 Pathophysiology of adhesion formation — 4 1.2.3 Approaches to adhesion prevention 5 1.2.3.1 Use of barriers to prevent adhesions 6 1.2.3.1.1 Oxidized regenerated cellulose barrier — 6 1.2.3.1.2 Expanded polytetrafluoroethylene barrier 7 1.2.3.1.3 Hyaluronic acid-carboxymethylcellulose barrier 8 1.2.3.1.4 Poloxamer 407 barrier 8 1.2.3.2 Use of drugs to prevent adhesions 9 1.2.3.2.1 Non-steroidal anti-inflammatory drugs 9 iv 1.2.3.2.2 Fibrinolytic drugs 9 1.2.3.2.3 Anti-coagulants 10 1.3 PACLITAXEL 10 1.3.1 Chemistry 10 1.3.2 Pharmacology 12 1.3.3 Toxicity 13 1.3.4 Pharmacokinetics 14 1.3.5 Paclitaxel in the treatment of surgical adhesions 16 1.4 POLYMER CHEMISTRY 16 1.4.1 Polymer structure - 16 1.4.2 Molecular weight and molecular weight distribution 17 1.4.2.1 Number-average molecular weight 18 1.4.2.2 Weight-average molecular weight ~ 19 1.4.2.3 Viscosity-average molecular weight 19 1.4.2.4 Polymer solution viscometry 20 1.4.2.5 Molecular weight distribution 22 1.4.2.6 Gel permeation chromatography 23 1.4.3 Morphology of solid polymers 24 1.4.4 Thermal transitions 25 1.4.5 Modification of polymer properties 26 1.4.5.1 Physical blends 26 1.4.5.2 Copolymers 27 1.4.5.3 Graft copolymers 28 v 1.4.5.4 Synthesis of graft copolymers 28 1.5 POLYMERIC DRUG DELIVERY SYSTEMS 29 1.5.1 Polymers as drug vehicles — 30 1.5.1.1 Biodegradability 30 1.5.1.2 Biocompatibility 35 1.5.1.3 Hydrogel as drug vehicles 36 1.5.2 Mechanism of drug release 37 1.5.2.1 Diffusion controlled systems 37 1.5.2.2 Diffusion and degradation controlled systems 38 1.5.2.3 Swelling-controlled systems 40 1.5.3 Factors influencing drug release 41 1.6 POLYMERIC DRUG DELIVERY SYSTEMS FOR PACLITAXEL 43 1.6.1 Microspheres 43 1.6.2 Injectable surgical pastes 44 1.6.3 Micellar formulations 44 1.6.4 Prodrugs 45 1.7 DEVELOPMENT OF PACLITAXEL LOADED FILMS FOR THE PREVENTION OF SURGICAL ADHESIONS 46 1.7.1 Hydrophilic polymeric films 46 1.7.2 Novel amphiphilic polymeric films 47 1.8 RESEARCH OBJECTIVES 47 CHAPTER 2. CHITOSAN BASED FILMS FOR PACLITAXEL DELIVERY --48 2.1 INTRODUCTION 48 vi 2.2 EXPERIMENTAL 51 2.2.1 Materials 51 2.2.2 General equipment and supplies 52 2.2.3 Preparation of chitosan and chitosan/PEOs films 53 2.2.4 Optical microscopy studies of paclitaxel loaded chitosan films 54 2.2.5 Thermal analysis 54 2.2.6 Determination of in vitro paclitaxel release rate 54 2.2.7 In vitro degradation of chitosan — 55 2.2.8 Statistical analysis 56 2.3 RESULTS 56 2.3.1 Effect of glycerol on the flexibility of chitosan films 56 2.3.2 In vitro degradation of chitosan 56 2.3.3 Nature of paclitaxel dispersion in chitosan film 58 2.3.4 In vitro paclitaxel release study— 58 2.3.4.1 Effect of crosslinking on the release rate of paclitaxel 60 2.3.4.2 Effect of paclitaxel loading on the release rate ~ 60 2.3.4.3 Effect of blending with PEOs on the release rate of PTX 64 2.3.5 Thermal analysis of the films of chitosan/PEOs blends 68 2.4 DISCUSSION 72 2.5 CONCLUSIONS 76 CHAPTER 3. SYNTHESIS AND CHARACTERIZATION OF HYDROXYPROPYLCELLULOSE- GRAFT-POLY(s-CAPROLACTONE) 77 vii 3.1 INTRODUCTION 77 3.2 EXPERIMENTAL 79 3.2.1 Materials- 79 3.2.2 General equipment and supplies 80 3.2.3 Synthesis of HPC-g-PCL 80 3.2.4 Conversion of CL 81 3.2.5 Purification and fractionation of HPC-g-PCL 81 3.2.6 Gel permeation chromatography 83 3.2.7 Nuclear magnetic resonance spectrometry - 83 3.2.8 Preparation of HPC-g-PCL films 85 3.2.9 Microscopic study of HPC-g-PCL films 85 3.2.10 Differential scanning calorimetry 85 3.2.11 X-ray diffraction 86 3.2.12 In vitro degradation of HPC-g-PCL films 86 3.3 RESULTS 86 3.3.1 Synthesis and purification of HPC-g-PCL 86 3.3.2 Evidence of graft copolymer formation using NMR spectroscopy 93 3.3.3 Molar substitution 107 3.3.4 Side chain length and distribution 109 3.3.5 Morphological study of HPC-g-PCL 110 3.3.6 Characterization of the fractions of HPC-g-PCL 113 3.3.7 In vitro degradation of HPC-g-PCL 117 3.4 DISCUSSION 121 viii 3.5 CONCLUSIONS 128 CHAPTER 4. SYNTHESIS AND CHARACTERIZATION OF DEXTRAN-GRAFT-POLY(e-CAPROLACTONE) 130 4.1 INTRODUCTION 130 4.2 EXPERIMENTAL 135 4.2.1 Materials 135 4.2.2 Synthesis of dextran-g-PCL 135 4.2.3 Conversion of dextrans and CL 136 4.2.4 Purification of dextran-g-PCLs 136 4.2.5 Gel permeation chromatography 137 4.2.6 Nuclear magnetic resonance spectrometry 137 4.2.7 Preparation of dextran-g-PCL films 137 4.2.8 Differential scanning calorimetry 138 4.3 RESULTS 138 4.3.1 Synthesis of dextran-g-PCL copolymers 138 4.3.2 Structural characterization 143 4.3.2.1 Characterization of the dextrans using NMR spectroscopy 143 4.3.2.2 Characterization of dextran-g-PCLs using NMR spectroscopy 146 4.3.2.3 Molar substitution of CL on dextrans 152 4.3.2.4 The length and the distribution of the side chains 153 4.3.3 Morphological study of dextran-g-PCLs 154 4.4 DISCUSSION 154 4.5 CONCLUSIONS 158 ix CHAPTER 5. FORMULATION AND CHARACTERIZATION OF PACLITAXEL LOADED AMPHIPHILIC GRAFT COPOLYMER FILMS 159 5.1 INTRODUCTION 159 5.2 EXPERIMENTAL 159 5.2.1 Materials and supplies 159 5.2.2 Preparation of paclitaxel loaded films 160 5.2.3 Determination of the degree of swelling of the films 160 5.2.4 In vitro study of paclitaxel release 161 5.2.5 Effect of paclitaxel loading on the thermal transition of HPC-g-PCL films 161 5.2.6 Morphological study of paclitaxel loaded dextran-g-PCL films ~ 162 5.2.7 Determination of residual THF in the copolymer films 162 5.3 RESULTS 163 5.3.1 Degree of swelling of the films 163 5.3.2 Paclitaxel loaded HPC-g-PCL films 165 5.3.3 Morphology of the paclitaxel loaded copolymer films 170 5.3.3.1 Paclitaxel loaded HPC-g-PCL films 170 5.3.3.2 Paclitaxel loaded Dx-g-PCL films 173 5.3.4 Residual THF in the copolymer films 176 5.4 DISCUSSION 181 5.5 CONCLUSIONS 186 CHAPTER 6. EVALUATION OF PACLITAXEL LOADED FILMS IN ANIMAL MODELS OF SURGICAL ADHESIONS 187 6.1 INTRODUCTION 187 6.2 EVALUATION OF THE FILMS IN A RAT SIDEWALL CECUM MODEL 188 6.2.1 Objectives 188 6.2.2 Experimental 188 6.2.2.1 Formulations 188 6.2.2.2 Animals and housing 189 6.2.2.3 Animal care and use committee approval 189 6.2.2.4 Experiment design 189 6.2.2.5 Surgical procedures 190 6.2.2.6 Evaluation 191 6.2.2.6.1 Biocompatibility and biodegradation- 191 6.2.2.6.2 Adhesion scoring system 192 6.2.2.6.3 Statistical analysis 192 6.2.2.7 Estimation of pacitaxel release in rats - 192 6.2.3 Results 193 6.2.3.1 Biocompatibility and biodegradation 193 6.2.3.2 Adhesion scores 195 6.2.3.3 In vivo paclitaxel release 196 6.2.4 Discussion 197 6.2.5 Conclusions 198 6.3 EFFICACY OF PACLITAXEL LOADED DX500-G-PCL FILMS IN A RAT CECAL ABRASION MODEL 198 6.3.1 Objectives 198 xi 6.3.2 Experimental 199 6.3.2.1 Formulations 199 6.3.2.2 Animal and housing 199 6.3.2.3 Animal Care and Use Committee approval 199 6.3.2.4 Randomization of animals 199 6.3.2.5 Surgical procedures — 200 6.3.2.6 Evaluation of the adhesion formation- — 201 6.3.2.6.1 Adhesion score 201 6.3.2.6.2 Incidence of adhesions 201 6.3.2.7 Statistical analysis 202 6.3.3 Results 202 6.3.4 DISCUSSION 203 6.3.5 CONCLUSIONS 205 CHAPTER 7. SUMMARIZING DISCUSSION, CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK 206 7.1 SUMMARIZING DISCUSSION — 206 7.2 CONCLUSIONS 210 7.3 SUGGESTIONS FOR FUTURE WORK 211 REFERENCES 213 xii LIST OF TABLES Table 1.1 Terminologies used in the viscometry of polymer solution 21 Table 1.2 Biodegradable polymers used in drug delivery systems 31 Table 2.1 Total amount of paclitaxel released within 7 days from chitosan films crosslinked by different amounts of glutaraldehyde and loaded with different amounts of paclitaxel 62 Table 2.2 Cumulative paclitaxel release within 7 days from the films of Chitosan/PEOs at different blending ratios. Paclitaxel loading was 5% (w/w) based on the weight of the dry film 67 Table 3.1 Results for the synthesis of four batches of HPC-g-PCL with a feeding ratio of 1:4 of the starting materials HPC and CL 88 Table 3.2 Observed chemical shifts (5) in the NMR spectra of HPC and PCL homopolymer (component P) formed during the reaction 99 Table 3.3 Observed chemical shifts (5) in the 1 3 C and *H NMR spectra of purified HPC-g-PCL 106 Table 3.4 Physicochemical data on HPC-g-PCL and its four fractions produced by precipitating from the THF solution using different molar fractions of hexanes 115 Table 4.1 Results of the syntheses Dx-g-PCLs with a molar feed ratio of 1:3 (w/w) of dextran and CL 140 Table 4.2 Observed chemical shifts (8) in the 1 3 C and ! H N M R spectra of dextran 148 xiii Table 4.3 Observed chemical shifts (8) in the NMR spectra of purified Dx-g-PCL and PCL homopolymer formed during the reactions 151 Table 5.1 The amount of paclitaxel released over the first 7 days (M7) from the films of the three copolymers and PCL homopolymer with different paclitaxel loadings. Values represent mean ± SD (n = 4) 171 . Table 5.2 Effect of paclitaxel loading on the thermal transitions of HPC-g-PCL films measured using DSC 174 Table 5.3 X-ray diffraction peak locations (°20) and d-spacings of paclitaxel powder precipitated from A C N solution by water, and the 10% paclitaxel loaded Dx500-g-PCL and Dx70-g-PCL films after being incubated in 10 mM PBS at 37 °C for 5 days 180 Table 6.1 Group assignments, paclitaxel dose, number of rats, and schedule of sacrificing- 190 Table 6.2 Summary of the observations on biocompatibility and biodegradation — 194 Table 6.3 Adhesion formations in the groups A and B after one week 196 Table 6.4 In vivo paclitaxel release after one week 196 Table 6.5 Group assignments, number of rats and paclitaxel doses for the efficacy study 200 Table 6.6 Percentage of the animals with no adhesions and with grade 2 or higher adhesions, and incidence of adhesions in the efficacy study 203 xiv LIST OF FIGURES Figure 1.1 Chemical structure of paclitaxel 11 Figure 1.2 A typical plot of extrapolation methods using Huggins and Kraemer equations 22 Figure 1.3 A typical molecular weight distribution curve 23 Figure 1.4 Schematic representation of the fringed micelle model 24 Figure 1.5 Theoretical drug release from a polymer slab by diffusion and by combined degradation and diffusion 40 Figure 2.1 Chemical structure of chitosan 48 Figure 2.2 In vitro degradation profile of chitosan film incubated in 10 mM PBS (pH7.4) at 37 °C. [x\] was determined at 30 ± 0.1°C in 0.1 M acetate buffer (pH4.5) using an Ubbelohde viscometer. Data represent mean ± SD (n=3) 57 Figure 2.3 A representative optical micrograph of paclitaxel-loaded chitosan film - 59 Figure 2.4 Paclitaxel release from chitosan films crosslinked by different amounts of glutaraldehyde or treated with NaOH: (•) no crosslinking, (x) chitosan-NaOH, (•) chitosan-0.1%GA, ( A ) chitosan- 1%GA, and (•) chitosan-10%GA. The thickness of the films was 50 urn and the initial paclitaxel loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The xv represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity 61 Figure 2.5 Paclitaxel release from chitosan-0.1%GA films with different initial loadings (w/w): (•) 10%, (•) 5%, and (•) 1%. The thickness of the films was 50 um. The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount released was normalized to a surface area of 1 cm2. Data represent mean + SD (n = 4) with the error bars shown only in the positive direction for clarity 63 Figure 2.6 Paclitaxel release from the films of chitosan/PEO200 blended at different ratios: (•) 40/60, (•) 60/40, (•) 80/20, and (x) 100/0. Initial Paclitaxel loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount of paclitaxel released was normalized to a surface area of 1 cm2. Data represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity 65 Figure 2.7 Paclitaxel release from the films of chitosan/PEO900 blended at different ratios: (•) 40/60, (•) 60/40, (•) 80/20, and (x) 100/0. Initial paclitaxel loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end fumbling tubes. The amount released was normalized to a surface area of 1 cm . Data represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity 66 xvi Figure 2.8 DSC scans of chitosan/PEO200 films with blending ratios of (A) 100/0, (B) 80/20, (C) 60/40, (D) 40/60, and (E) 0/100. Heating rate was 10°C/min 69 Figure 2.9 DSC scans of chitosan/PEO900 films with blending ratios of (A) 100/0, (B) 80/20, (C) 60/40, (D) 40/60, and (E) 0/100. Heating rate was 10°C/min 70 Figure 2.10 Effect of blending ratios of chitosan/PEOs on the melting point of PEOs. ( A ) Chitosan/PEO200, (•) Chitosan/PEO900 - 71 Figure 3.1 Standard curve of CL in water. Chromatographic conditions: column, Aqua C18; mobile phase, methanol/water at 35/65 (v/v); flow rate, 0.5 mL/min; injection volume, 10 pL; detection wavelength, 234 nm 82 Figure 3.2 Molecular weight calibration curve using polystyrene standards. Chromatographic conditions: columns, (A) Styragel HR3 and HR4 in series, (B) PLgel® Mixed-D; mobile phase, THF; flow rate, 1 mL/min; detector, differential refractive index detector 84 Figure 3.3 GPC chromatograms of (a) backbone HPC, (b) reaction product, in which G represents the component of HPC-g-PCL, P represents the PCL homopolymer formed during the reaction, and (c) purified HPC-g-PCL 90 Figure 3.4 GPC chromatograms of the precipitates and the supernatants when the reaction product was purified using different hexanes/THF ratios. (A) 0/1; (B) 0.8/1; (C) 1/1; (D) 2/1; and (E) 3/1. The subscript 1 denotes precipitate and 2 supernatant 92 xvii Figure 3.5 *H NMR spectra of HPC in (A) CDC13 and (B) DMSO-d 6 recorded at 25 °C using a Bruker WH400 spectrometer 94 Figure 3.6 *H NMR spectrum of HPC in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 96 Figure 3.7a H M Q C spectrum of HPC in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 97 Figure 3.7b Expanded view of Figure 3.7a indicating the connection between C l -H l 98 Figure 3.8 ' H NMR spectrum of PCL homopolymer (component P) formed during the graft reaction in CDCI3 recorded at 25°C using a Bruker WH400 spectrometer 101 Figure 3.9 ' H NMR spectrum of HPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 102 Figure 3.10 COSY-45 spectrum of HPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 104 Figure 3.11 H M Q C spectrum of HPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 105 Figure 3.12 Optical microscopy of HPC-g-PCL film under polarizing lens 111 Figure 3.13 Representative X-ray diffraction pattern of HPC-g-PCL film cast from THF solution 112 Figure 3.14 Representative DSC scans of HPC and purified HPC-g-PCL. Heating rate was 10°C/min 114 xviii Figure 3.15 Representative DSC scans for HPC-g-PCL (FO) and the four fractions: F l , F2, F3, and F4. The samples were films cast from THF solutions. The heating rate was 10°C/min 116 Figure 3.16 Changes in weight-average molecular weight (•) and polydispersity index ( A ) of the HPC-g-PCL films incubated in 10 mM PBS (pH7.4) a t 3 7 ° C 118 Figure 3.17 Representative DSC scans of HPC-g-PCL films incubated in 10 mM PBS (pH7.4) at 37 °C for different length of time. (A) 0 week; (B) 1 week; (C) 4 weeks; and (D) 12 weeks. Heating rate was 10 °C/min 119 Figure 3.18 Changes of the peak position in the DSC scans of HPC-g-PCL films incubated in 10 mM PBS (pH7.4) at 37 °C. (•) peak 1; (•) peak 2; and ( A ) peak 3. Each datum point represents mean ± SD (n=3) 120 Figure 4.1 GPC chromatograms of (a) the reaction product of Dx500-g-PCL, in which D l and PI represent the components of Dx500-g-PCL and the PCL homopolymer formed during the reaction, respectively; (b) purified Dx500-g-PCL; (c) the reaction product of Dx70-g-PCL, in which D2 and P2 represent the components of Dx70-g-PCL and the PCL homopolymer formed during the reaction, respectively; (d) purified Dx70-g-PCL 142 Figure 4.2 H M Q C spectrum of dextran in DMSO-d6 recorded at 27 °C using a Bruker AMX500 spectrometer 144 Figure 4.3 COSY-45 spectrum of dextran in DMSO-d6 recorded at 27 °C using a Bruker AV400 spectrometer 145 xix Figure 4.4 Representative ' H NMR spectrum of Dx-g-PCLs in DMSO-d 6 recorded at 95°C using a Bruker AMX500 spectrometer 149 Figure 4.5 Representative HMQC of Dx-g-PCLs in DMSO-d 6 recorded at 95°C using a Bruker AMX500 spectrometer 150 Figure 4.6 Representative DSC scans of (A) Dx500; (B) Dx500-g-PCL; (C) Dx70; and (D) Dx70-g-PCL. The heating rate was 10 °C/min 155 Figure 5.1 Degree of swelling of the films as a function of the nature of the polymers and paclitaxel loading. The films were immersed in 10 mM PBS (pH7.4) at 37 °C for 12 hours. The data represent mean ± SD (n=3) 164 Figure 5.2 In vitro paclitaxel release profiles from the Dx500-g-PCL films with different paclitaxel loadings: (A) 1%, 5%, and 10%; (B) 0.1% and 0.5%. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 166 Figure 5.3 In vitro paclitaxel release profiles from the Dx70-g-PCL films with different paclitaxel loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 167 Figure 5.4 In vitro paclitaxel release profiles from the HPC-g-PCL films with different paclitaxel loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 xx samples with error bars (standard deviation) shown only in the positive direction for clarity 168 Figure 5.5 In vitro paclitaxel release profiles from the PCL films with different loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 169 Figure 5.6 The first heating scans of the paclitaxel loaded HPC-g-PCL films prepared by the solution casting method. The paclitaxel loadings were: (Al) 0%; (Bl) 1%; (CI) 5%; and (Dl) 10%. The heating rate was 10 °C/min 172 Figure 5.7 The second heating scans of the paclitaxel loaded HPC-g-PCL films loaded after the films were heated from -100 °C to 100 °C and then cooled from 100 °C to -100 °C both at 10 °C/min. The paclitaxel loadings were: (A2) 0%; (B2) 1%; (C2) 5%; and (D2) 10%. The heating rate was 10 °C/min -175 Figure 5.8 Representative micrographs the 5% paclitaxel loaded Dx500-g-PCL films after incubation in 10 mM PBS (pH7.4) at 37 °C for (A) 6 h; (B) 12 h; (C) 24 h; (D) 2 days; (E) 3 days; and (F) 5 days 177 Figure 5.9 Representative micrographs the 10% paclitaxel loaded Dx500-g-PCL films after incubation in 10 mM PBS (pH7.4) at 37 °C for (A) 6 h; (B) 12 h; (C) 24 h; (D) 2 days; (E) 3 days; and (F) 5 days 178 Figure 5.10 Representative XRD patterns of (A) 10% paclitaxel loaded Dx500-g-PCL films and (B) 10% paclitaxel loaded Dx70-g-PCL films after the incubation in lOmM PBS (pH7.4) at 37 °C for 5 days; and (C) xxi paclitaxel powder precipitated by adding a solution of paclitaxel in A C N into an excess amount of water 179 Figure 6.1 Diagrammatic representations of (A) sidewall cecum model and (B) cecal abrasion model in the rat 204 xxii LIST OF SCHEMES Scheme 3.1 Structure and synthesis of HPC-g-PCL 87 Scheme 3.2 Structure and synthesis of Dx-g-PCLs 139 xxiii LIST OF ABBREVIATIONS Mn Number-average molecular weight Mv Viscosity-average molecular weight Mw Weight-average molecular weight N Average number of side chains on each repeating unit of main chains L Average side chain length rj Viscosity p Density AHf Heat of fusion 7]inh Inherent viscosity rjr Relative viscosity rjred Reduced viscosity r]Sp Specific viscosity [ rj\ Intrinsic viscosity A C N Acetonitrile A N O V A Analysis of variance A Q Acquisition time (for each scan) Co Initial drug concentration Ci8 Octadecylsilane CaH2 Calcium hydride CDCI3 Deuterated chloroform C M C Carboxymethylcellulose xxiv COSY Homonuclear correlated spectroscopy Csm The solubility of a drug in a polymer C X Width of spectrum D Diffusion coefficient D l Delay between pulses D C M Dichloromethane DMSO Dimethylsulfoxide DMSO-d6 Deuterated dimethylsulfoxide DS Dummy scan DSC Differential scanning calorimetry DW Dwell time Dx Dextrans Dx500 Dextran with a nominal molecular weight of 500,000 Dx70 Dextran with a nominal molecular weight of 70,000 Dx-g-PCL Dextran-graft-poly(e-caprolactone) ePTFE Expanded polytetrafluoroethylene FIDRES Free-induction decay resolution GA Glutaraldehyde GPC Gel permeation chromatography HA Hyaluronic acid HAc Acetic acid HL1 Power used for high pulse H M Q C Heteronuclear multiple quantum coherence X X V HPC Hydroxypropylcellulose HPC-g-PCL Hydroxypropylcellulose-graft-poly(s-caprolactone) HPLC High performance liquid chromatography i.p. Intraperitoneal i.v. Intravenous ICH International Conference on Harmonization ICH3 Peak intensity of methyl groups I H a Peak intensity of a protons I H E Peak intensity of s protons IHI Peak intensity of the anomeric protons at C l / Thickness of the diffusion device LB Line broadening Mo Total amount of drug loaded initially MeOH Methanol MePEG Methoxy poly(ethylene glycol) M S C L Molar substitution of s-caprolactone on hydroxypropylcellulose M S H P Molar substitution of hydroxypropyl group on cellulose M t Amount of drug release at time t NMR Nuclear magnetic resonance spectroscopy NS Number of scans NSAIDs Non-steroidal anti-inflammatory drugs ORC Oxidized regenerated cellulose PI Pulse actually used xxvi P 2 0 5 Phosphorus pentoxide PBS Phosphate buffered saline PC Integral, peak picking frequency PCL Poly(e-caprolactone) PDI Polydispersity index PDLLA Poly(D,L-lactide) PEG Poly(ethylene glycol) PEO Poly(ethylene oxide) PEO200 Poly(ethylene oxide) with molecular weight of 200,000 PEO900 Polyethylene oxide) with molecular weight of 900,000 PGA Poly(glycolide) PLA Poly(lactide) PLGA Poly(lactide-co-glycolide) PMN Polymorphonuclear leukocytes PROBHO Probe PTX Paclitaxel PULPROG Pulse program RG Receive gain rpm Revolutions per minute SD Standard deviation SF Setting of the frequency SF01 Transmitter frequency SI Size of spectrum xxvii Sn(0ct)2 Stannous 2-ethylhexanoate SSB Sign shift SWH Sweep width in Hertz t Time TC7 Interceed®, oxidized regenerated cellulose barrier TD Time domain size T E Temperature Tg Glass transition temperature THF Tetrahydrofuran Tm Melting point tPA Plasminogen activator USP United States Pharmacopoeia U V Ultraviolet WDW Window function Wgraft The weight of graft copolymer WHPC The initial weight of hydroxypropylcellulose X H Molar fraction of hexanes in the mixture of hexanes and tetrahydrofuran XRD X-ray diffraction xxviii ACKNOWLEDGEMENTS I would like to thank Dr. Helen Burt, my supervisor, for her guidance, help and support throughout the past five years. Her faith in my abilities has been crucial in enabling me to surmount cultural barriers and other difficulties during this program. Her dedication to high academic standards has had and will continue to have an influence on me. I would also like to thank my committee members for their direction and advice on my research project: Drs. Frank Abbott, Colin Fyfe, Goran Fernlund, and Marc Levine. Thank you to Mr. John Jackson for his technical support and his invaluable opinion on life. I am grateful to Dr. Richard Liggins for his help with the gas chromatography experiment and his advice on one of the manuscripts. I would also like to express my appreciation to my colleagues in the laboratory for their friendship, wisdom and help: Chuck Winternitz, Chris Springate, Wesley Wong, Jingfang Wang, Linda Liang, Jason Zastre, Kelvin Letchford, Tobi Higo, and Vivian Lee. I am sincerely grateful to Ms. Danielle Wenkstern for teaching me the surgical techniques and for her friendship. Also, I am grateful to Dr. Philip Toleikis and Dr. Laurette Burgess for their generous help with the animal experiments. Thank you to Drs. Christopher Tudan and Christine Allen for their friendship and encouragement. I have thoroughly enjoyed the discussions we had on scientific research and career. I also wish to acknowledge the help with NMR experiments by Ms. Liane Darge, Ms. Marietta Austria, Mr. Ziwei Xiao, and Mr. Lu Yang, and the help with in vitro degradation and drug release studies by Ms. Agatha Ng. xxix I would like to thank my parents for always believing in me. Their encouragement has been a key factor keeping me motivated and inspiring me to strive to greater heights in my life. Thank you to my wife, Jing, for her tireless and continuous support. I am grateful to University of British Columbia and British Columbia Science Council for financial assistance, and Angiotech Pharmaceuticals for funding the research project. xxx CHAPTER 1 PROJECT OVERVIEW AND BACKGROUND 1.1 INTRODUCTION Postsurgical adhesions are a common and serious surgical problem resulting in considerable morbidity and economic loss. Despite an increased understanding of the mechanisms of normal and abnormal wound healing, progress in the prevention of the formation of postsurgical adhesions has been limited. To date, two major strategies, physical barriers and drug treatment, have been used for prevention of adhesion formation. The concept behind the use of physical barriers is to mechanically separate the traumatized surfaces. Several types of barrier based on absorbable polymers have been developed and evaluated both in human and animal models. Drug treatment of postsurgical adhesions is based on the fact that some drugs can modulate various aspects of adhesion formation. The drugs that have been evaluated include nonsteroidal anti-inflammatory drugs such as tolmetin and ibuprofen, fibrinolytic drugs such as tissue plasminogen activator (tPA), and anti-coagulants such as heparin. However, none of these approaches have been proved to be uniformly effective in preventing postsurgical adhesion formation. Paclitaxel (PTX), a potent anticancer drug, has been shown to possess anti-angiogenic and anti-proliferative activities. Angiogenesis and fibroblast proliferation are two processes involved in the pathophysiology of postsurgical adhesion formation. Paclitaxel, administered either by intraperitoneal injection of a Cremophor formulation 1 (Taxol ) or by implantation of paclitaxel loaded, crosslinked hyaluronic acid (HA) films, was recently shown to be an effective inhibitor of adhesion formation in a rat cecal sidewall abrasion model (Jackson et al. 2002). However, this study found that there was no barrier effect exerted by the control (paclitaxel absent) crosslinked H A films and due to its hydrophobicity, paclitaxel was present as large crystals and aggregates within the HA matrix (Jackson et al. 2002). This work aimed at developing more suitable paclitaxel loaded polymeric film formulations to prevent surgical adhesions. Early work focused on the use of chitosan and blends of chitosan and poly(ethylene glycol) as polymeric carriers of paclitaxel. The major part of this thesis involved the synthesis and characterization of novel amphiphilic graft copolymers for the incorporation of paclitaxel. The graft copolymers were based on the hydrophilic polysaccharides, hydroxypropylcellulose (HPC) and dextrans (Dx), as the main chain polymers and hydrophobic poly(s-caprolactone) (PCL) as the side chain polymer. HPC, Dx, and PCL are all biocompatible and biodegradable and have been extensively used in the drug delivery systems and in many other biomedical applications. It was hypothesized that the grafting of PCL onto the main chain would confer increased hydrophobicity on the hydrophilic polysaccharides and make these novel biomaterials more suitable for the incorporation of hydrophobic paclitaxel. 1.2 SURGICAL ADHESIONS 1.2.1 The problem and incidence Postsurgical adhesions are abnormal attachments between tissues and organs that form after surgery (Wiseman 1994). Any trauma involved in the surgical procedures brings about inflammation and bleeding which may eventually result in adhesion 2 formation. Trauma can be caused by abrasion, desiccation ischemia, infection, exposure to foreign materials, overheating by lamps or irrigation fluid (Risberg 1997). Even when doctors follow strict surgical protocols, postsurgical adhesions can occur at 40% to 70 % of surgical sites. In the United States, 70% to 90% of the 1.8 million abdominal surgeries and 55% to 100% of 1.7 million gynecological surgeries performed annually result in adhesion formation (O'Reilly 1997). The problems caused by adhesions are varied and depend on the surgical site in question. These include: • Abdominal surgery: Between 67% and 93% of all laparotomies result in adhesion formation (Weibel and Majno 1973). In intestinal surgery, adhesions involving intestinal loops account for about one third of all cases of intestinal obstruction (Ellis 1971; Ellis 1982; Mcentee et al. 1987). • Gynecological surgery: Adhesion formation between pelvic structures is one of the major causes of infertility (Bronson and Wallach 1977; Trimbos-Kemper et al. 1985). Peritoneal adhesions are also a significant cause of pelvic pain. Adhesion formation after Cesarean section may cause problems on subsequent Cesarean deliveries. These include longer time to delivery, longer operation time, and increased risk of injury to bladder and uterus on re-entry (Kirkinen 1988; Seifer et al. 1990; Pados and Devroey 1992). • Cardiac surgery: In repeat cardiac surgery, adhesions between the heart or aorta and the sternum make sternotomy a hazardous, sometimes catastrophic procedure. Fibrous tissue may also fill the pericardial space, impairing the normal movement of the heart. Right ventricular function may be impaired because of adhesions between the right ventricle and the chest wall (Bailey et al. 1984). 3 • Spinal surgery: After lumber laminectomy, fibrous scar tissue adheres the dura and nerve roots to the disk, erector spine muscles and the vertebral body. This scarring may be the source of low back pain, and sphincter disturbance. The movement of the nerve root is restricted and vulnerable to stenosis and the disk is prone to recurrent protrusion (Pheasant 1975). • Cranial surgery: Many cranial surgical procedures require more than one operation. Adhesions involving the skull, dura, and cortex will complicate these procedures, increase intra-operative bleeding, and the risk of cerebral injury (Meddings et al. 1992). • Ocular surgery: Adhesions involving extraocular structures, especially the muscles, usually cause abnormalities in eye movement (strabismus) and therefore vision (Yaacobi a/. 1992). 1.2.2 Pathophysiology of adhesion formation in the peritoneal cavity Injury to the intact peritoneum is the initial step in the cascade of events that leads to the formation of adhesions. Following damage to very sensitive mesothelial cells, there is a flow of mediators, such as serotonin, bradykinin, histamine, and prostaglandin, which leads to an increase in vessel permeability and results in the extravasation of serosanguine liquid in the abdominal cavity. Through the concomitantly extravasated fibrinogen, fibrin develops via the influence of released tissue thrombokinase, which goes on to form a loose three-dimensional network. This finally causes the adhesion of the adjacent serosal surfaces (Ellis et al. 1965; Buckman et al. 1976). Adhesion formation is a variant of the normal physiologic healing process in which fibrin is broken down into fibrin degradation products by triggering the fibrinolytic 4 system. Plasminogen is activated by endogenous activators as well as by the cytokinase of the mesothelial cells themselves. This step is the prerequisite for rapid reserosing and thus for adhesion-free healing. In abnormal healing and adhesion formation, the dynamic balance between fibrin formation and fibrinolysis is altered. The spontaneous fibrinolytic activity of the peritoneum is reduced by trauma and above all by ischemia (Ellis et al. 1965). The plasminogen activator activity is decreased. As a result, fibrinous adhesions are not lysed, and a tighter fibrin network forms. The unlysed fibrin provides a matrix that is invaded by fibroblasts and eventually blood vessels, resulting in its organization and permanent adhesion formation (Gomel et al 1996). 1.2.3 Approaches to adhesion prevention Theoretically, any measure that limits surgical trauma and fibrin deposition will reduce adhesion formation. This can be accomplished by minimizing the invasiveness of the surgery, for example, choosing laparoscopy over open surgery. The use of talc-free gloves, non-linting cotton, swabs and less reactive suture materials can reduce tissue reactions and granulomatous responses, and thus reduce the probability of adhesion formation. In addition, handling tissue carefully, avoiding desiccation and the use of overheated irrigation solutions can minimize ischemia, which is a very important cause of adhesion formation. However, improving surgical techniques alone cannot effectively eliminate adhesion formation (Gomel et al. 1996). A variety of physical barriers and drugs have been adopted for this purpose. 5 1.2.3.1 Use of barriers to prevent adhesions Generally, adhesion formation occurs only when one traumatized peritoneal surface attaches to another or to the omentum (Haney and Doty 1994; Risberg 1997). The concept behind the use of barriers is to mechanically separate the traumatized surfaces. Once re-epithelialization has occurred, the barrier is no longer needed, because the neomesothelium itself is no longer susceptible to adhesions. The barriers can be in the form of a liquid or solid. When a liquid barrier is used, the organ is floated and separated in the liquid, consequently reducing the attachment of tissues. The interposition of a solid barrier, such as a film or a sheet, between two attaching surfaces is the most direct approach for the prevention of adhesions. Some successfully used barriers are listed below: 1.2.3.1.1 Oxidized regenerated cellulose barrier Interceed®(TC7) (Johnson & Johnson Medical Inc., New Brunswick, New Jersey, US) absorbable adhesion barrier is composed of oxidized regenerated cellulose (ORC) that has been knitted in a special pattern so as to leave only small interstices between the strands (Diamond et al. 1994). ORC is produced by the oxidation of rayon fabric by nitrogen tetroxide. The oxidation converts 50-75% of the primary hydroxyl groups of cellulose into carboxylic acids and renders the beta 1,4-glycosidic linkages susceptible to enzymatic and aqueous hydrolysis. When applied to tissue or organ, the Interceed barrier adheres to the surface without requiring suturing and subsequently gelates over a short period of time. Gelation begins to occur within 2 hours and forms a continuous hydrophilic film after approximately 8 hours. Macroscopically, the Interceed® barrier 6 disappears from a peritoneal implantation site between 3 and 10 days. There is no evidence of the barrier material after 28 days (Diamond and Nezhat 1993). The Interceed® barrier has been evaluated in several animal abdominal or pelvic models with both positive (Steinleitner et al. 1992; Montz et al. 1993) and negative (Best et al. 1992; Haney and Doty 1992; Pagidas and Tulandi 1992; Rice et al. 1993) results. However, it was shown to be effective in several clinical trials. Patients undergoing pelvic, tubal and ovarian surgery were treated with Interceed barriers (Sekiba 1992) which were effective in reducing postsurgical adhesions (Diamond et al. 1994). 1.2.3.1.2 Expanded polytetrafluoroethylene (ePTFE) barrier Because of its excellent biocompatibility, ePTFE has been used clinically for many products, including sutures, vascular grafts, soft tissue reinforcements and prosthetic ligaments. Gore-Tex surgical membrane (Preclude™, W. L. Gore & Associates, Flagstaff, AZ, USA) is ePTFE woven in sheets with a thickness of 0.1 mm and pore size of less than 1 micron. The membrane has been successfully used in one multicenter, non-controlled study (Rodgers et al. 1994) and is indicated for use in gynecological surgery. It is a non-degradable, inert material that causes little or no foreign body reaction even eight years after implantation (group 1992). However, it has several drawbacks that have limited its application. These include its hydrophobic nature, poor tissue adherence and nonbiodegradability. For these reasons, the barrier must be sutured in place as a permanent foreign material. This increases the risks of adhesions at sites of suturing or stapling and of infection. Also, the Gore-Tex membrane is difficult and time-consuming to apply, especially when it is applied laparoscopically. 7 1.2.3.1.3 Hyaluronic acid-carboxymethylcellulose (HA-CMC) barrier HA and C M C have been chemically modified and combined into a new bioabsorbable barrier (Seprafilm™, Genzyme Corporation, Cambridge, M A , USA), which is nontoxic, nonimmunogenic and biocompatible (Burns et al. 1997). H A - C M C barrier turns to a hydrophilic gel within one day of placement, and effectively coats the traumatized tissues for up to seven days during re-epithelialization. It disappears from the implant site after 28 days. It does not require suturing and has significantly reduced the incidence, extent and severity of postsurgical adhesions in both animal models and humans. 1.2.3.1.4 Poloxamer 407 barrier Poloxamer 407 (Pluronic F127) is a nonionic surfactant composed of a triblock copolymer of ethylene oxide and propylene oxide. Poloxamer is used as a barrier because of its 'reverse thermal gelation' where an aqueous solution is liquid at, or below, ambient temperature, and a gel is formed at higher temperatures (body temperature). Gels of poloxamer 407 at concentrations above 25% have reduced the incidence and/or extent of adhesions in animal models (Leach and Henry 1990; Steinleitner et al. 1991). The effectiveness of the barrier may, however, be reduced by the presence of blood and Ringer's solution, because it is readily removed by washing. Other barriers used in the prevention of postsurgical adhesions include: photopolymerized hydrogel, hyaluronic acid solution, dextran 70 solution (Wiseman 1994). 8 1.2.3.2 Use of drugs to prevent adhesions Adhesion formation has been treated by several types of drugs. These drugs can be directed against various causes and components of the inflammatory response and adhesion formation, such as: infection, exudation, coagulation, fibrin deposition and fibroblast proliferation. 1.2.3.2.1 Non-steroidal anti-inflammatory drugs (NSAIDs) NSAIDs modulate several aspects of inflammation, such as vascular permeability, fibrinolysis and macrophage function and have been used to prevent surgical adhesions. These drugs, such as tolmetin (Rodgers et al. 1990; Rodgers et al. 1990; LeGrand et al. 1994) and ibuprofen (Bateman et al. 1982; De Leon et al. 1984; Jarrett and Dawood 1986), reduce the formation of surgical adhesions in animal models. For example, tolmetin sodium has been shown to reduce adhesion formation after a single intraperitoneal dose (LeGrand et al. 1994) delivered at the time of surgery, as well as after continuous administration via an osmotic minipump (Rodgers et al. 1990). 1.2.3.2.2 Fibrinolytic drugs A number of fibrinolytic drugs have been evaluated for prevention of postsurgical adhesions in animal models. These drugs, such as tissue plasminogen activator (tPA) (Doody et al. 1989), streptokinase (Wiseman et al. 1992) and urokinase (Rivkind et al. 1985) have been delivered intraperitoneally and shown to reduce adhesion formation. In a rabbit model, tPA was delivered directly to the site of adhesion formation using an osmotic minipump and adhesions were eliminated after two days of treatment (Orita et al. 1991). A very low dose was required for this effect compared with the doses used after intraperitoneal administration (Doody et al. 1989; Wiseman et al. 1992). 9 Fibrin deposition is not only an important component of adhesion formation, but is also critical in wound healing and hemostasis. Therefore, the use of fibrinolytic drugs carries a risk of impairing wound healing and/or inducing bleeding (Wiseman 1994). 1.2.3.2.3 Anti-coagulants Heparin has long been used as an anti-coagulant, and is believed to reduce adhesions by altering the deposition and clearance of fibrin from a surgical site. It exerts a variety of effects on the clotting and fibrinolytic systems as well as on cellular function. Given three doses daily, heparin reduced adhesion formation in an animal model (Wiseman 1994). However, a single dose into the abdomen after surgery failed to reduce adhesions in either humans or rabbits (Wiseman 1994). Bleeding is a potential problem with heparin therapy. The use of an efficacious, low-dose, and site-specific delivery system would help to alleviate these concerns. It is apparent that drug treatment of postsurgical adhesions must overcome several problems. Firstly, delivery of the drug to the specific site to increase the local drug concentration, enhance efficacy and avoid the side effects associated with systemic administration. Secondly, protection of the drugs against rapid clearance and thirdly, control of drug concentration at the site such that adhesions are prevented but normal wound healing can proceed. 1.3 PACLITAXEL 1.3.1 Chemistry Paclitaxel is a natural alkaloid, isolated from the pacific yew tree, Taxus brevifolia (Wani et al. 1971). The structure of paclitaxel (shown in Figure 1.1) consists of a complex taxane ring system linked to an oxetane ring at position C-4 and C-5, and to an 10 ester side chain at C-13. Both the oxetane ring and the ester side chain are essential for cytotoxicity and paclitaxel's unique microtubule effects (Kingston et al. 1990). Analogues prepared with an opened oxetane ring show a greatly reduced activity. The oxetane ring is believed to serve as a hydrogen-bond acceptor and a rigid "lock" on the taxoid skeleton (Kingston 2000). Loss of the ester side chain occurs by mild basic hydrolysis, the process being slowed when paclitaxel is bound to proteins or microtubules (Kingston et al. 1990). Figure 1.1. Chemical structure of paclitaxel Paclitaxel can be readily dissolved in organic solvents, such as methanol, ethanol, methylene chloride and acetonitrile. However, solubility of paclitaxel in aqueous media is very low because of its hydrophobic nature. Aqueous solubility values ranging from 0.7 pg/ml to 30 pg/ml have been reported (Straubinger 1996). Studies in this laboratory have shown the equilibrium water solubility of paclitaxel to be around 1 pg/ml (Liggins et al. 1997). In its commercial formulation, paclitaxel is dissolved in a nonaqueous vehicle composed of 50:50 Cremophor EL and alcohol (Taxol ®). Cremophor EL, a nonionic surfactant, is a polyoxyethylated castor oil that has been used to dissolve several other water-insoluble drugs (Dorr 1994). 11 It has been reported that paclitaxel converts primarily to 7-epitaxol, a thermodynamically more stable isomer (Huang et al. 1986), upon heating in the dry state (Richheimer et al. 1992) or dissolving in organic solvents such as dimethyl sulfoxide, methanol, isobutanol, and chloroform (MaacEachem-Keith et al. 1997). 7-epitaxol retains over 90% of the activity of paclitaxel (Ringel et al. 1994). The stability of paclitaxel in aqueous media was found to be pH dependent. The most stable region is pH3-5 (Dordunoo and Burt 1996). At constant temperature, the degradation of paclitaxel follows pseudo first order kinetics. Complexaion with cyclodextrin can improve the stability of paclitaxel in aqueous solution (Dordunoo and Burt 1996). 1.3.2 Pharmacology Paclitaxel is indicated for the treatment of ovarian (McGuire et al. 1989) and breast (Rowinsky et al. 1990) cancer. Substantial antitumor activity of paclitaxel has also been demonstrated in women with metastatic breast cancer (Holmes et al. 1991; Reichman et al. 1993). It acts by a concentration-dependent and reversible binding to microtubules, specifically the beta subunit of tubulin at the N-terminal domain (Rao et al. 1994). The binding results in the formation of extraordinarily stable and dysfunctional microtubules which disrupts the normal microtubule dynamics required for cell division and blocks cells in the late G2 and M phase of the cell cycle, thereby causing the death of cells (Rowinsky and Donehower 1995). Paclitaxel has also been evaluated and shown antitumor activity in patients with other tumor types, such as lung cancer (Murphy et al. 1993), head and neck cancer (Forastiere et al. 1993), and esophageal cancer (Ajani et al. 1994). 