Citation
Synthesis, Properties, and in vivo Evaluations of Protein Meso/Microsphere Compositions for Intratumoral Chemotherapy

Material Information

Title:
Synthesis, Properties, and in vivo Evaluations of Protein Meso/Microsphere Compositions for Intratumoral Chemotherapy
Creator:
YORK, AMANDA M.
Copyright Date:
2008

Subjects

Subjects / Keywords:
Chemotherapy ( jstor )
Composite particles ( jstor )
Diameters ( jstor )
DNA ( jstor )
Dosage ( jstor )
Gelatins ( jstor )
Particle density ( jstor )
Particle size classes ( jstor )
Swelling ( jstor )
Tumors ( jstor )
City of Gainesville ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Amanda M. York. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
5/31/2007
Resource Identifier:
496626893 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

SYNTHESIS, PROPERTIES, AND IN VIVO EVALUATIONS OF PROTEIN MESO/MICROSPHERE COMPOSITIONS FOR INTRATUMORAL CHEMOTHERAPY By AMANDA M. YORK A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2006

PAGE 2

Copyright 2006 by Amanda M. York

PAGE 3

To my parents, Dennis and Paula York, my si ster, Sarah York, and to Jason Ely for all their love, support, understanding, and unwavering belief in me.

PAGE 4

iv ACKNOWLEDGMENTS I would like to thank my faculty advisor and committee chair, Dr. Eugene P. Goldberg, for his support, guidance, and patien ce with me on this re search. Additionally, I would like to thank my graduate co mmittee: Dr. Christophe r Batich, Dr. Anthony Brennan, and Dr. Wolfgang “Jake” Streit. I wi sh to also acknowledge Dr. Carol Detrisac for her help with histologi cal analysis and Dr. Lynn Peck for her extensive knowledge and experience in animal research and husbandr y. My thanks extend to the following for their expertise in various areas of this research: Paul Mar tin, Dr. Brett Almond, Dr. Brian Cuevas, Dr. Josh Stopek, and Dr. Ahmad “Bob” Hadba. Special thanks are extended to severa l other colleagues for their moral and technical support, humor, and friendship duri ng my years at UF: Dr . Margaret Kayo, Dr. Daniel Urbaniak, and Iris Enriquez. I would also like to th ank Jennifer Wrighton, Samesha Barnes, Michelle Carman, Dr. Les lie Wilson, Amin Elachchabi, Dr. Clay Bohn, and Dr. Chris Widenhouse for their support. I would like to finally thank my family and friends whose love and support I could not have made it without. Special thanks go to my parents, Denni s and Paula York, and my sister, Sarah York, for their unconditional support during all my years before and during graduate school. A special thank you also goes to Jason Ely, whose love and support carried me throughout this journey. Than k you also to the rest of my friends and extended family-your support did not go unnoticed or unappreciated.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES.............................................................................................................ix LIST OF FIGURES..........................................................................................................xii 1 INTRODUCTION........................................................................................................1 1.1 Specific Aims........................................................................................................5 1.1.1. Specific Aim 1: Develop Synt hesis for Mesospheres with Novel Protein Compositions.........................................................................................5 1.1.2. Specific Aim 2: Develop Synthesis for In Situ Mitoxantrone (MXN) Loaded Mesospheres with Novel Protein Compositions...................................6 1.1.3. Specific Aim 3: Determine the In Vitro MXN Release Properties of MXN-Loaded Mesospheres...............................................................................6 1.1.4. Specific Aim 4: Determine the In Vitro Cytotoxic Properties of In Situ MXN-Loaded Meso/microspheres.............................................................6 1.1.5. Specific Aim 5: In Vivo Evaluations of Neoadjuvant IT Chemotherapy and Scheduled Multiple IT Injections.......................................7 1.2 Related Intratumoral Chemotherapy Work..........................................................7 2 BACKGROUND..........................................................................................................9 2.1 Cancer...................................................................................................................9 2.2 Chemotherapy.......................................................................................................9 2.2 Bovine Serum Albumin......................................................................................14 2.4 Gelatin.................................................................................................................1 5 2.6 Deoxyribonucleic Acid.......................................................................................16 2.6 Microsphere Synthesis........................................................................................17 2.7 Experimental Approach......................................................................................18 2.7.1 Previous Meso/Microsphere Research....................................................18 2.7.2 Research Goals.......................................................................................19 3 PROTEIN MESO/MICROSPHERE SYNTHESIS AND PROPERTIES.................21 3.1 Introduction.........................................................................................................21 3.2 Materials.............................................................................................................21 3.3 Methods..............................................................................................................22