12 Paclitaxel has been demonstrated to possess antiangiogenic activity (Burt et al. 1995; Belotti et al. 1996). Angiogenesis is the proliferation of blood vessels and capillaries within human body and is mediated by several angiogenic factors such as fibroblast growth factor and endothelial cell growth factor (Folkman and Klagsbrun 1987). Angiogenic factors bind to endothelial cell receptors in capillaries and initiate the process of angiogenesis by stimulating endothelial cell division, secretion of proteases which digest the basement membrane surrounding the vessel and promotion of endothelial cell migration to form the growing "sprout". Paclitaxel inhibits several steps involved in angiogenesis such as cell division, migration and collagenase secretion (Stearns and Wang 1992). Paclitaxel has been applied to the chick chorioallantoic membrane (CAM), an immunologically protected site in the living egg where it is possible to observe the development of blood vessels. An avascular region was formed within 48 hours (Winternitz et al. 1996). Paclitaxel has been proven to be a potent antiproliferative agent (Schiff and Horwitz 1980; Gloushankova et al. 1994; Matsuoka et al. 1994; Axel et al. 1997). The antiproliferative effects of paclitaxel on mouse fibroblast cells were first reported by Schiff and Horwitz (Schiff and Horwitz 1980). The antiproliferative effects have also been evaluated in tumor cell lines (Matsuoka et al. 1994), epithelial cells (Gloushankova et al. 1994), and arterial smooth muscle cells (Axel et al. 1997). 1.3.3 Toxicity Neutropenia is the principal toxic effect of paclitaxel (Rowinsky et al. 1993). At doses of 200 to 250 mg of paclitaxel per square meter given over a period of 24 hours, neutropenia is usually severe even in previously untreated patients. The onset of 13 neutropenia is usually at day 8 to 10 after treatment, and recovery is usually complete by day 15 to 21. Neutropenia is not cumulative, suggesting that paclitaxel does not irreversibly damage immature hematopoietic cells (Rowinsky and Donehower 1995). Hypersensitivity reactions were encountered during the early clinical studies (Rowinsky and Donehower 1995). The incidence was high, approaching 25 to 30% in some studies (Kris et al. 1986; Grem et al. 1987; Wiernik et al. 1987). Most affected patients had type I hypersensitivity reactions, including dyspnea with bronchospasm, urticaria, and hypertension. The polyoxyethylated castor oil vehicle (Cremophor EL) was thought to be responsible for the hypersensitivity reaction, since other drugs formulated in polyoxyethylated castor oil, such as cyclosporine and vitamin K, have been associated with similar reactions (Weiss et al. 1990). Other toxic effects induced by paclitaxel include neurotoxicity (Rowinsky et al. 1993; Rowinsky et al. 1993), cardiac toxicity (Rowinsky et al. 1991), and gastrointestinal toxicity (Rowinsky et al. 1993). 1.3.4 Pharmacokinetics Upon intravenous administration, paclitaxel binds rapidly and extensively to plasma proteins due to its hydrophobic nature (Jamis-Dow et al. 1993). Over 90% of paclitaxel transported in blood is bound by plasma proteins, primarily albumin (Wiernik et al. 1987; Straubinger 1996). The generally accepted clinical dose is 200-250 mg/m and is given as 3 and 24 hours infusion (Singla et al. 2002). Peak plasma concentration (Cmax) and area under the concentration-time curve (AUC) show marked interpatient variability. The plasma concentration decreases rapidly as paclitaxel distributes to the tissues with a distribution-phase half-life (ti/2a) of the order of 15 to 30 min. A mean i\aa 14 of 20.4 min was calculated based on the data from human clinical trials. The terminal elimination half-life (ti/2 p) ranges from 1.3 to 8.6 hours with a mean of 4,9,hours (Straubinger 1996). Initial plasma clearance is rapid despite extensive plasma protein binding. The apparent volume of distribution is large and increases with increased duration of infusion (Huizing et al. 1993). When paclitaxel is administered intraperitoneally, peak plasma concentration correlates directly with the dose. Apparent volume of distribution is low, suggesting that paclitaxel is initially confined to the intraperitoneal cavity (Markman et al. 1992). Clearance from the peritoneal cavity is slow with significant intraperitoneal concentrations persisting for at least 72 hours (Spencer and Faulds 1994). The peak peritoneal concentration (Cm a x ) occurred 1 hour following paclitaxel administration. A comparison of plasma and intraperitoneal A U C values suggested that peritoneal exposure to paclitaxel was 336 to 2890 times greater than systemic exposure (Markman et al. 1992). Paclitaxel does not appear to easily cross the blood-brain barrier. In rodents, paclitaxel distributed predominantly to the liver, lung, spleen, pancreas, adrenal and salivary glands, stomach, intestine, heart muscles and kidneys, but was not found in the nervous system or testes (Spencer and Faulds 1994). Urinary clearance of paclitaxel is minimal. Hepatic metabolism and biliary clearance appear to be the major means of elimination (Spencer and Faulds 1994). Paclitaxel is metabolized mainly by cytochrome P450 isozyme CYP2C and CYP3A. The major metabolites include 6a-hydroxytaxol and other monohydroxylated metabolites (Cresteil et al. 1994; Kumar et al. 1994). 15 1.3.5 Paclitaxel in the treatment of surgical adhesions Fibroblast proliferation and angiogenesis are two processes involved in the formation of surgical adhesions and since paclitaxel possesses antiproliferative and antiangiogenic activities, it was suggested that paclitaxel might be a suitable agent to prevent the formation of surgical adhesions. Studies were initiated by our group, using the cecal sidewall abrasion model in rats, to determine whether paclitaxel delivered either locally at the site of the abrasions or intraperitoneally (i.p.), would inhibit the formation of adhesions (Jackson et al. 2002). The results showed that Taxol® (Cremophor EL formulation of paclitaxel) administered by i.p. injection at 4 mg/kg on a daily basis for between 3 to 5 days, resulted in a significant reduction in adhesion formation. The application of 5% paclitaxel loaded H A films to the site of abrasions also significantly decreased adhesion formation (Jackson et al. 2002). 1.4 POLYMER CHEMISTRY 1.4.1 Polymer structure Polymers are macromolecules built up by the linking together of large numbers of much smaller molecules, which are termed monomers (Odian 1991). Polymers can be classified as linear, branched, or crosslinked polymers based on the structure of the molecular chains. In the linear polymers, the monomer molecules are linked together in one continuous length to form the polymer molecules. Branched polymer molecules are those in which there are side branches of linked monomer molecules protruding from points along the main polymer chain. When polymer molecules are linked to each other at points other than their ends, the polymers are said to be crosslinked. 16 Polymers can also be classified as homopolymers or copolymers. Homopolymers have only one type of repeating unit, whereas copolymers have two or more different repeating units on their chains. According to the sequence distribution of the two repeating units (A and B) and the linearity of the molecular chain, copolymers are classified as: • Statistical (random) copolymers, in which the two monomers appear in irregular, unspecified sequences along the chain: A B B A B A A A B A B • Alternating copolymers, in which the two monomers occur in an alternating pattern: A B A B A B A B A B A B • Block copolymers, in which long linear sequences of the monomer A are joined to long linear sequences of the monomer B: A A A A A A B B B B B B • Graft copolymers, in which chains of one monomer are pendent (side chain) from a backbone of the other (main chain): A A A A A A A A A A A A A A B B B B B B B B 1.4.2 Molecular weight and molecular weight distribution The molecular weight of a polymer is of prime importance in its application and has fundamental impacts on its properties. For example, the mechanical properties of a polymer depend on, and vary considerably with, molecular weight. When the molecular weight is below a certain value, usually about a thousand, the polymer will not have any 17 significant mechanical strength. Above this value, strength increases with molecular weight and eventually reaches a limiting value (Odian 1991). Properties related to pharmaceutical applications, such as drug permeability and biodegradation rate, also show a significant dependence on molecular weight. Polymers differ from low-molecular weight compounds in that they are polydisperse in molecular weight. Even in their purest form, polymers are mixtures of molecules of different molecular weights. The molecular weight of a polymer actually refers to its average molecular weight. Both the average molecular weight and the distribution of different molecular weights within a polymer are essential in its characterization. Various methods based on solution properties are used to determine the average molecular weight of a polymer sample. These include the methods based on colligative properties, light scattering, and viscometry (Billingham 1977). The various methods yield different average molecular weights., 1.4.2.1 Number-average molecular weight (Mn) Mn is defined as the total weight of all the molecules in a polymer sample divided by the total number of moles present, i.e., _ w YNM Mn- — = -^ =j Equation 1.1 N £jV,. where W is the total weight of polymer sample, N is the total numbers of moles, and Nj is the number of moles of molecular weight M;. Mn may be determined by one of the colligative property methods based on counting the number of polymer molecules in a sample of the polymer. These methods 18 include vapor pressure osmometry, membrane osmometry, freezing point depression, and boiling point elevation (Cowie 1991; Odian 1991). The most commonly used methods are vapor pressure osmometry and membrane osmometry since reliable commercial instruments are available. Another method for determining Mn is end-group analysis, which is useful for polymers having end groups that can be readily quantified by titration, infrared or NMR methods (Odian 1991). The use of a colligative property method leads to determination of absolute molecular weight. The disadvantage of using these methods to determine Mn is that they can only be used to measure the samples with Mn below a certain limit. 1.4.2.2 Weight-average molecular weight (Mw ) _ _ YWiMl y/v,.M,.2 Mw is defined as: M w = — = ^ Equation 1.2 w Y,NiMi where W; is the weight of molecules of molecular weight Mi. Mw may be determined by light scattering techniques (Cowie 1991). The method is more accurate for polymers with higher molecular weight, since the amount of light scattered by a polymer solution increases with molecular weight. Monomers and other low-molecular-weight impurities do not influence the measurement unless they are present at levels sufficient to dilute the solution or to change the refractive index of the solution. There is no upper limit to the molecular weight that can be accurately measured except the limit imposed by insolubility of the polymer. 1.4.2.3 Viscosity-average molecular weight (MJ M„ is defined as: M„ -Ma Ma w I>,.M,. Equation 1.3 19 where a is a constant, which is dependent on the hydrodynamic volume of the polymer and the effective volume of the solvated polymer molecule in solution, and varies with polymer, solvent and temperature. Its value is usually in the range of 0.5-0.9 (Odian 1991). Mv is determined by measuring the viscosity of a polymer solution. 1.4.2.4 Polymer solution viscometry The viscosity of a fluid is a measure of its resistance to flow when a shearing force is applied. Based on Poiseuille's equation, the viscosity (rj) measured using a capillary viscometer can be described by Equation 1.4 (Lovell 1989): rj = Apt- Bp/1 Equation 1.4 where A and B are viscometer constants, p is the density of the fluid, and t is the time taken for the fluid to flow through the capillary. The second term on the right-hand side of Equation 1.4 is a kinetic energy correction. When relative measurements are made, it is common practice to assume that the solution density p is equal to the solvent density p 0 and that the kinetic energy correction is negligible (Lovell 1989). The terminologies used in polymer solution viscometry are defined and summarized in Table 1.1, in which r/0 is the viscosity of pure solvent, and TJ is the viscosity of a polymer solution of concentration c. By definition, intrinsic viscosity [TJ\ is the jjred at infinite dilution, which eliminates the effects of intermolecular polymer-polymer interaction. Therefore, theories of the solution behavior of isolated polymer molecules can be applied, and [ rj\ can be described in terms of molecular weight and hydrodynamic volume of the polymer molecules, [rj] and Mv are related by the Mark-Houwink equation (Lovell 1989): [q]=K>Mva Equation 1.5 20 where K is a constant for a given polymer/solvent/temperature system. Table 1.1 Terminologies used in the viscometry of polymer solution Name Symbol and definition Relative viscosity 7r = 7]/TJo Specific viscosity Usp z = (V-Vo)/ Vo = TJr-\ Reduced viscosity rjred = fjsp/c Inherent viscosity TJinh = (ln/7r)/c Intrinsic viscosity [>?]-~ ( Tjred)c-,a ~ ( ?]inh)c^>0 Two methods have been used to determine [rj\: the extrapolation method and the single point method. The extrapolation method involves measuring Tjr, giving rjsp, of a polymer solution at a series of concentrations. The experimental data are plotted according to the Huggins and Kraemer equations (Collins et al. 1973): Huggins: rjsp /c = [7] + k i [rjf c Equation 1.6 Kraemer: (Inrjr) / c = [ 7] - k2 [ rjf c Equation 1.7 where k i and k2 are the Huggins and Kraemer constants, respectively, and in most of the cases, ki+k2 « 0.5. A typical plot is shown in Figure 1.2. The intercepts of the two lines give the best estimation of the [7] value. The inconvenience of the extrapolation method for routine analysis has given rise to considerable interest in the estimation of [7] using single point methods. One of the frequently used equations (Solomon and Ciuta 1962; Rao and Yaseen 1986) is: 21 This equation is only valid for solutions of polymer in a good solvent, and the concentration should be chosen such that rjsp < 0.2. c (g/cm3) Figure 1.2 A typical plot of extrapolation methods using Huggins and Kraemer equations c is the concentration of the polymer solution, rjred is reduced viscosity, and rjinh is the inherent viscosity 1.4.2.5 Molecular weight distribution A polymer is said to be monodisperse when its molecules all have the same molecular weight, i.e., Mn - Mw = Mv. However, for practical purposes, all polymers are polydisperse. The relationship among the three average molecular weights is: M >M> M„ with the differences between the various average molecular weights increasing as the molecular weight distribution broadens. A typical molecular weight distribution curve is shown in Figure 1.3. The ratio of weight- and number-average molecular weights Mw I Mn is used as a measure of the polydispersity of a polymer. The value of Mw I Mn would be unity for a perfectly monodisperse polymer. It is greater than unity for all real polymer systems and increases with increasing polydispersity. The 22 polydispersity of a polymer is usually measured by gel permeation chromatography (Janca 1984). o bt) Molecular weight, M x Figure 1.3 A typical molecular weight distribution curve (adapted from: Odian 1991) 1.4.2.6 Gel permeation chromatography (GPC) GPC involves the permeation of a polymer solution through a column packed with microporous beads of crosslinked polymers such as polystyrene (Janca 1984). The packing contains beads of different-sized pore diameters. Molecules pass through the column by a combination of transport into and through the beads and through the volume between the beads. The smaller-sized molecules (smaller than the pore size) can penetrate the beads, thus it takes them longer time to pass through the column. The larger-sized molecules cannot penetrate the beads, and can only pass through the volume between the beads, therefore they pass through the column faster. The time for passage of the polymer molecules through the column decreases with increasing molecular weight. By this mechanism, polymer molecules with different molecular weights are separated in the column of the GPC. A detector (usually a refractive index detector) is used to measure the amount of polymer passing through the column as a function of time. A calibration 23 curve for a given column may be constructed by measuring the retention time of a series of monodisperse polymer standards of known molecular weight (MW) and plotting log(MW) versus retention time. After calibrating the column, the molecular weight (both Mn and Mw) and the polydispersity of polymer samples can be determined (Collins et al. 1973). 1.4.3 Morphology of solid polymers Most polymers show, simultaneously, the characteristics of both crystalline solids and highly viscous liquids. Polymers generally constitute a spectrum of materials from those that are completely amorphous, to semicrystalline polymers with crystallinity ranging from low to high. The nature of polymer crystallinity has been extensively investigated. The traditional model used to explain the properties of the semicrystalline polymers is the "fringed micelle model" (Krevelen 1976). This model assumes that polymers consist of Figure 1.4. Schematic representation of the fringed micelle model 24 small-sized, ordered crystalline regions embedded in a disordered, amorphous polymer matrix. Crystallites are formed when chain segments from different polymer chains are precisely aligned together and undergo crystallization. Each chain can pass through several different crystalline regions and contribute ordered segments to several crystallites. The segments of the chain in the regions between the crystallites make up the disordered amorphous matrix (Krevelen 1976). A schematic representation of the "fringed micelle model" is shown in Figure 1.4. 1.4.4 Thermal transitions Polymeric materials are characterized by two major types of thermal transitions, the glass transition and crystalline melting transitions. Completely amorphous polymers show only a glass transition. A completely crystalline polymer (rarely encountered) shows only a melting transition. Semicrystalline polymers exhibit both the crystalline melting and glass transitions (Odian 1991). Amorphous solid polymers are either in the glassy state or in the rubbery state. At low temperatures (below the glass transition temperature), the motion of the molecular chains is frozen. The polymer is rigid and glassy, characterized by a high modulus. With an increase in temperature, the chains gain energy, and the chain segments move resulting in a transition from the glass to the rubber state. This temperature is known as the glass transition temperature (Tg). Molecular structure has a significant effect on Tg. Chains based on Si-O, P-N, C-C, and C-0 are flexible and have low values of Tg (Pitt et al. 1980; Cowie 1991). Ring structures, e.g. p-phenylene groups, in the chain increase the Tg. Substituents, particularly rigid, polar, or branched substituents, will generally increase Tg by impeding intramolecular motion or increasing intermolecular interaction 25 by van der Waals forces, dipolar interactions, or hydrogen bonding (Cowie 1991). Long chain alkyl substituents can reduce Tg by self-plasticization. In addition, plasticizers will often reduce the Tg (Cowie 1991). The melting of a semicrystalline polymer takes place over a wider temperature range than that observed for small organic compounds due to the presence of crystalline regions of different sizes and the more complex process for melting of large molecules. The crystalline melting temperature, Tm, is generally reported as the point of melting of the highest melting crystallites, that is, the point of disappearance of the last traces of crystallinity (Collins et al. 1973). The effect of polymer structure on Tm is similar to that on Tg. Polymers with low Tg values usually have low Tm values. The two thermal transitions are generally affected in the same manner by the molecular symmetry, structural rigidity, and secondary forces of polymer chains. The Tm values of crystalline polymers produced from rigid chains would be high. High secondary forces lead to strong crystalline forces requiring high temperatures for melting. Polymers such as chitosan and cellulose have high Tm values due to the high rigidity and strong hydrogen bonding in their molecules. Symmetrical and regular structure generally yields tightly packed crystallites with a high Tm value (Cowie 1991). 1.4.5 Modification of polymer properties 1.4.5.1 Physical blends The most direct and versatile method for modifying polymer properties is the physical blending of two or more polymers. The properties and, therefore, the utility of physical blends are strongly dependent upon the degree of compatibility of the components. Only a very small number of amorphous polymer-polymer pairs are 26 thermodynamically compatible, i.e., truly soluble in each other. The great majority of the physical blends are incompatible (Noshay and McGrath 1977). The incompatibility of the blend components provides a driving force for each to aggregate and form separate phases. These two-phase systems are coarse dispersions in which the particles are large, non-homogeneous, and characterized by poor interphase adhesion (Noshay and McGrath 1977). 1.4.5.2 Copolymers Copolymerization is the most general and powerful method of effecting modifications in polymer properties, in which two components are covalently connected to each other. Copolymerization modifies the symmetry of the polymer and modulates both intramolecular and intermolecular forces, so properties such as Tg, Tm, crystallinity, solubility and permeability may be varied within wide limits (Eastmond et al. 1989). By using different polymerization methods, four types of copolymers can be synthesized. These are random copolymers, alternating copolymers, block copolymers and graft copolymers. Random copolymers and alternating copolymers are the most common types of copolymer. These copolymers display single-phase morphology and properties representing a weighed average of the two repeat units. Block copolymers and graft copolymers are similar in many basic characteristics. Usually, the materials of these types of copolymer exhibit two-phase morphology, but this occurs on a micro-scale rather than the macro-scale dimension of incompatible physical blends. This is due to the influence of the intersegmental linkage, which restricts the extent of phase separation (Noshay and McGrath 1977). 27 1.4.5.3 Graft copolymers All graft copolymers possess two general structural features: a supposedly linear main chain of polymer " A " to which a number of polymer "J3" side chains are grafted. The major justification for utilizing graft copolymers is that a graft copolymer is a single chemical species that displays the properties characteristic of each of the components, rather than an averaging of their properties. The properties exhibited by the side chain have led to a number of applications as emulsifiers, surface-modifying agents, and coating materials in polymer blends (Ceresa 1973). Amphiphilic graft copolymers, in which the solubility behavior of the main chains and side chains are quite different, play an important role as emulsifiers or solubilizing agents. Although a single-phase morphology is possible in graft copolymers, two-phase morphology is much more commonly observed (Stannett 1970). The morphology is greatly dependent on the volume fraction of the graft and backbone species. The component present in the higher concentration will normally form the continuous phase and thereby greatly influence the physical properties of the copolymer. In compositions containing nearly equal concentrations of both components, phase continuity can be dramatically altered by varying the formulation fabrication conditions, e.g., type of casting solvent (Ceresa 1973). This behavior can be used to control properties such as flexibility and permeability of the copolymers. 1.4.5.4 Synthesis of graft copolymers The methods used to synthesize graft copolymers have been classified into three main categories (Dreyfuss and Quirk 1987): the 'graft from' process, the 'graft onto' process, and the 'graft through' process. In the 'graft from' process, a polymer chain 28 (main chain) has active sites attached to it. The polymerization of a second monomer can then be initiated from the active sites to yield the grafts. The active sites on the main chain can be those of free radical, anionic, cationic or Ziegler-Natta type (Dreyfuss and Quirk 1987). In the 'graft onto' process, grafting results from the reaction between a polymer chain carrying randomly distributed reactive functions and another polymer molecule carrying antagonist functions located selectively at its chain ends. Grafting may also result from a growing polymer chain incorporating a pendant unsaturation belonging to another polymer chain. This has been called the 'graft through' process. In this project, the 'graft from' process was used for the synthesis of the graft copolymers. Graft copolymers of polysaccharides with various unsaturated monomers have been described in U.S. Pat. Nos. 5,286,770 to Bastioli et al., 5,254,607 to McBride et al., and 5,268,422 to Yalpani. Graft copolymers of polysaccharides produced from vinyl monomers (such as: styrene, methyl methacrylate) have been prepared by a solution process in which the grafting was initiated by radiation or cerium ions (Bagley et al. 1977). Polysaccharides grafted with aliphatic polyesters was disclosed in U.S. Pat No. 5,540,929 to Narayan et al., in which novel synthetic methods to derive the graft copolymers were described. 1.5 POLYMERIC DRUG DELIVERY SYSTEMS The development of new drug delivery systems is driven by the continuous emergence of new drugs and the need to maximize therapeutic activity while minimizing side effects. A critical component in the design of delivery systems is the selection of appropriate biomaterials used as drug carriers. 29 1.5.1 Polymers as drug carriers The development of biomaterials suitable for drug delivery is an important and growing area and a wide variety of polymers are available as drug carriers. Generally, selection of an appropriate biomaterial requires the consideration of the physicochemical, mechanical, and biological properties of the polymers. Biocompatibility and biodegradability are two important factors in the selection of a suitable drug carrier. Table 1.2 gives the most frequently used polymers in drug delivery systems. 1.5.1.1 Biodegradability Biodegradation is a broad term that refers to hydrolytic, enzymatic or bacteriological degradation processes occurring in a polymer matrix (Piskin 1994). Hydrolytic and enzymatic degradation processes have been considered to be the major mechanisms of in vivo polymer biodegradation. The biodegradability of polymers is mainly dependent on the molecular composition and morphology. To be biodegradable, the polymer chain should contain some susceptible chemical bond. It is well known that the ester bond is hydrolytically unstable. Many aliphatic polyesters have been successfully used in biodegradable polymeric drug delivery systems. Crystalline regions in a semicrystalline polymer are usually degraded more slowly than amorphous regions, because water molecules cannot easily penetrate crystalline regions (Piskin 1994). Polymers with a Tg lower than body temperature exhibit rubbery characteristics in vivo, and degrade faster than glassy polymers with a Tg higher than body temperature. The degradation products can also have an effect on degradation rate. The carboxylic acid end groups generated during hydrolysis of ester bonds can accelerate the degradation process by an autocatalytic mechanism (Piskin 1994). Enzymes usually enhance the degradation 30 Table 1.2 Biodegradable polymers used in drug delivery systems Classification Polymer Chemical structure Notes and Reference Aliphatic polyesters Poly(s-caprolactone) (PCL) Membrane for progesterone (Pitt et al. 1979); Injectable paste for paclitaxel (Winternitz et al. 1996) Polylactide (PLA) Microspheres for prednisolone (Smith 1986) and paclitaxel (Liggins and Burt 2001) Polyglycolide (PGA) Endovascular coils for human vascular endothelial growth factor (Abrahams et al. 2001) Poly(lactide-co-glycolide) (PLGA) Poly(P-hydroxybutyrate) (PHB) Poly(hydroxybutyrate-co-hydroxyvalerate) o Microspheres for proteins (Wang et al. 1999); film for mitomycin-C (Gumusderelioglu and Deniz 2000) Microspheres for rifampicin (Kassabefa/. 1997); implant and microparticulates Koosha, 1989 #297] Microcapsules for tetracycline (Sendil et al. 1999); microspheres for progesterone (Chen et al. 1999) 31 Table 1.2 Biodegradable polymers used in drug delivery systems (continued) Poly anhydrides Poly(sebacic anhydride) Nanoparticles (Fu and Wu ; 2001) Poly(adipic anhydride) Microspheres for ocular drug delivery (Albertsson et al. 1996) Poly(ortho esters) 3,9-bis(ethylidene-2,4,8,10-tetraoxaspiro[5,5]undecan e) based poly(ortho esters) -o o O O R -Injectable semisolid for tetracycline (Roskos et al. 1995) Phosphorus containing polymers Polyphosphazenes R - N = P -R' Microspheres for insulin (Caliceti et al. 2000); film for ethacrynic acid (Allcock etal. 1994) Polyphosphoesters - O R O P O -O R ' Microspheres for PTX (DePalma et al. 2000) Polysaccharides and derivatives Chitosan C H 2 O H -ho- 1 O H \ N H , Microspheres for mitoxantrone (Jameela and Jayakrishnan 1995); films for ketoprofen (Bodek 2000) 32 Table 1.2 Biodegradable polymers used in drug delivery systems (continued) Dextrans (Dx) - O C H , O H J Colon-specific drug delivery system (Chiu et al. 1999); conjugates for antibodies (Rowland 1977) Hydroxypropylcellulose (HPC) O R - , C H 2 -f-o-1 O R i R. \ OR, Tablet formulations of acetaminophen (Harcum et al. 1998) and ascorbic acid (Ishikawa et al. 2001) Hyaluronic acid (HA) C H 2 O H Prodrug of sodium butyrate (Coradini et al. 1999); films for NSAIDs (Luo et al. 2000) Proteins Collagen Protein Antibiotics (Kind et al. 1984; Yeh et al. 2001) Gelatin Protein Microspheres for clonidine hydrochloride (Vandelli et al. 2001) and amoxicillin (Wang et al. 2000) Albumin Protein Microspheres for 5-fluorouracil (Sugibayashi et al. 1979) 33 Table 1.2 Biodegradable polymers used as drug vehicles (continued) Polypeptides c ) Polylysine H N H 2 n Conjugates for extracellular matrix-targeted local drug delivery (Sakharov et al. 2001) Poly(aspartic acid) 0 H — .( O H 0 H n Prodrug for colon-specific delivery of dexamethasone (Leopold and Friend 1995) and for doxorubicin (Pratesi etal. 1985) Poly(y-benzyl glutamate) } - n 0 Slabs for procainamide hydrochloride and protamine sulfate (Markland et al. 1999) of polymers, particularly naturally occurring polymers such as polysaccharides, by enzymatic mechanisms (Piskin 1994). Environment factors, such as the site of implantation or injection, local pH value, and temperature, can also have marked effects on the polymer degradation rate. 34 1.5.1.2 Biocompatibility Biocompatibility is the ability of a material to perform with an appropriate host response in a specific application (Williams 1987). Appropriate host responses vary depending on the type of materials and their intended use. Based on the environment surrounding most implanted polymers, biocompatibility mainly includes the interaction of polymers with blood and tissues, referred as blood compatibility and tissue compatibility, respectively. A desirable host response would be total inertness and no interaction with blood and/or tissues (Park and Park 1996). Inflammation is a typical tissue reaction to the injury caused by the implantation of foreign materials, which is indicated by the symptoms of reddening, swelling, pain, and fever (Park and Park 1996). These symptoms are accompanied by a series of defensive reactions by neutrophils, macrophages and foreign body giant cells. Macrophage cells initiate the repair of the damaged tissue by forming the scaffold for repair, which is called granulation tissue (Park and Park 1996). The granulation tissue starts to surround the implant and foreign body giant cells attach to the surface of the implant. If the implant is not phagocytosed by the cells, the body tends to completely isolate the foreign implant by forming a sheath-like fibrous capsule around the implant (Park 1984). Fibrous encapsulation is a major tissue reaction to a polymeric implant, especially for the long-acting implants (Ward et al. 1995). The fibrous capsule often contracts and causes pain and deformation of the implant. For a drug delivery implant, the fibrous capsule may alter the drug release kinetics. 35 1.5.1.3 Hydrogels as drug carriers Hydrogels are a three dimensional network of hydrophilic polymers crosslinked either by chemical bonds or physical cohesive forces such as ionic interactions, hydrogen bonding, or hydrophobic interactions. The hydrogels crosslinked by covalent bonds are referred to as chemical gels, and those by physical cohesive forces are referred to as physical gels (Park et al. 1993). Because of the high water content, the surface and interfacial properties of hydrogels are similar to those of natural biological gels and tissues and are recognized to be highly biocompatible (Andrade 1976). Protein adsorption and cell adhesion are minimized due to the low interfacial tension with surrounding biological fluids and tissues (Jhon and Andrade 1973) and the high mobility of the polymer chains at the hydrogel surface (Horbett 1986; Ratner 1986). The soft and low friction surfaces result in delivery systems that do not cause pain, damage to tissues, infection or thrombus formation (Nagaoka and Akashi 1990). Hydrogels have been extensively used as drug carriers in various delivery systems. Proteins such as albumin and gelatin have been used to make microspheres for the delivery of drugs such as steroids (Burgess and Davis 1988), insulin (Goosen et al. 1982) and interferon (Tabata and Ikada 1989). Microspheres made of dextran were developed for the delivery of protein drugs (Edman et al. 1980). Due to the presence of a large number of hydroxyl groups, dextrans were frequently used to attach low molecular weight drugs to make prodrugs(Larsen 1989). Hydrogels based on other polysaccharides, such as chitosan (Jameela and Jayakrishnan 1995; Bodek 2000), hyaluronic acid 36 (Coradini et al. 1999; Luo et al. 2000), starch (Baillie et al. 1987), and cellulose ether (Harcum et al. 1998; Ishikawa et al. 2001), are also widely used as drug carriers. 1.5.2 Mechanisms of drug release Polymeric drug delivery systems are generally classified into two types: capsule-type and matrix-type. In a capsule-type of delivery system, the drug is encapsulated inside a polymeric membrane; while in a matrix-type of delivery system, the drug is dispersed throughout a polymer matrix (Chien 1982; Baker 1987). In this thesis, only the matrix-type of delivery systems will be discussed. 1.5.2.1 Diffusion controlled systems In diffusion-controlled systems, a drug is released from a polymer device by permeation from its interior to the surrounding medium. When the drug to be released is dispersed uniformly throughout the rate-controlling polymer, the system is called a monolithic diffusion system (Baker 1987). There are two principal categories of monolithic systems. If the drug is dissolved in the polymer, the system is called a monolithic solution; if the drug has a limited solubility in the polymer, then only a portion of the drug is dissolved in the polymer, and the remainder is dispersed as small particles throughout the polymer, and such a system is called a monolithic dispersion (Baker 1987). A monolithic solution system can be a device with different geometries such as a slab, cylinder, and sphere. The kinetics of the drug release from a slab device has been described by the following equations (Baker 1987; Park et al. 1993): for 0 <M,/M0 < 0.6 Equation 1.9 and, 37 M, , 8 f-x2Dt^ L=l =- exp — for 0.4 < Mt/M0 < 1.0 Equation 1.10 M0 x \ I J where Mo is the total amount of drug loaded in the polymer, Mt is the amount of drug released at time t, D is the diffusion coefficient, and / is the thickness of the device. In monolithic dispersion systems, the amount of initial drug loading has an effect on the release mechanism (Baker 1987). At low loading levels (1-5%), the release of the drug involves dissolution of the drug in the polymer followed by diffusion to the surface of the device. The release kinetics from a slab device can be predicted by the Higuchi model and described by the following equation (Higuchi 1961; Baker 1987): M, = A(2DtCsmCof2 for C 0 » C S m Equation 1.11 where A is the total area of the slab, Csm is the solubility of the drug in the polymer, and Co is the total concentration of the drug (dissolved plus dispersed) initially present in the polymer. At higher loading levels (5-10%), the release mechanism becomes complex, since the cavities remaining from the loss of drug near the surface are filled with fluid from the external environment and this will result in an enhanced drug release. The release kinetics can then be described by a modified form of Equation 1.11 (Baker 1987), M=A Equation 1.12 u \-CJp) where p is the density of the drug. 1.5.2.2 Diffusion and degradation controlled systems The mechanism of drug release from a biodegradable polymeric device is more complex than that from a nondegradable polymeric device. Depending on the degradation 38 mechanism of the polymer and the relative rate of drug diffusion and polymer degradation, the drug release can be governed by mainly degradation, diffusion, or most frequently by the combined diffusion and degradation mechanisms (Baker 1987; Ranade and Hollinger 1996). In degradation-controlled monolithic systems, the drug is dispersed uniformly throughout the polymer matrix, and diffusion is slow compared with degradation. When the polymer degrades by a homogeneous (bulk) degradation mechanism, drug release is quite slow initially. With an increase in the degradation rate, the drug release rate increases considerably. When the polymer degrades by a heterogeneous (surface) mechanism, the drug release is confined to the surface of the device, and release is therefore affected by the surface-to-volume ratio and the geometry of the device (Baker 1987). If the drug diffuses from the device rapidly relative to the degradation of the polymer, the drug release is governed by a simple diffusion mechanism at the initial stage. However, subsequent degradation of the polymer can affect the release significantly, causing the permeability of the polymer matrix to increase significantly, and drug release rates to increase. Theoretical drug release profiles from a biodegradable polymer device are shown in Figure 1.5. 39 15 0 1 2 3 4 5 Time Figure 1.5. Theoretical drug release from a polymer slab by diffusion and by combined degradation and diffusion (adapted from: Baker 1987) 1.5.2.3 Swelling-controlled systems A swelling-controlled system consists of a dispersion of a drug in a hydrophilic polymer matrix. In a swelling-controlled system, absorption of water from the environment changes the dimensions and/or physical properties of the polymer matrix and thus the drug release kinetics. Drug release from such systems is a function of the rate of uptake of water from the surrounding media and the rate of drug diffusion (Baker 1987; Park et al. 1993). A model has been developed that describes the swelling polymer matrix in terms of three zones (Alfrey et al. 1966). Adjacent to the bulk water is a layer of completely swollen gel. Then there is a fairly thin swelling zone in which the polymer chains are slowly hydrating and relaxing. The third zone is a matrix of unswollen, completely dried, rigid polymer. 40 When the drug is loaded into the hydrogel by equilibrium swelling in the drug solution, drug release from the swollen gel follows Fick's law. Thus, the rate of drug release from the equilibrated slab device can be described by Equation 1.9, indicating that drug release from this system is dependent on t (Baker 1987; Park et al. 1993). When the drug is loaded into an unswollen, completely dehydrated gel, the rate of drug release depends on both the rate of water uptake into the polymer and the rate of drug diffusion. Diffusion of drug through the swollen gel is rapid compared with the rate of water uptake, which is controlled by the relaxation of polymer chains. Such a system has been called "Case II diffusion". For case II diffusion, drug release is a linear function of the time (0 (Park et al. 1993). In many practical swelling-controlled systems, the diffusion rate and the polymer relaxation rate are comparable. Such systems are called non-Fickian diffusion. For non-Fickian diffusion systems, the following semi-empirical equation (Alfrey et al. 1966) was suggested, Ms=kxt + k2t111 Equation 1.13 where Ms is water weight gain, k; and kj are constants. Equation 1.13 includes expressions for both Fickian and Case II transport mechanisms, and was adapted to describe some drug release processes from swellable polymers (Park et al. 1993). 1.5.3 Factors influencing drug release from biodegradable polymers Drug release from a biodegradable polymeric delivery system can be influenced by a number of factors including physicochemical properties of both the drug and the polymer, initial drug loading and environment conditions surrounding the delivery system, such as temperature, pH, and agitation. 41 Drug release from a polymeric matrix generally involves four steps: drug molecules dissociate from their crystal lattice structure, dissolve into the polymer matrix, diffuse through the matrix, and finally partition into the medium surrounding the delivery system (Chien 1982). The solubility of the drug in the polymer directly affects the drug release rate. The importance of this effect can be appreciated by examining Equations 1.9 to 1.12. Drug diffusion through a polymeric matrix is a result of random motion of drug molecules through the free volume between the polymer chains. In a semicrystalline polymer, drug diffusion usually occurs in the amorphous region rather than in the crystalline region, since the ordered alignment of polymer chains in the crystalline region lowers the free volume, thus impeding the diffusion of drug and water molecules (Ranade and Hollinger 1996). Consequently, lowering the crystallinity of a polymer will increase its permeability to drugs (Pitt et al. 1979). In the amorphous region, flexibility of the polymer chains plays an important role in drug diffusion. High flexibility (low Tg) allows the polymer chains to move easily, thus increasing the permeability of the polymer (Pitt and Schindler 1980). Crosslinking leads to the reduction in the mobility of the polymer chains and consequently a decrease in permeability of the polymer matrix. The reduction in polymer permeability is a linear function of the reciprocal of the extent of crosslinking (Chien 1982). Molecular weight of polymers also affects drug release. Generally, the permeability of a polymer increases as its molecular weight decreases. This has been attributed to increased free volume caused by the chain ends, therefore lowered Tg. On the other hand, a decrease in the molecular weight of polymers may lead to an increase in crystallinity, thereby, decreasing the permeability of the polymer matrix. This was 42 observed for biodegradable semi-crystalline polymers such as poly(s-caprolactone) (Pitt etal. 1979). Drug release rate can be affected by the properties of the drug. It was found that the diffusion coefficient decreases as the molecular weight of the drug increases (Pitt and Schindler 1980). Increase of the drug solubility in both the polymer matrix and the release medium leads to an increase in the drug release rate (Chien 1982). For a matrix-type drug delivery system, drug release rate is influenced by the initial drug loading as indicated by Equations 1.9 to 1.12. An increase in drug loading results in an increased concentration gradient, thus increasing release rate. Increasing temperature has a marked effect, enhancing the mobility of the polymer chains, and therefore increasing the diffusion rate of the drug molecules through the polymer matrix. The pH of the release media can affect the drug release rate of the delivery systems when pH-responsive polymers, such as chitosan, are used as matrices (Vazquez-Duhalt et al. 2001). 1.6 POLYMERIC DRUG DELIVERY SYSTEMS FOR PACLITAXEL A variety of polymeric delivery systems for paclitaxel have been developed, including microspheres (Burt et al. 1995; Liggins and Burt 2001), injectable surgical pastes (Winternitz et al. 1996), micelles (Zhang et al. 1996), and prodrugs (Dosio et al. 2001). 1.6.1 Microspheres Targeted delivery of an anticancer drug to a tumor using arterial chemoembolization with drug-loaded microspheres has been extensively studied (Fujimoto et al. 1985; Okamoto et al. 1986). Paclitaxel loaded microsphere formulations 43 have been developed using poly(L-lactide) (PLLA) (Liggins et al. 2000) and the blends of poly(D,L-lactide) (PDLLA) and poly(ethylene-co-vinyl acetate) (EVA) (Burt et al. 1995). Antiangiogenic activity and efficacy in the inhibition of tumor growth were demonstrated using the chicken chorioallantoic membrane (CAM) model and using a rat model of a tumor cell spill after a cecotomy repair, respectively (Burt et al. 1995; Liggins et al. 2000). 1.6.2 Injectable surgical pastes Paclitaxel loaded biodegradable surgical pastes were developed for application to a tumor resection site to eliminate remaining tumor cells and prevent local recurrence of tumor. Two types of polymers were investigated for this application, poly(e-caprolactone) (PCL) (Winternitz et al. 1996) and a triblock copolymer PDLLA-PEG-PDLLA) (Zhang et al. 1996). The pastes could be delivered via a syringe following gentle warming. Efficacy of the pastes was evaluated in a mice tumor model and significant inhibition in tumor growth was achieved (Zhang et al. 1996). 1.6.3 Micellar formulations Micellization is an effective approach of solubilizing paclitaxel in hydrophilic environments. An amphiphilic diblock copolymer poly(DL-lactide)-block-methoxy poly(ethylene glycol) (PDLLA-b-MePEG) was synthesized and used as a vehicle for intravenous administration of paclitaxel (Zhang et al. 1996). The copolymer formed micelles in aqueous media with a water-soluble MePEG shell and a PDLLA core. Micelles with up to 25% paclitaxel loading were prepared using a solution casting method (Zhang et al. 1996). Micellar paclitaxel was shown to be efficacious in a mouse tumor model (Zhang et al. 1997). 44 1.6.4 Prodrugs Water soluble prodrugs of paclitaxel have been another approach to the improvement of the solubility of paclitaxel. A prodrug is defined as the chemically modified form of a drug, which undergoes in vivo transformation to its active form (Bundgaard 1985). Two accessible sites for chemical modification of the paclitaxel molecule are 7-hydroxyl group and 2'-hydroxyl group (Panchagnula 1998). Most of the prodrugs of paclitaxel have been prepared through modifications at these two sites. The modifications include two major types: paclitaxel bonded to water-soluble small molecules and macromolecules. The small molecule prodrugs were designed by introducing ionisable groups or solubilizing moieties into the molecule of paclitaxel. These include paclitaxel succinate (Deutsch et al. 1989), glutarate (Deutsch et al. 1989), sulfonic acid derivatives (Zhao and Kingston 1991), amino acid derivatives (Mathew et al. 1992) and sialic acid derivatives (Takahashi et al. 1998). The macromolecular prodrugs involve covalent attachment of paclitaxel to water-soluble polymers such as PEG (Li et al. 1996) and albumin (Dosio et al. 2001). Physico-chemical properties and anti-tumor activity of these prodrugs were evaluated both in vitro and in vivo. It was found that the prodrugs can continuously release paclitaxel in the plasma over prolonged periods of time and the high cytotoxicity of paclitaxel was maintained (Singla et al. 2002). 45 1.7 DEVELOPMENT OF PACLITAXEL LOADED POLYMERIC FILMS FOR SURGICAL ADHESIONS An ideal formulation for the prevention of surgical adhesions should be one that combines the functions of both physical barrier and drug delivery system, and possesses the following properties: • The biomaterial used is biocompatible and biodegradable; • Uniform dispersion of drug in the biomaterial matrix; • Controlled drug release within about 10 days; • In vivo degradation lifetime of less than one month; • Flexible and easy to handle; • Sterilizable. 1.7.1 Hydrophilic polymeric films Paclitaxel loaded hyaluronic acid (HA) films were developed by our group for the prevention of surgical adhesions. The films containing 0%, 1% and 5% paclitaxel were prepared and evaluated both in vitro and in a rat cecal sidewall model (Jackson et al. 2002). The in vitro release rate of paclitaxel from the films was rapid. Almost all of the paclitaxel loaded in the films was released within 2 days. Following intraperitoneal implantation in rats, both 1% and 5% paclitaxel loaded films were found to inhibit adhesion formation. However, there was no barrier effect exerted by the use of the blank HA films (with no paclitaxel). Paclitaxel was found to be present as large crystals in the HA matrix (Jackson et al. 2002). In the early work of this project, chitosan was selected as a barrier material due to its excellent biocompatibility, biodegradability and mucoadhesive properties (Li et al. 46 1992). Chitosan-based polymeric films loaded with paclitaxel were developed, including crosslinked chitosan films and films composed of chitosan and poly(ethylene oxide) (Shi et al. 1999). In contrast to the HA films, paclitaxel release from the chitosan films was extremely slow, with less than 15% of the initially loaded paclitaxel being released within 10 days (Shi et al. 1999). Paclitaxel was also present as large crystals in the chitosan matrix. 1.7.2 Novel amphiphilic polymeric films In this project, paclitaxel loaded films based on novel amphiphilic graft copolymers of polysaccharides have been developed for the prevention of surgical adhesions. The main chain polymers of the graft copolymers are hydrophilic polysaccharides such as HPC and dextrans, and the side chain polymer is hydrophobic PCL. It is hypothesized that grafting of hydrophobic PCL onto the hydrophilic polysaccharides would confer increased hydrophobicity on the hydrophilic polysaccharides and make these novel biomaterials more suitable for the incorporation of hydrophobic drugs such as paclitaxel. 1.8 R E S E A R C H OBJECTIVES (1) Develop and characterize paclitaxel loaded polymeric films based on chitosan; (2) Synthesize and characterize the novel amphiphilic graft copolymers: HPC-g-PCL and dextrans-g-PCL; (3) Develop and characterize paclitaxel loaded polymeric films based on these novel graft copolymers; (4) Evaluate the paclitaxel loaded graft copolymer films in rat models of surgical adhesions. 47 Chapter 2 CHITOSAN BASED FILMS FOR PACLITAXEL DELIVERY 2.1 INTRODUCTION Chitosan is a biocompatible and biodegradable polysaccharide, which is widely used in medical and pharmaceutical applications (Li et al. 1992; Sabnis and Block 2000). It is derived from chitin (poly[p-(l->4)-Af-acetyl-D-glucosamine]), an abundant polysaccharide present in the exoskeleton of crustaceans, by deacylation. A typical deacylation process involves treating chitin with an aqueous solution of sodium hydroxide (concentration 40-50%, w/w) at 110-120°C for several hours to hydrolyze N-acetyl linkages. The product is then rinsed and dehydrated. This treatment converts chitin into chitosan (Hon 1996). The structure of chitosan is shown in Figure 2.1. Chitosan in its free amine form (-NH2) is insoluble in water, alkali and organic solvents, but soluble in most solutions of organic acids such as formic, adipic, or acetic acid. Some dilute inorganic acids can also be used to prepare a chitosan solution, however, prolonged stirring and sometimes heating are required (Li et al. 1992). Under Figure 2.1 Chemical structure of chitosan 48 acidic conditions, the free amino group becomes protonated to form cationic amine groups (-NFf/), and the chitosan molecule possesses a high positive charge density. This makes it different to most high-molecular-weight polysaccharides, which are primarily neutral or polyanionic in nature. Because most living tissues carry negative charges (e.g., proteins, anionic polysaccharides, nucleic acids), the positive charge of chitosan interacts strongly with negatively charged mucosal surfaces and therefore, chitosan films can readily adhere to the surfaces of organs, bone and skin. Chitosan has been used as a drug carrier in a variety of formulations (Paul and Sharma 2000) and as a non-viral vector for gene delivery (Erbacher et al. 1998). Due to its unique polycationic properties, chitosan formed polyelectrolyte complexes with negatively charged DNA, and successfully transfected HeLa cells. Chitosan and its derivatives may be used as excipients in tablet manufacturing. The drug release rate from the tablets was found to depend on the amount and type of chitosan used, and a zero order drug release profile could be obtained (Nagai et al. 1984). Chitosan has also been investigated for making microsphere and microcapsule formulations of drugs such as cisplatin (Nishioka et al. 1990), mitoxantrone (Jameela and Jayakrishnan 1995), salmon calcitonin (Aydin and Akbuga 1996), and insulin (Aiedeh et al. 1997). Crosslinking has a significant effect on the drug release rate from chitosan matrices. Two different approaches have been used in crosslinking chitosan: ionic complexation and chemical crosslinking (Janes et al. 2001). Polyanions such as tripolyphosphate have been employed to produce ionic complexation with cationic chitosan (Bodmeier et al. 1989). Glutaraldehyde is a frequently used chemical crosslinking agent and reacts primarily with the amine groups in chitosan (Genta et al. 49 1998) . Studies have shown that drug release rates from chitosan matrices could be effectively controlled by altering the extent of crosslinking (Thanoo et al. 1992; Jameela and Jayakrishnan 1995). Chitosan is used to promote wound-healing in veterinary medicine (Ueno et al. 2001). It enhances several functions of inflammatory cells such as polymorphonuclear leukocytes including phagocytosis, production of osteopontin and leukotriene B4. In macrophages, phagocytosis and production of interleukin (IL)-l, transforming growth factor pi and platelet derived growth factor are increased and in fibroblasts, production of IL-8 is increased (Usami et al. 1998; Ueno et al. 2001). As a result, chitosan promotes granulation and organization, and is beneficial for wound healing. A Japanese company has developed and marketed a chitosan-cotton (Chitopak TMC) and a chitosan suspension (Chitofme TMS) for veterinary wound healing. An US company has marketed a product (Tegasorb™), in which chitosan is an excipient, for human wound healing (Ilium 1998). The success of chitosan as a wound-healing accelerator has encouraged an interest in exploring chitosan as a potential barrier material for surgical adhesions. Physical blending of chitosan with other polymers is a convenient and effective approach to modify the physicochemical and/or biological properties of chitosan. Poly(ethylene oxide) (PEO), also referred to as poly(ethylene glycol) at low molecular weight, is a biocompatible polymer which has been widely used as a component of drug vehicles (Uhrich et al. 1999). Studies of blends of chitosan and PEO have indicated that the two materials are partially miscible and the miscibility is facilitated by the interaction (hydrogen bonding) between chitosan and PEO molecules (Zhao et al. 1995; Mucha et al. 1999) . 50 The objectives of this study were to develop and characterize paclitaxel loaded film formulations of crosslinked chitosan and blends of chitosan and PEO and to determine the potential suitability of the films for the prevention of post-surgical adhesions. 2.2 E X P E R I M E T N A L 2.2.1 Materials Paclitaxel was obtained from Hauser Chemicals, Inc. (Boulder, CO.). The chitosan used for in vitro studies, which had a nominal molecular weight of 70,000 was purchased from Fluka. Biomedical grade chitosan used for the in vivo evaluation was purchased from Carbomer, Inc.. Acetic acid (HAc), glycerol and glutaraldehyde were from Fluka, and poly(ethylene oxide)s (PEOs) with molecular weights of 200,000 (PEO200) and 900,000 (PEO900) were from Union Carbide Corporation (Danbury, CT). Acetonitrile (ACN), methanol (MeOH), dichloromethane (DCM) of HPLC grade, sodium dihydrogen orthophosphate, sodium phosphate, sodium chloride (NaCl), and sodium hydroxide (NaOH) of A.C.S grade were purchased from Fisher Scientific (Napean, Ontario, Canada). Albumin was purchased from Boeringer Mannheim, Germany. Nitrogen and prepurified helium gases were supplied by Praxair (Burnaby, BC, Canada). Distilled water was used throughout the studies. Phosphate buffered saline-albumin (PBSA, pH7.4) was prepared by dissolving 0.32 g sodium dihydrogen orthophosphate, 2.15 g sodium phosphate, 8.22 g NaCl, and 0.40 g albumin in one liter of distilled water. 51 2.2.2 General equipment and supplies Balances, Mettler model A E 163, AJ 100, and PJ 300 (Mettler Instruments, Zurich, Switzerland) Acumet Model 230 pH meter (Fisher Scientific, Fairlawn, NJ) Olympus BH-2 microscope (Olympus Optical Company Limited, Japan) Camera, Contax 167 M T (Tokyo, Japan) Corning hot plate/stirrer, model PC-351 (Corning Glass Works, Corning, NY) Centrifuges, Beckman GS-6 Centrifuge (Palo Alto, CA.) and an Eppendorf Centrifuge model 5415 C Ovens, Thelco oven, Precision Scientific Company (Chicago, IL) Shel Lab oven, Johns Scientific Company (Portland, OR.) Napco vacuum oven, Model 5831, Precision Scientific Company (Chicago, IL.) Reactiv-Therm III heating/stirring module (Pierce Inc., Rockford, IL) Reacti-Vap III gas drying module (Pierce Inc., Rockford, IL) Freezer (Caltec Scientific, Richmond, BC) Branson 2200 ultrasonic cleaner (Branson Ultrasoncis Corporation, Danbury, CT) VWR Vanlab vortex mixer, Scientific Industries, Inc., (Bohemia, NY) Kimax® brand 15 mL test tubes with Teflon® lined screw caps, Buchner funnels, and graduated cylinders, Pyrex® brand vacuum dessicator, beakers, Erlenmeyer flasks, (Fisher Scientific, Toronto, ON) Pipettes, variable volume Pipetman from Gilson Company Petri dishes, size 35x10 mm (Corning Glass Works, Corning NY) Micro-centrifuge tubes (1.5 mL) with caps, Elkay Products (Shrewsbury, MA) 52 2.2.3 Preparation of chitosan and chitosan/PEOs films Chitosan solutions were prepared by dissolving chitosan in 0.25 M acetic acid and then filtering using a Buchner funnel (with a coarse porosity) to remove insoluable particulate impurities. All the films prepared in this study contained 20% (w/w) of glycerol based on the dry weight of the films. The thickness of the films was 50 urn, and the paclitaxel loading levels were 0, 1%, 5%, and 10% (w/w). Chitosan films were prepared by dissolving chitosan in acetic acid solution (1.5% w/w) and then mixing with glycerol followed by addition of paclitaxel in A C N solution (1.5% w/v) under vigorous agitation. The mixture was cast into Petri dishes and dried at room temperature for 48 hours. NaOH treated chitosan films were prepared by treating the chitosan films with 0.1 M NaOH solution for 1 min, and then drying at room temperature for 48 hours. The films are referred to as chitosan-NaOH films in this thesis. Crosslinked chitosan films were prepared using glutaraldehyde as a crosslinking agent. Glutaraldehyde solution (1% w/v) was added to the mixture of the chitosan, glycerol and paclitaxel solutions while stirring. The mixture was then cast into Petri dishes and dried under the same conditions. Films containing 0.1%, 1% and 10% of glutaraldehyde were prepared and are referred to, in this thesis, as chitosan-0.1%GA, chitosan-1%GA, and chitosan-10%GA, respectively. For the preparation of films of chitosan/PEO blends, PEOs were dissolved in distilled water to form solutions with a concentration of 1.5% w/w. The solutions of chitosan and PEOs were mixed at ratios of 5:0, 4:1, 3:2, and 2:3 (v/v), and then glycerol 53 in water and paclitaxel in A C N solutions were added and mixed under vigorous agitation. The mixture was cast into Petri dishes and dried at room temperature for 48 hours. 2.2.4 Optical microscopy studies of paclitaxel loaded chitosan films Chitosan films with 10% paclitaxel were examined using an Olympus BH-2 optical microscope. Photographs were taken at a magnification of 150x using a Contax 167MT camera. 2.2.5 Thermal analysis Differential Scanning Calorimetry (DSC) was conducted using a Pyris 1 differential scanning calorimeter (Perkin Elmer Corp., Norwalk, CT) equipped with a Perkin Elmer Cryofill. The coolant was liquid nitrogen. The purge gas was prepurified helium at a flow rate of 30 mL/min. Three to ten milligrams of samples were loaded in crimped, but not hermetically sealed, aluminum pans (Perkin-Elmer sample pans) and scanned at a heating rate of 10°C/min. An empty pan was used as the reference. Results were analyzed using Perkin Elmer Pyris Data Analysis Software. 2.2.6 Determination of in vitro paclitaxel release rate Paclitaxel was assayed using a Waters HPLC (Waters Corporation, Milford, MA). The instrument was equipped with a model 600 controller, a model 486 U V detector, and a model 717+ autosampler. Millennium software was used for instrument control and data analysis. The column was a NovoPak Cig column, the mobile phase was acetonitrile/water/methanol (58:37:5), the flow rate was 1 mL/min, and the detection wavelength was 232 nm. In vitro release studies were carried out in PBSA (pH7.4) at 37 °C. Paclitaxel loaded films were placed in screw-capped test tubes containing 15 mL of PBSA. The 54 tubes were tumbled end-over-end at 30 rpm in a thermostatically controlled oven. At given time intervals, the release medium was replaced by fresh PBSA. Paclitaxel in the release medium was extracted into 1 mL of D C M , dried under a stream of nitrogen, reconstituted in 1 mL of 60% (v/v) A C N in water and analyzed by HPLC. 2.2.7 In vitro degradation of chitosan Chitosan films were prepared using the same method described in 2.2.3 except that no glycerol was added. The films were incubated in PBS (10 mM, pH7.4) at 37 °C and the release medium was replaced with fresh buffer every 24 hours. The films, removed at given time intervals, were dissolved in 0.25 M acetic acid to form a solution with a concentration of about 1.5% (w/v) approximately, and filtered through a Buchner funnel (coarse). The chitosan concentrations of the solutions were determined by drying known amounts of the solution to constant weight at 105 °C. Intrinsic viscosities [n] of the chitosan solutions were determined using a single point method (Solomon and Ciuta 1962; Rao and Yaseen 1986) in 0.1 M acetate buffer solution using an Ubbelohde viscometer (Technical Glass Products, Inc., Dover, NJ). The measurement was conducted at 30+0.1 °C in a Constant Temperature Water Bath (Cannon Instrument Company, State College and Boalsburg, PA). Two mL of the filtered chitosan solution were mixed with 14 mL of filtered acetate buffer, and the mixture was transferred into the viscometer. The viscometer was then placed into the constant-temperature bath, securely fastened and with the upright tubes vertical. After temperature equilibration had been achieved (about 10 min), the efflux time of the solution through the capillary was measured at least three times for every solution. The three readings agreeing within 0.2 second were used for the calculation of the mean of the efflux time 55 (f). The efflux time of 0.1 M acetate buffer (t0) was determined using the same procedures under the same conditions. The [r\] was calculated using Equation 2.1: 2 ( - i - l - ln( -^) ) tQ t0 — - - — Equation 2.1 c where c was the concentration of the chitosan solution in the viscometer. 2.2.8 Statistical analysis The paclitaxel release among the films with different paclitaxel loadings was compared by Tukey test. The Student's t-test was used to compare the release from the crosslinked films and the non-crosslinked films, the chitosan/PEO films and the chitosan-only films, and the chitosan/PEO200 films and chitosan/PEO900 films. In all cases, a p value less than 0.05 was considered statistically significant. 2.3 RESULTS 2.3.1 Effect of glycerol on the flexibility of chitosan films Films prepared using chitosan only were rigid and brittle. Bending the films resulted in break or fracture of the films. The incorporation of glycerol, a plasticizer, into the chitosan films produced films that were flexible and could be readily manipulated into and around the surgical sites. It was found that the addition of 20% glycerol (w/w) imparted desired flexibility to the chitosan films. 2.3.2 In vitro degradation of chitosan The relationship of molecular weight and intrinsic viscosity ([rj]) has been established by Mark-Houwink equation (Equation 1.5). Therefore, [r\] can be used to monitor the degradation of a polymer. In this study, the in vitro degradation profile of chitosan, shown in Figure 2.2, is illustrated by the change of [r|]t / [r\]o with time, where 56 1.2 0.4 -0.2 -0 i 1 1 1 1 1 1 1 1 0 5 10 15 20 25 30 35 40 45 Time (days) Figure 2.2 In vitro degradation profile of chitosan films incubated in 10 mM PBS (pH7.4) at 37 °C. Intrinsic viscosities were determined at 30 ± 0.1°C in 0.1 M acetate buffer (pH4.5) using an Ubbelohde viscometer. Data represent mean ± SD (n=3) 57 [r)]o is the intrinsic viscosity at time zero and [n]t the intrinsic viscosity at a sampling time. The degradation rate was more rapid during the first 12 days and [n] decreased to 62% of its initial value. The degradation rate then decreased dramatically. The films were still intact after 42 days' incubation in PBS at 37 °C despite the decrease in molecular weight. 2.3.3 Nature of paclitaxel dispersion in chitosan film Paclitaxel was found to precipitate out immediately when the chitosan in acetic acid solution and the paclitaxel in A C N solution were mixed during the preparation of the films. In the dried films, paclitaxel was present as crystals of different sizes. A representative optical microscopic image of a paclitaxel-loaded film is shown in Figure 2.3, in which large needle-shaped crystals (up to 100 pm in length) were observed. 2.3.4 In vitro paclitaxel release study In the release studies, an appropriate film size was chosen so that sink conditions would be met. Sink conditions were assumed to be maintained if the concentration of paclitaxel in the release media did not exceed 15% of paclitaxel solubility (Carstensen 1977). Over the time course of the release, a fraction of the released paclitaxel (less than 20%) was found to convert to 7-epitaxol, an isomer of paclitaxel that has been shown to exhibit biological activities comparable to those of paclitaxel (Ringel and Horwitz 1987). In a typical chromatogram for paclitaxel analysis, the paclitaxel peak appeared at a retention time of 2.6 min, and the 7-epitaxol peak at 3.6 min. In the quantitation of paclitaxel released, the amount of paclitaxel was taken to be the sum of both paclitaxel and 7-epitaxol peaks. In all cases, the amount of paclitaxel released was normalized to a film size of 1 cm . 58 Figure 2.3 A representative optical micrograph of paclitaxel loaded chitosan film 59 2.3.4.1 Effect of crosslinking on the release rate of paclitaxel The cumulative release profiles of paclitaxel from the films of chitosan, chitosan-GA, and chitosan-NaOH are shown in Figure 2.4. The thickness of the films was 50pm, and the initial paclitaxel loading level was 5% (w/w). Paclitaxel release was characterized by a small initial burst phase over one day followed by a zero-order release at a lower rate. The release profiles for the chitosan film and chitosan-NaOH film were almost identical indicating that treatment of chitosan film using NaOH did not change paclitaxel release rate. This indicates that whether the amino group in chitosan molecules is in its protonated form or deprotonated form does not have an effect on paclitaxel release rate. Crosslinking resulted in a slower rate of paclitaxel release, and the release rate decreased with the increase in the extent of crosslinking. Since surgical adhesions primarily occur during the first 7 days following surgeries, the total amounts of paclitaxel released from different films within the first 7 days are summarized in Table 2.1. For the non-crosslinked films, the amounts of paclitaxel released were significantly higher (p<0.05) than those from the crosslinked films. An increase in the extent of crosslinking from 1% to 10% glutaraldehyde significantly lowered release rate. 2.3.4.2 Effect of paclitaxel loading on the release rate The paclitaxel release profiles from chitosan-0.1%GA films with different loadings are shown in Figure 2.5. The films were crosslinked with 0.1% glutaraldehyde, and had a thickness of 50 um. The initial paclitaxel loading level had a significant effect on the release rate. For the 1% paclitaxel loaded films, the release profile is characterized by a continuously decelerating release rate. The total amount of paclitaxel released over a 60 o C O 13 1 - 1 o 4 6 8 Time (day) 10 12 Figure 2.4 Paclitaxel release from chitosan films crosslinked using different amounts of glutaraldehyde or treated with NaOH: (•) no crosslinking, (x) chitosan-NaOH, (•) chitosan-0.1%GA, ( A ) chitosan-1%GA, and (•) chitosan-10%GA. The thickness of the films was 50 pm and the initial PTX loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount released was normalized to a surface area of 1 cm2. Data represent mean + SD (n = 4) with the error bars shown only in the positive direction for clarity. 61 Table 2.1 Total amount of paclitaxel released within 7 days from chitosan films crosslinked by different amounts of glutaraldehyde and loaded with different amounts of paclitaxel. Film ID Paclitaxel released within 7 days (ug/cm2)a 1% loading 5% loading 10% loading Chitosan — 26.9+0.8 — Chitosan-0.1%GA 14.5+1.7 b 21.6+0.9 b , c 30.2+1.6 b Chitosan- 1%GA — 21.5+1.7° — Chitosan- 10%GA — 15.8+0.9 c — Chitosan-NaOH — 28.2+2.7 — a Data represent mean ± SD (n=4) b Statistically different (p<0.05) from each other (Tukey test) c Statistically different (p<0.05) from non-crosslinked chitosan film (t-test) 62 o CO CD ID > u 4 6 8 Time (day) 10 12 Figure 2.5 Paclitaxel release from chitosan-0.1%GA films with different initial loadings (w/w): (•) 10%, (•) 5%, and (•) 1%. The thickness of the films was 50 urn. The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount released was normalized to a surface area of 1 cm2. Data represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity. 63 7-day period was found to be significantly lower (p<0.05) than that for the 5% and 10% paclitaxel loaded films (Table 2.1). For the 5% and 10% paclitaxel loaded films, the profiles show a small initial burst phase lasting one day followed by a zero-order release. The release rate of 10% paclitaxel loaded films was significantly higher (p<0.05) than that of the 5% paclitaxel loaded films (Table 2.1). 2.3.4.3 Effect of blending with PEOs on the release rate of paclitaxel The paclitaxel release profiles from the chitosan/PEO200 and chitosan/PEO900 blended films are shown in Figure 2.6 and 2.7, respectively. The profiles show an almost linear relationship between cumulative paclitaxel release and time, indicating that paclitaxel release followed zero-order kinetics. No initial burst phase was observed. Total amounts of paclitaxel released within the first 7 days from the films of chitosan/PEOs blends are summarized in Table 2.2. The release rates of paclitaxel from the blended films were higher than those from the non-crosslinked chitosan films. For the chitosan/PEO900 films, the difference was statistically significant for the films at all the three blending ratios. For the chitosan/PEO200 films, the release rates from the films containing 60% and 40% of PEO200 were significantly higher than from non-crosslinked chitosan films, but the difference between those from non-crosslinked chitosan films and the films containing 20% of PEO200 was not statistically significant. At the same blending ratio, paclitaxel release rates were higher from the films of chitosan/PEO900 than from those of chitosan/PEO200, and the differences were statistically significant at blending ratios of 80/20 and 60/40. 64 4 6 Time (day) 10 12 Figure 2.6 Paclitaxel release from the films of chitosan/PEO200 blended at different ratios: (•) 40/60, (•) 60/40, (•) 80/20, and (X) 100/0. Initial PTX loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount of PTX released was normalized to a surface area of 1 cm2. Data represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity. 65 70 ^ 60 |> 50 H 40 g 30 a! 2 20 U 10 0 X X 4 6 8 Time (day) z 10 12 Figure 2.7 Paclitaxel release from the films of chitosan/PEO900 blended at different ratios: (•) 40/60, (•) 60/40, (•) 80/20, and (x) 100/0. Initial paclitaxel loading was 5% (w/w). The release was conducted in 10 mM PBSA (pH7.4) at 37 °C in end-over-end tumbling tubes. The amount released was normalized to a surface area of 1 cm2. Data represent mean ± SD (n = 4) with the error bars shown only in the positive direction for clarity. 66 Table 2.2 Cumulative paclitaxel release within 7 days from the films of Chitosan/PEOs at different blending ratios. Paclitaxel loading was 5% (w/w) based on the weight of the dry film. Blending ratio Chitosan/PEO200 (ug/cm2) Chitosan/PEO900 (ug/cm2) p valuea 100/0 26.9±0.8 26.9±0.8 — 80/20 28.8±2.1 33.5+0.5 b 0.002 60/40 31.9+3.1b 39.2±2.6 b 0.005 40/60 35.6±5.8 b 37.4±3.5 b 0.3 a Student's t-test, p < 0.05 was considered statistically significant b Statistically different (p<0.05) from chitosan only film (Student's t-test) 67 2.3.5 Thermal analysis of the films of chitosan/PEOs blends Chitosan/PEO200 and chitosan/PEO900 films with blending ratios of 100/0, 80/20, 60/40,40/60 and 0/100) (corresponding to PEO weight fractions of 0, 0.2, 0.4, 0.6, and 1) were examined by DSC, and the DSC scans are shown in Figure 2.8 and 2.9, respectively. All the films contained 5% paclitaxel and 20% glycerol. As indicated by curve (A) in the figures, no thermal events occurred for the films containing no PEOs in the temperature range of 30 to 90 °C. For the films containing PEOs, endothermic peaks were evident in the range of 45 to 75 °C, corresponding to the melting of PEO. From these curves, melting points (Tm) of PEOs in the films with different blending ratios were obtained, which were determined as the endpoints of the melting peaks in the DSC curves (Collins et al. 1973). Melting point depression was observed for both chitosan/PEO200 and chitosan/PEO900 blends indicating a partial compatibility of chitosan with PEO (Zhao et al. 1995). Dependence of Tm on PEO weight fraction in the blends is shown in Figure 2.10. Tm decreased with decreasing PEO weight fraction and levelled off when PEO weight fraction reached 0.2. 68 10 H 20 1 30 o Q o TJ C UJ o u. «S « X 40 50 70 A B C D \ / E I 1 1 1 1 1 / ' 30 35 40 45 50 55 60 65 70 75 80 85 90 Temperature (*C) Figure 2.8. DSC scans of chitosan/PEO200 films with blending ratios of (A) 100/0, (B) 80/20, (C) 60/40, (D) 40/60, and (E) 0/100. Heating rate was 10 °C/min. 69 -I 1 1 1 1 1 1 1 I 1 1 1 30 35 40 45 50 55 60 65 70 75 80 85 90 Temperature (*C) Figure 2.9. DSC scans of chitosan/PEO900 films with blending ratios of (A) 100/0, (B) 80/20, (C) 60/40, (D) 40/60, and (E) 0/100. Heating rate was 10 °C/min. 70 75 50 H 1 1 1 1 1 0 0.2 0.4 0.6 0.8 1 PEO weight fraction Figure 2.10 Effect of blending ratios of chitosan/PEOs on the melting point of PEOs. ( A ) Chitosan/PEO200, (•) Chitosan/PEO900. 71 2.4 DISCUSSION The amine groups in the chitosan molecule can be in either a protonated or deprotonated form depending on the pH, as shown by the equilibrium below. The chitosan films used in this study were prepared by casting from chitosan solutions that had a pH less than 4. Since the pKa of chitosan is 6.0 (Desbrieres 2002), over 99% of the amino groups in chitosan molecules should be in their protonated form according to the Henderson-Hasselbach equation. Due to the polyelectrolyte nature of protonated chitosan, the films absorbed water quickly and swelled extensively when in contact with water. The mechanical strength of the hydrated films was very poor. C H 2 O H Acid C H 2 O H \ Base _ | _ 0 J \ O H N H 2 J n L NH3 J n The tensile strength of chitosan films may be improved by either chemical crosslinking using glutaraldehyde (Thacharodi and Rao 1993; Kawase et al. 1997) or physical blending with PEO (Alexeev et al. 2000). In this study, we focused on the effect of these modifications on the properties and release behaviour of paclitaxel from the films. The handling characteristics of films intended for placement at surgical sites to prevent postsurgical adhesions are critical. The film should possess good mechanical strength and flexibility in both the dry and wet states, so that it may be manoeuvred readily into and around the surgical site. Ideally, a film with mucoadhesive properties and which conforms readily to the tissue surfaces is also desirable. Glycerol, which has been widely used as a plasticizer (Entwistle and Rowe 1979), was added to the chitosan 72 formulations to increase the flexibility of the films and improve handling characteristics. A plasticizer can be defined as a small molecule that is added to a polymer to lower its glass transition temperature (Odian 1991). The fundamental principle associated with a plasticizer is to interact with the polymer chains on a molecular level so as to increase the chain mobility and speed up the viscoelastic response of the polymer (Paul and Newman 1978). In this case, glycerol was believed to interact with chitosan chains by forming hydrogen bonds and act as a "lubricant" to reduce the interaction between the chitosan chains. In vitro degradation studies of chitosan films were conducted in PBS (pH7.4) at 37 °C. The biphasic degradation profile may have been caused by the hydrolysis of chitosan chains occurring initially at the surface of the films. Since the majority of the chitosan molecules are in the deprotonated form at pH7.4, the hydrophobicity of the deprotonated chitosan has become a major hindrance for the penetration of water molecules through the films (Gupta and Kumar 2001). Chitosan chains at the surface of the film would be more susceptible to hydrolysis, thus degrading faster than those in the bulk. The loading of paclitaxel into all films formulations resulted in the immediate precipitation of crystalline paclitaxel in the film (as shown in Figure 2.3). Paclitaxel is an extremely hydrophobic drug and the solubility of paclitaxel in the hydrophilic chitosan and chitosan/PEO matrices would be expected to be very low. The particle size of the precipitated paclitaxel crystals was variable but there were large needle-shaped crystals up to 100 urn in length deposited in the film matrix. 73 Crosslinking was employed to increase the mechanical strength of the chitosan films. Chemical crosslinking was accomplished using glutaraldehyde. Sodium hydroxide treatment of the dry chitosan films was intended to increase the strength of the films by converting the protonated form of chitosan to deprotonated form. Blending with PEO was also used as an approach to the improvement of mechanical properties of chitosan films (Alexeev et al. 2000). Both chemical crosslinking of the films and the preparation of chitosan/PEO blend films provided a means of controlling drug release rate from the films. For the non-crosslinked chitosan films with 5% paclitaxel loading, the total amount of paclitaxel released within 7 days was 26.9 pg/cm , representing 9.1% of the initial loading. The slow release rate could be explained by the nature of paclitaxel dispersion in chitosan matrices. Due to the presence of particulate paclitaxel, release from the film was controlled not only by diffusion but also by a much slower dissolution rate of paclitaxel. Crosslinking of the chitosan films resulted in a decrease in the paclitaxel release rate compared to non-crosslinked. This was due to the fact that the chemical bonds formed between chitosan chains reduced the mobility of the chains and decreased the diffusion of paclitaxel molecules. The crosslinking made the diffusion of paclitaxel the rate-limiting step. The release rate did not change when the extent of crosslinking increased from 0.1% to 1% glutaraldehyde, presumably because an increase in crosslinking within this range was not significant enough to affect paclitaxel diffusion. Physical blending of chitosan with PEOs increased paclitaxel release rate (Figure 2.6 and 2.7). This could be explained by the incorporation of PEO molecules weakening 74 the interaction between chitosan molecules, thus increasing the diffusivity of paclitaxel molecules. At a fixed blending ratio, the release rate from chitosan/PEO900 film was higher than from chitosan/PEO200. The differences were statistically significant for the 80/20 and 60/40 blends, but not for the 40/60 blends (see Table 2.2). The differences in paclitaxel release rate may have resulted from the differences in the degree of the miscibility of these two polymer pairs. Melting point depression has been widely used to explain polymer/polymer interactions and miscibility or compatibility in a blend (Nishi and Wang 1975; Rim and Runt 1984). Figure 2.10 shows that the magnitude of melting point depression in chitosan/PEO900 was greater than in chitosan/PEO200, indicating that chitosan/PEO900 was more miscible than chitosan/PEO200. The more miscible the polymer pair, the weaker the interaction between chitosan chains, and this should lead to enhanced diffusivity of paclitaxel through the matrices. Assuming that paclitaxel diffusivity in the matrix is controlling the release kinetics of paclitaxel, this would explain why the release rates from the films of chitosan/PEO900 were higher than from those of chitosan/PEO200. Following the development and characterization of films composed of chitosan alone, or chitosan modified by sodium hydroxide treatment and glutaraldehyde crosslinking, the biocompatibility of the films was evaluated in a contract study (Wiseman et al. 1998). The rat sidewall abrasion model was used and films were placed over the abraded area. Between 2-28 days later, the abdomen was examined for remnants of material and gross tissue reaction. All chitosan films curled on contact with moisture and tissue fluids and did not remain in place once positioned. The films were still intact 75 after 28 days and had dislodged from the implantation sites. The films produced mild to moderate tissue reactions (Wiseman et al. 1998). 2.5 CONCLUSIONS Paclitaxel precipitated in the films as large crystals and drug release profiles showed prolonged paclitaxel release and substantial amounts of paclitaxel remaining in the films after 10 days. It has been reported that an ideal barrier or drug delivery system for post surgical adhesions should release its drug and be removed from the implantation site in less than about 10 days. Based on the biocompatibility data and the slow drug release and degradation rates of the films, these formulations were considered unsuitable drug delivery systems for the prevention of post surgical adhesions. 