PAGE 6

vi 3.3.1 Solution Preparation................................................................................22 3.3.2. Microsphere Synthesis.............................................................................24 3.3.2.1 Gelatin-poly(glutamic ) acid mesosphere synthesis...................26 3.3.2.2. Gelatin-bovine serum al bumin mesosphere synthesis...............27 3.3.2.3 Bovine serum albumin-deoxyribonucleic acid mesosphere synthesis....................................................................................27 3.3.3 Particle Characterization..........................................................................29 3.3.3.1 Particle morphology.....................................................................29 3.3.3.2 Particle size analysis....................................................................30 3.3.3.3. Particle swelling..........................................................................30 3.3.4. Statistical Analysis...................................................................................31 3.4 Results and Discussion.......................................................................................32 3.4.1 Gelatin – Poly(glutamic) Acid Mesospheres.............................................32 3.4.2. Gelatin – Bovine Serum Albumin Mesospheres......................................38 3.4.3. Bovine Serum Albumin – Deoxyr ibonucleic Acid Mesospheres.............46 4 EVALUATION OF IN SITU DRUG LOADED PROTEIN MESO/MICROSPHERE COMPOSITIONS..............................................................62 4.1 Materials.............................................................................................................62 4.2. Methods..............................................................................................................63 4.2.1. Solution Preparation................................................................................63 4.2.2. In Situ Mitoxantrone-Loaded Mi crosphere Synthesis.............................64 4.2.2.1. In situ MXN-loaded gelatin – poly(glutamic) acid microspheres...............................................................................64 4.2.2.2. In situ MXN-loaded bovine serum albumin – deoxyribonucleic acid microspheres...........................................65 4.2.3. Particle Characterization.........................................................................66 4.2.4. Drug Content...........................................................................................67 4.2.5. In Vitro Mitoxantrone Release................................................................68 4.2.6. RG-2 Cell Culture....................................................................................68 4.3 Results and Discussion.......................................................................................69 4.3.1. In situ MXN-loaded Gelatin – Poly(glu tamic) Acid Microspheres........69 4.3.2. In situ MXN-loaded Bovine Seru m Albumin – Deoxyribonucleic Acid Microspheres...................................................................................81 5 IN VIVO EVALUATION OF BOVINE SERUM ALBUMIN MICROSPHERES FOR INTRATUMORAL CHEMOTHERAPY........................................................101 5.1 Introduction....................................................................................................101 5.2 Materials........................................................................................................102 5.3 Methods..........................................................................................................102 5.3.1 Microsphere Synthesis & Characterization........................................102 5.3.2 In Vivo Animal Study Protocols..........................................................105 5.3.2.1 Tumor passage.........................................................................105 5.3.2.2 Standard protoc ol for in vivo 16/C MMAC studies................105 5.3.2.3 Blood collection.......................................................................106