76 CHAPTER 3 SYNTHESIS AND CHARACTERIZATION OF HYDROXYPROPYLCELLULOSE-GRAFT-POLY(e-CAPROLACTONE) 3.1 INTRODUCTION Amphiphilic diblock and triblock copolymers have been recognized as an important type of pharmaceutical carrier due to their ability to solubilize hydrophobic drugs (Torchilin 2001). Since poly(ethylene glycol) (PEG) is biocompatible and highly water soluble, block copolymers with PEG as the hydrophilic component have been the most extensively investigated amphiphilic copolymers (Torchilin 2001). A variety of polymers have been used to build the hydrophobic blocks. These include polymers of e-caprolactone (CL) (Allen et al. 2000; Yuan et al. 2000), lactide (Li and Kissel 1993; Zhang et al. 1996), glycolide (Li and Kissel 1993), P-benzyl-aspartate (Kwon et al. 1993), and propylene oxide (Miller et al. 1997). These copolymers have been widely investigated as biodegradable carriers for the controlled release of drugs and as amphiphilic micellizing agents for the solubilization of hydrophobic drugs. During the past few years, increased interest has been shown in amphiphilic graft copolymers of hydrophilic polysaccharides. Water-soluble polysaccharides such as starch (Choi et al. 1999), dextran (Ydens et al. 2000) and pullulan (Donabedian and McCarthy 1998; Ohya et al. 1998) have been used as the main chain polymers with hydrophobic poly(e-caprolactone) (PCL), polylactide, and polyglycolide as side chains. The graft copolymerizations were initiated by the hydroxyl groups of the polysaccharides and 77 catalyzed by stannous 2-ethylhexanoate. Enzyme catalyzed graft copolymerization of CL onto hydroxyethylcellulose films has also been reported (Li et al. 1999). Hydroxypropylcellulose (HPC) is a derivative of cellulose which has a backbone of cellulose and short side chains of poly(propylene oxide). It is usually synthesized by the reaction of propylene oxide with cellulose under alkaline conditions (Klug 1971). The structure of HPC is shown in Table 1.2 (page 33). It is water soluble when at least two propylene oxide molecules reacted with each anhydroglucose unit (Ho et al. 1972). HPC is widely used as a pharmaceutical excipient in oral and topical formulations. In oral formulations, HPC is primarily used in tableting as a binder, film-coating, and extended release-matrix former (Kibbe 2000). HPC coated diltiazem hydrochloride tablets were prepared using a rotary tableting machine and a delayed drug release was achieved by altering the thickness of the HPC coating (Fukui et al. 2000). Acetaminophen tablets were prepared by a moist granulation technique using HPC as a release rate-controlling agent (Railkar and Schwartz 2001). HPC is also used in microencapsulation processes (Calanchi et al. 1986; Levy 1994). In topical formulations, HPC has been used in transdermal patches (Irifune et al. 1999; Yanagimoto et al. 1999) and ophthalmic preparations (Saettone et al. 1984). PCL is an aliphatic polyester synthesized by ring-opening polymerization of e-caprolactone (Schindler et al. 1977). Its structure is shown in Table 1.2. PCL is a semicrystalline polymer with a melting point (Tm) of 63 °C and a Tg of -60 °C (Pitt 1990). It crystallizes readily, and the crystallinity varies inversely with the molecular weight. For PCL with molecular weights in excess of 100,000, the crystallinity is about 40%, rising to about 80% as the molecular weight decreases to 5,000 (Pitt 1990). Both 78 the Tg and crystallinity of PCL can be modified by copolymerization (Schindler et al. 1977). The low Tg of the polymer imparts good drug permeability characteristics, and the high crystallinity results in good mechanical strength. A two-stage in vivo degradation process was reported (Pitt 1990). In the first stage, degradation began with random hydrolytic chain scission of the ester linkages, manifested by a reduction in the molecular weight without loss of mass until the molecular weight decreased to approximately 5000; the second stage was characterized by a decrease in the rate of chain scission and the onset of mass loss. PCL possesses excellent biocompatibility and has been used in many polymeric drug delivery systems (Pitt et al. 1979; Pitt et al. 1980; Cha and Pitt 1988). The synthesis of a graft copolymer of HPC with PCL as side chains should provide a copolymer combining the elements of biocompatibility of the hydrophilic HPC and the hydrophobicity of PCL side chains for enhanced solubilization of paclitaxel. The objectives of this study were to synthesize a graft copolymer of HPC-g-PCL by ring opening polymerization without using a catalyst and to characterize the HPC-g-PCL. To our knowledge, this is the first report both of the synthesis without using a catalyst and of the characterization of this novel graft copolymer of HPC-g-PCL. 3.2 E X P E R I M E N T A L 3.2.1 Materials HPC with a nominal molecular weight of 100,000 was purchased from Aldrich and was dried at room temperature under reduced pressure (0.3 mmHg) for 16 hours. e-caprolactone was purchased from Fluka. It was dried over calcium hydride (CaH2, from Fluka) for 24 hours and distilled twice under a reduced pressure. Tetrahydrofuran (THF) (HPLC grade, Fisher Scientific), hexanes (certified A.C.S, Fisher Scientific), methanol 79 (HPLC grade, Fisher Scientific), acetonitrile (HPLC grade, Fisher Scientific), deuterated dimethyl sulfoxide (DMSO-d6) (D, 99.9%, Cambridge Isotope Laboratories, Inc.) and deuterated chloroform (CDCI3) (D, 99.8%, Cambridge Isotope Laboratories, Inc.), were used as received. 3.2.2 General equipment and supplies Precision® Direct Drive Vacuum Pump, model PC200 (Precision Scientific, Winchester, VA) Edwards® Active Pirani Gauge, model D021-73-000; and Active Gauge Controller, model D386-55-000 (AGC-single disp, no RS232, 3 heads) (Edwards High Vacuum International, Manor Royal, Crawley, West Sussex, UK) Pump oil: Welch DirecTorr® Gold Oil (Fisher Scientific, Toronto, ON) Glassware: Pyrex® Liebig Condenser, round bottom flask, vacuum-type distillation trap, distilling head, and distillation receiver (Fisher Scientific, Toronto, ON) 3.2.3 Synthesis of HPC-g-PCL Graft copolymer HPC-g-PCL was synthesized by a bulk polymerization method. HPC and CL were added into a dry, two-neck, round bottom flask at a ratio of 1 to 4 (wt/wt), mixed thoroughly and left at room temperature overnight until HPC was fully swelled by CL. The mixture was placed under vacuum (0.3 mmHg) for 30 minutes and purged with nitrogen gas. The purging process was repeated three times. The flask was then sealed under a nitrogen blanket and placed in an oil bath at 150°C for 24 hours. 80 3.2.4 Conversion of CL The unreacted C L was isolated from the original reaction product and assayed by HPLC as follows: The reaction product was accurately weighed and dissolved in THF. An excess amount of hexanes was added to the solution to precipitate the polymers. The supernatant was collected, dried down using nitrogen gas, reconstituted using distilled water and injected onto the HPLC (Waters Corporate, Milford, MA). The instrument was equipped with a model 600 controller and pump module, a model 4 8 6 U V detector, a model 717+ autosampler, and Millennium32 software for instrument control and data analysis. The column was an Aqua CI8 (Phenomenex®), the mobile phase was methanol/water (35/65) at a flow rate of 0.5 mL/min. Pure CL was used as the standard to construct the calibration curve for the quantitation. The detection wavelength was 2 3 4 nm. A representative standard curve is shown in Figure 3 .1. 3.2.5 Purification and fractionation of HPC-g-PCL. Homopolymer PCL formed during the reaction and unreacted C L were removed by fractional precipitation using THF as the solvent and hexanes as the precipitant. The reaction product was dissolved in THF at a concentration of 5 % (w/v). Hexanes were added under agitation to precipitate HPC-g-PCL. The mixture was left at room temperature for 3 0 minutes to allow for complete precipitation. The dissolution and precipitation were repeated 3 times. A solution of HPC-g-PCL, purified using the method described above, was prepared in THF with a concentration of 3 % (w/v). Hexanes were added to the solution to precipitate higher molecular weight fractions (hexanes molar fraction XH=0 .327) . The system was allowed to stand at room temperature until the two phases could be physically 81 4.00E+06 T 3.50E+06 -3.00E+06 -^ 2.50E+06 -TO Z 2.00E+06 -ca ^ 1.50E+06 -1.00E+06 -5.00E+05 -O.OOE+00 -0 2 4 6 8 10 12 14 16 18 20 22 CL concentration (pg/mL) Figure 3.1 Standard curve of CL in water. Chromatographic conditions: column, Aqua CI8; mobile phase, methanol/water at 35/65 (v/v); flow rate, 0.5 mL/min; injection volume, 10 uL; detection wavelength, 234 nm. Regression equation: Peak area = 1.74x 105C - 3.20xl03, R 2 = 1 82 separated (approximately 30 minutes were required). The supernatant was collected and dried to give the first fraction (Fl). The precipitate was dissolved in THF, and hexanes with a lower molar fraction were added. The procedures were repeated to yield three more fractions (F2, F3, and F4). The fraction F4 was the precipitate in the last cycle. All fractions were dried at room temperature. The hexanes molar fractions used for the fractionation of HPC-g-PCL are given in Table 3.4 on page 115. 3.2.6 Gel permeation chromatography (GPC) HPC-g-PCL purity and its molecular weight and molecular weight distribution were determined using a Waters GPC system that was equipped with the same components as the HPLC instrument except that a model 2410 refractive index detector was used in this study. A PLgel® Mixed-D column (Polymer Laboratories Inc, Amherst, MA) was used alone or Styragel® HR3 and HR4 columns (Waters Corporate, Milford, MA) were used in series for qualitatively examining and quantitatively determining the molecular weight, molecular weight distribution and the purity of the copolymers. The mobile phase was THF at a flow rate of 1.0 mL/min. Millennium32® software was used for instrument control and data analysis. Polystyrene standards were used to calibrate the columns. Calibration curves are shown in Figure 3.2. 3.2.7 Nuclear magnetic resonance spectrometry (NMR) ' H NMR, 1 3 C NMR spectra and two-dimensional heteronuclear multiple quantum coherence (HMQC) spectra of HPC and HPC-g-PCL were recorded using a Bruker AMX500 spectrometer operating at 95°C. Two-dimensional correlation spectroscopy studies were performed with a 90°-ti-45°-t2 pulse sequence (COSY-45) using a Bruker AV400 spectrometer operating at 95°C. Samples were dissolved in DMSO-d 6 . *H NMR 83 Figure 3.2 Molecular weight calibration curve using polystyrene standards. Chromatographic conditions: columns, (A) Styragel® HR3 and HR4 in series, (B) PLgel Mixed-D; mobile phase, THF; flow rate, 1 mL/min; detector, differential refractive index detector. 84 spectra of PCL homopolymer and the fractions of HPC-g-PCL were acquired using a WH400 spectrometer operating at 25°C. The samples were dissolved in CDCI3. Chemical shifts are reported in parts per million (ppm) from tetramethylsilane with CDCI3 and DMSO-d6 as the internal references. For ! H NMR spectra, residual CHCI3 was taken as 7.24 ppm and residual DMSO as 2.49 ppm. For 1 3 C NMR the central peak of DMSO-d6 was taken as 39.5 ppm. Sample concentrations were around 50 mg/mL. 3.2.8 Preparation of HPC-g-PCL films HPC-g-PCL films were prepared by a solution casting method. A solution of HPC-g-PCL in THF with a concentration of 30 mg/mL was cast into Teflon dishes and dried at room temperature in the fume hood for 48 hours. The films were further dried under reduced pressure for another 48 hours. 3.2.9 Microscopic study of HPC-g-PCL fdms Optical microscopy studies of HPC-g-PCL films were conducted using an Olympus BH-2 optical microscope and photographs were taken at a magnification of 150x using a Contax 167MT camera. 3.2.10 Differential scanning calorimetry (DSC) DSC was conducted using the same instrument as in 2.2.5. The coolant was liquid nitrogen. The purge gas was prepurified helium at a flow rate of 30 mL/min. HPC powder (as received) and HPC-g-PCL films were weighed (3-8 mg) into crimpled, but not hermetically sealed, aluminum pans (Perkin-Elmer sample pans) and scanned at 10°C/min. An empty pan was used as the reference. 85 3.2.11 X-ray diffraction (XRD) A Rigaku Rotaflex X-ray diffractometer equipped with a rotating target X-ray tube and a wide angle goniometer was used to obtain the diffraction patterns of HPC-g-PCL films. The X-ray source was K a radiation from a copper target with a graphite monochrometer. The X-ray tube was operated at a potential of 50 kV and a current of 150 mA. The range (20) of scans was from 10° to 30° and the scan speed was 2 degrees per minute at increments of 0.02°. 3.2.12 In vitro degradation of HPC-g-PCL films HPC-g-PCL films were incubated in 10 mM PBS (pH7.4) at 37 °C and sampled every week for 12 weeks. The incubation medium was replaced with fresh PBS weekly following the sampling. Changes in molecular weight and molecular weight distribution were monitored by GPC using a PLgel® Mixed-D column. The mobile phase was THF at a flow rate of 1.0 mL /min. Calculations of molecular weight were based on the polystyrene standard curve shown in Figure 3.2B. Thermal events in the films were examined using DSC and a heating rate of 10 °C/min. 3.3 RESULTS 3.3.1 Synthesis and purification of HPC-g-PCL Graft copolymer HPC-g-PCL was synthesized via a bulk polymerization method without using a catalyst. As shown in Scheme 3.1, HPC is a hydroxyl group rich polymer. On each repeat unit, there are three hydroxyl groups. The graft reaction was initiated by these hydroxyl groups. Four batches of HPC-g-PCL were synthesized and the results are summarized in Table 3.1. In batch 1, 2 and 3, one gram of HPC and four grams of CL were added into the flasks and the syntheses were carried out under the 86 R 2 = H, Rj or z y z * * y * * r a p Y 8 e i a* P* Y* 6* e* ^ C H 2 C H O ^ C H 2 C H ( > 4 - C C H 2 C H 2 C H 2 C H 2 C H 2 0 4 — C C H 2 C H 2 C H 2 C H 2 C H 2 0 H I C H 3 J m ' C H 3 L o J m O m, m' = 0, 1, 2, Scheme 3.1 Structure and synthesis of HPC-g-PCL 87 _1 u C co U PL, X .2 'C co £ GO c CO CD X5 O o o CO *o co 43 - 4 - * U O H G O i o PL X o co a> O +-» co x> t-, o o 'co <U XS O 00 so CO o~-C o • J S Co c o o • J U PL, u PL, I u PL, X 00 C N VO p OO C N V O V O O N C N vo C N vo oo 00 V O C N C N C N C N S - 6 o co oo p co O N C N 00 V O C N c 00 00 CO 00 co O co O N O N C N .S a c co CD CD u co X i <2 c ™ •2 .oo <L> a> co K ° o B E ? 3 co CD C O x> CO o x : o CO OQ — i C N CO CD CO X> O " O CO 5 b b ^ L- ~ •zi o <u oo x : 88 same conditions. The reproducibility of the reaction was demonstrated by the small variance among the three batches in the conversions of CL, the yields, the molecular weights, the polydispersity indices of HPC-g-PCL and the molar substitution of CL on HPC (MSCL)- In batch 4, the synthesis was carried out on a much larger scale (30-fold) under the same conditions. As expected, both conversion and yield decreased, and HPC-g-PCL with a lower molecular weight and M S C L resulted. Table 3.1 also shows that PCL homopolymer formed during the reaction had weight-average molecular weights around 3000 g/mol and a polydispersity index of 1.2. GPC chromatograms of the starting material HPC and the reaction product are shown in Figure 3.3a and 3.3b, respectively. Before grafting, HPC had one peak at a retention time of 6.1 minutes. After grafting, the peak shifted to a lower retention time region (the component represented by the peak denoted as G). The shift shows an increase in molecular weight, indicative of graft copolymer formation. The chromatogram also showed another peak between 7.5-9.5 minutes (the component represented by the broad peak denoted as P). This peak was thought to be due to PCL homopolymer formed during the graft reaction. The component represented by peak P could no longer be detected after the purification process, as shown in Figure 3.3c. A reaction product resulting from the graft copolymerization could possibly contain three impurities including unreacted monomer CL, unreacted backbone HPC, and PCL homopolymer formed during the reaction. Unreacted C L monomer was identified and quantified by HPLC using an Aqua CI8 column and CL conversion data are given in Table 3.1. Unreacted HPC was not found based on the N M R results of the fractions of component G in the reaction product, which will be discussed in 3.3.6. The purification 89 Figure 3.3 GPC chromatograms of (a) backbone HPC, (b) reaction product, in which G represents the component of HPC-g-PCL, P represents the PCL homopolymer formed during the reaction, and (c) purified HPC-g-PCL 90 process was designed to separate HPC-g-PCL from unreacted C L and PCL homopolymer. THF was selected as the solvent since it could dissolve all the components in the reaction product. Hexanes were chosen as the precipitant because it could precipitate HPC-g-PCL and PCL homopolymer at different hexanes/THF ratios. It was found that by adding hexanes into the solution of the reaction product in THF, HPC-g-PCL started to precipitate out when the hexanes/THF ratio reached a certain value, and was completely precipitated when a sufficient amount of hexanes was added. Continuous adding of hexanes into the system resulted in the precipitation of PCL. CL stayed in supernatant since it was miscible with both hexanes and THF. The effect of the ratio of hexanes/THF (v/v) on the purification of HPC-g-PCL was examined and GPC was used to monitor the efficiency of the purification process. An efficient process was defined as the precipitated HPC-g-PCL contained all the graft copolymer in the reaction product but none of the PCL homopolymer. Any incomplete precipitation of HPC-g-PCL or precipitation of PCL was considered inefficient. The hexanes/THF ratios of 0.8/1,1/1,2/1, and 3/1 (v/v) were examined. At each hexanes/THF ratio, both precipitate and supernatant were collected and analyzed using GPC. The chromatograms are shown in Figure 3.4. The chromatogram of the original reaction product (Figure 3.4A) was included for comparison. When the hexanes/THF ratio of 3/1 was used, only part of the PCL homopolymer (component P) was removed, a large part of it remained in the precipitate as indicated by the peak corresponding to component P in Figure 3.4Ei. More PCL homopolymer was removed by decreasing the ratio to 2/1, but a small part of it still remained in the precipitate (Figure 3.4Di). When the ratio equalled 1, the peak for component P in Figure 3.4Q completely disappeared 91 ' 6.00 ' ' ' 8.00' ' 6.00" ' 8.00 Retention time (min) Retention time (min) Figure 3.4 GPC chromatograms of the precipitates and the superaatants when the reaction product was purified using different hexanes/THF ratios. (A) 0/1; (B) 0.8/1; (C) 1/1; (D) 2/1; and (E) 3/1. The subscript 1 denotes precipitate and 2 supernatant. 92 indicating that all PCL homopolymer in the reaction product was removed, and all HPC-g-PCL precipitated since no component G was detected in the supernatant (Figure 3.4C2). Further decrease of the ratio (0.8/1) resulted in incomplete precipitation of HPC-g-PCL. A small amount of HPC-g-PCL remained in the supernatant, as is indicated by the small peak corresponding to component G in Figure 3.4FJ2. Therefore, the hexanes/THF ratio of 1/1 (v/v) was determined to be the most efficient for the purification of the HPC-g-PCL. In order to ensure the purity of the HPC-g-PCL, the dissolving/precipitating process was repeated three times for all samples. 3.3.2 Evidence of graft copolymer formation using NMR spectroscopy ' H N M R spectroscopy is one of the most powerful tools for both qualitative and quantitative analysis of polymer structure. H M Q C spectroscopy was used to correlate the 1 3 C NMR spectra with *H NMR spectra and to obtain information on C-H connectivity. 13 In a typical H M Q C spectrum, the longitudinal axis corresponds to one-dimensional C NMR spectrum, and the transverse axis corresponds to one-dimensional *H NMR spectrum. COSY was used to correlate the coupled nuclei so that the changes in chemical shifts caused by the graft copolymerization could be identified. A typical COSY spectrum contains two identical proton chemical shifts scales on the longitudinal and transverse axes. In this work, COSY-45 was used so that the cross-peaks would occur only between the directly connected transitions. The acquisition parameters, such as solvent and temperature, had a large impact on the resolution of the NMR spectra. The effect of solvent is illustrated by the spectra in Figure 3.5, which shows the spectra of HPC in CDC13 (A) and DMSO-d 6 (B) recorded at 25 °C. The peak at 1.06 ppm (corresponding to methyl protons) was clearly resolved in 93 B ORi 6 C H 2 ORi R! = H or -f-CH 2 CHO-|-CH2CHOH L CH 3 Jm" CH 3 -CH 3 -CH 3 111111111111111111111111111111111111111111111111111111 n 1111111111111111111 n 11 70 6.0 5.0 4.0 3.0 2.0 1.0 (Ppm) Figure 3.5 ' H N M R spectra of HPC in (A) CDC13 and (B) DMSO-d 6 recorded at 25 °C using a Bruker WH400 spectrometer 94 both spectra. However, the peak for anomeric proton HI was resolved only in the spectrum using DMSO-d6 as the solvent. The peak was important for the determination of the molar substitution of both HPC and HPC-g-PCL. Therefore, DMSO-d 6 was selected as the solvent for the two polymers in this study. The effect of temperature is illustrated by comparing Figure 3.5B and Figure 3.6. The strong peak at 3.3 ppm in Figure 3.5B originated from water in the sample and overlapped with the peaks from the sample. This overlap made it impossible to obtain an accurate integration of the sample peaks in that region. By elevating the temperature to 95 °C, the water peak moved to 3.0 ppm and no longer interfered with the sample peaks. For this reason, all the spectra used for quantitation in this study were acquired at 95 °C. *H NMR and H M Q C spectra of HPC are shown in Figure 3.6 and 3.7, respectively. The spectra were recorded at 95°C using DMSO-d6 as a solvent in order to improve the resolution (Aden et al. 1984; Robitaille et al. 1991). In the ; H N M R spectrum (Figure 3.6), the small peak at 2.49 ppm was from the solvent DMSO, and the peak at 2.99 ppm originated from the hydroxy groups in HPC molecules and water absorbed by DMSO. Two peaks at 1.06 and 4.44 ppm were resolved, however, the peaks in the region of 3.2 to 4.2 ppm were overlapped. The peaks in lH N M R spectrum were partially assigned through H M Q C spectra (Figure 3.7) based on a previous study on 1 3 C NMR of HPC (Lee and Perlin 1982). The assignment is given in Table 3.2. The two peaks at the chemical shifts of 19.59 and 16.74 ppm in the 1 3 C N M R spectrum (the longitudinal axis of Figure 3.7a) were assigned to the methyl carbons in the hydroxypropyl side chains. The peak at 16.74 ppm corresponded to the carbons in the inner methyl groups (Cx) and the one at 19.59 ppm to the carbons in the methyl groups at 95 n Figure 3.6 *H NMR spectrum of HPC in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 96 X O <o ^ O — O * CM O I I * o X O H o II 97 HI Ppm 4 .6 4.4 4.2 Figure 3.7b. Expanded view of Figure 3.7a indicating the connection between CI-HI 98 Table 3.2 Observed chemical shifts (5) in the NMR spectra of HPC and PCL homopolymer (component P) formed during the reaction H P C a P C L b 1 3 c a 5 (ppm) l H a 5 (ppm) i H . 5 (ppm) CI 101.06 HI 4.44 H« 2.23 c x 16.74 H x 1.06 Hp,8 1.53-1.61 c x * 19.59 H x * 1.06 Hy 1.31 Cy (Cy*) 64.71 Hy (Hy*) 3.7-3.8 H e 3.99 C z (CZ*) 74.06 H z (H2*) 3.2-3.6 H E * 3.56 a The spectra were recorded in DMSO-d 6 at 95 °C using a Bruker AMX500 spectrometer b The spectrum was recorded in CDC13 at 25 °C using a Bruker WH400 spectrometer 99 the end of the side chains (Cx*) (Lee and Perlin 1982). The cross-peak 1 in the H M Q C spectrum revealed a connectivity between Cx (or Cx*) and the proton at 1.06 ppm. Therefore, the peak at 1.06 ppm in the J H NMR was assigned to the methyl protons (Hx) of the hydroxypropyl groups. The cross-peak in Figure 3.7b clearly indicated the connectivity between C l at 101.06 ppm and the proton at 4.44 ppm. Therefore, the peak at 4.44 ppm in the *H NMR spectrum was assigned to the anomeric proton HI. The assignments of Hx and HI was consistent with those reported in the literature (Robitaille et al. 1991) and allowed the determination of molar substitution (MS) of hydroxypropyl groups on cellulose backbone from the areas of the two peaks. Integrations of the peaks were shown in Figure 3.6. The peaks corresponding to the methine (Cy) and methylene (Cz) carbons in the hydroxypropyl groups were well resolved in the C N M R spectrum. Therefore, the peak positions of the methine (Hy) and methylene (Hz) protons could be located through the H M Q C spectrum. The peak for Hy was located at 3.7-3.8 ppm by cross-peak 2, which indicated the connectivity between Cy and the protons at 3.7-3.8 ppm. Similarly, the peak for Hz was located at 3.2-3.6 ppm. The determination of the peak position of Hy was important in proving the formation of HPC-g-PCL later. The *H NMR spectrum for component P is shown in Figure 3.8. The spectrum matches previously reported spectra for PCL (Kricheldorf and Kreiser 1987; Jacquier et al. 1996) and therefore, the component P was determined to be PCL. The peak assignments are given in Table 3.2. The ! H NMR spectrum of component G is shown in Figure 3.9. The spectrum contained peaks from both HPC and PCL indicating HPC-g-PCL formation. Based on the assignments of the spectra of HPC and PCL, the major peaks in Figure 3.9 could be 100 f a P Y 8 e C C H 2 C H 2 C H 2 C H 2 C H 2 O O t a P,8 M 1 ' ' 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 (ppm) Figure 3.8 ' H N M R spectrum of PCL homopolymer (component P) formed during the graft reaction in CDC13 recorded at 25°C using a Bruker WH400 spectrometer 101 r z y -i z* y* R 2 = H , 4 - C H 2 C H O - h C H 2 C H O H , or I C H , W C H 3 z y z * * y * * r cc p Y 8 e 1 a* p* Y* 8* e* 4 - C H 2 C H O - r - CH 2 CH04" C C H 2 C H 2 C H 2 C H 2 C H 2 0 + - C C H 2 C H 2 C H 2 C H 2 C H 2 O H I C H , J m ' C H 3 1 O m, m' = 0, 1, 2, P,8 Figure 3.9 *H N M R spectrum of HPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 102 assigned. The COSY-45 spectrum shown in Figure 3.10 provided more information about the interactions between protons within HPC-g-PCL and enabled the assignment of the small peaks in the ' H NMR spectrum (Figure 3.9). Through the H M Q C spectrum shown in Figure 3.11, major peaks in 1 3 C NMR spectrum were assigned. The peak assignments are given in Table 3.3. The labelling of the protons and carbons is shown in Scheme 3.1. The methyl peak at 1.06 ppm in the ! H NMR spectrum of HPC (Figure 3.6) split into two in the ! H NMR spectrum of HPC-g-PCL, with one remaining at the chemical shift of 1.06 ppm and another shifting to a lower field (1.16 ppm) (Figure 3.9). The shifted peak corresponded to the methyl protons in hydroxypropyl groups connecting directly to PCL chains (Hx**). The shift is due to the de-shielding effect of ester groups formed during the graft reaction. In the 1 3 C NMR spectrum, the peak corresponding to Cx»* shifted from 19.42 to 15.99 ppm. In addition, there was another small broad peak at 4.9 ppm in the [ H NMR spectrum of HPC-g-PCL. Neither HPC nor PCL had this peak in their spectra. The cross-peak 1 in the COSY-45 spectrum (Figure 3.10) indicated that the protons corresponding to this peak (at 4.9 ppm) coupled with the methyl protons having a chemical shift of 1.16 ppm. Therefore, these are methine protons (Hy**) in hydroxypropyl groups connecting directly to PCL chains. Before grafting, the methine protons (H y and Hy*) were located in the range of 3.7-3.8 ppm, as shown in Table 3.2. This was also supported by cross-peaks 2 and 3 in Figure 3.10, which revealed the coupling between H x * -H y * and H x - H y , respectively. The shift from 3.7 to 4.9 ppm is a typical change for an acylation reaction (Simons and Zanger 1972). This was considered to be additional evidence for graft copolymer formation. 103 Figure 3.10 COSY-45 spectrum of rIPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AV400 spectrometer 104 Figure 3.11 H M Q C spectrum of HPC-g-PCL (component G) in DMSO-d 6 recorded at 95 °C using a Bruker AMX500 spectrometer 105 Table 3.3. Observed chemical shifts (8) in the 1 3 C and ! H N M R spectra of purified HPC-g-PCL a 13cb 5 (ppm) 8 (ppm) c x 16.56 H x 1.06 C x * 19.42 H x * 1.06 cx** 15.99 Hx*» 1.16 Cy (Cy*) 64.97 Hy (Hy.) -3.7 Cy** 68.56 Hy** 4.90 C z (Cz*) 74.08 H z (Hz*) 3.25-3.35 Ca 32.97 (33.18) H a 2.27 Cp 23.52 (23.86) Hp 1.56 Cy 24.41 (24.55) Hy 1.35 C 5 27.31 H 8 1.59 Cg* 31.55 Hg* 1.44 c6 62.91 H E 4.01 ce* 60.18 H E * 3.42 c=o 171.95 a The spectra were recorded in DMSO-d6 at 95 °C using a Bruker AMX500 spectrometer b The labelling of the protons and carbons is shown in Scheme 3.1 106 The cross-peak 4 in the COSY spectrum in Figure 3.10 indicates that the protons at the chemical shift of 1.44 ppm coupled with s* protons (s protons in the repeating units at PCL chain ends). It is known from the structure of PCL that s protons only couple with 8 protons. Therefore, the small peak at 1.44 ppm was assigned to 8 protons in the repeating units at PCL chain ends (8*). In the 1 3 C NMR spectrum of HPC-g-PCL, each of the peaks for C a , Cp, and C y had a small peak at a chemical shift (the value in brackets in Table 3.3) slightly higher than that of the corresponding main peak. A similar situation was also reported for starch-g-PCL (Choi et al. 1999). Based on the peak intensity, we speculate that these peaks corresponded to C a», Cp*, and CY* in the repeating units at the chain ends. 3.3.3 Molar substitution The way in which substituents are bonded to a polysaccharide molecule may be described in terms of molar substitution (MS), which is the average number of substituent molecules that have reacted with each anhydroglucose unit (Ho et al. 1972). In HPC-g-PCL molecules, there are two different substituents on the cellulose chains: hydroxypropyl groups and PCL chains. In order to calculate the MS of C L on HPC chains (MSCL), it is necessary to firstly determine the MS of hydroxypropyl groups on cellulose chains (MSHP). M S H P was estimated using the ' H NMR spectroscopy method (Ho et al. 1972; Robitaille et al. 1991). The method is based on the fact that each different proton in different structural groups gives rise to peaks at a characteristic magnetic field strength, and the peak intensity is directly proportional to the concentration of the protons. Peak intensity in a ! H N M R spectrum is affected by experimental parameters such as pulse 107 angle and relaxation delay. The ! H N M R spectra for both HPC and HPC-g-PCL were acquired using a pulse angle of 3 6 ° and a relaxation delay of 2 seconds. To ensure reliable peak intensities, spectra were also acquired at relaxation delays of 4 and 8 seconds, respectively (data not shown). The spectra obtained using the three different relaxation delays gave identical peak intensities indicating that the peak intensities used in the following calculations were quantitatively reliable. As shown in the structure of HPC in Scheme 3 .1, each hydroxypropyl group brings to the cellulose skeleton 3 methyl protons, and each anhydroglucose unit has one anomeric proton connecting to CI. Therefore, M S H P can be determined using the peak intensities of these two types of protons in the HPC spectrum. M S H P = (Icm / 3) / Im Equation 3.1 where Too is the peak intensity of the methyl group and Im is that of the anomeric proton at CI. The molar substitution of the hydroxypropyl group on the cellulose chain was calculated to be 6.5, which means on average there are 6.5 hydroxypropyl groups connected to each anhydroglucose unit. Based on this value, the average formula weight of the repeating unit of HPC was calculated to be 539. MSCL can be estimated in a similar way using the relative intensities of peaks for PCL and HPC in Figure 3.8. The calculation is given as follows: MSCL= ^ Equation 3.2 ICHJ3MSHP where Iua is the peak intensity for H a of PCL, Icm is the sum of the intensities of peaks at both 1.06 ppm and 1.16 ppm. Using this method, the MSCL for batch 4 was determined to be 22 .7 , which meant on average there were 22.7 CL monomers reacted with each HPC repeating unit. This value is even higher than the initial feed molar ratio 108 (18.9). Since the reliability of peak intensity was verified, the errors are considered to be mainly from peak integration. As shown in Figure 3.9, the peaks from backbone HPC are broad and small compared to those from PCL. This can bring about significant errors in the peak integration, which subsequently affect the results of MSCL-MSCL was also determined using a gravimetric method. The calculation is given as follows: = KAFT-WHPC)IUA Equation 3.3 »W/539 where Wgraft is the weight of HPC-g-PCL after purification; WHPC is the initial weight of HPC; the number 114 is the formula weight of repeating unit of PCL; and 539 is the average formula weight of the repeating unit of HPC. In this calculation, HPC was assumed to be 100% converted to HPC-g-PCL. M S C L values ranging from 8.6 to 10.1 were obtained for different batches (shown in Table 3.1), indicating that on average there were 8.6-10.1 CL monomers reacted with each HPC repeating unit. 3.3.4 Side chain length and distribution Based on the assignment and integration of [ H N M R spectrum of HPC-g-PCL, the average length of PCL side chains and the number of the side chains on each HPC repeating unit can be estimated. The average length of PCL chains (L ) can be calculated from the peak intensity of e protons at the chain ends (e*) and the total peak intensities of all e protons in PCL chains. However, an accurate peak intensity for 8* proton cannot be obtained due to the peak overlap. Theoretically, the number of s protons equals that of a protons. Therefore, the peak intensity for a protons is used as an equivalent of the total intensity for all s protons and the calculation is given by: 109 where IHE represents the peak intensities of the s protons that are not at the chain end and Ina the peak intensity of a protons. The L of HPC-g-PCL synthesized in batch 4 was determined to be 7.3. The average number of PCL chains on each HPC repeating unit (N) was then calculated by: N = MSCL l l Equation 3.5 Based on the MSCL determined using gravimetric method, TV for HPC-g-PCL in batch 4 is 1.2. Hence, on average, there are 12 PCL side chains attached to 10 HPC repeating units. 3.3.5 Morphological study of HPC-g-PCL Two-phase morphology is commonly observed for graft copolymers with components differing extensively in their properties (Noshay and McGrath 1977). In HPC-g-PCL, the backbone HPC is water-soluble, while the side chain PCL is very hydrophobic. An HPC-g-PCL film cast from THF solution was examined using an optical microscope with polarizing lens. The regular pattern shown in Figure 3.12 suggests heterogeneity in the film. An X-ray diffraction pattern of the HPC-g-PCL film is shown in Figure 3.13. Diffraction peaks were located at the diffraction angles 29 of 21.66°, and 23.96°, in agreement with those reported for PCL in the literature (Ong and Price 1978). These results suggested that the partially crystalline phase in the film of HPC-g-PCL could be attributed to PCL. 110 Figure 3.12 Optical microscopy of HPC-g-PCL film under polarizing lens 111 9000 21.66° 10 15 20 25 30 2 theta (degrees) Figure 3.13 Representative X-ray diffraction pattern of HPC-g-PCL film cast from THF solution 112 DSC analysis provided further evidence of the existence of a partially crystalline PCL phase in the HPC-g-PCL film. In Figure 3.14, the scan for HPC-g-PCL shows one glass transition (Tg) and three melting peaks. The Tg occurred at -72 °C (onset), in agreement with the Tg for PCL homopolymer. The broad melting peak at 17.8°C (peak 3) may originate from the backbone HPC, which has a similar broad melting peak at the same position (peak 4). There is a shoulder ("peak" 2) appearing on the main endothermic peak 1 or what will be referred to as a double endotherm in the range of 20-50°C. The double endotherm likely corresponds to the melting of PCL side chains. 3.3.6 Characterization of the fractions of HPC-g-PCL Purified HPC-g-PCL was fractionated and the fractions were characterized by GPC, DSC and ' H NMR. The molar fractions of hexanes in the precipitating solvent, the molecular weights and molecular weight distributions of the fractions are given in Table 3.4. [ H NMR spectra of the fractions (data not shown) were obtained at 25°C and all four fractions had identical ! H NMR spectra indicating that the four fractions were graft copolymer HPC-g-PCL. Hence, the assumption that 100% of HPC was converted to HPC-g-PCL was likely appropriate. DSC scans of the four fractions are shown in Figure 3.15. Thermal transition temperatures (Tg and Tm) and heat of fusion (AHf) data are given in Table 3.4. The Tg's for the four fractions were similar at approximately -72°C, and all four DSC scans possessed a double-melting endotherm in the temperature range of 20 to 50 °C. However, the relative sizes of peak 1 and peak 2 in the double-melting endotherm were quite different. In the scan for F l , peak 1 constituted the major part of the melting endotherm. 113 H P C -100 -50 0 50 100 Temperature (°C) Figure 3.14 Representative DSC scans of HPC and purified HPC-g-PCL. Heating rate was 10 °C/min. 114 Table 3.4. Physicochemical data on HPC-g-PCL and its four fractions produced by precipitating from the THF solution using different molar fractions of hexanes Fraction x H a M wb kg.mol"1 M w / M n b T g c ( ° C ) Tm d (°C) AH f e(J/g) FO — 299 1.8 -72.311.8 47.112.4 28.312.1 F l 0.327 133 1.3 -71.610.7 46.211.5 36.511.8 F2 0.275 184 1.3 -72.411.5 44.111.3 30.411.7 F3 0.249 330 1.3 -72.810.5 42.010.9 19.811.1 F4 0.240 554 1.1 -72.710.6 41.510.9 11.211.7 a Molar fractions of hexanes in the precipitating solvent b Weight average of molecular weight and polydispersity index, measured by GPC in THF at 1 mL/min. The columns used were Styragel® HR3 and HR4 in series, c Onset values of the glass transitions d Endpoint of the melting endotherm e Heat of fusion of the double melting endotherm The values in c, d and e represent mean 1 SD (n=3). 115 -100 -50 0 50 100 Temperature (°Q Figure 3.15 Representative DSC scans for HPC-g-PCL (F0) and the four fractions: F l , F2, F3, and F4. The samples were films cast from THF solutions. The heating rate was 10°C/min. 116 The relative sizes of peak 2 increased with increasing molecular weight of the fractions. In the scan for F4, the size of peak 2 exceeded that of peak 1 and became a major component of the melting endotherm. These results suggest that the two peaks in the HPC-g-PCL double-melting endotherm originated from different molecular weight fractions in the copolymer. The high molecular weight fractions had lower melting ranges and were largely responsible for the lower temperature part of the melting endotherm. The low molecular weight fractions had higher melting ranges and were mainly responsible for the higher temperature part of the melting endotherm. 3.3.7 In vitro degradation of HPC-g-PCL Figure 3.16 shows the changes in weight-average molecular weight (M w , left y-axis) and polydispersity index (PDI, right y-axis) of HPC-g-PCL during incubation in PBS (lOmM, pH7.4) at 37°C. At week 0, HPC-g-PCL had a Mw of 294 kg/mol and a PDI of 1.95. At week 12, the Mw and PDI were 296 kg/mol and 1.96, respectively. There was a negligible change in both Mw and PDI during the 12 weeks incubation indicating that HPC-g-PCL did not undergo significant degradation under these experimental conditions. Changes in thermal events in the films during the incubation was examined using DSC. The changes in DSC scans following incubation for different times are shown in Figure 3.17, in which the curve A was the scan of the film before the incubation (week 0), and the curves B, C, and D were the scans of the films incubated for 1 week, 4 weeks, and 12 weeks, respectively. Changes in the scans were observed particularly during the first week of incubation. The double endotherm at week 0 was separated into two distinct peaks during the first week. Peak 1 shifted from 42.1 °C to 45.9 °C, and peak 2 stayed at 117 400 360 h o S |? 320 280 240 2.2 1.8 Q P H 1.4 1 2 4 6 8 10 12 Time (week) Figure 3.16 Changes in weight-average molecular weight (•) and polydispersity index ( A ) of the HPC-g-PCL films incubated in 10 mM PBS (pH7.4) at 37 °C 118 -100 -50 0 50 100 Temperature (°C) Figure 3.17 Representative DSC scans of HPC-g-PCL films incubated in 10 mM PBS (pH7.4) at 37 °C for different lengths of time. (A) 0 week; (B) 1 week; (C) 4 weeks; and (D) 12 weeks. Heating rate was 10 °C/min. 119 60 Figure 3.18 Changes of the peak position in the DSC scans of HPC-g-PCL films incubated in 10 mM PBS (pH7.4) at 37 °C. (•) peak 1; (•) peak 2; and ( A ) peak 3. Each datum point represents mean ± SD (n=3). 120 the original position. The position of peak 3 remained unchanged. No significant change was observed for all three peaks after the first week. A plot of the peak positions of the three peaks versus the incubation time is shown in Figure 3.18, in which each datum point represents mean ± SD (n=3). 3.4 DISCUSSION Ring-opening polymerization of CL either to form a homopolymer or copolymers has been investigated extensively. In most of the studies, the polymerization was initiated by compounds containing hydroxyl groups, such as alcohols, and Lewis acids, such as stannous 2-ethylhexanoate and various mechanisms were proposed. In some studies, it was suggested that Lewis acids were acting as catalysts and that the acids were not covalently bonded to the polymer chains (Schindler et al. 1982; Du et al. 1995; Kricheldorf et al. 1995; Schwach et al. 1997; Veld et al. 1997). In other works (Kowalski et al. 2000; Kowalski et al. 2000; Ryner et al. 2001), Lewis acids were considered to be co-initiators, which covalently bonded to the polymer chains. However, reports clearly show that hydroxyl group-containing compounds alone can initiate CL polymerization. Without using Lewis acids, polymerizations are restricted to the synthesis of low molecular weight polymers (Brode and Koleske 1972; Schindler et al. 1982). In this work, the purpose of the graft copolymerisation of C L was to confer increased hydrophobicity on the HPC and make it more suitable for the incorporation of hydrophobic paclitaxel. The hydrophobicity of the final copolymer should be appropriate to ensure that the copolymer biomaterial retains some hydrophilic characteristics, which generally confers a greater degree of biocompatibility in the biomaterial. Therefore, PCL 121 side chains were intended to be short, and the restriction caused by not using a catalyst was not considered a problem. Generally, the initiation step of lactone polymerization involves ring opening with the formation of an ©-hydroxy ester (Schindler et al. 1982): The propagation reaction then proceeds by the stepwise addition of the lactone monomer to the terminal hydroxyl group. As shown in Scheme 3.1, HPC is a hydroxyl group rich polymer. On each repeat unit, there are three hydroxyl groups. The graft copolymerization of CL was initiated by these hydroxyl groups. It was reported that, without using a catalyst, the polymerization rate was slow even at elevated temperature (160-180 °C) (Brode and Koleske 1972). In this study, the reaction was carried out at 150 °C and CL conversion reached over 70% for the small batches and over 60% for the large batches within 24 hours. This was considered to be an acceptable rate of polymerization. The key factors affecting the graft reaction were water (moisture) and oxygen in the reaction system. Water was reported to play an important role in the polymerization of CL. Similar to an alcoholic hydroxyl group, it can initiate the polymerization and form linear homopolymers (Schindler et al. 1982). However, since the goal in this study was to synthesize graft copolymer, the formation of PCL homopolymer was considered to be a side reaction. Efforts were made to remove water/moisture in the reaction system and minimize the side reaction. Both the starting materials HPC and monomer CL were dried before being used for the reaction. HPC was dried under vacuum (0.3 mmHg) for 16 R O H + 122 hours. CL was dried over CaEk and distilled twice under vacuum (0.3 mmHg). Although the homopolymer still formed, the yields were quite low (less than 16%). Oxygen is a powerful inhibitor for most polymerization reactions. The mechanism of inhibition is not fully understood. The presence of oxygen in a polymerization system results in a decreased reaction rate, lowered molecular weight and even complete inhibition of the reaction. In this study, oxygen was removed by applying high vacuum to the system and then purging with nitrogen gas. The vacuum/purging procedures were repeated three times to ensure a minimum oxygen level. From a practical point of view, it is not possible to obtain absolutely water and oxygen free systems. However, the procedures used in this study minimized their influence to such an extent that the reactions were well controlled and highly reproducible. The product of a graft copolymerisation is most likely a mixture of the intended graft copolymer, unreacted backbone, homopolymer of the graft and unreacted monomer. It is always a challenge to purify graft copolymers since they often act as emulsifiers to reduce incompatibility of chemically different polymeric species (Dreyfuss and Quirk 1987). Two different types of methods have been used for the purification of graft copolymers, these being dialysis (Donabedian and McCarthy 1998) and fractional precipitation (Robitaille et al. 1991; Ydens et al. 2000). The dialysis method is based on the difference in molecular sizes of graft copolymer and the impurities, while the fractional precipitation method is based on solubility or polarity differences. In this study, both methods could have been used for the separation of HPC-g-PCL and PCL since both the molecular weights and the solubility of the two polymers are different from each 123 other. Fractional precipitation was used because it was less time-consuming than dialysis. The reaction product was dissolved in THF and the addition of hexanes resulted first in the precipitation of HPC-g-PCL. This solubility difference between HPC-g-PCL and PCL homopolymer was considered to be mainly due to the dramatic differences in molecular weights rather than differences in polarity of the polymers. Graft copolymers have been regarded as the most difficult type of polymer to accurately characterize (Noshay and McGrath 1977). There are a number of complex structural issues to address, including the number of side chains on the backbone, the length of side chains and the polydispersity of the side chains. Many of these issues remain largely unsolved despite the advancement in polymer characterization techniques. In the case of HPC-g-PCL, structural characterization was complicated because the backbone HPC itself was a graft copolymer. The structure of HPC-g-PCL was partially elucidated using NMR spectroscopy. The structural characterization of HPC using *H NMR (Ho et al. 1972) and 1 3 C NMR (Lee and Perlin 1982) was reported previously. However, no complete assignment 13 of the spectra was achieved due to the complexity of the molecular structure. C-chemical shifts of the hydroxypropyl substituents were assigned experimentally and those of the cellulosic skeleton were calculated based on the substituent effects obtained from the spectra of hydroxypropyl derivatives of D-glucose (Lee and Perlin 1982). The ' H NMR spectrum recorded in CDCI3 by Ho et. al. was poorly resolved and M S H P was determined based on the spectrum using the following equation: MSUD = HP 3(B-A) Equation 3.6 124 where A designates the peak area of methyl protons and B the total area of the peaks (not resolved) of the rest of the protons ranging from 2.5 to 6.0 ppm. In this study, the ' H NMR spectrum recorded in DMSO-d6 at 95 °C was better resolved, and in particular the HI peak was clearly identified. This allowed the M S H P to be determined according to Equation 3.1 using the peak areas of HI and methyl protons. This method was also used by other groups (Aden et al. 1984; Robitaille et al. 1991). An M S H P value of 6.5 was obtained, indicating that on average each anhydroglucose unit in HPC molecules was substituted with 6.5 hydroxypropyl units. The value is in good agreement with the literature value (Fortin and Charlet 1989). The exact position and length of the side chains are not known. Studies on the reactivity of the hydroxyl groups on the cellulose backbone suggested that their rates of etherification were in an order of 1:1:0.7 for the C6-OH, C2-OH and C3-OH, respectively, when they reacted with the first propylene oxide, and after that, all three hydroxyl groups became equally reactive (Lee and Perlin 1982). Therefore, it is believed that there might exist unsubstituted hydroxyl groups on the cellulose backbone and that the hydroxypropyl substituents are likely bonded to C6 and C2. Estimation of the molar substitution of CL on HPC (MSCL) using ' H NMR spectroscopy resulted in a value higher than the initial feed ratio, and therefore was an unsuitable method in this case. A similar situation has been documented by Ydens et. al., who concluded that ! H NMR method usually resulted in higher M S C L values than the gravimetric method, especially when M S C L was high (Ydens et al. 2000). The error was considered to originate mainly from the determination of peak areas of H a and methyl protons. In the ! H N M R spectrum of HPC-g-PCL, peaks corresponding to HPC backbone 125 were broad and small. The spectrum was dominated by the strong peaks from PCL side chains because of the large number and the high mobility of PCL chains in the copolymer. This made it very difficult to obtain accurate integration of the peaks from HPC. The sources of error could also include the error carried over by M S H P according to Equation 3.2. In the gravimetric method, the error might originate mainly from the determination of Wgraft in Equation 3.5, which is the weight of purified HPC-g-PCL. The purification process involved several procedures. Any loss of the copolymer in these procedures might affect the accurate determination of Wgmfu and lead to an underestimation of M S C L - The accuracy of M S H P could also have an effect on the determination of M S C L since the formula weight of HPC repeating unit (the number "539" in Equation 3.5) was calculated based on the value of M S H P . The result of the average side chain length was considered to be accurate since the estimation was solely based on the intensities of PCL peaks, which are narrow and strong. The average number of PCL chains on each HPC repeating unit was estimated to be 1.2, indicating that most of the HPC repeating units have only one side chain. There was no direct evidence of the exact position at which PCL side chains were attached. However, it is likely that PCL side chains attached equally to any of the ends of the hydroxypropyl substituents, since the hydroxyl groups on all these substitutents have equal reactivity (Lee and Perlin 1982). The possibility of any unsubstituted hydroxyl groups on cellulosic units, if there were any, reacting with CL to form PCL side chains was considered to be negligible due to the significant steric hindrance. 126 The HPC-g-PCL film did not undergo any significant degradation over a period of 12 weeks in 10 mM PBS (pH7.4). PCL homopolymer undergoes hydrolytic chain scission of the ester linkages in an aqueous environment, but this process is slow and PCL homopolymer possesses a long degradation lifetime (Pitt 1990). The complete lack of a change in molecular weight might be an effect of the special initiation process in the graft reaction. It has been reported that the nature of the initiator has a dramatic impact on the thermal and/or hydrolytic stability of PCL (Brode and Koleske 1972). In this study, CL polymerization was initiated by HPC, a high molecular weight polymer. HPC is stable in aqueous media with pH between 6.0-8.0 and is quite resistant to degradation by bacteria and molds (Kibbe 2000). The stability of HPC-g-PCL may be a result of this "stable macromolecular initiator". Heterogeneity in the films of HPC-g-PCL was suggested by the regular pattern (Figure 3.12) observed using optical microscopy and the bright spots in the regular pattern were determined to be PCL-rich regions. The existence of PCL-rich regions in the films indicates that the hydrophobicity of the HPC-g-PCL molecules exerted by the PCL side chains was sufficient to drive the HPC-g-PCL molecules to align themselves in such a way that PCL side chains formed microcrystalline regions. The double endotherm observed in DSC scans was found to originate from different molecular weight fractions of HPC-g-PCL (Figure 3.15). The low melting endotherm was associated with the high molecular weight fractions and the high melting endotherm with the low molecular weight fractions. This is an unexpected finding based on the normal relationship between melting point and molecular weight, and is likely a result of the preparation method of the films. In this work, the films for the DSC measurements were prepared by the solution 127 casting method. The solutions of high molecular weight fractions had a higher viscosity than those of low molecular weight fractions. The higher viscosity solutions might hinder the movement of the polymer chains during solidification, thus resulting in crystallites with a lower degree of perfection. The high molecular weight fractions would then have lower melting points than the low molecular weight fractions. The significant difference in AHf for the different fractions (Table 3.4) may be a function of crystallinity and/or degree of perfection of crystallites. The changes in DSC scans of HPC-g-PCL films incubated in PBS shown in Figure 3.17, revealed that the copolymer chains may have undergone a relaxation process during incubation at 37 °C. During incubation, peak 1 (low molecular weight fraction) shifted to a higher temperature, while the position of peak 2 (high molecular weight fraction) remained unchanged. The shifting of peak 1 may indicate that the process of film formation by solvent evaporation and precipitation of polymer occurred rapidly and did not allow the copolymer chains to fully relax to their lowest energy states. When the films were incubated at a higher temperature (37 °C), the chains gained energy and gradually re-aligned themselves into a more ordered structure. Therefore, the melting temperature increased. The absence of a change in the position of peak 2 may be a result of low mobility and possible entanglement of the high molecular weight fractions. 3.5 CONCLUSIONS An amphiphilic graft copolymer, HPC-g-PCL, was reproducibly synthesized by bulk polymerization without using a catalyst. The evidence for the graft reaction was obtained using both one- and two-dimensional NMR spectroscopy. The downfield shift of the *H NMR peaks corresponding to the tertiary and methyl protons in the 128 hydroxypropyl groups was considered to be direct evidence for the graft polymerization. Molar substitution of CL on HPC was determined to be in the range of 8.6 to 10.1 by a gravimetric method and the average PCL side chain length was estimated to be 7.3 by NMR spectroscopy. Heterogeneity was observed in the HPC-g-PCL films cast from THF solutions. DSC and X-ray diffraction data indicated that the partially crystalline phase was PCL. The double-melting endotherm observed in DSC scans was found to originate from different molecular weight fractions of HPC-g-PCL. The low melting endotherm was associated with the high molecular weight fractions and the high melting endotherm with the low molecular weight fractions. No degradation was observed during the incubation of the HPC-g-PCL films over a three-month period. 129 CHAPTER 4 SYNTHESIS AND CHARACTERIZATION OF DEXTRAN-GRAFT-POLY(e-CAPROLACTONE) 4.1 INTRODUCTION Dextrans are a class of polysaccharides synthesized by bacteria, when grown on sucrose-containing media, to give D-glucans with contiguous a(l—»6) glucosidic linkage in the main chains and a variable amount of a(l->2), a ( l - » 3 ) and a(l->4) branch linkages (Robyt 1985). Dextrans have been produced by two principal genera of bacteria, namely, Leuconostoc and Streptococcus . The chemical and physical properties of dextrans depend upon the strain of bacteria employed and the conditions for the bacterium growing. Leuconostoc mesenteroides B-512F is one of the most widely used strains for producing dextrans. The dextran produced from this strain is composed of about 95% a(l->6) linkages and 5% a ( l - » 3 ) linkages (Robyt 1985; Tsuchiya 1992) and has molecular weights ranging from 5 x 107 to 5 x 108 (Bovey 1959). The structure of dextran is shown in Scheme 4.1. It is highly water-soluble and is also soluble in DMSO, formamide, ethylene glycol and glycerol, but is not soluble in monohydric alcohol and other organic solvents (Belder 1996). In vitro degradation of dextran under different conditions has been well documented. Dry dextran was found to begin degrading when heated in a vacuum at 100 °C. Discoloration occurred when the dextran was heated at 150 °C (Stacey and Pautard 1952). Heating in hydrochloric acid solution has been used to prepare low molecular 130 weight dextran (Bixler et al. 1953). Enzymatic degradation of dextran was observed in both animals and humans. Dextran was found to be degraded in vivo by dextranase, an enzyme found in liver, spleen, and kidneys, and excreted mainly via the kidneys (Terry et al. 1953; Gruber 1976). After administration by intravenous infusion, dextran with a molecular weight lower than 7,000 can freely pass across the glomerulus. The elimination rate decreases with the increase of molecular weight. Chemical modification of dextran impairs enzyme-substrate interactions and results in a lower degradation rate. The higher the degree of modification, the lower the degradation rate. The type of chemical modification had a minor influence on the rate of degradation (Park et al. 1993). Biomedical applications of dextran have been extensively investigated since it was used as a plasma volume expander in the 1940s, including blood flow improvement, thrombosis treatment, organ perfusion and preservation, and distending media for hysteroscopy (Belder 1996). The physicochemical properties of dextran, such as molecular weight and molecular weight distribution, have a major impact on the outcome of these applications. Commercial dextrans with different molecular weights are produced by controlled hydrolysis of the natural dextran with hydrochloric acid at 100-105 °C followed by repeated fractionation (Bixler et al. 1953). A plasma volume expander is an aqueous solution capable of restoring and maintaining the plasma volume and the colloid osmotic pressure after an acute hemorrhage or more generally, in shock treatments (Dellacherie 1996). Clinically used dextran for plasma volume expansion has a molecular weight of 50,000 to 100,000 (Robyt 1985). Dextran with a molecular weight of 40,000 was used for the improvement of blood flow by reducing blood viscosity and inhibiting erythrocyte aggregation (Belder 1996). Dextran was also shown to be effective 131 as a prophylactic agent against postoperative thrombo-embolism (Ljungstrom 1988). Dextran has been one of the ingredients in solutions for the preservation of lung (Chien et al. 2000; Struber et al. 2000), liver (Jamieson et al. 1989; Sumimoto et al. 1989), and kidney (Schlumpf et al. 1991). The solutions serve to maintain cell integrity, minimize hypothermia-induced swelling, and protect tissues from free radical injury and intracellular acidosis. A 32% solution of dextran (molecular weight of 70,000) with 10% glucose has been used in diagnostic and operative hysteroscopy as a distending medium (Amin and Neuwirth 1983; Pellicer and Diamond 1988). Dextran (molecular weight of 70,000) has also been used in the prevention of surgical adhesions. Data obtained from both animal and human studies showed that a concentrated dextran solution (32%) was of significant benefit in the minimization of surgical adhesion formation and reformation after lysis (Rosenberg 1990). A "hydroflotation theory" has been proposed to explain its mechanism of action. When the solution is placed in the abdomen, the abdominal contents are "floated" apart, reducing the likelihood of adhesion formation. Dextran was found to be able to modulate platelet function, inhibit thrombin formation, and render a clot more susceptible to fibrinolysis (Wiseman 1994). These properties also contributed to the efficacy in the prevention of surgical adhesions. Dextran biomaterials have been used as carriers for a variety of bioactive agents such as drugs, peptides, proteins and enzymes. These include antimicrobial drugs (e.g. ampicillin, kanamycin, and tetracycline), cytostatic drugs (e.g. mitomycin, methotrexate, and bleomycin), cardiac drugs (e.g. alprenolol and procainamide), insulin, immunoglobulin G, bovine carbonic anhydrase, and catalase (Park et al. 1993; Kumar 132 and Banker 1996). These agents can be either physically entrapped into dextran matrices or form conjugates with dextran. Due to the presence of a large number of hydroxyl groups, many studies have focused on dextran-drug/agent conjugates. Low molecular weight drugs can be conjugated to dextran by direct esterification, periodate oxidation, or cyanogens bromide activation (Schacht et al. 1988; Vansteenkiste et al. 1990). Frequently, a spacer is used between the activated dextran and the drugs. For instance, conjugation of mitomycin C, an antitumor drug, to dextran involved three steps: activation of dextran with cyanogen bromide, coupling the activated dextran with a spacer (e-aminocaproic), and then attaching the drug. The conjugated mitomycin C showed improved efficacy after intravenous injection (Kojima et al. 1980). Dextran-enzyme conjugates are generally prepared through cyanogen bromide or periodate activation. Enzymes such as lysozyme, chymotrypsin, and (3-glucosidase have been conjugated to dextran. The conjugation protected the enzymes from being inactivated and significantly increased the in vivo lifetime of enzymes (Vegarud and Christnsen 1975). Hydrophobic drugs, such as paclitaxel, precipitate out when they are directly loaded into dextran matrices due to the highly hydrophilic nature of dextran. Similar to the graft copolymer HPC-g-PCL discussed in Chapter 3, an amphiphilic copolymer of dextran with PCL as side chains should combine properties of the hydrophilic dextran with hydrophobic PCL to produce a copolymer with good biocompatibility and enhanced capacity for solubilizing paclitaxel. The syntheses of amphiphilic graft copolymers with dextran as the backbone have been reported previously. Dextran-graft-poly(D,L-lactic acid) was synthesized via a "graft onto" process using an U V photopolymerization method (Zhang et al. 1999). 133 Dextran and poly(D,L -lactic acid) (PDLLA) were modified firstly with unsaturated vinyl groups to form dextran-acrylate and PDLLA-diacrylate macromers, respectively. The macromers were then photocrosslinked by low-intensity U V light. The copolymer has been used as a carrier for the controlled release of albumin and indomethacin (Zhang and Chu 2000; Zhang and Chu 2002). Using a "graft onto" process, Jong et al. also synthesized dextran graft copolymers with both poly(L-lactic acid) and poly(D-lactic acid) as the side chains. The hydrophilic dextran and hydrophobic poly(lactic acid) (PLA) were coupled using a chemical method instead of U V light. The hydroxyl groups of PLA were firstly activated with N,N'-carbonyldiimidazole in THF solution and then the coupling reaction was carried out in DMSO solution using 4-N,N-dimethylaminopyridine as the catalyst (de Jong et al. 2001). Graft copolymers of dextran (molecular weights of 6,600 and 21,300) with PCL as the side chains were synthesized via a "graft from" process by solution polymerization in toluene (Ydens et al. 2000). The dextrans were modified by partial silylation by reacting with hexamethyldisilazane to make them soluble in toluene. The silylated dextrans were then reacted with C L to form the graft copolymers. After the formation of the graft copolymers, the trimethylsilyl groups were removed by hydrolysis using hydrochloric acid (1 M). The objectives of this study were to synthesize graft copolymers of dextrans with short PCL side chains by ring opening polymerization and to characterize the copolymers. Dextrans with two different molecular weights were used as the main chain polymers. 134 4.2 E X P E R I M E N T A L 4.2.1 Materials Dextran70 (Dx70) (Mw = 69,000, Mw/Mn = 1.6) and dextran500 (Dx500) (Mw = 473,000, Mw/Mn = 2.4) were purchased from Pharmacia (Sweden) and were dried under vacuum (0.3 mmHg) at 60 °C for 16 hours before being used. Dimethyl sulfoxide (DMSO, Fisher) was distilled over molecular sieves (4A type, Fisher) under vacuum with the first 20% of the distillate discarded. The distillate was sequentially dried with two batches of 4A molecular sieves for a total of 96 hours (48 hours per batch). The dried DMSO was stored over 4A molecular sieves. s-caprolactone (CL) was purchased from Fluka. It was dried over calcium hydride (CaF^, from Fluka) for 24 hours and distilled twice under a reduced pressure. Stannous 2-ethylhexanoate (95%, Sigma), tetrahydrofuran (THF) (HPLC grade, Fisher Scientific), hexanes (certified A.C.S, Fisher Scientific), methanol (HPLC grade, Fisher Scientific), acetonitrile (HPLC grade, Fisher Scientific), phosphorus pentoxide (P2O5) (Laboratory grade, Fisher Scientific), deuterated dimethyl sulfoxide (DMSO-d6) (D, 99.9%, Cambridge Isotope Laboratories, Inc.) and deuterated chloroform (CDCI3) (D, 99.8%, Cambridge Isotope Laboratories, Inc.), were used as received. 4.2.2 Synthesis of Dextran-g-PCL Graft copolymers with Dx500 and Dx70 as main chains and PCL as side chains (Dx-g-PCLs) were synthesized by solution polymerization using stannous 2-ethylhexanoate as the catalyst. The glass reactors were flamed and then cooled down to room temperature under vacuum and purged with nitrogen gas. Dextran and DMSO were added into the reactors at a ratio of 1:16 (w/v) and placed in a 60 °C oven until the 135 dextran was dissolved. CL and stannous 2-ethylhexanoate (Sn(Oct)2) were then added into the dextran solution with agitation. The feed ratio of dextran:CL:Sn(Oct)2 was 1:3:0.05 (w/w/w). The reaction mixtures were placed under vacuum (0.3 mmHg) at room temperature for 30 minutes, and then purged with nitrogen gas. The vacuuming/purging procedures were repeated three times, and then the tubes were sealed under a nitrogen blanket and placed in an oil bath at 122 °C for 76 hours. 4.2.3 Conversion of dextrans and CL The conversion of dextran was examined by adding the reaction products (200 uL) into THF (1 mL). If any unreacted dextran was present in the products, they would precipitate out or the solution would become cloudy since dextran is not soluble in THF. A clear solution would indicate that dextran had been 100% converted. The unreacted CL was extracted from the original reaction product and assayed by HPLC. The reaction product (-100 mg) was accurately weighed. THF (200 pL) was added to dilute the product. Distilled water (1 mL) was added to the solution to precipitate the graft copolymer and PCL homopolymer formed during the reaction. The mixture was centrifuged at 14,000 rpm for 20 minutes, and the supernatant was collected and injected onto the HPLC. The instrument settings and the assay conditions were the same as those in 3.2.4. 4.2.4 Purification of Dx-g-PCL The reaction products were precipitated by an excess amount (10-fold volume) of water to remove unreacted dextrans (if present), unreacted CL and the solvent DMSO. The PCL homopolymer formed during the reaction was removed by fractional precipitation using THF as the solvent and hexanes as the precipitant. The reaction 136 mixture precipitated by water was dissolved in THF at a concentration of 1% (w/v). Hexanes were added with agitation to precipitate the Dx-g-PCLs. The hexanes/THF ratio used in this study was 3/8 (v/v). The mixture was left at room temperature for 30 minutes to allow for complete precipitation. The dissolving/precipitating procedures were repeated 3 times. 4.2.5 Gel permeation chromatography (GPC) The purity, the molecular weight and molecular weight distribution were determined by GPC using the instrumentation and assay conditions as in 3.2.6. The columns used were Styragel® HR3 and HR4 in series. 4.2.6 Nuclear magnetic resonance spectrometry (NMR) H M Q C and COSY-45 (with a pulse sequence of 90°-ti-45°-t 2) spectra of the dextran were recorded at 27 °C using Bruker AMX500 and AV400 spectrometers, respectively. *H NMR, 1 3 C NMR and H M Q C spectra of the graft copolymers were recorded at 95 °C using a Bruker AMX500 spectrometer. Both dextran and the graft-copolymers were dissolved in DMSO-d6. ' H NMR spectra of PCL homopolymers formed during the reaction were acquired using a WH400 spectrometer in CDCI3 operating at 25°C. Chemical shifts are reported in parts per million (ppm) from tetramethylsilane with CDCI3 and DMSO-d 6 as the internal references. For ! H NMR spectra, residual CHC13 was taken as 7.24 ppm and residual DMSO as 2.49 ppm. For 1 3 C NMR, the central peak of DMSO-d6 was taken as 39.5 ppm. Sample concentrations were about 50 mg/mL. 4.2.7 Preparation of Dx-g-PCL films Dx-g-PCL films were prepared by a solution casting method. Solutions of Dx-g-PCLs in THF with concentrations of 20-30 mg/mL were cast into Teflon dishes and dried 137 at room temperature in the fume hood for 48 hours. The films were further dried under reduced pressure with P2O5 as a desiccant for a week. 4.2.8 Differential Scanning Calorimetry (DSC) DSC was conducted using the same instrumentation as in 2.2.5. The coolant was liquid nitrogen. The purge gas was prepurified helium at a flow rate of 30 mL/min. The powders of the dextrans (as received) or the films of Dx-g-PCLs were weighed (3-8 mg) into crimpled, but not hermetically sealed, aluminum pans (Perkin-Elmer sample pans) and scanned at 10°C/min. An empty pan was used as the reference. 4.3 RESULTS 4.3.1 Syntheses of Dx-g-PCL copolymers The syntheses of Dx-g-PCL copolymers were carried out in DMSO solutions using SnOct as a catalyst. The reaction process is shown in Scheme 4.1. Direct evidence of the formation of Dx-g-PCLs is the change in the solubility of the dextrans. After the grafting reaction, the products were readily dissolved in THF and formed clear solutions indicating dextrans were all converted into the graft copolymers. The results of the syntheses of Dx-g-PCLs are summarized in Table 4.1. For both Dx500-g-PCL and Dx70-g-PCL, the feed ratio of dextrans:CL:Sn(Oct)2 was 1:3:0.05 and the syntheses were carried out under the same conditions. The results represent averages of four batches for each copolymer. The reproducibility of the reaction was demonstrated by the small variance among the four batches in the conversion of CL, the yields, the molecular weights, and the polydispersity indices of Dx-g-PCLs, and the molar substitution of CL on the dextrans (MSCL)- The synthesis was also conducted on different reaction scales, which is defined as the sum of the weights of dextrans and CL added at 138 Dextran-g-PCL where R = H or t a P Y 5 e , I? a* p* Y* 5* s* C C H 2 C H 2 C H 2 C H 2 C H 2 0 ) - C C H 2 C H 2 C H 2 C H 2 C H 2 - O H O •'m m = 0, 1, 2, 3, Scheme 4.1 Structure and synthesis of Dx-g-PCL 139 Table 4.1 Results of the syntheses of Dx-g-PCLs with a molar feed ratio of 1:3 (w/w) of dextran and CL a Graft copolymer Dx500-g-PCL Dx70-g-PCL M w x l ( r 4 b 6.06 ± 0.47 3.60 ± 0.22 Mw / M n 0 2.2 ± 0 . 1 1.6 ± 0 . 0 CL conversion (%) d 74.5 ± 1.9 79.8 ± 2 . 8 Copolymer Yield (%)e 44.4 ± 0.2 46.4 ± 1.1 M S C L F 1.2 ± 0 . 0 1.3 ± 0 . 1 a The results represent mean ± SD (n=4) b Weight-average molecular weight, measured by GPC and calculated by Millennium32 software based on polystyrene standard curve. The mobile phase was THF at 1 mL/min. The columns used were Styragel® HR3 and HR4 in series. c Polydispersity index, determined using the same method as for weight-average molecular weight d Calculated by the amount of reacted CL divided by the feed amount of CL e Calculated by the weights of the purified Dx-g-PCLs divided by the total weight of dextrans and CL initially added f Molar substitution of CL on dextran chains determined using the gravimetric method 140 the beginning of the reactions. The results of large-scale (40 g) reactions showed no difference from those of small-scale (4 g) reactions, indicating that the reaction was reproducible in scales ranging from 4 g to 40 g. GPC chromatograms of the reaction product and purified Dx-g-PCLs are shown in Figure 4.1. A GPC chromatogram of dextran is not included in the figure since it cannot be analyzed using the same chromatographic conditions due to its solubility. Figure 4.1a shows the chromatogram of the reaction product of Dx500-g-PCL, which includes peaks in two retention time regions. The bimodal peak (the component represented by the peak denoted as Dl) in the range of 9.5 to 17.5 minutes was presumed to be the graft copolymer Dx500-g-PCL. The multiple peaks (the component represented by the peaks denoted as PI) in the range of 17.5 to 21.5 minutes could be due to PCL homopolymer formed during the reaction. Similar to the purification of HPC-g-PCL, the fractional precipitation method was used to separate Dx500-g-PCL from the PCL homopolymer. The component represented by peaks PI could no longer be detected after the purification process, as shown in Figure 4.1b. Figure 4.1c shows the chromatogram of the reaction product of Dx70-g-PCL. Similar to that of Dx500-g-PCL, the chromatogram also includes peaks in two retention time regions. The peak (the component represented by the peak denoted as D2), which occurred approximately in the range of 9.8 to 17.5 minutes, was presumed to be the graft copolymer Dx70-g-PCL. The multiple peaks (the component represented by the peaks denoted as P2) in the range of 17.5 to 21.5 minutes may have originated from PCL homopolymer formed during the reaction. The component represented by peaks P2 was removed after the purification process, as indicated by Figure 4.Id. 141 io!oo ' ' i5!oo ' 2o!oo 10.00 15.00 20.00 Retention time (minute) Retention time (minute) Figure 4.1 GPC chromatograms of (a) the reaction product of Dx500-g-PCL, in which D l and PI represent the components of Dx500-g-PCL and the PCL homopolymer formed during the reaction, respectively; (b) purified Dx500-g-PCL; (c) the reaction product of Dx70-g-PCL, in which D2 and P2 represent the components of Dx70-g-PCL and the PCL homopolymer formed during the reaction, respectively; (d) purified Dx70-g-PCL 142 4.3.2 Structural characterization 4.3.2.1 Characterization of the dextrans using N M R spectroscopy H M Q C and COSY-45 spectra of the dextrans are shown in Figure 4.2 and Figure 4.3, respectively. The cross-peaks in the H M Q C spectrum revealed the connectivity between C and H, while those in the COSY-45 spectrum shows the couplings between neighbouring protons. The [ H and 1 3 C chemical shifts of dextran ring were assigned based on the combination of the information from both the H M Q C and COSY-45 spectra. According to the structure of dextran (shown in Scheme 4.1), there are a total of 6 13 carbons and 9 hydrogens in each anhydroglucose unit. Six carbon peaks in the C NMR spectrum were clearly resolved. However, the ' H NMR spectrum showed only 8 peaks. Based on the peak integration, the broad peak at 3.17-3.19 ppm should contain two peaks. The peak at 3.39 ppm was superimposed with the peak from water. Since there was no cross-peak corresponding to the peaks at 4.89,4.82 and 4.47 ppm in the H M Q C spectrum (Figure 4.2), the three peaks must originate from the three hydroxyl groups in the anhydroglucose units of dextran. It is well known that the anomeric CI in polysaccharides has a characteristic chemical shift at around 100 ppm. In Figure 4.2, the chemical shift of CI for dextran was 98.26 ppm. The cross-peak 1 indicated the connectivity between CI-HI, which led to the assignment of the peak at 4.66 ppm in 1H NMR spectrum to be HI. The cross-peaks 6 and 6' indicated that the two protons at 3.72 and 3.48 ppm connected to the same carbon. C6 is the only carbon that has two protons in dextran. Therefore, the two peaks at 3.72 and 3.48 ppm were assigned to H6 and H6' and the peak at 66.14 ppm in the 1 3 C spectrum to C6. Cross-peak 5-6 in Figure 4.3 revealed the coupling between H5 and H6, 143 OH4 0H3H10H2 „ A H3 H2 H4 u • • • A A i • 1 • 1 1—!—i—i—i—i—i—r— p p m 5 Figure 4.2 H M Q C spectrum of dextran in DMSO-d 6 recorded at 27 °C using a Bruker AMX500 spectrometer. 144 Figure 4.3 COSY-45 spectrum of dextran in DMSO-d 6 recorded at 27 °C using a Bruker AV400 spectrometer 145 therefore the peak at 3.61 ppm in the ' H spectrum was assigned to H5. This led to the assignment of the peak at 70.43 ppm in the 1 3 C spectrum to C5 because of the cross-peak 5 in Figure 4.2. The coupling between H4 and H5 was shown by the cross-peak 4-5 in Figure 4.3, which led to the assignment of the higher field side of the broad peak at 3.17-3.19 ppm to H4. Therefore, C4 (70.18 ppm) could be located by the cross-peak 4 in Figure 4.2. The peak for H2 was located by the cross-peak 1-2 in Figure 4.3, which was at the lower field side of the broad peak at 3.17-3.19 ppm. The peak for C2 was then 13 assigned by the cross-peak 2 in Figure 4.2. The leftover peaks at 73.40 ppm in the C spectrum and 3.39 ppm in the lH spectrum should correspond to C3 and H3, respectively. The peaks at 4.47, 4.82, and 4.89 ppm in the ' H spectrum were assigned to OH2, OH3 and OH4, respectively, by the cross-peaks 2-2, 3-3, and 4-4 in Figure 4.3. A summary of the assignments is given in Table 4.2. The 1 3 C assignments are in good agreement with those reported in the literature (Friebolin et al. 1979; Arranz and Sanchez-Chaves 1988). 4.3.2.2 Characterization of Dx-g-PCLs using N M R spectroscopy A representative *H NMR spectrum of purified Dx-g-PCLs is shown in Figure 4.4. The spectrum contained peaks from both the dextran and PCL indicating the formation of the graft copolymers Dx-g-PCLs. Based on the side chain assignments for the spectra of HPC-g-PCL in 3.3.2, the peaks from side chain PCL could be readily assigned. Similar to HPC-g-PCL, the peaks corresponding to H E * and H5* in the repeating units at the chain ends shifted to higher fields compared to the internal H E and H5. H a * also had a different chemical shift from H a , however, the peak shifted to a lower field. In 1 3 C NMR spectra of Dx-g-PCLs, there was a "side peak" located very close to each of the peaks corresponding to Cp and C y with slightly higher chemical shifts (values in brackets 146 in Table 4.3). H M Q C spectra indicated that both the "main peaks" and the "side peaks" of Cp and C y were connected to Hp and H y . The differences in chemical shifts between the "main peaks" and the "side peaks" were 0.33 ppm and 0.14 ppm for Cp and C Y , respectively, which were almost identical to the differences observed in HPC-g-PCL. We speculate that the "side peaks" corresponded to Cp* and CY* in the repeating units at the chain ends. Most of the peaks from the main chain (dextran) were highly overlapped, and therefore could not be assigned. The broad peak at 4.65-4.85 ppm in Figure 4.4 was assigned to HI based on the H M Q C spectrum shown in Figure 4.5. The H M Q C spectrum indicated that this broad peak contained three peaks corresponding to the protons at three different chemical shifts, and all of them connected to C l . The slightly different chemical shifts of HI might originate from different substitution patterns on the anhydroglucose units, which resulted in different chemical environments for HI. Particularly, the grafting reaction could occur at three different hydroxyl groups (OH2, OH3, and OH4), and some of the anhydroglucose units might have no side chain grafted. The assignments are summarized in Table 4.3. The labels of the carbons and the protons are shown in Scheme 4.1. The major peaks in the *H NMR spectra (not shown) of the components represented by the peaks PI and P2 in Figure 4.1 could be readily assigned. The components were shown to contain mainly PCL homopolymers. The assignments are summarized in Table 4.3. There were some small peaks that could not be assigned. These peaks were considered to correspond to the side products formed during the reactions. 147 Table 4.2 Observed chemical shifts (8) in the 1 3 C and ' H NMR spectra of dextrana 1 3c 8 (ppm) ' H 8 (ppm) C l 98.26 HI 4.66 C2 71.90 H2 3.19 C3 73.40 H3 3.39 C4 70.18 H4 3.17 C5 70.43 H5 3.61 C6 66.14 H6, H6' 3.48, 3.72 OH2 4.47 OH3 4.82 OH4 4.89 a The spectra were recorded in DMSO-d6 at 95 °C using a Bruker AMX500 spectrometer 148 6 C H 2 5 OR OR^ o-OR - in R = H or f o . . . . . a p Y 5 e t II a p Y 8 e C C H 2 C H 2 C H 2 C H 2 C H 2 0 f - C C H 2 C H 2 C H 2 C H 2 C H 2 - O H O rn m = 0, 1,2,3, -r • r ' I < 1 ' 1 ' 1 1 1 • 1 1 1 ' i W» 50 *i *.0 3.5 3.0 2.5 J.O 1.5 1.0 Figure 4.4 Representative lH NMR spectrum of Dx-g-PCLs in DMSO-d 6 recorded at 95°C using a Bruker AMX500 spectrometer 149 to * CO . to O ll o HI UU • '1111 • i1111 11 ppm A . . . 7 rrrr | , , , r 2 Figure 4.5 Representative H M Q C of Dx-g-PCLs in DMSO-d* recorded at 95°C using a Bruker AMX500 spectrometer 150 Table 4.3 Observed chemical shifts (8) in the NMR spectra of purified Dx-g-PCL and PCL homopolymer formed during the reactions Dx-g-PCL a P C L b 1 3 C 8 (ppm) ' H 8 (ppm) 1H 8 (ppm) CI 98.21 HI 4.65-4.85 ~ — c„ 32.99 H a 2.26 H a 2.17 Ca* 33.21 H a * 2.32 — — Cp 23.55 (23.88) Hp 1.56 Hp,8 1.48-1.53 Cy 24.43 (24.57) Hy 1.34 Hy 1.26 c 8 27.34 H 5 1.60 — — C5* 31.56 Hg* 1.44 ~ — cE 62.93 H 6 4.01 H 6 3.92 ce* 60.21 H E * 3.41 H E * 3.47 c=o 172.04 a The spectra were recorded in DMSO-d6 at 95 °C using a Bruker AMX500 spectrometer b The spectrum was recorded in CDCI3 at 27 °C using a Bruker AV400 spectrometer 151 4.3.2.3 Molar substitution of C L on dextrans (MSCL) MScl was determined using the gravimetric method. The calculation is given as follows: = KAFL-WDX)/\U Equation 4.1 where Wgmft is the weight of purified Dx-g-PCL; Wdx is the initial weight of dextran; the number 114 is the formula weight of the repeating unit of PCL; and 162 is the formula weight of the repeating unit of dextran. In this calculation, dextran was assumed to be 100% converted to Dx-g-PCL. As shown in Table 3.1, the M S C L values for Dx500-g-PCL and Dx70-g-PCL were determined to be L2 and 1.3, respectively, which means on average there were 12 and 13 CL monomers reacted with every 10 anhydroglucose units in Dx500 and Dx70, respectively. M S C L was also estimated using the ' H N M R spectroscopy method (Ho et al. 1972; Ydens et al. 2000) based on the same principle described in 3.3.3. A pulse angle of 36° and a relaxation delay of 2 second were used for the acquisition of the spectra. To ensure reliable peak intensities, spectra were also acquired at relaxation delays of 4 and 8 seconds, respectively. The spectra obtained using the three different relaxation delays gave identical peak intensities indicating that the peak intensities used in the following calculations were quantitatively reliable. As shown in the structure of Dx-g-PCL in Scheme 4.1, each anhydroglucose unit of dextran has one proton (HI) connecting to C l , and each PCL repeating unit has two a protons. Therefore, M S C L could be estimated using the peak intensities of these two types of protons in the *H N M R spectra of Dx-g-PCLs. The calculation is given as follows: 152 I, H\ M S a = -jj2- Equation 4.2 where Ina is the peak intensity of H a in PCL side chains and Ihi is the peak intensity of HI in dextran. Using this method, the M S C L was estimated to be 1.4 for Dx500-g-PCL and 1.6 for Dx70-g-PCL. These estimations were close to those determined by the gravimetric method. 4.3.2.4 The length and the distribution of the side chains As shown in Figure 4.4, the peaks for Hs and Hs* occurred far apart in the spectrum and the peak for Hs was resolved fairly well. Therefore, the average side chain length (L) could be estimated using the same method described in 3.3.4 and calculated by the Equation 3.4: L = — ^ 2 — I Ha ~ I He where Ihs represents the peak intensities of the internal s protons and Iua the peak intensity of the a protons. The L obtained was 2.4 for Dx500-g-PCL, and 3.1 for Dx70-g-PCL. The average number of PCL side chains on each anhydroglucose unit of dextran (N) was then calculated by the Equation 3.5: N = MSCL/L Based on the M S C L estimated using *H N M R method, N was estimated to be 0.6 for Dx500-g-PCL and 0.5 for Dx70-g-PCL. Hence, in both of the copolymers, approximately every two anhydroglucose units has one side chain, i.e., only half of the anhydroglucose units in dextran were grafted. The reactivity of the hydroxyl groups in dextrans was 153 reported to be in an order of OH2 > O H 4 >OH3 (Arranz and Sanchez-Chaves 1988; Zhang et al. 2000). Therefore, the PCL side chains were most likely grafted at C2. 4.3.3 Morphology and thermal analysis of Dx-g-PCLs The fdms of Dx500-g-PCL and Dx70-g-PCL were both transparent indicating a single-phase morphology. DSC scans of the dextrans and the copolymers are shown in Figure 4.6. As indicated by the scans A and C, Dx500 and Dx70 showed no thermal events in the temperature range of-100 to 150 °C except the very broad endotherms centered at around 50 °C, which were due to the loss of absorbed water (Scandola et al. 1991; Stenekes et al. 2001). The scans for Dx500-g-PCL (B) and Dx70-g-PCL (D) showed only similar very broad endotherms before 130 °C, and both the fdms started to decompose after 130 °C. No glass transition and melting of PCL were observed in the scans, indicating that no PCL-rich phase formed in the films of both the copolymers. 4.4 DISCUSSION The synthesis of Dx-g-PCL directly from dextrans and C L monomer was a challenge due to the difficulty of finding a suitable reaction system. Unlike the synthesis of HPC-g-PCL, the bulk polymerization method was not feasible, due to the negligible yield of the graft copolymer caused by the complete immiscibility of dextran and monomer CL. A solution polymerization method was reported by Ydens et. al. for synthesis of Dx-g-PCL, in which dextran was modified by partial silylation before graft reaction to make it soluble in the solvent toluene (Ydens et al. 2000). The silyl substituents must be removed after the reaction. The silylation and de-silylation procedures increased the complexity of the synthesis and the uncertainty of the structure of the final product. In this work, the graft reactions were carried out in DMSO. The 154 -100 -50 0 50 100 150 Temperature (°C) Figure 4.6 Representative DSC scans of (A) Dx500; (B) Dx500-g-PCL; (C) Dx70; and (D) Dx70-g-PCL. The heating rate was 10 °C/min. 155 selection of the solvent was based on fulfilling two criteria: (1) the selected solvent should be a common solvent for all the reactants; (2) the solvent should not interfere with the graft reaction. As mentioned in 4.1, dextrans are soluble in only a few solvents including water, DMSO, formamide, ethylene glycol, and glycerol. Although these solvents were all miscible with CL, DMSO was the only solvent that was selected for this reaction because the other four contained either hydroxyl groups or active hydrogens in their molecules that would result in only PCL homopolymer and dramatically suppress the formation of the graft copolymers. Sn(Ocf)2 has been a catalyst for a variety of lactone polymerizations (Zhang et al. 1994; Zhang et al. 1996; Kricheldorf et al. 2000; Storey and Sherman 2002). A coordination-insertion mechanism of the catalysis has been well established, which includes two major steps: formation of stannous alkoxide initiation species with hydroxyl group-containing compounds, and coordination-insertion of CL monomer into the stannous alkoxide bond to initiate the polymerization (Kowalski et al. 2000; Kricheldorf et al. 2000; Storey and Sherman 2002). In this study, the hydroxyl group-containing compounds were dextrans and traces of water in the reaction system. Graft copolymers would be formed if the stannous alkoxide initiation species were formed with dextrans. However, if the initiation species were formed with water, the polymerization would result in PCL homopolymer, which was considered to be a side product. To minimize the influence of water, vigorous drying procedures were performed in the syntheses as described in the experimental section. Nevertheless, PCL homopolymer formed with yields in the range of 35-40%, which was much higher than those in the syntheses of HPC-g-PCL. The higher yield of PCL homopolymer was a consequence of higher water 156 content in the system. The solution polymerization method involved using a large amount of DMSO as the solvent. Given the hygroscopic nature of DMSO, it was very difficult to completely eliminate water/moisture in the system. The Dx-g-PCLs were also purified by the fractional precipitation method using THF as the solvent and hexanes as the precipitant. The difference between the purification of HPC-g-PCL and Dx-g-PCLs was the ratio of hexanes/THF. The ratio used in the purification of Dx-g-PCLs was 3/8 (v/v), which was much lower compared to that used for HPC-g-PCL. This was primarily because the Dx-g-PCLs had lower M S C L than HPC-g-PCL, thus the overall polarity of Dx-g-PCL molecules was much higher than HPC-g-PCL. As a result, less hexanes were needed to precipitate Dx-g-PCLs. In Figure 4.2 and 4.3, the peaks in the *H NMR spectra of the dextran were strong, narrow and well resolved. However, grafting of PCL side chains decreased the mobility of the chains of dextran molecules and caused significant band broadening and overlapping. The peaks corresponding to the main chains were highly overlapped in the *H NMR spectra of Dx-g-PCLs recorded at 27 °C (not shown). The resolution was improved by increasing the temperature to 95 °C. The improvement was manifested by the peak corresponding to HI, which was much better resolved than that at 27 °C. M S C L was estimated using both gravimetric and *H NMR methods. The values obtained from the [ H NMR method were only slightly higher than those from the gravimetric method. In the gravimetric method, there may be loss of the graft copolymers in the multiple steps of the purification process, leading to an underestimation of the M S C L - The errors in the *H NMR method mainly originated from the peak integration, particularly the integration of the peak HI, which was broad. 