PAGE 7

vii 5.3.2.4 Tumor histology.......................................................................107 5.3.3 Study Design for In Vivo Animal Studies............................................107 5.3.3.1 Neoadjuvant intratumoral ch emotherapy and tumor histology107 5.3.3.2 Scheduled intratumoral chemotherapy injections....................108 5.3.4 Statistical Analysis...............................................................................109 5.4 Results & Discussion.....................................................................................109 5.4.1 Neoadjuvant Intratumoral Chem otherapy and Tumor Histology........109 5.4.1.1 Surgical observations...............................................................110 5.4.1.2 Histology results......................................................................111 5.4.1.3 Blood cell counts......................................................................111 5.4.1.4 Survival....................................................................................113 5.4.2 Scheduled Intratumoral Chemotherapy Injections...............................118 5.4.2.1 Animal weight..........................................................................118 5.4.2.2 Survival....................................................................................120 6 RELATED INTRATUMORAL CHEMOTHERAPY WORK................................122 6.1 Introduction.......................................................................................................122 6.2 Intratumoral Chemotherapy with 5fluorouracil for Palliation of Bronchial Cancer in Patients with Severe Airway Obstruction.........................................123 6.2.1 Study Summary......................................................................................123 6.2.2 Results....................................................................................................124 6.3 Intratumoral Administration of Cisplatin through a Bronchoscope Followed by Irradiation for Treatment of Inoperable Non-Small Cell Lung Cancer.......125 6.3.1 Study Summary......................................................................................125 6.3.2 Results..................................................................................................126 6.4 Discussion.........................................................................................................128 7 CONCLUSIONS......................................................................................................130 7.1. Unloaded Meso/Microsphere Synthesis...........................................................130 7.1.1. Gelatin-PGA Mesosphere Synthesis.....................................................130 7.1.2. Gelatin-Bovine Serum Albu min Mesosphere Synthesis.......................131 7.1.3. Bovine Serum Albumin-De oxyribonucleic Acid Mesosphere Synthesis...............................................................................................131 7.2. In situ Mitoxantrone-Loaded Mi crosphere Synthesis and In Vitro MXN Release..................................................................................................................132 7.2.1. In situ MXN-loaded Gelatin-PGA Mesospheres Synthesis & Release.132 7.2.2. In situ MXN-loaded Gelatin-PGA Mesospheres Synthesis & Release.133 7.3. In Vitro Cytotoxic Properties of In Situ MXL-Loaded Mesospheres..............133 7.3.1. Gelatin-PGA MS In Vitro Cell Culture Studies....................................133 7.3.2. BSA-DNA MS In Vitro Cell Culture Studies........................................134 7.4. In vivo Evaluation of Protein Mesospheres.....................................................134 7.4.1. Neoadjuvant IT Chemotherapy.............................................................134 7.4.2. Scheduled Multiple IT Injections..........................................................135 8 FUTURE WORK......................................................................................................136

PAGE 8

viii LIST OF REFERENCES.................................................................................................138 BIOGRAPHICAL SKETCH...........................................................................................143

PAGE 9

ix LIST OF TABLES Table page 3.1: Synthesis condition summary for G-PGA mesospheres.............................................26 3.2: Selected gelatin and PGA aqueous solu tion concentrations for synthesis of GPGA MS...................................................................................................................27 3.3: Aqueous phase concentrations for G-BSA compositions...........................................27 3.4: Summary of steric stabilizati on synthesis parameters for BDMS..............................28 3.5: Aqueous phase concentrations for BSA-DNA compositions. X-Y: X refers to the BSA:DNA composition and Y refers to the crosslink density.................................29 3.6: Particle size results fo r G-PGA compositions synthesized on mixer 2. Only data from particles in the size range i ndicated in Figure 3.11 are included.....................38 3.7: Mean particle diameter for G-PGA 1 from swelling study........................................39 3.8: Mean particle size for G-BSA compositions as determined by C oulter LS particle size analysis..............................................................................................................44 3.9: Percent increase in mean particle diameter (% IPD) for G-BSA compositions based on particle swelling studies............................................................................46 3.10: Mean particle size data for BSA-DNA MS determined by Coulter LS particle sizing........................................................................................................................5 6 3.11: Mean dry and swollen particle si ze measurement from optical microscopy imaging. All BSA-DNA compositions swell to less than 80% IPD, much less than that observed for G-C (494% IPD)...................................................................56 4.12: Solution parameters for G-PGA-M compositions....................................................64 4.13: Synthesis conditions for G-PGA-M compositions...................................................65 4.14: Summary of synthesis conditions for BSA-DNA-M................................................66 4.15: Solution and synthesis parameters for the BSA-DNA-M compositions..................66

PAGE 10

x 4.16: Percent increase in mean particle di ameter for G-PGA-M compositions based on particle swelling studies. Results from G-PGA 1 are included here for reference.72 4.17: In situ MXN-loading results for G-P GA-M and related compositions....................74 4.18: MXN release results for G-PGA M and related compositions.................................76 4.19: In vitro RG-2 cell culture treatment groups and MXN doses and resulting normalized MTT absorbance values on day 4.........................................................77 4.20: Summary of mean particle size from C oulter LS particle sizing. Note the large SD of BSA-DNA M2-2 and BSA-DNA M33, likely caused by agglomeration....86 4.21: Percent increase in mean particle di ameter for BSA-DNA M compositions from particle swelling studies...........................................................................................89 4.22: In situ MXN-loading results for BSA-DNA M compositions..................................94 4.23: In vitro RG-2 cell culture MTT assay treat ment groups and mean normalized absorbance values of day 4.......................................................................................98 5.24: Synthesis parameters for BSA-M microspheres.....................................................103 5.25: Design for neoadjuvant IT chemothera py study. CBC values we re also analyzed from mice with no tumor (NT)...............................................................................108 5.26: Design for scheduled IT injection study.................................................................109 5.27: CBC data for mice enrolled in this study. MS L refers to lo w crosslink density BSA-M and MS H refers to high crossli nk density BSA-M; NT refers to mice with no tumors. (n=3 per group)...........................................................................113 5.28: Survival summary for neoadjuvant IT chemotherapy study. All groups with 100% survival showed significantly increas ed survival time versus controls.......116 5.29: Body weight summary for the schedul ed IT injections study at selected timepoints (n=10 per group)...................................................................................119 5.30: Survival summary for scheduled IT injection study...............................................120 6.31: Semi-quantitative numeri cal value assignment for degree of increase in lumen diameter (ILD)........................................................................................................124 6.32: Number of patients in each ILD category by tumor location.................................124 6.33: Results of Wilcoxon Matched-Pairs Signed-Rank tests using lumen opening values from individual patient examina tions before and after IT treatment..........125

PAGE 11

xi 6.35: Patient survival by efficacy of IT chemotherapy debulking using IT cisplatin injections................................................................................................................127 6.36: Patient survival following IT cisplatin treatment by tumor localization on bronchial tree..........................................................................................................127

PAGE 12

xii LIST OF FIGURES Figure page 2.1: Mitoxantrone structure (MW=517.4 g/mol)...............................................................12 2.2: Illustration on intercalation of a planar structure between base pairs in DNA. The exact mechanism through which MXN intercalates DNA is not understood, but there is an apparent preference for G-C base pairs..................................................13 2.3: The cell cycle and phases. G1 involves RNA and protein sythesis and is a period of normal metabolism. S phase involves DNA synthesis, G2 involves growth and further protein synthesis prior to cell division, and M phase is the mitotic phase, where the cell splits into 2 new cells. G0 denotes a prolonged or permanent G1 phase. (Figure adapted from Wilkes, et al.29)..................................14 2.4: Schematic representation of DNA struct ure. Shading highlights the phosphate backbone of the structure.........................................................................................17 3.5: Illustration of suspension crosslinki ng technique for microsphere synthesis (diagram adapted from Cuevas 11)...........................................................................25 3.6: SEM image of G-PGA 4 at 1000x magnifi cation. Note the large non-spherical particles and the rough surface texture.....................................................................33 3.7: Particle size dist ribution for G-PGA 4........................................................................33 3.8: Comparison of SEM images by high-speed mixer for G-PGA 1 composition. All images at 2000x magnification. G-PGA show n in A) and C) were synthesized using mixer 1 and those shown in B) a nd D) were synthesized using mixer 2........35 3.9: Particle size distribution compar ison by high-speed mixer for G-PGA 1 composition. (M#a #b): #a denotes mixer number and #b denotes the batch number on that mixer...............................................................................................35 3.10: SEM comparison of G-PGA compos itions. All images taken at 3000x magnfication.............................................................................................................37 3.11: Particle size distribution comparis on by batch and G-PGA composition. An upper limit was set to ~50 m to avoid inclusion of agglomerates in analysis........38