157 4.5 CONCLUSIONS Amphiphilic graft copolymers of dextran-g-PCL were reproducibly synthesized by solution polymerization in DMSO using stannous 2-ethylhexanoate as the catalyst. Evidence of the formation of Dx-g-PCLs was obtained from both the solubility change and NMR spectroscopy. Molar substitution was determined to be in the range of 1.2 to 1.4 for Dx500-g-PCL and 1.3 to 1.6 for Dx70-g-PCL. The gravimetric and *H N M R methods were considered to be equally suitable for the determination of the M S C L of Dx-g-PCLs synthesized in this study. The films of both Dx500-g-PCL and Dx70-g-PCL were transparent and showed a single-phase morphology. 158 CHAPTER 5 FORMULATION AND CHARACTERIZATION OF PACLITAXEL LOADED AMPHIPHILIC GRAFT COPOLYMER FILMS 5.1 INTRODUCTION Following the characterization of the three graft copolymers, HPC-g-PCL, Dx500-g-PCL, and Dx70-g-PCL, we were interested in investigating the application of these copolymer biomaterials as paclitaxel loaded films for the prevention of surgical adhesions. Ideally, these films should possess the following properties. They should be biocompatible and cause little or no inflammatory reactions. The films should also be homogeneous with no paclitaxel precipitation. Since adhesions are believed to form during the first few days following surgeries, it is desirable for the films to to release all loaded paclitaxel within about 10 days and the films completely eroded from the site in less than one month. The films should be sterilizable, flexible, and easy to handle with sufficient mechanical strength to be manipulated into and around the implantation site. The objectives of this study were: (1) to prepare paclitaxel loaded films based on amphiphilic graft copolymers, HPC-g-PCL, Dx500-g-PCL, and Dx70-g-PCL; (2) to determine the swelling behaviour and the morphology of the films; (3) to determine the in vitro paclitaxel release kinetics. 5.2 EXPERIMENTAL 5.2.1 Materials and supplies HPC-g-PCL with a M S C L of 8.6 was synthesized using the method described in 3.2.3. Dx500-g-PCL with a M S C L of 1.2 and Dx70-g-PCL with a M S C L of 1.3 were 159 synthesized using the method described in 4.2.2. PCL homopolymer (nominal molecular weight 80,000) used for making the control films for the in vitro drug release study was purchased from Polysciences (Warrington, PA). Paclitaxel was purchased from Hauser Chemicals, Inc. (Boulder, CO.). THF, A C N , D C M , and methanol, all HPLC grade, were from Fisher Scientific (Napean, Ontario, Canada). 5.2.2 Preparation of paclitaxel loaded films Paclitaxel loaded films of HPC-g-PCL, Dx500-g-PCL, Dx70-g-PCL, and PCL were prepared by the solution casting method. The films of the graft copolymers were cast in PTFE dishes, which were made of Bytac® VF-81 PTFE film (Norton Performance Plastics Corporation, Wayne, NJ). The PCL films were cast in aluminum dishes (Fisher Scientific, Toronto, ON). THF solutions of the polymers with concentrations in the range of 20-30 mg/mL and THF solutions of paclitaxel with a concentration of 5 mg/mL were mixed at a given ratio with vigorous agitation. The mixtures were transferred to the dishes, and dried in a fume hood at room temperature for 48 hours. The films of PCL and HPC-g-PCL were further dried under reduced pressure in vacuum desiccators with P2O5 as a desiccant for 48 hours, and the films of Dx-g-PCLs similarly dried for one week. Films with 0%, 1%, 5%, and 10% (w/w) paclitaxel loading were prepared. For Dx500-g-PCL, 0.1% and 0.5% paclitaxel loaded films were also prepared. 5.2.3 Determination of the degree of swelling of the films The swelling studies of the films of HPC-g-PCL, Dx500-g-PCL, Dx70-g-PCL, and PCL with or without paclitaxel were conducted in 10 mM PBS (pH7.4) at 37 °C. The dry films were accurately weighed and immersed into the PBS. At given time intervals, 160 the swollen films were withdrawn from the buffer solution. The excess surface liquid was blotted using a filter paper and the films were weighed. The swelling was continued until a constant weight for each of the films was obtained. The degree of swelling was defined (Zhang etal. 1999) as: W-W Degree of swelling (%) = -^x 100 ^o where IF is the weight of the swollen film and W0 the weight of the dry film. 5.2.4 In vitro paclitaxel release The in vitro studies of paclitaxel release from the films of HPC-g-PCL, Dx500-g-PCL, Dx70-g-PCL, and PCL were carried out at 37 °C in an Innova™ 4000 Shaker (New Brunswick Scientific, Edison, NJ). Paclitaxel loaded films of the same size were accurately weighed and placed in Erlenmeyer flasks. PBS (10 mM, pH7.4) was used as the release medium. The volume of the release medium was adjusted to ensure sink conditions. Films of each of the polymers without paclitaxel were used as controls. The flasks were shaken at 100 rpm. The release medium was replaced by fresh PBS every 24 hours. Paclitaxel in the release medium was extracted into 1 mL of D C M , dried under a stream of nitrogen, and reconstituted in 1 mL of 60% (v/v) A C N in water solution. The concentration of this solution was analyzed by HPLC, as described in 2.2.6. The cumulative paclitaxel release between different drug loadings and different polymers was compared using Student's t-test. A p value less than 0.05 was considered statistically significant. 5.2.5 Thermal transitions of paclitaxel loaded HPC-g-PCL films The effect of paclitaxel loading on the thermal transitions of HPC-g-PCL films was examined using DSC as described in 2.2.5. The films were accurately weighed (3-8 161 mg) into the aluminum sample pans and analyzed using the following temperature program: (1) Hold for 5 minutes at -100 °C; (2) Heat from -100 °C to 100 °C at 10 °C/min; (3) Hold for 3 minutes at 100 °C; . (4) Cool from 100 °C to -100 °C at 10 °C/min; (5) Hold for 3 minutes at -100 °C (6) Heat from -100 °C to 100 °C at 10 °C/min. 5.2.6 Morphology change of paclitaxel loaded Dx-g-PCL films The films were incubated in 10 mM PBS (pH7.4) and subjected to the same conditions as in the release study (described in 5.2.3). The incubation medium was replaced by fresh PBS every 24 hours. The morphology of the films before and after the incubation was studied using optical microscopy and XRD, as described in 3.2.9 and 3.2.11, respectively. Paclitaxel dihydrate was prepared using a modified method of Liggins et. al. (Liggins et al. 1997) as follows: commercial paclitaxel was dissolved in A C N to form a solution with a concentration of 1.5% w/v. The solution was added into an excess amount of water with stirring and a turbid solution resulted. The turbid solution was centrifuged at 14,000 rpm for 10 min. The precipitate was collected and dried at room temperature for 15 hours in a vacuum oven (150 mmHg), and the sample analyzed by XRD. 5.2.7 Determination of residual THF in the copolymer films Residual THF in the films of the three graft copolymers: HPC-g-PCL, Dx70-g-PCL, and Dx500-g-PCL was determined using gas chromatography (GC) by Dr. R. 162 Liggins (Angiotech Pharmaceuticals, Inc.). The method was set up according to the USP <467>, method for Organic Volatile Impurities. The films (300mg) were accurately weighed and placed in test tubes. Distilled water (10 mL) was added to each tube to extract THF. The extracts were then analyzed using an HP6890 GC system, which was equipped with a HP7694 Headspace Autosampler, a Supelcowax capillary column (30mx0.53mm ID, and 1.0 pm film thickness), and a flame ionization detector (FID). The carrier gas was helium at a flow rate of 30 mL/min. Hydrogen and air flow rates were 30 mL/min and 400 mL/min, respectively. The samples were injected in a split mode. The temperature of the injector was 140 °C and the detector 260 °C. The oven temperature was programmed as follows: 50 °C for 9 minutes, ramp to 160 °C at rate of 12 °C/min, and hold at 160 °C for 4 minutes. 5.3 RESULTS Both paclitaxel and the three copolymers could be readily dissolved in the common solvent THF and formed clear solutions. No paclitaxel crystallization was observed in all three copolymer films within the loading range of 0.1% to 10% (w/w). Dx500-g-PCL and Dx70-g-PCL films with or without paclitaxel were transparent and were fragile when completely dried. The dextran-based copolymer films were hygroscopic, became very flexible following moisture uptake and when immersed in aqueous media, the films swelled rapidly. HPC-g-PCL films with or without paclitaxel were opaque and flexible and immersing into aqueous media resulted in limited swelling. 5.3.1 Degree of swelling of the films The swelling of the films loaded with 0%, 1%, 5%, and 10% of paclitaxel is shown in Figure 5.1. The swelling was found to be rapid for all the graft copolymer films 163 PCL HPC-g-PCL Dx500-g-PCL Dx70-g-PCL Figure 5.1 Degree of swelling of the films as a function of the nature of the polymers and paclitaxel loading. The films were immersed in 10 mM PBS (pH7.4) at 37 °C for 12 hours. The data represent mean ± SD (n=3). The significance of the difference of the swelling between the films was analyzed using Student t-test, and p<0.05 was considered statistically significant. ">" denotes significantly greater and " ~ " not significantly different: Dx500-g-PCL ~ Dx70-g-PCL > HPC-g-PCL > PCL 164 reaching the maximum extent of swelling within 1.5 hours. The films were placed in the buffer solution for 12 hours to ensure equilibrium was reached. For the films without paclitaxel, the degrees of swelling of the Dx500-g-PCL and Dx70-g-PCL films were 36.2% and 37.7%, respectively, which were considerably higher than that of the HPC-g-PCL film (2.8%). Very little swelling (0.9%) was observed for the PCL film. An increase in paclitaxel loading resulted in decreased swelling for both the Dx500-g-PCL and Dx70-g-PCL films. However, paclitaxel loading showed no effect on the swelling of either PCL or HPC-g-PCL films. 5.3.2 In vitro paclitaxel release The cumulative paclitaxel release profiles for the films of Dx500-g-PCL, Dx70-g-PCL, HPC-g-PCL, and PCL homopolymer are shown in Figure 5.2, 5.3, 5.4, and 5.5, respectively. The size of the films was normalized to 1 cm . In general, the release of paclitaxel from the films was characterized by an initial rapid release phase lasting about 2 to 5 days followed by a slower release phase. The initial paclitaxel loading had marked effects on the release rate of paclitaxel from the films of all four polymers. The nature of the polymers also influenced paclitaxel release profiles. For the films of HPC-g-PCL and PCL, the release rate increased when the paclitaxel loadings increased from 1% to 10%. However, for Dx500-g-PCL and Dx70-g-PCL films, the release rate increased with an increase in paclitaxel loading only when the loadings were lower than 5%. The release rates at 10% loading were lower than those at 5% loading during the initial release phase (up to about 5 days for Dx500-g-PCL films and up to about 2-3 days for Dx70-g-PCL films). After that, paclitaxel was released from the 10% paclitaxel loaded films at almost constant rates while the release rate of the 5% loaded films declined gradually. 165 1800 \ 1600 lb 1400 f 1200 | 1000 -\ 1 =3 U 800 600 H 400 200 4. • * 0 A 10%PTX • 5% PTX • 1% PTX !£•••••••••••• 10 15 Time (day) 20 25 B o c/3 O u _> '^ -> S u 100 90 80 70 60 50 40 30 20 10 0 x 0.5% PTX • 0.1% PTX x 4 6 8 Time (day) 10 12 Figure 5.2 In vitro paclitaxel release profiles from the Dx500-g-PCL films with different loadings: (A) 1%, 5%, and 10%; (B) 0.1% and 0.5%. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 166 1400 1200 1000 I 800 <D 15 » 600 | 400 H u 200 0 A 10% PTX • 5% PTX • 1%PTX s 1 i t * • • • • • • • • • • • • • o 10 15 Time (day) 20 25 Figure 5.3 In vitro paclitaxel release profiles from the Dx70-g-PCL films with different loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 167 1400 1200 1a> 1 0 0 0 1 800 IS 8> 600 H S u 400 200 0 A 10% PTX • 5% PTX • 1% PTX lH 1 • • • • • • • • • • • • • • • • • 10 15 20 Time (day) 25 30 Figure 5.4 In vitro paclitaxel release profiles from the HPC-g-PCL films with different loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 168 B o CD C/3 CO CD "S S-i CD > +-> 3 O 500 450 H 400 350 300 250 200 H 150 100 50 0 A 10% PTX • 5% PTX • 1%PTX •II n 1 u s i H 1 i i i[\ 10 15 20 25 30 Time (day) Figure 5.5 In vitro paclitaxel release profiles from the PCL films with different loadings. The films were incubated in 10 mM PBS (pH7.4) and shaken at 100 rpm at 37°C. Data represent the mean of 4 samples with error bars (standard deviation) shown only in the positive direction for clarity 169 The amounts of paclitaxel released over the first 7 days are summarized in Table 5.1. It is evident that at all loading levels, the amounts of paclitaxel released over the first 7 days from the graft copolymer films were significantly higher than those from the PCL homopolymer films (p<0.05, t-test). The amounts released from the Dx-g-PCL films were significantly higher than those from the HPC-g-PCL films (p<0.05, t-test) only at 1% and 5% paclitaxel loadings. At 1% loading, there was no difference between the release from the Dx500-g-PCL films and that from the Dx70-g-PCL films. At 5% loading, the release from the Dx500-g-PCL films was significantly higher than that from the Dx70-g-PCL films. At 10% loading, the release from Dx500-g-PCL films was very close to that from the HPC-g-PCL films, and was higher than from the Dx70-g-PCL films. 5.3.3 Morphology of the paclitaxel loaded copolymer films 5.3.3.1 Paclitaxel loaded HPC-g-PCL fdms The HPC-g-PCL films loaded with 1%, 5% and 10% paclitaxel were examined using XRD. There was no evidence of crystalline paclitaxel in the films at any loading levels, indicating that either paclitaxel was dissolved in the films or the paclitaxel was present as an amorphous dispersion. The effect of paclitaxel loading on the thermal transitions of HPC-g-PCL films was investigated using DSC. The films prepared by the solution casting method were heated from -100 °C to 100 °C at 10 °C/min and the DSC scans are shown in Figure 5.6. Each of the four scans showed a glass transition between -50 to -75 °C and a double endotherm between 20-50 °C. The double endotherms of the films with different paclitaxel loadings appeared in the same temperature range. However, the glass transition 170 Table 5.1 The amount of paclitaxel released over the first 7 days (M7) from the films of the three copolymers and PCL homopolymer with different paclitaxel loadings. Values represent mean ± SD (n = 4) Polymer M7 (ug) from the films with different paclitaxel loadings * 0.1 % 0.5% 1% 5% 10% Dx500-g-PCL 17.1+1.1 72.9±3.5 168.6+4.5 a 609.4+4.7 b 472.2+17.3 c Dx70-g-PCL — — 171.3+2.0a 578.3+4.6 b 376.1+19.0 c HPC-g-PCL — — 105.1+4.6 a 371.1+7.0 b 474.1+43.5 c PCL ~ — 30.5+2.5 a 136.8+11.lb 173.5+29.3 c The significance of the difference of the paclitaxel release between different films at each loading level were analyzed using Student t-test, and p<0.05 was considered statistically significant. ">" denotes significantly greater and " ~ " not significantly different. a 1 % loading: Dx500-g-PCL ~ Dx70-g-PCL > HPC-g-PCL > PCL b 5% loading: Dx500-g-PCL > Dx70-g-PCL > HPC-g-PCL > PCL c 10% loading: Dx500-g-PCL ~ HPC-g-PCL > Dx70-g-PCL > PCL 171 ] j Tg i 1 1 — i 1 -100 -50 0 50 100 Temperature (°C) Figure 5.6 The first heating scans of the paclitaxel loaded HPC-g-PCL films prepared by the solution casting method. The paclitaxel loadings were: (Al) 0%; (Bl) 1%; (Cl) 5%; and (Dl) 10%. The heating rate was 10 °C/min. 172 temperatures (Tg) increased with the increase in paclitaxel loading. A difference of 11.1 °C was observed between the Tg of the film without paclitaxel and that of the 10% paclitaxel loaded film. A summary of the transition temperatures is given in Table 5.2. The films were held at 100 °C for 3 minutes and then cooled down to -100 °C at 10 °C/min. After being held at -100 °C for 3 minutes, the films were heated again at 10 °C/min to 100 °C. The scans of the second heating are shown in Figure 5.7. The second heating scans differed from the first heating scans in two major aspects. Firstly, an exothermic peak occurred between the glass transition and the melting endotherm in each of the scans, which corresponded to the recrystallization of the side chains. The peak temperature (Tc) of the exotherms increased from -26.9 °C for the film without paclitaxel to -0.8 °C for the films loaded with 10% paclitaxel as shown in Table 5.2. Another difference was that the double endotherms in the first heating scans became single endotherms and the endpoint of the endotherms (Tm) decreased by 11-12 °C (Table 5.2). 5.3.3.2 Paclitaxel loaded Dx-g-PCL films Morphological changes in the 1%, 5% and 10% paclitaxel loaded films of both Dx500-g-PCL and Dx70-g-PCL following incubation in PBS at 37 °C were examined by optical microscopy, DSC and XRD. Before incubation, all the films were transparent at all paclitaxel loadings. No crystalline structure was detected in the films using both DSC and XRD. The 1% paclitaxel loaded films remained transparent, and no paclitaxel crystallization was observed throughout the period of the release study. However, crystals in the form of spherulites formed in both the 5% and 10% paclitaxel loaded films during the incubation, and both the size and number of the spherulites were found to grow with the increase of the incubation time. The morphological changes were observed by optical 173 Table 5.2 Effect of PTX loading on the thermal transitions of HPC-g-PCL films measured using DSC a ' e PTX loading (w/w) First heating Second heating T g b T m c T g b T c d T m c 0% -67.4 46.4 -67.4 -26.9 34.7 1% -63.8 46.5 -67.1 -24.5 35.4 5% -63.4 45.7 -65.4 -13.9 33.4 10% -56.3 45.3 -61.4 -0.8 34.0 a Temperature program: (1) Hold for 5 minutes at -100 °C; (2) Heat from -100 °C to 100 °C at 10 °C/min; (3) Hold for 3 minutes at 100 °C; (4) Cool from 100 °C to -100 °C at 10 °C/min; (5) Hold for 3 minutes at -100 °C; (6) Heat from -100 °C to 100 °C at 10 °C/min b Onset values of the glass transitions c Endpoint of the melting endotherms d Peak values of the recrystallization exotherms e The values of Tg, Tm, and Tc represent averages of two measurements. 174 1 -100 -50 0 50 100 Temperature (°C) Figure 5.7 The second heating scans of the paclitaxel loaded HPC-g-PCL films loaded after the films were heated from -100 °C to 100 °C and then cooled from 100 °C to -100 °C both at 10 °C/min. The paclitaxel loadings were: (A2) 0%; (B2) 1%; (C2) 5%; and (D2) 10%. The heating rate was 10 °C/min. 175 microscopy, and representative micrographs are shown in Figure 5.8 and 5.9 for the 5% and 10% paclitaxel loaded films, respectively. For 5% paclitaxel loaded films (Figure 5.8), the spherulites were small (approximately 20-30 um) at 6 hours and increased in both size and number to a maximum size in about 5 days. For 10% paclitaxel loaded films, the size of the spherulites at 6 hours was approximately 50-80 pm. Both the size and the number of the spherulites grew rapidly. They started to impinge on each other in less than 12 hours. The growth of the spherulites reached a maximum extent within 48 hours. After being incubated for 5 days, the films were dried and examined using X-ray diffraction. Scans A and B in Figure 5.10 show representative X-ray diffraction patterns of 10% paclitaxel loaded Dx500-g-PCL films and Dx70-g-PCL films, respectively, after incubation for 5 days. The peaks in the patterns are at the same 2-theta angles as those in the pattern of paclitaxel dihydrate shown in Figure 5.10, indicating that the spherulites formed in the films are crystals of paclitaxel dihydrate. The 2-theta angles and d-spacings of both paclitaxel dihydrate and the crystals in the films are given in Table 5.3. 5.3.4 Residual T H F in the copolymer films The residual THF was removed by drying the films under vacuum (0.3 mmHg) for 48 hours for the HPC-g-PCL films and 1 week for the Dx-g-PCL films. GC results indicated that the content of THF in the 10 mL extracts from all the films was lower than the detection limit of the instrument, which was 1 ppm. Therefore, the residual THF was less than 33 ppm in the films. 176 i # > i t iB9 B D Figure 5.8 Representative micrographs the 5% paclitaxel loaded Dx500-g-PCL films after incubation in 10 mM PBS (pH7.4) at 37 °C for (A) 6 h; (B) 12 h; (C) 24 h; (D) 2 days; (E) 3 days; and (F) 5 days 177 Figure 5.9 Representative micrographs the 10% PTX loaded Dx500-g-PCL films after incubation in 10 mM PBS (pH7.4) at 37 °C for (A) 6 h; (B) 12 h; (C) 24 h; (D) 2 days; (E) 3 days; and (F) 5 days 178 0 10 20 30 40 2 theta Figure 5.10 Representative XRD patterns of (A) 10% paclitaxel loaded Dx500-g-PCL films and (B) 10% paclitaxel loaded Dx70-g-PCL films after the incubation in lOmM PBS (pH7.4) at 37 °C for 5 days; and (C) paclitaxel powder precipitated by adding a solution of paclitaxel in A C N into an excess amount of water. 179 Table 5.3 X-ray diffraction peak locations (°20) and d-spacings of PTX powder precipitated from A C N solution by water, and the 10% PTX loaded Dx500-g-PCL and Dx70-g-PCL films after being incubated in 10 mM PBS at 37 °C for 5 days PTX dihydrate Dx500-g -PCL films Dx70-g- PCL films °20 d-spacing (A) °20 d-spacing (A) °29 d-spacing (A) 3.96 22.29 — — ~ — 4.68 18.87 — ~ — — 5.16 17.11 5.14 17.18 5.14 17.18 6.04 14.62 — — — ~ 9.60 9.21 9.58 9.22 9.58 9.22 10.90 8.11 10.86 8.14 10.86 8.14 12.44 7.11 12.42 7.12 12.44 7.11 13.48 6.56 13.50 6.55 13.50 6.55 14.02 6.31 — — — — 16.34 5.42 — — ~ ~ 18.78 4.72 — — — — 20.22 4.39 20.16 4.40 20.16 4.40 180 5.4 DISCUSSION Paclitaxel loaded films based on the three graft copolymers were prepared using the solution casting method and both paclitaxel and the copolymers could be dissolved in the common solvent THF. Homogeneous films were prepared and no crystalline paclitaxel was detected in the films of all three graft copolymers with paclitaxel loading up to 10%. The HPC-g-PCL films exhibited limited swelling in water (2-3% as shown in Figure 5.1) because of the relatively high M S C L and long side chains, which likely resulted in strong hydrophobic interactions between the side chains. The interaction was further enhanced by the formation of a microcrystalline phase of PCL during the film-casting process, as discussed in chapter 3 and shown in Figure 3.12. Due to these interactions, the degree of swelling was very low and the films did not form a hydrogel matrix in an aqueous environment. In contrast to HPC-g-PCL, the side chains in Dx-g-PCLs were short, no phase separation was detected and no PCL-rich microcrystalline phase was formed in the films. The films were transparent and amorphous. After loading with paclitaxel, the films remained transparent and no crystalline paclitaxel was detected, indicating that the amount of PCL side chains in the copolymers was sufficient to confer amphiphilicity to the copolymers and make them suitable matrices for a molecular level dispersion of paclitaxel. Due to the low M S C L , the degree of swelling in water was between 30-40% for the films of both Dx500-g-PCL and Dx70-g-PCL, indicating that the overall hydrophilicity of the copolymers was high. Loading of paclitaxel into the films lowered the extent of swelling of the films, probably because the addition of the hydrophobic paclitaxel increased the overall hydrophobicity of the films. 181 DSC data for the HPC-g-PCL films illustrate that the thermal transitions of the films were influenced by paclitaxel loading. In the first heating scans (Figure 5.6), the Tg's increased with an increase in paclitaxel loading. This might be a consequence of reduced free volume caused by the presence of paclitaxel molecules. In the second heating scans (Figure 5.7), an increase in paclitaxel loading resulted in an increase in the recrystallization temperature (Tc), which was thought to be due to the reduced mobility of the copolymer chains caused by the enhanced interaction between the side chains in the presence of paclitaxel. This would have the effect of inhibiting the recrystallization process. In the first heating scans, melting transitions of HPC-g-PCL films were double endotherms in agreement with data from chapter 3 (Figure 3.15) where it was suggested that the low molecular weight fractions of HPC-g-PCL corresponded to the higher melting component of the double endotherm. The endpoint of the endotherms (Tm) occurred at higher values in the first heating scan compared to the second heating scan (Table 5.2). This may be explained in terms of side chain relaxation processes as follows. In the solution casting process, the solidification process took approximately 24 hours, which was considerably longer than the cooling/re-heating cycle in the DSC studies. Therefore, the side chains of both low and high molecular weight fractions had sufficient time to align into a more ordered crystalline structure leading to a higher Tm. The contribution of slightly different melting events due to higher and lower molecular weight fractions of HPC-g-PCL were erased on cooling and second heating and a single endotherm was obtained. Paclitaxel release from Dx-g-PCL films was significantly higher than from HPC-g-PCL films. This was likely due to the much higher diffusivity of paclitaxel in Dx-g-182 PCL matrices compared to HPC-g-PCL film matrices, arising from the greater extent of swelling of the films. The PCL crystalline phase in the HPC-g-PCL matrix probably also further impeded the diffusion of paclitaxel. As a comparison, paclitaxel release from pure PCL films was much slower than paclitaxel release from any of the copolymer films. PCL is a semicrystalline and hydrophobic polymer and paclitaxel has been shown to be released very slowly from PCL matrices (Winternitz et al. 1996). The 1% paclitaxel loaded Dx-g-PCL films remained transparent throughout the release study. At this loading, the release profiles for paclitaxel from Dx500-g-PCL and Dx70-g-PCL films were very similar indicating that the difference in the molecular weights of the main chain polymer did not result in a significant difference in the paclitaxel release behaviour. However, films loaded with 5% and 10% paclitaxel showed evidence of paclitaxel crystallization during the release studies, indicating that the amount of paclitaxel in the 5% and 10% loaded films exceeded the paclitaxel solubilization capacity of the copolymers following water uptake. It is apparent that a paclitaxel loading of 1% was lower than the paclitaxel solubilization capacity of the copolymers and that the paclitaxel molecules were likely associated via hydrophobic interactions with PCL side chains and solubilized by the copolymers following water uptake. However, when the loadings (5% and 10%) exceeded the solubilization capacity, it is speculated that sufficient concentrations of paclitaxel molecules were dispersed in aqueous "compartments" of the hydrated films, that crystallization of paclitaxel occurred within these "compartments". More crystallization was evident as the loading was increased from 5% to 10%. 183 There was no evidence of paclitaxel crystallization in the HPC-g-PCL films during the release study. This could be explained by the presence of PCL microcrystalline regions in the films which would greatly lower the mobility of the copolymer chains and lead to a very limited water uptake and swelling of the films. Under these conditions, crystallization of paclitaxel would be unlikely. It is suggested that the crystallization events in the films of Dx-g-PCLs during the release study may provide an explanation for the unusual release profiles of 5% and 10% paclitaxel loaded films. The mechanism of paclitaxel release from the copolymer matrices would involve: (1) water diffusion into the matrices; (2) paclitaxel dissolution/partitioning; (3) paclitaxel diffusion through the matrices and release into the media. For 0.1%, 0.5%, and 1% paclitaxel loaded Dx-g-PCL films, paclitaxel was completely solubilized by the copolymers and dispersed in the matrices at a molecular level throughout the period of the release study. Hence, the rate-limiting step for the release of paclitaxel from films at these low loadings was likely paclitaxel diffusion (step 3). However, in 5% and 10% paclitaxel loaded Dx-g-PCL films, paclitaxel was dispersed at the molecular level (solubilized) and in the crystalline state. For solubilized paclitaxel, its release would be rate-limited by step 3 (paclitaxel diffusion). XRD data indicated that crystalline paclitaxel was present in the dihydrate form. It has been reported that the dissolution of paclitaxel dihydrate is a very slow process, and it takes 20 hours to obtain a saturated solution (Liggins et al. 1997). Therefore, for crystalline paclitaxel, the rate-limiting step for its release from films at higher loadings was 184 dissolution of paclitaxel within the film matrix. In the 10% loaded films, a large fraction of the loaded paclitaxel formed crystals during the first 12 hours of the release study and the crystallization reached a maximum in 2 days. Hence, the rapid release phase corresponding to paclitaxel diffusion as rate limiting, only lasted for 2 days. Subsequently, the release rate was mainly controlled by the dissolution rate of paclitaxel. The crystallization in the 5% loaded films was slower than in the 10% loaded films, and did not reach a maximum until the fifth day. During the first 5 days, the release rate from the 5% loaded films was probably controlled by a mixed mechanism including paclitaxel diffusion (fast) and dissolution (slow), and overall release was therefore higher than from the 10% loaded films. Similar observation was reported for the release of diclofenac sodium (water insoluble) from polyfhydroxyethyl methacrylate) hydrogel, where diclofenac sodium precipitated out from the hydrogel at a high loading level and the release was then controlled by the slow dissolution of the drug (Lee 1985). Residual solvents in pharmaceuticals are defined as organic volatile chemicals that are used in the manufacture of active substances or excipients. THF was the solvent used for the preparation of the films. Since there is no therapeutic benefit from residual THF, it should be removed to the extent that meets certain quality-based requirements. According to the Guideline for Residual Solvents recommended by International Conference on Harmonisation (ICH), the Permitted Daily Exposures (PDEs) for THF is 7.2 mg/day (ICH 2002). In this study, the residual THF was determined to be less than 33 ppm in the films, equivalent to less than 13.2 pg of THF in a film with a size of 3.8 cm x 5.1 cm, which is the size of the commercialized adhesion barrier Interceed ® marketed by 185 Ethicon, Inc. (Somerville, NJ). This level is substantially less than the PDE specified by ICH. 5.5 CONCLUSIONS Homogeneous film formulations with molecular level dispersed paclitaxel were successfully developed using the three novel amphiphilic graft copolymers HPC-g-PCL, Dx500-g-PCL, and Dx70-g-PCL. The degree of swelling of the Dx-g-PCL films in PBS (10 mM, pH7.4) was significantly higher than that of HPC-g-PCL films and the loading of paclitaxel into the films lowered the degree of swelling. The release rates from the Dx-g-PCL films were higher than from the HPC-g-PCL films, and the release profiles were influenced by crystallization events occurring in the Dx-g-PCL films during the release study. 186 CHAPTER 6 EVALUATION OF PACLITAXEL LOADED DEXTRAN-GRAFT-PCL FILMS IN ANIMAL MODELS OF SURGICAL ADHESIONS 6.1 INTRODUCTION Following the in vitro characterization of the paclitaxel loaded graft copolymer films, we were interested in evaluating the films in animal models. Paclitaxel was recently shown to be effective in preventing surgical adhesions when administered either as an intraperitoneal injection or loaded in hyaluronic acid films. However, the films without paclitaxel did not show a barrier effect. Intraveneous (i.v.) and intraperitoneal (i.p.) injections have been two major administration routes in the drug treatments of surgical adhesions. The administration regimen was found to have a profound impact on the outcome of the treatments. A single i.p. injection of heparin irrigating solution (5000 IU/Liter) failed to reduce adhesions in operations on the pelvis for infertility performed upon 92 patients (Jansen 1988). Failures were also reported in a rabbit uterine horn model by single i.v. injection of heparin (Diamond et al. 1991). The efficacy was improved by giving multiple doses of heparin (Knightly et al. 1962; Parker et al. 1987). Improved efficacy by using multiple dose regimens was also reported for other drugs such as fibrinolytic drugs (Knightly et al. 1962). Using an osmotic minipump to deliver drugs directly to the site of adhesions led to a reduction in adhesion formation at lower doses and showed the potential of the 187 application of controlled release systems in the treatment of surgical adhesions (Fukasawa et al. 1991; Orita et al. 1991). In this study, the paclitaxel loaded graft copolymer films were designed such that they could act both as a physical barrier and a controlled paclitaxel delivery system. Based on the results of the in vitro characterization, Dx-g-PCL films were found to be more hydrophilic with greater flexibility after swelling in aqueous media than HPC-g-PCL films. In addition, paclitaxel release from the Dx-g-PCL films was faster than from HPC-g-PCL films. Therefore, paclitaxel loaded Dx-g-PCL films were selected as suitable candidates for in vivo evaluation in the rat model of surgical adhesions. 6.2 EVALUATION OF THE FILMS IN A RAT SIDEWALL CECUM MODEL 6.2.1 Objectives The objectives of this study were (1) to assess the biocompatibility and the biodegradation lifetime of both Dx500-g-PCL and Dx70-g-PCL films; (2) to determine in vivo paclitaxel release rate; (3) to evaluate the efficacy of the films in preventing surgical adhesions. 6.2.2 Experimental 6.2.2.1 Formulations Dx500-g-PCL films with paclitaxel loadings of 0%, 1%, and 5%, and Dx70-g-PCL with 1% paclitaxel loading were selected as the formulations for this study. The dimensions of the films were 1.2 cm x 1.8 cm, and the thickness was around 140 pm. The films were sealed in plastic packages and sterilized by y-irradiation (cobalt-60) at a total dose of 2.5 Mrad. 188 6.2.2.2 Animals and housing Sixty Sprague-Dawley rats (weight 250-350g) were obtained from U B C Animal Care. The experiments were conducted at Jack Bell Research Center and the animals were housed in plastic cages. Before surgery, each cage housed two rats. After surgery, each cage contained one rat. The room was maintained at a temperature of approximately 20°C with 30-70% relative humidity and a light/dark cycle of 12 hours/12 hours. Rodent chow and tap water were provided ad libitum to the animals for the duration of the study. 6.2.2.3 Animal Care and Use Committee approval This study fell under the scope of protocol No. AO 1-0012 approved by the Committee on Animal Care of the University of British Columbia. The experiments were conducted in accordance with the principles contained in Care of Experimental Animals—A guide for Canada, published by the Canadian Council on Animal Care. 6.2.2.4 Experiment design The sixty rats were randomly divided into 9 groups. Table 6.1 gives the group assignments, paclitaxel dose, number of rats, and time of sacrifice of each group. Groups A, B, C, D, and E were assigned to the efficacy study, and groups F, G, H, and I were assigned to the degradation study. Tissue reactions were observed in all animals with a film implanted. 189 Table 6.1 Group assignments, paclitaxel dose, number of rats, and schedule of sacrifice Group Treatment Amount of PTX in films (pg) Numbers of animals Sacrifice time (week) A Surgical control (no film) 0 9 1 B Dx500-g-PCL films with no PTX 0 9 1 C 1% PTX loaded Dx500-g-PCL films 397±30 9 1 D 5% PTX loaded Dx500-g-PCL films 1873+195 9 1 E 1% PTX loaded Dx70-g-PCL films 457±38 9 1 F Dx500-g-PCL films with no PTX 0 4 2 G Dx500-g-PCL films with no PTX 0 4 4 H 5% PTX loaded Dx500-g-PCL films 1873+195 3 2 I 5% PTX loaded Dx500-g-PCL films 1873+195 4 4 6.2.2.5 Surgical procedures The rats were prepared for surgery by anaesthetic induction with 5% halothane in an enclosed chamber, then transferred to the surgical table, and anaesthesia maintained by a nose cone with halothane throughout the procedure. The abdomen was shaved, sterilized, draped and entered via a midline incision. The cecum was lifted from the abdomen and placed on sterile gauze dampened with saline. Dorsal and ventral aspects of the cecum were successively scraped a total of 45 times over the terminal 1.5 cm using a No. 10 scalpel blade. (Blade angle and pressure are controlled to produce punctuated bleeding, while avoiding severe tissue damage or tearing.) The left side of the abdominal cavity was then retracted and everted to expose a section of the peritoneal wall nearest the natural resting cecal location. The exposed superficial layer of muscle (transverses 190 abdominis) was excised over an area of 1x1.5 cm2. Excision included portions of the underlying internal oblique muscle, leaving behind some intact and some torn fibers from the second layer. Minor local bleeding was tamponaded until controlled. The wound areas were overlaid with the assigned films. The abraded cecum was positioned over the film covered sidewall wound and sutured at four points immediately beyond the dorsal corners of the 1.5 cm edge. The large intestine was replaced in a natural orientation continuous with the cecum. The abdominal incision was then closed in two layers with 4-0 vicryl sutures. Subjects were collared while still anaesthetized, placed on clean bedding, and covered with towels to maintain body temperature during recovery. When completely mobile, rats were returned to their cages, provided with food and water ad libitum, and examined daily for signs of wound infection. Collars were removed 24 hours following surgery (the time period found to be sufficient for prevention of subject interference with the abdominal closure). At the endpoints of the experiments, animals were sacrificed by lethal injection and examined post mortem by two investigators who independently scored the tissue reactions and adhesions using established scoring systems. 6.2.2.6 Evaluation 6.2.2.6.1 Biocompatibility and biodegradation A composite subjective assessment of overall tissue reaction was made and categorized as None, Mild, Moderate, and Severe, which were defined as follows: None: Virgin appearance Mild: Slight erythema/blanching Moderate: Erythema, or minor encapsulation 191 Severe: Severe encapsulation, necrosis, or organ anomalies including liver lesion/mottled and/or spleen blistered/darkened. Biodegradation lifetime of the films was assessed based on the integrity and general appearance of the films after two and four weeks. 6.2.2.6.2 Adhesion scoring system Adhesions at abraded sites were assessed according to the following scale: 0-0.5 No adhesions or only fat attachment 0.5-1.5 Slight adhesions, separable by blunt dissection 1.5-2.0 Adhesions, not easily separable in single area 2.0-3.0 Cohesive adhesions, aggressive blunt or sharp dissection required >3.0 Cohesive adhesions, perforation or tearing of cecum unavoidable 6.2.2.6.3 Statistical analysis The mean adhesion scores was compared by Student t-test. The percentage of animals with no adhesion and the percentage of animals with adhesion scores of 1.5 or higher between each pair of groups were compared by Chi-Square analysis. In all cases, a p value less than 0.05 was considered statistically significant. 6.2.2.7 Estimation of paclitaxel release in rats The amount of paclitaxel released during the one-week implantation was estimated by subtracting the residual paclitaxel in the films from the initial loading. The residual paclitaxel was determined as follows. The films withdrawn from the animals were dried under vacuum for 48 hours. A portion of each film was accurately weighed and put into a test tube. 2 mL of DMSO was added to the tube to swell the film. After 24 hour swelling, the DMSO solution (200uL 192 for 1% paclitaxel loaded films, lOOpL for 5% films) was transferred to another tube, dried under vacuum to remove DMSO, and reconstituted using 1 mL of ACN/H2O (60/40 v/v). The amount of paclitaxel was measured using HPLC. 6.2.3 Results 6.2.3.1 Biocompatibility and biodegradation A summary of tissue reactions and film degradation is given in Table 6.2. Dx500-g-PCL films with no paclitaxel induced the least tissue reaction. None of the animals in group (B) developed necrosis. Minor encapsulation was observed in 4 of the animals. Tissues and organs possessed a good color and normal appearance. In group (C), animals implanted with the 1% paclitaxel loaded Dx500-g-PCL films, fibrotic encapsulation was observed in 4 of the 9 animals. Borderline necrotic reactions including white spots or paleness were also found in 4 of the animals. White spots appeared on the sidewall wound areas where a film had been applied. Occasionally they were seen on the other areas, such as the abdominal incision line. The white spots might originate from a precipitation of paclitaxel or be associated with spots of necrosis in tissue fibers. For the animals implanted with 1% paclitaxel loaded Dx70-g-PCL films (group E), severe necrosis with bad odor was found in one animal, which was apparently caused by leakage of cecal contents. Fibrotic encapsulation of the films occurred in the other 8 animals. The density of the fibrotic tissue was greater than for Dx500-g-PCL films. The 5% loaded films (group D) induced severe necrotic reactions and organ anomalies due to the high local concentration of paclitaxel. The necrosis might also result from small amounts of cecal contents leaked from the tiny holes where tacking sutures were inserted. These punctures appeared to seal themselves in the case of the no treatment control group and 193 control film (no paclitaxel) group, but not in the groups treated with paclitaxel loaded films. This suggested that paclitaxel may have played a role in preventing the healing of the holes. As the loading increased from 1% to 5%, the inhibition of wound healing became more significant. Tissue reactions and necrosis were much more pronounced over an extended period of exposure to the formulations than those over one week. All the animals implanted with 5% paclitaxel loaded Dx500-g-PCL films developed massive necrosis compared to none of those with control Dx500-g-PCL films after two or four weeks. The above observations clearly indicated that the paclitaxel doses in the groups C, D and E were toxic to tissues. Table 6.2 Summary of the observations of biocompatibility and biodegradation of the implanted films in the rat sidewall cecum model of surgical adhesions Group Overall tissue reaction % With necrotic reaction % With film encapsulated Film integrity A — 0 — — B None-mild 0 44 Intact (1 week) C Moderate plus 44 44 Intact (1 week) D Severe 67 33 Fragile (1 week) E Moderate plus 11 89 Gelatinous (1 week) F None-mild 0 0 Intact (2 weeks) G Moderate 0 50 Gelatinous (4 weeks) H Severe 100 — Intact (2 weeks) I Severe 100 -- Intact (4 weeks) The degradation of the films was judged by the integrity of the films. A change of the film from intact to gelatinous was considered to be evidence of film degradation. As 194 shown in Table 6.2, after one-week implantation, Dx500-g-PCL films remained intact while Dx70-g-PCL films became gelatinous and could no longer be withdrawn as an intact film. This suggested that Dx70-g-PCL degraded faster than Dx500-g-PCL. Four weeks implantation was required for Dx500-g-PCL to become gelatinous. The films in Group I were not gelatinous after four weeks, possibly due to the massive necrosis developed, which may have inhibited enzyme activity in the wound areas, thus inhibiting any enzymatic degradation process. 6.2.3.2 Adhesion scores The scoring system was designed to assess the density and tenacity of the adhesions formed between the tissues and/or organs. It became non-applicable when necrosis occurred, because decomposition of connective tissue in the presence of necrosis would dramatically lower the tenacity of the adhesions and thus yield faulty scores. In this study, all the groups (C, D, and E) treated with paclitaxel were overdosed and developed necrotic reactions. Therefore, only groups A and B were assessed in terms of adhesion scores and the results are shown in Table 6.3. In group A (sham), adhesions developed in all the animals, and all had adhesion scores equal to or greater than 1.5. In group B, in which the animals were implanted with control (paclitaxel absent) Dx500-g-PCL films, only 3 of the 9 animals had adhesions, and the adhesion scores were all lower than 1.5. The percentage of animals with no adhesion in group B was significantly higher than that in group A (Chi-Square). The mean of the adhesion scores of group A was 2.1, which was significantly higher than that of group B (Student t-test). This evidence indicated that, as a barrier, Dx500-g-PCL films could effectively reduce the incidence of adhesion formation. 195 Table 6.3 Adhesion formations in the groups A and B after one week Group Adhesion score % With no adhesion % With score > 1.5 A 2.1 ± 0 . 4 0 100 B 0.3 + 0.3 67 0 p value 1.1 x 10"8* 0.01 ** 0.001 ** * Student t-test, p<0.05 was considered statistically significant; ** Chi-square analysis, p<0.05 was considered statistically significant. 6.2.3.3 In vivo paclitaxel release The amount of paclitaxel released during the one-week implantation is given in Table 6.4. Approximately 330-350 ug of paclitaxel was released in the week from both the 1% paclitaxel loaded Dx500-g-PCL (group C) and Dx70-g-PCL (group E) films, and over 1.4 mg released from the 5% paclitaxel loaded Dx500-g-PCL films (group D). The percentage paclitaxel released was in the range of 74% to 88%, indicating that the majority of the initially loaded paclitaxel was released within a week. Table 6.4. In vivo paclitaxel release after one week Group Amount released (pg) % Release C 348 ± 72 87.2 ± 8 . 1 D 1431 ± 3 6 5 77.3 ± 9 . 1 E 338 ± 3 9 74.2 ± 6.9 196 6.2.4 Discussion Adhesion formation is a complex phenomenon that has only been replicated in whole animal systems. There is no in vitro model available to date. Rabbits and rats have been the species of choice for the study of surgical adhesions. In the present study, rats were used since rats are the smallest and most easily managed animal subjects suited to the proposed procedures, and most importantly, recognized rat models for surgical adhesions are available. The sidewall cecum model involved injuries of both the surfaces of the cecum and a section of the sidewall. The films were sandwiched between the injured surfaces which were held together by suturing at the four corners of the wounds. The advantage of this model was that it allowed the investigators to look at specifically the development of adhesions between two target surfaces. The drawbacks of the model included damages to the cecum caused by the suturing needle and anchoring of the cecum. Punctures in the cecum by the suturing needle might cause leakage of the cecal contents, containing bacteria. Fecal induction had apparently caused necrosis in at least 8 of the rats. Since all these rats were treated with paclitaxel loaded films, it was believed that paclitaxel played a role in preventing the needle wound from healing. In addition, the anchoring of the cecum to the sidewall limited the natural movement of the cecum and might increase the likelihood of adhesion formation. The Dx500-g-PCL films without paclitaxel significantly reduced the incidence and the severity of adhesion formation. The effect of paclitaxel on the inhibition of adhesion formation was not shown due to the tissue reactions caused by the high paclitaxel dose. The paclitaxel loading levels in the films were 1% and 5%, 197 corresponding to total amounts of 397 jag and 1873 jag paclitaxel, respectively. All films caused necrotic reactions, and the reaction in the 5% paclitaxel group was more severe than in the 1% paclitaxel group. This suggested that the necrotic reactions were paclitaxel related and were exacerbated with the increase of paclitaxel dose. 6.2.5 Conclusions The Dx500-g-PCL and Dx70-g-PCL films could be readily manipulated into and around the implantation sites. The Dx500-g-PCL films were considered more suitable for adhesion barriers due to the lower incidence of fibrotic encapsulation compared to the Dx70-g-PCL films. The Dx500-g-PCL films with no paclitaxel significantly reduced adhesion formation indicating a barrier effect. The paclitaxel loaded films at both 1% and 5% loading levels resulted in necrotic reactions. It was thus concluded that paclitaxel doses used in this study were toxic. 6.3 EFFICACY OF PACLITAXEL LOADED DX-g-PCL FILMS WITH LOWER DOSES IN A RAT CECAL ABRASION MODEL The study was conducted using an agreed-upon protocol (Genzyme Study Number 02-0064) by the group of Biomaterials and Surgical Products Research at Genzyme Corporation, one of the collaborating groups of this laboratory. 6.3.1 Objectives The objective of this study was to evaluate the efficacy of paclitaxel loaded Dx500-g-PCL films at lowered dosing levels in the prevention of surgical adhesions in a rat cecal abrasion model. 198 6.3.2 Experimental 6.3.2.1 Formulations Based on the results in 6.2, paclitaxel loading level in the films was reduced and Dx500-g-PCL films with paclitaxel loadings of 0.1% and 0.5% were used in this study. The dimensions of the films were 3.5 cm x 3.5 cm, and the thickness was around 140 pm. The films were sealed in plastic packages and sterilized by y-irradiation (cobalt-60) at a total dose of 2.7 Mrad. 6.3.2.2 Animal and housing Forty Sprague Dawley rats (rattus norvegicus) (weight 250 g) were obtained from Charles River Labs, Wilminton, M A . Only animals which appeared grossly normal (i.e., healthy appetite, bright clear eyes, no unusual exudates from any body orifice, alert and active posture) were used. The animals were housed in polycarbonate cages. The room temperature was maintained at 18-25 °C with 30-50% humidity and a light/dark cycle of 12 hours/12 hours. Rodent chow and tap water were provided ad libitum. 6.3.2.3 Animal Care and Use Committee approval The study protocol fell under the scope of protocol number 99-1021-7 approved by the Genzyme Institutional Animal Care and Use Committee. The study was performed in accordance with the Guide for the Care and Use of Laboratory Animals, National Academy Press, 1996. 6.3.2.4 Randomization of animals The forty rats were equally randomized into four groups prior to the initiation of the study. Table 6.5 gives the group assignments, number of rats in each group, and PTX dose. The animals were sacrificed after one week following surgeries. 199 Table 6.5 Group assignments, number of rats and PTX doses for the efficacy study Group ID Number of rats Film PTX(pg) Surgical control 10 No film 0 Film control 10 Dx500-g-PCL without PTX 0 0.1% PTX 10 Dx500-g-PCL with 0.1% PTX 223 ± 22 0.5% PTX 10 Dx500-g-PCL with 0.5% PTX 952 ± 98 6.3.2.5 Surgical procedures: At the time of surgery, the animals were weighed and anesthetized with a single injection of ketamine hydrochloride (85 mg/kg body weight) and xylazine hydrochloride (6 mg/kg), administered into the large muscles of the thigh. A bland ophthalmic ointment was placed in each eye to protect it from corneal desiccation and ulceration. The abdomen was shaved with #40 veterinary clippers and prepped with povidone/iodine scrub and successive alcohol wipes. Sterile towels, drapes, and instruments were used and procedures were performed in a room reserved for aseptic survival surgery. A 4 cm incision was made through the skin with a #10 scalpel blade. A #11 scalpel blade was used to pierce the lima alba while the muscle was tented with forceps. Iris scissors were used to extend the laparotomy. The cecum was removed from the abdomen with sterile cotton swabs. The contents of the cecum were expressed manually into the ascending colon. The cecum was abraded a total of four times on the ventral and dorsal surfaces with a mechanical abrading device which permitted operator independent, controlled abrasion over a defined area. The films were placed on the cecum over the abraded area before returning the cecum to the abdominal cavity. 200 The incisions of all animals were closed in two layers: the muscle and peritoneum were closed in a simple continuous pattern with 3.0 Dexon suturing material, while the skin and attending fascia were closed with 9 mm stainless steel staples. Animals were allowed to recover completely in an incubator, and then provided with food and water ad libitum. 6.3.2.6 Evaluation of the adhesion formation 6.3.2.6.1 Adhesion score Animals were evaluated seven days post surgery immediately following euthanasia. The evaluators were blind to group assignment. Adhesions were evaluated and scored from 0 to 4 according to the following scoring system: 0 No adhesions 1 Filmy adhesions with easily identifiable plane 2 Mild adhesions with freely dissectable plane 3 Moderate adhesions with difficult dissection of plane 4 Dense adhesions with non-dissectable plane 6.3.2.6.2 Incidence of adhesions The incidence of adhesions was recorded based on the anatomical location using the following code: CC cecum to cecum adhesion CO cecum to omentum adhesion C M cecum to mesentery adhesion CB cecum to bowel adhesion CSI cecum to small intestine adhesion 201 The mean incidence of adhesions of all types in each group was determined by dividing the total number of adhesions by the total number of animals. 6.3.2.7 Statistical analysis The mean incidence of adhesions between each pair of groups was compared by Wilcoxon Rank-Sum analysis. The percentages of animals with significant (grade 2 or higher) adhesions and with no adhesions between each pair of groups were compared by Chi-Square analysis. In all cases, a p value less than 0.05 was considered statistically significant. 6.3.3 Results The efficacy results in terms of percentage of animals with no adhesions and with grade 2 or higher adhesions, and the mean incidence adhesions are given in Table 6.6. In the surgical control group, 70% of the animals developed grade 2 or higher adhesions and only 20% of them had no adhesion. In both the groups implanted with 0.1% and 0.5% paclitaxel loaded films, 80% of the animals formed no adhesions and only 10% of them developed grade 2 or higher adhesions, indicating that these films significantly reduced adhesion formation compared to surgical control group. The mean incidence data also showed that the paclitaxel treated groups had significantly lower values (0.2 for the 0.1% paclitaxel group and 0.5 for the 0.5% group) than the surgical control group, which had a mean incidence of 1.4. The blank Dx500-g-PCL films (film control group) also reduced adhesion formation compared to the surgical control group. The percentage of animals with no adhesions increased from 20% to 40%, and the percentage with grade 2 or higher adhesions decreased from 70% to 40%. Mean incidence was also lowered from 1.4 to 1.0. Although these differences were not statistically significant, the reductions in adhesion 202 formation and in mean incidence were considered an indication of a barrier effect of the films. Table 6.6. Percentage of the animals with no adhesions and with grade 2 or higher adhesions, and incidence of adhesions in the efficacy study Group % With no adhesions a % With adhesions > 2 a Incidence b Surgical control 20 70 1.4 ± 0 . 3 Film control 40 40 1.0 ± 0 . 3 0.1% PTX 80* 10* 0.2 ± 0 . 1 * + 0.5% PTX 80* 10* 0.5 ± 0.4* a, Difference between the groups compared by Chi-Square analysis b, Difference between the groups compared by Wilcoxon Rank-Sum analysis * p < 0.05 compared to the surgical control group + p < 0.05 compared to the film control group The ceca in the film treated groups appeared normal with no evidence of impaired wound healing or trauma. Excess fluid in the abdominal cavity was observed in 5 of the 10 animals in the 0.5% PTX group, and 1 of the 10 animals in the 0.1% PTX group. No encapsulation and other abnormalities were noted. 6.3.4 Discussion The use of paclitaxel to prevent surgical adhesion formation in the rat cecal abrasion model has already been established by Jackson et. al. (2002). Intraperitoneal paclitaxel injection served as a positive control for our studies, demonstrating that at a dose of 4 mg/kg daily for 4 days, paclitaxel significantly inhibited adhesion formation. 203 In the cecal abrasion model, only the surfaces of the cecum were damaged, and the films were placed over the damaged surfaces. The ceca were not anchored and could move naturally in the abdominal cavity. The control Dx500-g-PCL films (paclitaxel absent) reduced the incidence of adhesion formation and showed a barrier effect. The amounts of paclitaxel in unit area of the films used in this model were about one tenth of those used in the sidewall cecum model. Both 0.1% and 0.5% paclitaxel loaded films significantly inhibited adhesion formation suggesting that paclitaxel is an effective inhibitor of surgical adhesions. A B Figure 6.1 Diagrammatic representations of (A) sidewall cecum model and (B) cecal abrasion model in the rat Localization of paclitaxel in the cecal abrasion model was different from that in the sidewall cecum model. Diagramatic representations of the two models are shown in Figure 6.1. In the sidewall cecum model, the paclitaxel released was highly localized in the sutured area. However, in the cecal abrasion model, one side of the film was attached to the damaged cecal surface and the other side was open to the peritoneal cavity, which 204 would allow a large fraction of paclitaxel to be released into the peritoneal fluid. This would likely substantially reduce the local paclitaxel concentration at the wound area compared to that in the sidewall cecum model. No tissue reaction and abnormalities were observed in the film control group indicating that Dx500-g-PCL is a biocompatible polymer. Excess fluid in the abdominal cavity was noted in 50% of the animals implanted with the 0.5% paclitaxel loaded films and in 10% of the animals with the 0.1% paclitaxel loaded films. It appeared that the excess fluid was paclitaxel related and the percentage of the animals with the excess fluid was increased with the increase of paclitaxel dose. 6.3.5 Conclusions Dx500-g-PCL is a biocompatible material. The control films (paclitaxel absent) showed a barrier effect. Adhesion formation was effectively inhibited by applying a paclitaxel loaded Dx500-g-PCL film on the damaged surface of ceca. An appropriate loading level of paclitaxel in the films was 0.1% to achieve efficacy in this model. 205 CHAPTER 7 SUMMARIZING DISCUSSION, CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK 7.1 SUMMARIZING DISCUSSION The goal of this work was the development and characterization of paclitaxel loaded polymeric films for the prevention of surgical adhesions. The ideal properties of formulations for this purpose should include: biocompatible and biodegradable polymeric materials with an in vivo degradation time of less than one month; the formulation should be uniform in terms of paclitaxel dispersion and the loaded paclitaxel should be released within about 10 days. In addition, the formulation should be easy to handle and sterilizable. The film formulations developed in this work were based on biocompatible polysaccharides including chitosan and the novel amphiphilic graft copolymers HPC-g-PCL and Dx-g-PCL. In chapter 2, paclitaxel loaded chitosan films were prepared and characterized. Due to its hydrophobicity, paclitaxel precipitated out and formed crystals in the chitosan matrices. Needle-shaped paclitaxel crystals with lengths up to 100 pm were observed. In vitro paclitaxel release rates from the films decreased with an increase in the degree of crosslinking and increased when chitosan was blended with PEO. The films were found intact after being implanted for 28 days in a rat model of surgical adhesions. Given these properties of long degradation times, slow paclitaxel release rates and heterogeneity of paclitaxel dispersions, chitosan based films were considered unsuitable drug delivery 206 systems for potential application in the prevention of surgical adhesions and were not subjected to further evaluation. In general, hydrophilic polymers and hydrogels possess optimal biocompatibility due to their rapid water uptake properties and subsequent high water content following placement in aqueous media (Andrade 1976). However, hydrophobic drugs such as paclitaxel are not miscible with hydrophilic polymers and the loading and precipitation of paclitaxel in the matrix results in non-uniform, heterogeneous formulations (Shi et al. 1999; Jackson et al. 2002). In this work, amphiphilic biomaterials were synthesized with balanced hydrophilicity and hydrophobicity, such that they would retain sufficient hydrophilic properties to be biocompatible and also be suitable for loading paclitaxel and maintain good handling properties. In chapters 3 and 4, the synthesis and characterization of novel amphiphilic graft copolymers of HPC-g-PCL and Dx-g-PCLs was achieved. To our knowledge, this represents the first synthesis and characterization of HPC-g-PCL (Shi and Burt 2003). Ydens et al. reported the synthesis of Dx-g-PCLs with low molecular weight dextrans ( M W : 6,600 and 21,300) as the main chains, in which dextran was silylated before the graft polymerization (Ydens et al. 2000). In this work, Dx-g-PCLs with high molecular weight dextrans ( M W : 70,000 and 500,000) as the main chains were synthesized without silylation of dextran. The application of HPC-g-PCL and Dx-g-PCL copolymers as biomaterials for drug delivery has not been previously reported. The evidence for the formation of the graft copolymers was obtained using GPC and N M R spectroscopy. HPC-g-PCL with M S C L ranging from 8.6 to 10.1 was synthesized. The average side chain length was estimated to be 7.3. The HPC-g-PCL 207 copolymer exhibited excellent film-forming ability. However, due to the length of the PCL side chains, heterogeneity and PCL-rich microcrystalline regions were observed in the films cast from THF solutions. Compared to HPC-g-PCL, the side chains of Dx-g-PCLs were much shorter. The M S C L was determined to be in the range of 1.2 to 1.3 and the side chain length in the range of 2.4 to 3.1. The M S C L and the side chain length showed a marked effect on the overall hydrophilicity of the films. The results of the swelling study of the films presented in chapter 5 demonstrated that the longer side chains of HPC-g-PCL resulted in a relatively higher hydrophobicity of the copolymer and thus very limited swelling of the films (less than 3%). In contrast, with shorter PCL side chains, Dx-g-PCL remained hydrophilic and the films became hydrogels when incubated in 10 mM PBS (pH7.4) at 37 °C. The degree of swelling was in the range of 36-38%. In chapter 5, paclitaxel loaded HPC-g-PCL and Dx-g-PCL films were prepared and characterized. Uniform paclitaxel dispersions were obtained for both types of films with paclitaxel loadings up to 10% (w/w) and controlled paclitaxel release rates were achieved. Handling characteristics of the films were good in that the films were flexible and had sufficient mechanical strength after being hydrated. Given that the release rates from the HPC-g-PCL films were significantly lower than from the Dx-g-PCL films, they were considered less suitable paclitaxel delivery systems for the prevention of surgical adhesions. The Dx-g-PCL films with paclitaxel loading in the range of 0.1% to 1% released over 70% of the initial loaded drug within one week, which was considered appropriate for a surgical adhesion formulation. The rapid release of drug in a timeframe of less than about 10 days was believed necessary because adhesion-related events, such as migration of fibroblasts and angiogenesis, are known to be initiated 2 to 3 days 208 following injury and reach a significant level at day 6 (Treutner and Schumpelick 1997). Hence, it has been estimated that the inhibitory drug paclitaxel would need to be released at the site of injury over a period of about 10 days. The phenomenon of paclitaxel crystallization within the Dx-g-PCL matrices at loadings of 5% and 10% was observed during release studies and was because the loading exceeded the solubilization capacity of the copolymers. In chapter 6, paclitaxel loaded Dx500-g-PCL and Dx70-g-PCl films were evaluated in two animal models of surgical adhesions: a rat sidewall cecum model and a rat cecal abrasion model. The films were first evaluated in the rat sidewall cecum model, in which both the sidewall and cecal surfaces were damaged and the films were sandwiched between the sidewall and the cecum and anchored using sutures. Based on the results obtained using this model, Dx500-g-PCL films were considered more suitable for adhesion barriers due to the lower incidence of fibrotic encapsulation compared to the Dx70-g-PCL films. It is not clear why Dx70-g-PCL films induced a higher incidence of fibrotic encapsulation. A major drawback of the sidewall cecum model is that the punctures in the cecum caused by the suturing needle caused leakage of the cecal contents, which contained bacteria and resulted in necrosis in some animals. In the cecal abrasion model, only the cecal surfaces were damaged and the ceca were not anchored and could move naturally in the abdominal cavity. The control Dx500-g-PCL films (paclitaxel absent) reduced the incidence of adhesion formation and showed a barrier effect in both the models. In the rat cecal abrasion model, 0.1% and 0.5% paclitaxel loaded Dx500-g-PCL films led to a much more effective inhibition of adhesion formation compared to control films, indicating that paclitaxel as an inhibitor of surgical adhesions 209 played an important role. Animals implanted with films with the lower loading level of 0.1% paclitaxel showed less evidence of an inflammatory reaction than the animals implanted with 0.5% paclitaxel loaded films. The latter animals showed excess fluid in the abdominal cavity, which was considered to be caused by a mild toxicity effect of paclitaxel. The in vivo degradation of the films was evidenced by the fact that the films became gelatinous after being implanted in rats for 4 weeks. The degradation rate was slower than an ideal barrier, which should stay in place during the critical phases of adhesion formation (normally about 7 days) and should then undergo rapid biodegradation and absorption (Gomel et al. 1996). The results of this work demonstrate that a 0.1% paclitaxel loaded film formulation based on the novel amphiphilic graft copolymer Dx500-g-PCL possessed suitable characteristics as a potential drug delivery system for placement at surgical sites for the prevention of surgical adhesions. One of the significant limitations of the formulation includes the slow biodegradation rate of the films following drug release. 7.2 CONCLUSIONS Novel amphiphilic graft copolymers with balanced hydrophilicity and hydrophobicity were successively synthesized using biocompatible and biodegradable HPC and dextrans as the main chains and PCL as the side chains. These copolymers were able to effectively solubilize hydrophobic paclitaxel and resulted in homogeneous film formulations with up to 10% (w/w) paclitaxel loading. Handling characteristics of the films were good in that the films were flexible and had sufficient mechanical strength after being hydrated. The degradation rate of HPC-g-PCL was very slow and the copolymer was considered an unsuitable carrier for a surgical 210 adhesion formulation. Significant degradation of Dx-g-PCL films was observed after being implanted in rats for one month. Controlled paclitaxel release was achieved using the copolymers as carriers. The Dx500-g-PCL films with paclitaxel loadings of 0.1% and 0.5% released over 70% of the initial loaded drug within one week and were considered appropriate for a surgical adhesion formulation. The Dx500-g-PCL films were biocompatible and showed a barrier effect in the rat models of surgical adhesions. Adhesion formation was significantly inhibited by applying 0.1% paclitaxel loaded Dx500-g-PCL films to a surgical site in a rat cecal abrasion model. 7.3 SUGGESTIONS F O R F U T U R E W O R K The films developed in this work were sterilized with gamma irradiation prior to the in vivo studies. Typically, the irradiation of biodegradable polymeric drug delivery systems with a minimum dose of gamma irradiation results in slight reduction in molecular weight of polymers (Mohr et al. 1999; Montanari et al. 2001). Limited influences of the irradiation on the drug release rates are described in the literature (Mohr et al. 1999). Complete characterization of possible changes in the copolymer properties remains to be done in future work. In this thesis, the synthesis and characterization of the three graft copolymers, HPC-g-PCL, Dx500-g-PCL, and Dx70-g-PCL, were focused on one side chain length for each copolymer. In future studies, copolymers with a series length of side chains could be synthesized and the properties of these copolymers compared. For instance, copolymers with different side chains length should have different solubilization capacities. It may be 211 possible to establish a quantitative relation between the side chain length and the solubilization capacity. In addition, it would be interesting to synthesize graft copolymers with different side chain polymers such as PLA and PLGA and compare the properties of these copolymers as well as the properties of formulations. It would also be worthwhile to investigate the effect of structural parameters such as main chain length, side chain length, and molar substitution on the properties of the graft copolymer molecules in solution, and the possibility that these amphiphilic graft copolymers may form self-assembly structures, such as micelles. As to the in vivo evaluation of the formulations, the future work should include studies of pharmacokinetics and toxicology. Issues such as blood concentration, tissue distribution, major metabolites and excretion pathways should be addressed. Toxic effects of both paclitaxel and the copolymers should also be investigated. Commercialized barriers such as Interceed® and/or Seprafilm™ should be included in the study for comparison. 212 R E F E R E N C E : Abrahams, J. M . , M . S. Forman, M . S. Grady and S. L. Diamond (2001). "Biodegradable polyglycolide endovascular coils promote wall thickening and drug delivery in a rat aneurysm model." Neurosurgery 49(5): 1187-1195. Aden, M . A., E. Bianchi, A. Ciferri, G. Conio and A. Tealdi (1984). "Mesophase Formation and Chain Rigidity in Cellulose and Derivatives. 2. (Hydroxypropylcellulose in Dichloroacetic Acid." Macromolecules 17: 2010-2015. Aiedeh, K., E. Gianasi, I. Orienti and V. Zecchi (1997). "Chitosan microcapsules as controlled release systems for insulin." Journal of Microencapsulation 14(5): 567-576. Ajani, J. A., D. H. Ilson, K. Daugherty, R. Pazdur, P. M . Lynch and D. P. Kelsen (1994). "Activity of taxol in patients with squamous cell carcinoma and adenocarcinoma of the esophagus." Journal of the National Cancer Institute 86(14): 1086-1091. Albertsson, A., J. Carlfors and C. Sturesson (1996). "Preparation and characterisation of poly(adipic anhydride) microspheres for ocular drug delivery." Journal of Applied Polymer Science 62(4): 695-705. Alexeev, V. L., E. A. Kelberg, G. A. Evmenenko and S. V. Bronnikov (2000). "Improvement of the mechanical properties of chitosan films by the addition of poly(ethylene oxide)." Polymer Engineering and Science 40(5): 1211-1215. Alfrey, J. T., E. F. Gurnee and W. G. Lloyd (1966). "Diffusion in glassy polymers." Journal of Polymer Science. Part C, Polymer Symposia No. 12: 249-261. 213 Allcock, H. R., S. R. Pucher and A. G. Scopelianos (1994). "Poly[(amino acid ester)phosphazenes] as substrates for the controlled release of small molecules." Biomaterials 15(8): 563-569. Allen, C., J. Han, Y. Yu, D. Maysinger and A. Eisenberg (2000). "Polycaprolactone-b-poly(ethylene-oxide) copolymer micelles as a delivery vehicle for dihydrotestosterone." Journal of Controlled Release 63: 275-286. Amin, H. K. and R. S. Neuwirth (1983). "Operative hysteroscopy utilizing dextran as distending medium." Clinical Obstetrics and Gynecology 26(2): 277-284. Andrade, J. D. (1976). Hydrogels for Medical and Related Applications. Washington, D. C , American Chemical Society. Arranz, F. and M . Sanchez-Chaves (1988). "Carbon-13 nuclear magnetic resonance spectral study on the distribution of substituents in relation to the preparation method of partially acetylated dextrans." Polymer 29(3): 507-512. Axel, D. I., W. Kunert, C. Goggelmann, M . Oberhoff, C. Herdeg, A. Kuttner, D. H. Wild, B. R. Brehm, R. Riessen, G. Koveker and K. R. Karsch (1997). "Paclitaxel inhibits arterial smooth muscle cell proliferation and migration in vitro and in vivo using local drug delivery." Circulation 96(2): 636-645. Aydin, Z. and J. Akbuga (1996). "Chitosan beads for the delivery of salmon calcitonin: preparation and release characteristics." International Journal of Pharmaceutics 131(1): 101-103. Bagley, E. B., G. F. Fanta, R. C. Burr, W. M . Doane and C. R. Russell (1977). "Graft copolymers of polysaccharides with thermoplastic polymers. A new type of filled plastic." Polymer Engineering and Science 17(5): 311-316. 214 Bailey, L. L., L. Ze-jian, E. Schulz, H. Roost and P. Yahiku (1984). " A cause of right ventricular dysfunction after cardiac operations." The Journal of Thoracic and Cardiovascular Surgery 87(4): 539-542. Baillie, A. J., G. H. Coombs, T. F. Dolan, C. A. Hunter, T. Laakso, I. Sjoholm and P. Stjarnkvist (1987). "Biodegradable microspheres: polyacryl starch microparticles as a delivery system for the antileishmanial drug, sodium stibogluconate." The Journal of Pharmacy and Pharmacology 39(10): 832-835. Baker, R. W. (1987). Controlled Release of Biologically Acitve Agents. New York, John Wiley & Sons. Bateman, B. G., W. C. Nunley, Jr. and J. D. Kitchin, 3rd (1982). "Prevention of postoperative peritoneal adhesions with ibuprofen." Fertility and Sterility 38(1): 107-108. Belder, A. N. d. (1996). Medical applications of dextran and its derivatives. Polysaccharides in Medicinal Applications. S. Dumitriu. New York, M . Dekker: 505-523. Belotti, D., V. Vergani, T. Drudis, P. Borsotti, M . R. Pirelli, G. Viale, R. Giavazzi and G. Taraboletti (1996). "The microtubule-affecting drug paclitaxel has antiangiogenic activity." Clinical Cancer Research 2(11): 1843-1849. Best, C. L., D. Rittenhouse, C. Vasquez, T. Norng, E. Subias and C. E. Sueldo (1992). "Evaluation of interceed(TC7) for reduction of postoperative adhesions in rabbits." Fertility and Sterility 58(4): 817-820. Billingham, N. C. (1977). Molar Mass Measurements in Polymer Science. London, Kogan Page. 215 Bixler, G. H. , G. E. Hines, R. M . McGhee and R. A. Shurter (1953). "Dextran." Industrial and Engineering Chemistry 45: 692-705. Bodek, K. H. (2000). "Evaluation of properties microcrystalline chitosan as a drug carrier. Part 1. In vitro release of diclofenac from mictocrystalline chitosan hydrogel." Acta Poloniae Pharmaceutica 57(6): 431-440. Bodmeier, R., H. G. Chen and O. Paeratakul (1989). " A novel approach to the oral delivery of micro- or nanoparticles." Pharmaceutical Research 6(5): 413-417. Bovey, F. A. (1959). "Enzymic polymerization. I. Molecular weight and branching during the formation of dextran." Journal of Polymer Science 35: 167-182. Brode, G. L. and J. V. Koleske (1972). "Lactone polymerization and polymer properties." Journal of Macromolecular Science. Part A: Chemistry A6(61: 1109-1144. Bronson, R. A. and E. E. Wallach (1977). "Lysis of periadnexal adhesions for correction of infertility." Fertility and Sterility 28(6): 613-619. Buckman, R. F., M . Woods, L. Sargent and A. S. Gervin (1976). " A unifying pathogenetic mechanism in the etiology of intraperitoneal adhesions." The Journal of Surgical Research 20(1): 1-5. Bundgaard, H. (1985). Design of Prodrugs. New York, Elsevier. Burgess, D. J. and S. S. Davis (1988). "Potential use of albumin microspheres as a drug delivery system. II. In vivo deposition and release of steroids." International Journal of Pharmaceutics 46(1-2): 69-76. Burns, J. W., M . J. Colt, L. S. Burgees and K. C. Skinner (1997). "Preclinical evaluation of Seprafilm bioresorbable membrane." The European Journal of Surgery. Supplement 577: 40-48. 216 Burt, H. M . , J. K. Jackson, S. K. Bains, R. T. Liggins, A. M . Oktaba, A. L. Arsenault and W. L. Hunter (1995). "Controlled delivery of taxol from microspheres composed of a blend of ethylene-vinyl acetate copolymer and poly (d,l-lactic acid)." Cancer Letters 88(1): 73-79. Calanchi, M . , L. Gentilini and M . Marconi (1986). Preparing an extemporaneous homogenous microcapsule susepension. PCT International Application, Eurand Italia S.p.A., Italy.: 16 pp. Caliceti, P., F. M . Veronese and S. Lora (2000). "Polyphosphazene microspheres for insulin delivery." International Journal of Pharmaceutics 211(1-2): 57-65. Carstensen, J. T. (1977). Pharmaceutics of Solids and Solid Dosage Forms. New York, Wiley. Ceresa, R. J. (1973). Block and Graft Copolymerization. London ; New York, Wiley-Interscience. Cha, Y. and C. G. Pitt (1988). " A one-week subdermal delivery system for L-methadone based on biodegradable microcapsules." Journal of Controlled Release 7(1): 69-78. Chen, Y., J. Chen, Z. Wang, K. Wu, G. Du, D. Jing and S. Lun (1999). "Preparation of microspheres using PHBV as carrier for drug slow release." Gaoxiao Huaxue Gongcheng Xuebao 13(2): 129-134. Chien, S., F. Zhang, W. Niu, M . T. Tseng and L. Gray, Jr. (2000). "Comparison of University of Wisconsin, euro-collins, low-potassium dextran, and krebs-henseleit solutions for hypothermic lung preservation." The Journal of Thoracic and Cardiovascular Surgery 119(5): 921-930. 217 Chien, Y. W. (1982). Novel Drug Delivery Systems. New York, Marcel Dekker, Inc. Chiu, H. C , G. H. Hsiue, Y. P. Lee and L. W. Huang (1999). "Synthesis and characterization of pH-sensitive dextran hydrogels as a potential colon-specific drug delivery system." Journal of Biomaterials Science, Polymer Edition 10(5): 591-608. Choi, E.-J., C.-H. Kim and J.-K. Park (1999). "Synthesis and characterization of starch-g-polycaprolactone copolymer." Macromolecules 32: 7402-7408. Collins, E. A., J. Bares and J. Fred W. Billmeyer (1973). Experiments in Polymer Science. New York, John Wiley & Sons. Coradini, D., C. Pellizzaro, G. Miglierini, M . G. Daidone and A. Perbellini (1999). "Hyaluronic acid as drug delivery for sodium butyrate: improvement of the anti-proliferative activity on a breast-cancer cell line." International Journal of Cancer 81(3): 411-416. Cowie, J. M . G. (1991). Polymers: Chemistry and Physics of Modern Materials. New York, Chapman and Hall. Crested, T., B. Monsarrat, P. Alvinerie, J. M . Treluyer, I. Vieira and M . Wright (1994). "Taxol metabolism by human liver microsomes: identification of cytochrome P450 isozymes involved in its biotransformation." Cancer Research 54(2): 386-392. de Jong, S. J., B. van Eerdenbrugh, C. F. van Nostrum, J. J. Kettenes-van den Bosch and W. E. Hennink (2001). "Physically crosslinked dextran hydrogels by stereocomplex formation of lactic acid oligomers: degradation and protein release behavior." Journal of Controlled Release 71(3): 261-275. 218 De Leon, F. D., A. A. Toledo, J. S. Sanfilippo and M . A. Yussman (1984). "The prevention of adhesion formation by nonsteroidal antiinflammatory drugs: an animal study comparing ibuprofen and indomethacin." Fertility and Sterility 41(4): 639-642. Dellacherie, E. (1996). Polysaccharides in oxygen-carrier blood substitutes. Polysaccharides in Medicinal Applications. S. Dumitriu. New York, M . Dekker: viii, 794. DePalma, P., S. Dhanesar, S. Dordunoo, Z. Zhao, G. Troiano, H. Howland, H. Wang, A. Kader, J. Sclafani and W. Dang (2000). "Polyphosphoester paclitaxel microspheres (Paclimer microspheres): In vitro characterization using HPLC and L C / M S . " Proceedings of the International Symposium on Controlled Release of Bioactive Materials 27th: 502-503. Desbrieres, J. (2002). "Viscosity of semiflexible chitosan solutions: influence of concentration, temperature, and role of intermolecular interactions." Biomacromolecules 3(2): 342-349. Deutsch, H. M . , J. A. Glinski, M . Hernandez, R. D. Haugwitz, V. L. Narayanan, M . Suffness and L. H. Zalkow (1989). "Synthesis of congeners and prodrugs. 3. Water-soluble prodrugs of taxol with potent antitumor activity." Journal of Medicinal Chemistry 32(4): 788-792. Diamond, M . P., C. B. Linsky, T. Cunningham, L. Kamp, E. Pines, A. H. DeCherney and G. S. diZerega (1991). "Synergistic effects of INTERCEED(TC7) and heparin in reducing adhesion formation in the rabbit uterine horn model." Fertility and Sterility 55(2): 389-394. 219 Diamond, M . P. and F. Nezhat (1993). "Adhesions after resection of ovarian endometriomas." Fertility and Sterility 59(4): 934-936. Diamond, M . P., D. M . Wiseman and C. Linsky (1994). "Interceed (TC7) absorbable adhesion barrier." Infertility and Reproductive Medicine Clinics of North America 5(3): 485-508. Donabedian, D. H. and S. P. McCarthy (1998). "Acylation of Pullulan by Ring-Opening of Lactones." Macromolecules 31(4): 1032-1039. Doody, K., R. Dunn and V. J. Buttram (1989). "Recombinant tissue plasminogen activator reduces adhesion formation in a rabbit uterine horn model." Fertility and Sterility 51(31: 509-512. Dordunoo, S. K. and H. M . Burt (1996). "Solubility and stability of taxol: effects of buffers and cyclodextrins." International Journal of Pharmaceutics 133: 191-201. Dorr, R. T. (1994). "Pharmacology and toxicology of Cremophor EL diluent." The Annals of Pharmacotherapy 28: SI 1-14. Dosio, F., S. Arpicco, P. Brusa, B. Stella and L. Cattel (2001). "Polyethylene glycol)-human serum albumin-paclitaxel conjugates: preparation, characterization and pharmacokinetics." Journal of Controlled Release 76(1-2): 107-117. Dreyfuss, P. and R. P. Quirk (1987). Graft copolymers. Encyclopedia of Polymer Science and Engineering. H. F. Mark and J. I. Kroschwitz. New York, Wiley. 7: 551-579. Du, Y. J., P. J. Lemstra, A. J. Nijenhuis, H. A. M . v. Aert and C. Bastiaansen (1995). " A B A type copolymers of lactide with poly(ethylene glycol). Kinetic, mechanistic and model studies." Macromolecules 28: 2124-2132. 220 Eastmond, G. C., A. Ledwith, S. Russo and P. Sigwalt, Eds. (1989). Polymer Reactions. Comprehensive Polymer Science: The Synthesis, Characterization, Reactions & Applications of polymers. Oxford ; New York, Pergamon Press. Edman, P., B. Ekman and I. Sjoholm (1980). "Immobilization of proteins in microspheres of biodegradable polyacryldextran." Journal of Pharmaceutical Sciences 69(7): 838-842. Ellis, H. (1971). "The causes and prevention of postoperative intraperitoneal adhesions." Surgery, Gynecology and Obstetrics: an international magazine 133: 497-510. Ellis, H. (1982). "The cause and prevention of intestinal adhesions." The British Journal of Surgery 69: 241-243. Ellis, H. , W. Harrison and T. B. Hugh (1965). "The healing of peritoneum under normal and pathological conditions." The British Journal of Surgery 52: 471-476. Entwistle, C. A. and R. C. Rowe (1979). "Plasticization of cellulose ethers used in the film coating of tablets." Journal of Pharmacy and Pharmacology 31(5): 269-272. Erbacher, P., S. Zou, T. Bettinger, A. M . Steffan and J. S. Remy (1998). "Chitosan-based vector/DNA complexes for gene delivery: biophysical characteristics and transfection ability." Pharmaceutical Research 15(9): 1332-1339. Folkman, J. and M . Klagsbrun (1987). "Angiogenic factors." Science 235(4787): 442-447. Forastiere, A. A., D. Neuberg, S. G. t. Taylor, R. DeConti and G. Adams (1993). "Phase II evaluation of Taxol in advanced head and neck cancer: an Eastern Cooperative Oncology group trial." Journal of National Cancer Institute, Monographs 15: 181-184. 