PAGE 13

xiii 3.12: Batch and composition comparison of m ean particle size using Coulter LS 230 particle sizer.............................................................................................................39 3.13: G-PGA 1 composition optic al microscopy images A) dry and B) swollen at 1 minute post-hydration. Particle measurements are indicated on each image..........40 3.14: Dry and swollen mean particle di ameter for G-PGA 1 and G-C. G-PGA 1 exhibited a 194% increase in particle diameter upon hydration, compared to 494% increase in size for gelatin controls . Raw data were not available for G-C MS; as a result no error bars are in cluded on those corresponding bars..................41 3.15: Batch comparison for G-BSA 1: A) batc h 1, B) batch 2, C) ba tch 3, and D) batch 4. As with G-PGA, mixer 1 produced me sospheres with much larger average particle size (C). This phenomenon was also observed for G-BSA 2 (not shown). All images taken at 2000x.........................................................................42 3.16: Batch comparison of particle size fo r A) G-BSA 1 and B) G-BSA 3. Both compositions show one batch with skewed particle size distributions curves relative to the remaining batches. Th ese batches were both synthesized on mixer 1, which produced G-PGA mesospheres with similar results........................43 3.17: Representative SEM images of A) G-BSA1, B) G-BSA 2, C) G-BSA 3, and D) BSA-C. All images taken at 2000x magification....................................................44 3.18: Particle size distribution curves for A) G-BSA 1, B) G-BSA 2, C) G-BSA 3, and D) BSA-C.................................................................................................................45 3.19: Representative optical microscopy im ages of G-BSA MS from swelling study. The image on left is of A) G-BSA 2 dry mesospheres; the image on the right is the corresponding image of swollen GBSA 2 mesospheres at 1 minute posthydration...................................................................................................................47 3.20: Particle swelling comparison by GBSA composition. Matching symbols on dry particle sizes indicate st atistically significant differe nces between those groups. No statistical differences were detected between the swollen sizes of any of the G-BSA compositions. All G-BSA compos itions exhibited less than 100% IPD, compared to almost 500% for G-C. Raw data for G-C were not available and could not, therefore, be included in the statistical analysis......................................48 3.21: SEM image of BSA-DNA 1 at 2000x magnification...............................................49 3.22: Particle size distribution for BSA-DNA 1................................................................49 3.23: Dry and swollen mean pa rticle diameters based on swelling measurements using optical microscopy. BSA-DNA 1 MS dry and swollen particle sizes were significantly different from one another (73% IPD)................................................50

PAGE 14

xiv 3.24: Dry (A) and swollen (B) BSA-DNA 1 me sospheres via optical microscopy at 40x magnification. Swollen image taken at 1 minute post-hydration with PBS.....51 3.25: Representative SEM images of BSADNA compositions. Each row represents one composition, with increasi ng crosslink density from left to right; A) – C) BSA-DNA 2, D) – F) BSADNA 3, G) – I) BSA-C................................................54 3.26: Comparison of representative part icle size distribution curves by MS composition and crosslink density: A) BSA-DNA 2, B) BSA-DNA 3, and C) BSA-C. Each of these contains curves for low (X-1), medium (X-2), and high (X-3) crosslink densities. Note the ch ange in scale on part A compared with parts B) and C).........................................................................................................55 3.27: Particle swelling charts comparing th e effect of DNA content on dry and swollen sizes for each composition by crosslink density; A) low crosslink density, B) medium crosslink density, and C) high crosslink density........................................57 3.28: Particle swelling charts comparing the effect of crosslink density on dry and swollen sizes for each composition by DNA content; A) 0% DNA, B) 2.5% DNA, and C) 5% DNA content................................................................................58 3.29: BSA-DNA 2 swelling study optical micros copy images (40 x). Image on the left are of dry BSA-DNA 2 mesospheres. Corresponding swollen images are on the right.......................................................................................................................... 59 3.30: BSA-DNA 3 swelling study images. Imag es on the left are of dry particles; those on the right are swollen, ta ken at 1 minute post-hydration.............................60 3.31: Optical microscopy images of BSA-C......................................................................61 4.32: SEM images of A) G-PGA M-1 and B) G-PGA M-2. All images taken at 1000x magnification............................................................................................................71 4.33: Particle size distribution for G-P GA M compositions. G-CMC-M is also included here for direct comp arison to previous studies..........................................71 4.34: Optical microscopy images of unloade d and MXN-loaded G-PGA compositions. All images at 40x magnification; those on the left are of dry microspheres and those on the right are of fully swolle n microspheres, 1 minute post-hydration with PBS. A) – B) are of unloaded G-PGA 1, C) – D) are of G-PGA M-1, and E) – F) are of G-PGA M-2.......................................................................................73 4.35: Dry and swollen particle size comparison for G-PGA-M compositions. G-PGA (unloaded) and G-C (unloaded) are incl uded for comparison. Raw data were not available, therefore G-C was not include d in statistical analysis. Matching symbols indicate significant differences..................................................................74

PAGE 15

xv 4.36: MXN loading comparison for G-PGA-M and related compositions. G-C-M and G-CMC-M were synthesized by Cuevas and are included for comparison only.....75 4.37: Release curve for G-PGA-M and relate d compositions. F-MXN is a free MXN control used to quantify and MXN degrad ation and the binding affinity of MXN to the filter membrane..............................................................................................76 4.38: Mean absorbance value comparis on for G-PGA M in vitro RG-2 cell proliferation study....................................................................................................78 4.39: Day 2 optical microscopy images of RG-2 cells by treatment type and MXN dose. Image A) is of non-treatment controls. All images taken at 40x..................80 4.40: SEM image of BSA-DNA M-1 micr ospheres at 1000x magnification....................82 4.41: Particle size distribution comparis on for MXN-loaded and unloaded BSA-DNA 1 compositions..........................................................................................................83 4.42: Optical microscopy images of BSADNA M-1 A) dry and B) 1 minute posthydration. Both images at 40x magnification.........................................................84 4.43: Comparison of particle swelling for unloaded and in situ MXN-loaded BSADNA 1 microspheres. Matching symbols indicate significant differences..............85 4.44: Representative particle size distributions curves for MXN-loaded BSA-DNA MS compositions.............................................................................................................86 4.45: SEM image comparison of BSA-DNA M-2 (A & B), BSA-DNA M-3 (C & D), and BSA-C M (E & F). All images were taken at 2000x magnification................88 4.46: Particle swelling charts comparing dry and swollen sizes for each composition by crosslink density; A) middle crossli nk density, B) high crosslink density.........89 4.47: Comparison of dry and swollen particle sizes by composition; A) BDMS MC, B) BDMS M2, and C) BDMS M3................................................................................90 4.48: Particle swelling images for BSA-DN A M-2; A & B middle crosslink density; C & D high crosslink density. All images at 40x.................................................91 4.49: Particle swelling images for BSA-DN A M-3; A & B middle crosslink density; C & D high crosslink density. All images at 40x.................................................92 4.50: Particle swelling images for BSA-C M; A & B middle crosslink density; C & D high crosslink density. All images at 40x.........................................................93 4.51: Optical microscopy image of BSA-DNA M1 microspheres after 20 days in EDB at 37 C (40x magnification).....................................................................................94

PAGE 16

xvi 4.52: MXN loading comparison by BSA-DNA composition and crosslink density. * BSA-DNA M1 loading data are likely gr ossly underestimated (see explanation above).......................................................................................................................95 4.53: MXN release concentra tion per milligam of microspheres for BSA-DNA M1-2. Gelatin, gelatin-CMC, and gelatin-PGA mi crosphere release curves are included for comparison..........................................................................................................96 4.54: Release profile for BSA-DNA M compos itions for A) the entire release study and B) the first 25 hours only...................................................................................97 4.55: Normalized mean absorbance values for BSA-DNA M1 in vitro RG-2 cellular viability study. Dosage key: 1 = 0.5 g/mL, 2 = 12.5 g/mL, and 3 = 25 g/mL..99 4.56: Optical microscopy images of RG-2 cells from in vitro RG-2 cell culture study. All images were taken at 40x.................................................................................100 5.57: SEM images of BSA-M used in these in vivo studies: A) 5-10 m BSA-M MS with high crosslink density (8% GT A; 500x magnificati on) and B) 5-10 m BSA-M MS with low crosslink de nsity (2% GTA; 750x magnification)..............103 5.58: Particle size distribution curves for unloaded BSA MS, in situ MXN-loaded BSA MS with low crosslink dens ity (2% GTA w/w; BSA-M L), and in situ MXN loaded BSA MS with high crossli nk density (8% GTA w/w; BSA-M H)..104 5.59: In vitro release of MXN from BSA-M20 (20-30 m) and BSA-M (5-10 m) with low and high crosslink densities under infinite sink conditions.............................104 5.60: Tumor sections from A) MS H (day 1) and B) MS L (day 14) treatment groups. Microspheres are visible in both images, with significant degradation of MS L evident (B) by increased porosity and reduced eosinophilic staining. Both images at 100x........................................................................................................112 5.61: Comparison of WBC and RBC counts by treatment group. No significant differences were detected between any groups......................................................114 5.62: Comparison of hematocrit values and platelet counts by treatment group. No significant differences were detected for hematocrit, though it appears that microsphere treatments may result in lower hematocrit values compared to controls. Large variability in these data likely prevented detection of significance for this parameter. Signi ficant differences were detected for platelet counts. Matching symbols indi cate differences between these groups....115 5.63: Comparison of neutrophil and lympohc tye counts by treatment group. Control group neutrophil percentage was found to be significantly lower than all groups except MXN-14 and MS H-1.................................................................................116