221 Fortin, S. and G. Charlet (1989). "Phase diagram of aqueous solutions of (hydroxypropyl)cellulose." Macromolecules 22(5): 2286-2292. Friebolin, H. , G. Keilich, N. Frank, U. Dabrowski and E. Siefert (1979). "Structure investigations on polysaccharides. Part 6. Carbon-13 N M R investigations on the structure elucidation of polysaccharides and their methyl derivatives." Organic Magnetic Resonance 12(4): 216-222. Fu, J. and C. Wu (2001). "Laser light scattering study of the degradation of poly(sebacic anhydride) nanoparticles." Journal of Polymer Science, Part B: Polymer Physics 39(6): 703-708. Fujimoto, S., M . Miyazaki, F. Endoh, O. Takahashi, K. Okui and Y. Morimoto (1985). "Biodegradable mitomycin C microspheres given intra-arterially for inoperable hepatic cancer. With particular reference to a comparison with continuous infusion of mitomycin C and 5-fluorouracil." Cancer 56(10): 2404-2410. Fukasawa, M . , W. Girgis and G. S. diZerega (1991). "Inhibition of postsurgical adhesions in a standardized rabbit model: II. Intraperitoneal treatment with heparin." International Journal of Fertility 36(51: 296-301. Fukui, E., K. Uemura and M . Kobayashi (2000). "Studies on applicability of press-coated tablets using hydroxypropylcellulose (HPC) in the outer shell for timed-release preparations." Journal of Controlled Release 68(2): 215-223. Genta, I., M . Costantini, A. Asti, B. Conti and L. Montanari (1998). "Influence of glutaraldehyde on drug release and mucoadhesive properties of chitosan microspheres." Carbohydrate Polymers 36(2/31: 81-88. 222 Gloushankova, N. A., A. V. Lyubimova, I. S. Tint, H. H. Feder, J. M . Vasiliev and I. M . Gelfand (1994). "Role of the microtubular system in morphological organization of normal and oncogene-transfected epithelial cells." Proceedings of the National Academy of Sciences of the United States of America 91(18): 8597-8601. Gomel, V., B. Urman and T. Gurgan (1996). "Pathophysiology of adhesion formation and strategies for prevention." Journal of Reproductive Medicine 41(1): 35-41. Goosen, M . F., Y. F. Leung, S. Chou and A. M . Sun (1982). "Insulin-albumin microbeads: an implantable, biodegradable system." Biomaterial Medical Devices and Artificial Organs 10(3): 205-218. Grem, J. L., K. D. Tutsch, K. J. Simon, D. B. Alberti, J. K. Willson, D. C. Tormey, S. Swaminathan and D. L. Trump (1987). "Phase I study of taxol administered as a short i.v. infusion daily for 5 days." Cancer Treatment Reports 71(12): 1179-1184. group, T. s. m. s. (1992). "Prophylaxis of pelvic sidewall adhesions with Gore-Tex surgical membrane: a multicenter clinical investigation." Fertility and Sterility 57(4): 921-923. Gruber, U. F. (1976). "Pharmacokinetics of dextran." Anesthesie, Analgesie, Reanimation 33(4): 505-508. Gumusderelioglu, M . and G. Deniz (2000). "Sustained release of mitomycin C from poly(DL-lactide)/poly(DL-lactide-co-glycolide) films." Journal of Biomaterials Science. Polymer Edition 11(10): 1039-1050. 223 Gupta, K. C. and M . Kumar (2001). "Studies on semi-interpenetrating polymer network beads of chitosan-poly(ethylene glycol) for the controlled release of drugs." Journal of Applied Polymer Science 80(4): 639-649. Haney, A. F. and E. Doty (1992). "Murine peritoneal injury and de novo adhesion formation caused by oxidized-regenerated cellulose (Interceed [TC7]) but not expanded polytetrafluoroethylene (Gore-Tex Surgical Membrane)." Fertility and Sterility 57(1): 202-208. Haney, A. F. and E. Doty (1994). "The formation of coalescing peritoneal adhesions requires injury to both contacting peritoneal surfaces." Fertility and Sterility 61(4): 767-775. Harcum, W. W., G. W. Skinner, M . Altekar, S. K. Joneja, P. E. Barnum and J. H. Guo (1998). "Modeling the effects of hydroxypropylcellulose in acetaminophen tablet formulation." Drug Development and Industrial Pharmacy 24(10): 911-918. Higuchi, T. (1961). "Rate of release of medicaments from ointment bases containing drugs in suspension." Journal of Pharmaceutical Sciences 50: 874-875. Ho, F. F.-L., R. R. Kohler and G. A. Ward (1972). "Determination of Molar Substitution and Degree of Substitution of Hydroxypropyl Cellulose by Nuclear Magnetic Resonance Spectrometry." Analytical Chemistry 44(1): 178-181. Holmes, F. A., R. S. Walters, R. L. Theriault, A. D. Forman, L. K. Newton, M . N. Raber, A. U. Buzdar, D. K. Frye and G. N. Hortobagyi (1991). "Phase II trial of taxol, an active drug in the treatment of metastatic breast cancer." Journal of National Cancer Institute 83(24): 1797-1805. 224 Hon, D. N. S. (1996). Chitin and chitosan: Medical applications. Polysaccharides in Medicinal Applications. S. Dumitriu. New York, Marcel Dekker, Inc.: 631-649. Horbett, T. A. (1986). Protein Adsorption to Hydrogels. Hydrogels in Medicine and Pharmacy. N. A. Peppas. Boca Raton, CRC Press. 1. Huang, C. H. O., D. G. I. Kingston, N. F. Magri, G. Samaranayake and F. E. Boettner (1986). "New Taxanes from Taxus Brevifolia." Journal of Natural Products 49(4): 665-669. Huizing, M . T., A. C. Keung, H. Rosing, V. van der Kuij, W. W. ten Bokkel Huinink, I. M . Mandjes, A. C. Dubbelman, H. M . Pinedo and J. H. Beijnen (1993). "Pharmacokinetics of paclitaxel and metabolites in a randomized comparative study in platinum-pretreated ovarian cancer patients." Journal of Clinical Oncology 11(11): 2127-2135. ICH, T. (2002). "ICH Topic Q3C (M): Maintenance of Note for Guidance on Impurities: Residual Solvents." CPMP/ICH/283/95: 1/18-18/18. Ilium, L. (1998). "Chitosan and its use as a pharmaceutical excipient." Pharmaceutical Research 15(9): 1326-1331. Irifune, M . , S. Ogino, T. Harada and Y. Abe (1999). "Topical treatment of nasal polyps with a beclomethasone dipropionate powder preparation." Auris. Nasus. Larynx 26(1): 49-55. Ishikawa, T., B. Mukai, S. Shiraishi, N. Utoguchi, M . Fujii, M . Matsumoto and Y. Watanabe (2001). "Preparation of rapidly disintegrating tablet using new types of microcrystalline cellulose (PH-M series) and low substituted-225 hydroxypropylcellulose or spherical sugar granules by direct compression method." Chemical & Pharmaceutical Bulletin 49(2): 134-139. Jackson, J., K. C. Skinner, L. Burgess, T. Sun, W. L. Hunter and H. M . Burt (2002). "Paclitaxel-loaded crosslinked hyaluronic acid films for the prevention of postsurgical adhesions." Pharmaceutical Research 19(4): 411-417. Jacquier, V., C. Miola, M.-F. Llauro, C. Monnet and T. Hamaide (1996). "Functionalized poly(e-caprolactone) and copolymers with ethylene oxide through heterogeneous anionic coordinated polymerization. NMR characterization and crystallinity." Macromolecular Chemistry and Physics 197: 1311-1324. Jameela, S. R. and A. Jayakrishnan (1995). "Glutaraldehyde cross-linked chitosan microspheres as a long acting biodegradable drug delivery vehicle: studies on the in vitro release of mitoxantrone and in vivo degradation of microspheres in rat muscle." Biomaterials 16(10): 769-775. Jamieson, N. V., S. Lindell, R. Sundberg, J. H. Southard and F. O. Belzer (1989). "Evaluation of simplified variants of the UW solution using the isolated perfused rabbit liver." Transplant Procedure 21(1 Pt 2): 1294-1295. Jamis-Dow, C. A., R. W. Klecker, G. Sarosy, E. Reed and J. M . Collins (1993). "Steady-state plasma concentrations and effects of taxol for a 250 mg/m2 dose in combination with granulocyte-colony stimulating factor in patients with ovarian cancer." Cancer Chemotherapy and Pharmacology 33(1): 48-52. Janca, J. (1984). Steric Exclusion Liquid Chromatography of Polymers. New York, Dekker. 226 Janes, K. A., P. Calvo and M . J. Alonso (2001). "Polysaccharide colloidal particles as delivery systems for macromolecules." Advanced Drug Delivery Reviews 47(1): 83-97. Jansen, R. P. (1988). "Failure of peritoneal irrigation with heparin during pelvic operations upon young women to reduce adhesions." Surgery. Gynecology and Obstetrics 166(2): 154-160. Jarrett, J. C , 2nd and M . Y. Dawood (1986). "Adhesion formation and uterine tube healing in the rabbit: a controlled study of the effect of ibuprofen and flurbiprofen." American Journal of Obstetrics and Gynecology 155(6): 1186-1192. Jhon, M . S. and J. D. Andrade (1973). "Water and hydrogels." Journal of Biomedical Materials Research 7(6): 509-522. Kassab, A. C , K. Xu, E. B. Denkbas, Y. Dou, S. Zhao and E. Piskin (1997). "Rifampicin carrying polyhydroxybutyrate microspheres as a potential chemoembolization agent." Journal of Biomaterials Science. Polymer Edition 8(12): 947-961. Kataoka, K., G. S. Kwon, M . Yokoyama, T. Okana and Y. Sakurai (1993). "Block copolymer micelles as vehicles for drug delivery." Journal of Controlled Release 24: 119-132. Kawase, M . , N. Michibayashi, Y. Nakashima, N. Kurikawa, K. Yagi and T. Mizoguchi (1997). "Application of glutaraldehyde-crosslinked chitosan as a scaffold for hepatocyte attachment." Biological and Pharmaceutical Bulletin 20(6): 708-710. Kibbe, A. H. (2000). Handbook of pharmaceutical excipients. Washington, DC, American Pharmaceutical Association. 227 Kind, F. A., L. A. Ciaccio and S. B. Henderson (1984). "Sustained release preparations, XVI: Collagen as a drug carrier." Archiv der Pharmazie 317(8): 657-661. Kingston, D. G., G. Samaranayake and C. A. Ivey (1990). "The chemistry of taxol, a clinically useful anticancer agent." Journal of Natural Products 53: 1-12. Kingston, D. G. I. (2000). "Recent advances in the chemistry of Taxol." Journal of Natural Products 63: 726-734. Kirkinen, P. (1988). "Multiple caesarean sections: outcomes and complications." British Journal of Obstetrics and Gynaecology 95(8): 778-782. Klug, E. D. (1971). "Hydroxypropyl cellulose." Encyclopedia of Polymer Science and Technology 15: 307-314. Knightly, J. J., D. Agostino and E. E. Cliffton (1962). "The effect of fibrinolysin and heparin on the formation of peritoneal adhesions." Surgery 52: 250-258. Kojima, T., M . Hashida, S. Muranishi and H. Sezaki (1980). "Mitomycin C-dextran conjugate: a novel high molecular weight pro-dur of mitomycin C. " Journal of Pharmacy and Pharmacology 32(1): 30-34. Kowalski, A., A. Duda and S. Penczek (2000). "Kinetics and Mechanism of Cyclic Esters Polymerization Initiated with Tin(II) Octoate. 3. Polymerization of L,L-Dilactide." Macromolecules 33: 7359-7370. Kowalski, A., A. Duda and S. Penczek (2000). "Mechanism of Cyclic Ester Polymerization Initiated with Tin(II) Octoate. 2. Macromolecules Fitted with Tin(II) Alkoxide Species Observed Directly in MALDI-TOF Spectra." Macromolecules 33(3): 689-695. Krevelen, D. W. v. (1976). Properties of Polymers. Amsterdam, New York, Elsevier. 228 Kricheldorf, H. R. and I. Kreiser (1987). "Polylactones. 13. Transesterification of poly(L-lactide) with poly(glycolide), poly(P-propiolactone), and poly(s-caprolactone)." Journal of Macromolecular Science. Part A: Chemistry A24(ll): 1345-1356. Kricheldorf, H. R., I. Kreiser-Saunders and C. Boettcher (1995). "Polylactones: 31. Sn(II)octoate-initiated polymerization of L-lactide: a mechanistic study." Polymer 36(6): 1253-1259. Kricheldorf, H. R., I. Kreiser-Saunders and A. Strieker (2000). "Polylactones 48. SnOct2-Initiated Polymerizations of Lactide: A Mechanistic Study." Macromolecules 33(3): 702-709. Kris, M . G., J. P. O'Connell, R. J. Gralla, M . S. Wertheim, R. M . Parente, P. B. Schiff and C. W. Young (1986). "Phase I trial of taxol given as a 3-hour infusion every 21 days." Cancer Treatment Reports 70(5): 605-607. Kumar, G. N. , U. K. Walle and T. Walle (1994). "Cytochrome P450 3A-mediated human liver microsomal taxol 6 alpha- hydroxylation." The Journal of Pharmacology and Experimental Therapeutics 268(3): 1160-1165. Kumar, V. and G. S. Banker (1996). Target-oriented drug delivery systems. Modern Pharmaceutics. G. S. Banker and C. T. Rhodes. New York, M . Dekker: 662-663. Kwon, G. S., M . Naito, M . Yokoyama, T. Okano, Y. Sakurai and K. Kataoka (1993). "Micelles based on AB block copolymers of polymer(ethylene oxide) and poly(beta-benzyl L-aspartate)." Langmuir 9: 945-949. Larsen, C. (1989). "Dextran prodrugs - structure and stability in relation to therapeutic activity." Advanced Drug Delivery Reviews 3(1): 103-154. 229 Leach, R. E. and R. L. Henry (1990). "Reduction of postoperative adhesions in the rat uterine horn model with poloxamer 407." American Journal of Obstetrics and Gynecology 162(5): 1317-1319. Lee, D.-S. and A. S. Perlin (1982). "13C-NMR spectral and related studies on the distribution of substituents in 0-(2-Hydroxypropyl)cellulose." Carbohydrate Research 106: 1-19. Lee, P. I. (1985). Kinetics of drug release from hydrogel matrices. The Second International Symposium on Recent Advances in Drug Delivery Systems, Salt Lake City, UT, USA, Elsevier. LeGrand, E. K., K. E. Rodgers, W. Girgis, K. Struck, J. D. Campeau, G. S. diZerega and T. C. Kiorpes (1994). "Efficacy of tolmetin sodium for adhesion prevention in rabbit and rat models." The Journal of Surgical Research 56(1): 67-71. Leopold, C. S. and D. R. Friend (1995). "In vivo pharmacokinetic study for the assessment of poly(L-aspartic acid) as a drug carrier for colon-specific drug delivery." Journal of Pharmacokinetics and Biopharmaceutics 23(4): 397-406. Levy, M . C. (1994). Transacylation reaction between an esterified polysaccharide and a polyaminated or polyhydroxylated substance for fabrication of microparticles. PCT International Application, (Coletica, Fr.).: 43 pp. Li , C , D. Yu, T. Inoue, D. J. Yang, L. Milas, N. R. Hunter, E. E. Kim and S. Wallace (1996). "Synthesis and evaluation of water-soluble polyethylene glycol- paclitaxel conjugate as a paclitaxel prodrug." Anticancer Drugs 7(6): 642-648. 230 Li , J., W. Xie, H. N. Cheng, R. G. Nickol and P. G. Wang (1999). "Polycaprolactone-modified hydroxyethylcellulose films prepared by lipase-catalyzed ring-opening polymerization." Macromolecules 32: 2789-2792. Li, Q., E. T. Dunn, E. W. Grandmaison and M . F. A. Goosen (1992). "Applications and properties of chitosan." Journal of Bioactive and Compatible Polymers 7: 370-397. Li, Y. and T. Kissel (1993). "Synthesis and properties of biodegradable A B A triblock copolymers consisiting of poly(L-lactic acid) or poly(L-lactic-co-glycolic acid) A-blocks attached to central poly(oxyethylene) B-blocks." Journal of Controlled Release 27: 247-157. Liggins, R., W. L. Hunter and H. M . Burt (1997). "Solid-state characterization of paclitaxel." Journal of Pharmaceutical Sciences 86(12): 1458-1463. Liggins, R. T. and H. M . Burt (2001). "Paclitaxel loaded poly(L-lactic acid) microspheres: properties of microspheres made with low molecular weight polymers." International Journal of Pharmaceutics 222(1): 19-33. Liggins, R. T., S. D'Amours, J. S. Demetrick, L. S. Machan and H. M . Burt (2000). "Paclitaxel loaded poly(L-lactic acid) microspheres for the prevention of intraperitoneal carcinomatosis after a surgical repair and tumor cell spill." Biomaterials 21(19): 1959-1969. Ljungstrom, K. G. (1988). "The antithrombotic efficacy of dextran." Acta Chirurgica Scandinavica 543: 26-30. 231 Lovell, P. A. (1989). Dilute solution viscometry. Comprehensive Polymer Science: The Synthesis, Characterization, Reactions & Application of Polymers. C. Booth and C. Price. Oxford, Pergamon Press. 2: 173-197. Luo, Y., K. R. Kirker and G. D. Prestwich (2000). "Cross-linked hyaluronic acid hydrogel films: new biomaterials for drug delivery." Journal of Controlled Release 69(1): 169-184. MaacEachem-Keith, G. J., L. J. W. Butterfield and M . J. I. Mattina (1997). "Paclitaxel Stability in Solution." Analytical Chemistry 69: 72-77. Markland, P., G. L. Amidon and V. C. Yang (1999). "Modified polypeptides containing gamma-benzyl glutamic acid as drug delivery platforms." International Journal of Pharmaceutics 178(2): 183-192. Markman, M . , E. Rowinsky, T. Hakes, B. Reichman, W. Jones, J. L. Lewis, Jr., S. Rubin, J. Curtin, R. Barakat and M . Phillips (1992). "Phase I trial of intraperitoneal taxol: a Gynecoloic Oncology Group study." Journal of Clinical Oncology 10(9): 1485-1491. Mathew, A. E., M . R. Mejillano, J. P. Nath, R. H. Himes and V. J. Stella (1992). "Synthesis and evaluation of some water-soluble prodrugs and derivatives of taxol with antitumor activity." Journal of Medicinal Chemistry 35(1): 145-151. Matsuoka, H. , M . Furusawa, H. Tomoda and Y. Seo (1994). "Difference in cytotoxicity of paclitaxel against neoplastic and normal cells." Anticancer Research 14(1 A): 163-167. 232 Mcentee, G., D. Pender, D. Mulvin, M . McCullough, S. Naeeder, S. Farah, M . S. Badurdeen, V. Ferraro, C. Cham, N. Gillham and P. Matthews (1987). "Current spectrum of intestinal obstruction." The British Journal of Surgery 74: 976-983: McGuire, W. P., E. K. Rowinsky, N. B. Rosenshein, F. C. Brumbine, D. S. Ettinger, D. K. Armstrong and R. C. Conehower (1989). "Taxol: a unique antineoplastic agent with significant activity in advanced ovarian epithelial neoplasms." Annals of Internal Medicine 111: 273-279. Meddings, N. , R. Scott, R. Bullock, D. A. French, T. A. Hide and S. D. Gorham (1992). "Collagen vicryl~a new dural prosthesis." Acta Neurochirurgica 117(1-2): 53-58. Miller, D. W., E. V. Batrakova, T. O. Waltner, V. Y. Alakhov and A. V. Kabanov (1997). "Interactions of Pluronic Block Copolymers with Brain Microvessel Endothelial Cells: Evidence of Two Potential Pathways for Drug Absorption." Bioconjugate Chemistry. 8(5): 649-657. Mohr, D., M . Wolff and T. Kissel (1999). "Gamma irradiation for terminal sterilization of 17beta-estradiol loaded poly-(D,L-lactide-co-glycolide) microparticles." Journal of Controlled Release 61(1-2): 203-217. Montanari, L., F. Cilurzo, L. Valvo, A. Faucitano, A. Buttafava, A. Groppo, I. Genta and B. Conti (2001). "Gamma irradiation effects on stability of poly(lactide-co-glycolide) microspheres containing clonazepam." Journal of Controlled Release 75(3): 317-330. Montz, F. J., B. J. Monk and S. M . Lacy (1993). "Effectiveness of two barriers at inhibiting post-radical pelvic surgery adhesions." Gynecologic Oncology 48(2): 247-251. 233 Mucha, M . , J. Piekielna and A. Wieczorek (1999). "Characterisation and morphology of biodegradable chitosan/synthetic polymer blends." Macromolecular Symposia 144:391-412. Murphy, W. K., F. V. Fossella, R. J. Winn, D. M . Shin, H. E. Hynes, H. M . Gross, E. Davilla, J. Leimert, H. Dhingra and M . N. Raber (1993). "Phase II study of taxol in patients with untreated advanced non-small- cell lung cancer." Journal of National Cancer Institute 85(5): 384-388. Nagai, T., Y. Sawayanagi and N. Nambu (1984). Application of chitin and chitosan to pharmaceutical preparations. Chitin, Chitosan and Related Enzymes. J. P. Zikakis. Orlando, Academic Press: 21-40. Nagaoka, S. and R. Akashi (1990). "Low-friction hydrophilic surface for medical devices." Biomaterials 11(6): 419-424. Nishi, T. and T. T. Wang (1975). "Melting point depression and kinetic effects of cooling on crystallization in poly(vinylidene fluoride)-poly(methyl methacrylate) mixtures." Macromolecules 8(6): 909-915. Nishioka, Y., S. Kyotani, M . Okamura, M . Miyazaki, K. Okazaki, S. Ohnishi, Y. Yamamoto and K. Ito (1990). "Release characteristics of cisplatin chitosan microspheres and effect of containing chitin." Chemical & Pharmaceutical Bulletin 38(101: 2871-2873. Noshay, A. and J. E. McGrath (1977). Block Copolymers. Overview and Critical Survey. New York, Academic Press. Odian, G. (1991). Principles of Polymerization. New York, John Wiley & Sons, Inc. 234 Ohya, Y., S. Maruhashi and T. Ouchi (1998). "Graft polymerization of L-lactide on pullulan through the trimethylsilyl protection method." Macromolecules 31(14): 4662-4665. Okamoto, Y., A. Konno, K. Togawa, T. Kato, Y. Tamakawa and Y. Amano (1986). "Arterial chemoembolization with cisplatin microcapsules." British Journal of Cancer 53(3): 369-375. Ong, C. J. and F. P. Price (1978). "Blends of Poly(e-caprolactone) with Poly(vinyl chloride). I. Morphology." Journal of Polymer Science: Polymer Symposium 63: 45-58. O'Reilly, S. (1997). The market for adhesion preventing materials. Postsurgical Adhesions: Linking Business and Science to Further the Development of Adhesion Prevention Devices, Hyatt Regency Baltimore, Baltimore, M D . Orita, H. , M . Fukasawa, W. Girgis and G. S. diZerega (1991). "Inhibition of postsurgical adhesions in a standardized rabbit model: intraperitoneal treatment with tissue plasminogen activator." International Journal of Fertility 36(3): 172-177. Pados, G. A. and P. Devroey (1992). "Adhesions." Current Opinion in Obstetrics & Gynecology 4(31:412-418. Pagidas, K. and T. Tulandi (1992). "Effects of Ringer's lactate, Interceed(TC7) and Gore-Tex Surgical Membrane on postsurgical adhesion formation." Fertility and Sterility 57(11: 199-201. Panchagnula, R. (1998). "Pharmaceutical aspects of paclitaxel." International Journal of Pharmaceutics 172(1-2): 1-15. 235 Park, H. and K. Park (1996). "Biocompatibility issues of implantable drug delivery systems." Pharmaceutical Research 13(12): 1770-1776. Park, J. B. (1984). Biomaterials Science and Engineering. New York, Plenum Press. Park, K., W. S. E. Shalaby and H. Park (1993). Biodegradable Hydrogels for Drug Delivery. Lancaster, Technomic Publishing Co., Inc. Parker, J. E., S. L. Fibini, B. D. Car and H. N. Erb (1987). "Prevention of intraabdominal adhesions in ponies by low-dose heparin therapy." Veterinary Surgery 16(6): 459-462. Paul, D. R. and S. Newman (1978). Polymer blends. New York, Academic Press. Paul, W. and C. P. Sharma (2000). "Chitosan, a drug carrier for the 21st century: a review." S.T.P. Pharma : Sciences 10(1): 5-22. Pellicer, A. and M . P. Diamond (1988). "Distending media for hysteroscopy." Obstetrics and Gynecology Clinics of North America 15(1): 23-28. Pheasant, H. C. (1975). "Sources of Failure in Laminectomies." Orthopedic Clinics of North America 6(1): 319-329. Piskin, E. (1994). "Biodegradable polymers as biomaterials." Journal of Biomaterials Science. Polymer Edition 6(9): 775-795. Pitt, C. G. (1990). Poly(s-caprolactone) and its copolymers. Biodegradable Polymers as Drug Delivery Systems. M . Chasin and R. Langer. New York, Marcel Dekker Inc.: 71-120. Pitt, C. G., M . M . Gratzl, A. R. Jeffcoat, R. Zweidinger and A. Schindler (1979). "Sustained drug delivery systems II: Factors affecting release rates from poly(s-236 caprolactone) and related biodegradable polyesters." Journal of Pharmaceutical Sciences 68(12): 1534-1538. Pitt, C. G., A. R. Jeffcoat, R. A. Zweidinger and A. Schindler (1979). "Sustained drug delivery systems. I. The permeability of poly(epsilon-caprolactone), poly(DL-lactic acid), and their copolymers." Journal of Biomedical Materials Research 13(3): 497-507. Pitt, C. G., T. A. Marks and A. Schindler (1980). Biodegradable drug delivery systems based on aliphatic polyesters: application to contraceptives and narcotic antagonists. Controlled Release of Bioactive Materials. R. Baker. New York, Academic Press: 19-43. Pitt, C. G. and A. Schindler (1980). The design of controlled drug delivery systems based on biodegradable polymers. Biodegradable Delivery Systems for Contraception. W. A. A. v. Os and E. S. E. Hafez. Lancaster, MTP Press: xi, 212. Pratesi, G., G. Savi, G. Pezzoni, O. Bellini, S. Penco, S. Tinelli and F. Zunino (1985). "Poly-L-aspartic acid as a carrier for doxorubicin: a comparative in vivo study of free and polymer-bound drug." British Journal of Cancer 52(6): 841-848. Railkar, A. M . and J. B. Schwartz (2001). "The effects of formulation factors on the moist granulation technique for controlled-release tablets." Drug Development and Industrial Pharmacy 27(9): 893-898. Ranade, V. V. and M . A. Hollinger (1996). Drug Delivery Systems. Boca Raton, CRC Press. 237 Rao, M . V. R. M . and M . Yaseen (1986). "Determination of intrinsic viscosity by single specific viscosity measurement." Journal of Applied Polymer Science 31(8): 2501-2508. Rao, S., N. E. Krauss, J. M . Heerding, C. S. Swindell, I. Ringel, G. A. Orr and S. B. Horwitz (1994). "3'-(p-azidobenzamido)taxol photolabels the N-terminal 31 amino acids of beta-tubulin." The Journal of Biological Chemistry 269(5): 3132-3134. Ratner, B. D. (1986). Hydrogel surfaces. Hydrogels in Medicine and Pharmacy. N. A. Peppas. Boca Raton, CRC Press. 1. Reichman, B. S., A. D. Seidman, J. P. Crown, R. Heelan, T. B. Hakes, D. E. Lebwohl, T. A! Gilewski, A. Surbone, V. Currie, C. A. Hudis and et al. (1993). "Paclitaxel and recombinant human granulocyte colony-stimulating factor as initial chemotherapy for metastatic breast cancer." Journal of Clinical Oncology 11(10): 1943-1951. Rice, V. M . , A. Shanti, K. S. Moghissi and R. E. Leach (1993). " A comparative evaluation of Poloxamer 407 and oxidized regenerated cellulose (Interceed [TC7]) to reduce postoperative adhesion formation in the rat uterine horn model." Fertility and Sterility 59(4): 901-906. Richheimer, S. L., D. M . Tinnermeier and D. W. Timmons (1992). "High-performance Liquid Chromatographic Assay of Taxol." Analytical Chemistry 64: 2323-2326. Rim, P. B. and J. P. Runt (1984). "Melting point depression in crystalline/compatible polymer blends." Macromolecules 17(8): 1520-1526. 238 Ringel, C , H. Niell, A. Miller and C. Cox (1994). "Taxol and taxotere in bladder cancer: in vitro activity and urine stability." Cancer Chemotherapeutics and Pharmacology 33: 460-464. Ringel, I. and S. B. Horwitz (1987). "Taxol is converted to 7-epitaxol, a biologically active isomer, in cell culture medium." The Journal of Pharmacology and Experimental Therapeutics 242(2): 692-698. Risberg, B. (1997). "Adhesions: preventive strategies." European Journal of Surgery 163(Suppl. 557): 32-39. Rivkind, A. I., N. Lieberman and A. L. Durst (1985). "Urokinase does not prevent abdominal adhesion formation in rats." European Journal of Surgery 17(4): 254-258. Robitaille, L., N. Turcotte, S. Fortin and G. Charlet (1991). "Calorimetric study of aqueous solutions of hydroxypropylcellulose." Macromolecules 24: 2413-2418. Robyt, J. F. (1985). Dextran. Encyclopedia of Polymer Science and Engineering. J. I. Kroschwitz. New York, Wiley. 4: 17 v. Rodgers, K., W. Girgis, G. diZerega, K. Bracken and L. Richer (1990). "Inhibition of postsurgical adhesions by liposomes containing nonsteroidal antiinflammatory drugs." International Journal of Fertility 35(5): 315-320. Rodgers, K., W. Girgis, G. diZerega and D. Johns (1990). "Intraperitoneal tolmetin prevents postsurgical adhesion formation in rabbits." International Journal of Fertility 35(1): 40-45. 239 Rodgers, K. E., D. D. Ellefson, W. Girgis and G. S. diZerega (1994). "Modulation of postsurgical cell infiltration and fibrinolytic activity by tolmetin in two species." The Journal of Surgical Research(4): 314-325. Rosenberg, S. M . (1990). "Dextran 70~encouraging early clinical studies." Progress in Clinical and Biological Research 358: 157-163. Roskos, K. V. , B. K. Fritzinger, S. S. Rao, G. C. Armitage and J. Heller (1995). "Development of a drug delivery system for the treatment of periodontal disease based on bioerodible poly(ortho esters)." Biomaterials 16(4): 313-317. Rowinsky, E. K., L. A. Cazenaue and R. C. Donehower (1990). "Taxol: a novel investigational antimicrotubule agent." Journal of the National Cancer Institute 82: 1247-1259. Rowinsky, E. K., V. Chaudhry, D. R. Cornblath and R. C. Donehower (1993). "Neurotoxicity of Taxol." Journal of National Cancer Institute, Monographs 15: 107-115. Rowinsky, E. K. and R. C. Donehower (1995). "Paclitaxel (Taxol)." The New England Journal of Medicine 332(15): 1004-1014. Rowinsky, E. K., E. A. Eisenhauer, V. Chaudhry, S. G. Arbuck and R. C. Donehower (1993). "Clinical toxicities encountered with paclitaxel (Taxol)." Seminars in Oncology 20(4 Suppl 3): 1-15. Rowinsky, E. K., W. P. McGuire, T. Guarnieri, J. S. Fisherman, M . C. Christian and R. C. Donehower (1991). "Cardiac disturbances during the administration of taxol." Journal of Clinical Oncology 9(9): 1704-1712. 240 Rowland, G. F. (1977). "Effective antitumour conjugates of alkylating drug arid antibody using dextran as the intermediate carrier." European Journal of Cancer 13(6): 593-596. Ryner, M . , K. Stridsberg and A. -C. Albertsson (2001). "Mechanism of ring-opening polymerization of l,5-dioxepan-2-one and L-lactide with stannous 2-ethylhexanoate. A theoretical study." Macromolecules 34: 3877-3881. Sabnis, S. and L. H. Block (2000). "Chitosan as an enabling excipient for drug delivery systems. I. Molecular modifications." International Journal of Biological Macromolecules 27(3): 181-186. Saettone, M . F., B. Giannaccini, P. Chetoni, G. Galli and E. Chiellini (1984). "Vehicle effects in ophthalmic bioavailability: an evaluation of polymeric inserts containing pilocarpine." Journal of Pharmacy and Pharmacology 36(4): 229-234. Sakharov, D. V., A. F. Jie, M . E. Bekkers, J. J. Emeis and D. C. Rijken (2001). "Polylysine as a vehicle for extracellular matrix-targeted local drug delivery, providing high accumulation and long-term retention within the vascular wall." Arteriosclerosis. Thrombosis, and Vascular Biology 21(6): 943-948. Scandola, M . , G. Ceccorulli and M . Pizzoli (1991). "Molecular motions of polysaccharides in the solid state: dextran, pullulan and amylose." International Journal of Biological Macromolecules 13(4): 254-260. Schacht, E., R. Vercauteren and S. Vansteenkiste (1988). "Some aspects of the application of dextran in prodrug design." Journal of Bioactive and Compatible Polymers 3(1): 72-80. 241 Schiff, P. B. and S. B. Horwitz (1980). "Taxol stabilizes microtubules in mouse fibroblast cells." Proceedings of the National Academy of Sciences of the United States of America 77(3): 1561-1565. Schindler, A. , Y. M . Hibionada and C. G. Pitt (1982). "Aliphatic polyesters. III. Molecular weight and molecular weight distribution in alcohol-initiated polymerization of s-caprolactone." Journal of Polymer Science: Part A: Polymer Chemistry 20: 319-326. Schindler, A., R. Jeffcoat, G. L. Kimmel, C. G. Pitt, M . E. Wall and R. Zweidinger (1977). "Biodegradable polymers for sustained drug delivery." Contemporary Topics in Polymer Science 2: 251-289. Schlumpf, R., P. Morel, J. J. Loveras, R. M . Condie, A. Matas, J. Kurle, C. G. Fasola, J. S. Najarian and D. E. Sutherland (1991). "Dextran 40 successfully replaces the non-essential hydroxyethylstarch in the University of Wisconsin solution for 72-hour simple cold storage of the canine kidney." Transplant Procedure 23(1 Pt 1): 657-659. Schwach, G., H. Coudane, R. Engel and M . Vert (1997). "More about the polymerization of lactides in the presence of stannous octoate." Journal of Polymer Science: Part A: Polymer Chemistry 35: 3431-3440. Seifer, D. B., M . P. Diamond and A. H. DeCherney (1990). " A n appraisal of barrier agents in the reduction of adhesion formation following surgery." Journal of Gynecologic Surgery 6(1): 3-10. Sekiba, K. (1992). "Use of Interceed(TC7) absorbable adhesion barrier to reduce postoperative adhesion reformation in infertility and endometriosis surgery. The 242 Obstetrics and Gynecology Adhesion Prevention Committee." Obstetrics and Gynaecology 79(4): 51SS522. Sendil, D., I. Gursel, D. L. Wise and V. Hasirci (1999). "Antibiotic release from biodegradable PHBV microparticles." Journal of Controlled Release 59(2): 207-217. Shi, R. and H. M . Burt (2003). "Synthesis and characterization of amphiphilic hydroxypropylcellulose-graft-poly(e-caprolactone)." Journal of Applied Polymer Science. In press. Shi, R., J. Jackson and H. M . Burt (1999). Characterization and factors influencing drug release from paclitaxel loaded chitosan films. Annual meeting of American Association of Pharmaceutical Scientists, New Orleans, L A , USA. Simons, W. W. and M . Zanger (1972). The Sadtler Guide to N M R Spectra. Philadelphia, Sadtler Reserch Laboratories, Inc. Singla, A. K., A. Garg and D. Aggarwal (2002). "Paclitaxel and its formulations." International Journal of Pharmaceutics 235(1-2): 179-192. Smith, A. H. , Ian M . (1986). "Evaluation of poly(lactic acid) as a biodegradable drug delivery system for parenteral administration." International Journal of Pharmaceutics 30(2-3): 215-220. Solomon, O. F. and I. Z. Ciuta (1962). "Detemination de la viscosite intrinseque de solution de polymeres par une simple determination de la viscosite." Journal of Applied Polymer Science 6: 683. 243 Spencer, C. M . and D. Faulds (1994). "Paclitaxel. A review of its pharmacodynamic and pharmacokinetic properties and therapeutic potential in the treatment of cancer." Drugs 48(5): 794-847. Stacey, M . and F. G. Pautard (1952). "Thermal degradation of dextran." Chemistry & Industry: 1058-1059. Stannett, V. (1970). "Recent advances in pure and applied aspects of graft copolymers." Journal of Macromolecular Science. Part A: Chemistry. 4(5): 1177-1192. Stearns, M . E. and M . Wang (1992). "Taxol blocks process essential for prostate tumor cell (PC-3 ML) invasion and metastases." Cancer Research 52(13): 3776-3781. Steinleitner, A. , H. Lambert, C. Kazensky and B. Cantor (1991). "Poloxamer 407 as an intraperitoneal barrier material for the prevention of postsurgical adhesion formation and reformation in rodent models for reproductive surgery." Obstetrics and Gynaecology 77(1): 48-52. Steinleitner, A. , G. Lopez, M . Suarez and H. Lambert (1992). " A n evaluation of Flowgel as an intraperitoneal barrier for prevention of postsurgical adhesion reformation." Fertility and Sterility 57(2): 305-308. Stenekes, R. J., H. Talsma and W. E. Hennink (2001). "Formation of dextran hydrogels by crystallization." Biomaterials 22(13): 1891-1898. Storey, R. F. and J. W. Sherman (2002). "Kinetics and Mechanism of the Stannous Octoate-Catalyzed Bulk Polymerization of epsilon-Caprolactone." Macromolecules 35(5): 1504-1512. 244 Straubinger, R. M . (1996). Biopharmaceutics of paclitaxel (Taxol): formulation, activity, and pharmacokinetics. Taxol®: Science and Applications. M . Suffness. Boca Raton, CRC Press: 237-258. Struber, M . , J. M . Hohlfeld, S. Fraund, P. Kim, G. Warnecke and A. Haverich (2000). "Low-potassium dextran solution ameliorates reperfusion injury of the lung and protects surfactant function." Journal of Thoracic and Cardiovascular Surgery 120(3): 566-572. Sugibayashi, K., Y. Morimoto, T. Nadai, Y. Kato, A. Hasegawa and T. Arita (1979). "Drug-carrier property of albumin microspheres in chemotherapy. II. Preparation and tissue distribution in mice of microsphere-entrapped 5- fluorouracil." Chemical and Pharmaceutical Bulletin (Tokyo) 27(1): 204-209. Sumimoto, R., N. V. Jamieson, K. Wake and N. Kamada (1989). "24-hour rat liver preservation using UW solution and some simplified variants." Transplantation 48(1): 1-5. Tabata, Y. and Y. Ikada (1989). "Synthesis of gelatin microspheres containing interferon." Pharmaceutical Research 6(5): 422-427. Takahashi, T., H. Tsukamoto and H. Yamada (1998). "Design and synthesis of a water-soluble taxol analogue: taxol-sialyl conjugate." Bioorganic & Medicinal Chemistry Letters 8(1): 113-116. Terry, R., C. L. Yuile, A. Golodetz, C. E. Phillips and R. R. White, III (1953). "Metabolism of dextran-a plasma-volume expander. Studies of radioactive carbon-labeled dextran in dogs." The Journal of Laboratory and Clinical Medicine 42: 6-15. 245 Thacharodi, D. and K. P. Rao (1993). "Propranolol hydrochloride release behaviour of crosslinked chitosan membranes." Journal of Chemical Technology and Biotechnology Field 58(2): 177-181. Thanoo, B. C , M . C. Sunny and A. Jayakrishnan (1992). "Cross-linked chitosan microspheres: preparation and evaluation as a matrix for the controlled release of pharmaceuticals." Journal of Pharmacy and Pharmacology 44(4): 283-286. Torchilin, V. P. (2001). "Structure and design of polymeric surfactant-based drug delivery systems." Journal of Controlled Release 73: 137-172. Treutner, K. -H. and V. Schumpelick (1997). Peritoneal Adhesions. New York :, Springer-Verlag. Trimbos-Kemper, T. C , J. B. Trimbos and E. V. van Hall (1985). "Adhesion formation after tubal surgery: results of the eighth-day laparoscopy in 188 patients." Fertility and Sterility 43(3): 395-400. Tsuchiya, H. M . (1992). Dextran. McGraw-Hill Encyclopedia of Science & Technology. New York, McGraw-Hill. 5: 170-171. Ueno, H. , T. Mori and T. Fujinaga (2001). "Topical formulations and wound healing applications of chitosan." Advanced Drug Delivery Reviews 52(2): 105-115. Uhrich, K. E., S. M . Cannizzaro, R. S. Langer and k. M . Shakesheff (1999). "Polymeric systems for controlled drug release." Chemical Reviews 99: 3181-3198. Usami, Y., Y. Okamoto, T. Takayama, Y. Shigemasa and S. Minami (1998). "Chitin and chitosan stimulate canine polymorphonuclear cells to release leukotriene B4 and prostaglandin E2." Journal of Biomedical Materials Research 42(4): 517-522. 246 Vandelli, M . A., F. Rivasi, P. Guerra, F. Forni and R. Arletti (2001). "Gelatin microspheres crosslinked with D,L-glyceraldehyde as a potential drug delivery system: preparation, characterisation, in vitro and in vivo studies." International Journal of Pharmaceutics 215(1-2): 175-184. Vansteenkiste, S., E. Schacht, I. Z. A. Pawluczyk and R. w. Baldwin (1990). "Synthesis and evaluation of glycosylated dextrans." Polymer Preprints 31(2): 719-720. Vazquez-Duhalt, R., R. Tinoco, P. D'Antonio, L. D. Topoleski and G. F. Payne (2001). "Enzyme conjugation to the polysaccharide chitosan: smart biocatalysts and biocatalytic hydrogels." Bioconjugate Chemistry 12(2): 301-306. Vegarud, G. and T. B. Christnsen (1975). "Glycosylation of Proteins: a new method of enzyme stabilization." Biotechnology and Bioengineering 17(9): 1391-1397. Veld, P. J. A. 1.1, E. M . Velner, P. V. D. Witte, J. Hamhuis and P. J. Dijkstra (1997). "Melt block copolymerization of e-caprolactone and L-lactide." Journal of Polymer Science: Part A: Polymer Chemistry 35: 219-226. Wang, J., Y. Tauchi, Y. Deguchi, K. Morimoto, Y. Tabata and Y. Ikada (2000). "Positively charged gelatin microspheres as gastric mucoadhesive drug delivery system for eradication of H. pylori." Drug Delivery 7(4): 237-243. Wang, N. , X. S. Wu and J. K. Li (1999). " A heterogeneously structured composite based on poly(lactic-co-glycolic acid) microspheres and poly(vinyl alcohol) hydrogel nanoparticles for long-term protein drug delivery." Pharmaceutical Research 16(9): 1430-1435. 247 Wani, M . C , H. L. Taylor, M . E. Wall, P. Coggon and A. T. McPhail (1971). "The isolation and structure of taxol, a novel antileukemic and antitumor agent from Taxus bravifolia." Journal of the American Chemical Society 93: 2325-2327. Ward, C. R., C. M . Peterson and H. H. Hatasaka (1995). " A hook-traction technique for Norplant removal." Obstetrics and Gynaecology 86(5): 848-850. Weibel, M . and G. Majno (1973). "Peritoneal adhesions and their relation to abdominal surgery. A postmortem study." American Journal of Surgery 126(3): 345-353. Weiss, R. J3., R. C. Donehower, P. H. Wiernik, T. Ohnuma, R. J. Gralla, D. L. Trump, J. R. Baker, Jr., D. A. Van Echo, D. D. Von Hoff and B. Leyland-Jones (1990). "Hypersensitivity reactions from taxol." Journal of Clinical Oncology 8(7): 1263-1268. Wiernik, P. H. , E. L. Schwartz, A. Einzig, J. J. Strauman, R. B. Lipton and J. P. Dutcher (1987). "Phase I trial of taxol given as a 24-hour infusion every 21 days: responses observed in metastatic melanoma." Journal of Clinical Oncology 5(8): 1232-1239. Wiernik, P. H. , E. L. Schwartz, J. J. Strauman, J. P. Dutcher, R. B. Lipton and E. Paietta (1987). "Phase I clinical and pharmacokinetic study of taxol." Cancer Research 47(9): 2486-2493. Williams, D. F. (1987). "Advanced applications for materials implanted within the human body." Materials Science and Technology 3(10): 797-806. Winternitz, C. I., J. K. Jackson, A. M . Oktaba and H. M . Burt (1996). "Development of a polymeric surgical paste formulation for taxol." Pharmaceutical Research 13(3): 368-375. 248 Wiseman, D. (1994). Polymers for the prevention of surgical adhesions. Polymeric Site-specific Pharmacotherapy. A. J. Domb. New York, John Wiley: 370-421. Wiseman, D., L. Kamp, C. Linsky, R. Jochen, R. Pang and P. Scholz (1992). "Fibrinolytic drugs prevent pericardial adhesions in the rabbit." The Journal of Surgical Research 53(4): 362-368. Wiseman, D., P. Toleikis and H. M . Burt (1998). "Biocompatibility of novel polymeric films to be used as drug delivery devices and/or adhesion prevention barriers." Study No. SYN9804AG. Yaacobi, Y., L. M . Hamed, K. S. Kaul and M . M . Fanous (1992). "Reduction of postoperative adhesions secondary to strabismus surgery in rabbits." Ophthalmic Surgery 23(2): 123-128. Yanagimoto, G., T. Hayashi, T. Hasegawa, T. Seki, K. Juni, K. Sugibayashi and Y. Morimoto (1999). "Skin disposition of drugs after topical application in hairless rats." Chemical and Pharmaceutical Bulletin 47(6): 749-754. Ydens, I., D. Rutot, P. Degee, J.-L. Six, E. Dellacherie and P. Dubois (2000). "Controlled Synthesis of Polycaprolactone-Graft Dextran Copolymers as Potential Environmentally Friendly Surfactants." Macromolecules 33: 6713-6721. Yeh, C. H. , P. J. Lin and J. J. Chu (2001). "Collagen as a drug carrier for deep sternal wound infection after open heart surgery." Chang-Keng i Hsueh Tsa Chih 24(7): 451-454. Yuan, M . , Y. Wang, X. Li , C. Xiong and X. Deng (2000). "Polymerization of lactides and lactones. 10. Synthesis, characterization, and application of amino-terminated 249 poly(ethylene glycol)-co-poly(e-caprolactone) block copolymer." Macromolecules 33: 1613-1617. Zhang, X., H. M . Burt, G. Mangold, D. Dexter, D. Von Hoff, L. Mayer and W. L. Hunter (1997). "Anti-tumor efficacy and biodistribution of intravenous polymeric micellar paclitaxel." Anticancer Drugs 8(7): 696-701. Zhang, X., J. K. Jackson and H. M . Burt (1996). "Development of amphiphilic diblock copolymers as micellar carriers of taxol." International Journal of Pharmaceutics 132: 195-206. Zhang, X., J. K. Jackson, W. Wong, W. Min, T. Cruz, W. L. Hunter and H. M . Burt (1996). "Development of biodegradable polymeric paste formulations for taxol: an in vitro and in vivo study." International Journal of Pharmaceutics 137: 199-208. Zhang, X., D. A. MacDonald, M . F. A. Goosen and K. B. McAuley (1994). "Mechanism of lactide polymerization in the presence of stannous octoate: the effect of hydroxy and carboxylic acid substances." Journal of Polymer Science, Part A: Polymer Chemistry 32(15): 2965-2970. Zhang, Y. and C.-C. Chu (2000). "Biodegradable dextran-polylactide hydrogel network and its controlled release of albumin." Journal of Biomedical Materials Research 54(1): 1-11. Zhang, Y. and C.-C. Chu (2002). "Biodegradable dextran-polylactide hydrogel networks: Their swelling, morphology and the controlled release of indomethacin." Journal of Biomedical Materials Research 59(2): 318-328. 250 Zhang, Y., C.-Y. Won and C.-C. Chu (1999). "Synthesis and characterization of biodegradable network hydrogels having both hydrophobic and hydrophilic components with controlled swelling behavior." Journal of Polymer Science, Part A: Polymer Chemistry 37(24): 4554-4569. Zhang, Y., C.-Y. Won and C.-C. Chu (2000). "Synthesis and characterization of biodegradable hydrophobic-hydrophilic hydrogel networks with a controlled swelling property." Journal of Polymer Science, Part A: Polymer Chemistry 38(13): 2392-2404. Zhao, W., L. Yu, X. Zhong, Y. Zhang and J. Sun (1995). "The compatibility and morphology of chitosan-poly(ethylene oxide) blends." Journal of Macromolecular Science. Part B: Physics B34(3): 231-237. Zhao, Z. and D. G. Kingston (1991). "Modified taxols. 6. Preparation of water-soluble prodrugs of taxol." Journal of Natural Products 54(6): 1607-1611. 251 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

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"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0091213/manifest

Comment

Related Items