PAGE 17

xvii 5.64: Survival chart for neoadjuvant IT chemotherapy study. Control (C) mice received no chemotherapy treatment but underwent surgical resection on Day 1 or 7. Mice that received IT MXN treatm ents on Day 0 received either MXN at 8 mg/kg dose or 24 mg/kg of BSA-M MS delivered in 4 mg/kg MXN, where BSA-M MS were of either low or high crosslink density (2% or 8% GTA w/w, respectively). Surgical resection was then performed on either Day 1, 7, or 14 following IT treatment...........................................................................................117 5.65: Normalized TABW over time by treatm ent group. Gray lines represent times selected for statistical analysis...............................................................................119 5.66: Animal survival by treatment gr oup for scheduled IT injection study...................121

PAGE 18

xviii Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy SYNTHESIS, PROPERTIES, AND IN VIVO EVALUATION OF PROTEIN MESO/MICROSPHERE COMPOSITIONS FOR INTRATUMORAL CHEMOTHERAPY By Amanda M. York May 2006 Chair: Eugene P. Goldberg Major Department: Biomedical Engineering Conventional systemic chemotherapy dist ributes cytotoxic drugs throughout the body, killing both cancerous and non-cancerous ce lls. This dose-limiting toxicity reduces efficacy of chemotherapy treatments. New c oncepts for localized therapies such as intratumoral chemotherapy need investiga tion. Intratumoral in jection can greatly increase tumor dose and prolong exposure of cancer cells to cytotoxic drugs while minimizing systemic toxicity. Much of th is research was therefore devoted to the synthesis and evaluation of novel mitoxa ntrone (MXN) loaded protein mesomicrospheres (MS) for intr atumoral (IT) injection. Gelatin – polyglutamic acid (G-PGA), gela tin – bovine serum albumin (G-BSA), and bovine serum albumin – deoxyribonucle ic acid mesospheres (BSA-DNA) were synthesized with a mean particle diameter of 1-10 m using a steric stabilization process. In situ MXN-loaded G-PGA (G-PGA-M) a nd BSA-DNA (BSA-DNA-M) mesospheres

PAGE 19

xix were also synthesized. G-PGA-M had lo wer MXN loading and slower drug release compared with gelatin MS c ontrols (Gel-M), with < 20% release complete at 1000 hours compared to > 50% at 24 hours for GelM. BSA-DNA-M mesospheres using DNA from a herring source released 20% more MXN th an Gel-M for > 1000 hours. However, BSADNA-M MS incorporating salmon DNA comple ted MXN release within a matter of 12 hours. In vitro RG-2 studies revealed G-PGA-M form ulations inhibited cell proliferation more compared with non-treatment controls, but not with free MXN positive controls due to the slow release profile of G-PGA-M MS. Repeated courses of either IT injecti on of MXN alone or MXN-loaded BSA MS (BSA-M) achieved 70% survival of animals treated in an in vivo study using a murine 16/C murine mammary adenocarcinoma model. When a single IT injection of BSA-M MS was followed by surgery 14 days later, 100% survival was achieved in the same model, with no local recurrence for at least forty days. Complete blood counts revealed acute inflammation in tumor tissue associated with MXN injection using either modality. Analysis of in vivo clinical data from patients with inoperable lung cancer receiving IT chemotherapy showed that this treatment is effective at increas ing airway lumen and, when coupled with local irradiation, is also potentially effective at prolonging survival of these patients, particularly in th e absence of immediate treatment.

PAGE 20

1 CHAPTER 1 INTRODUCTION Despite recent cancer research adva nces suggesting improved chemotherapy, especially for preoperative (neoadjuvant) modalities, c onventional cancer treatment largely continues to be surgery followed by radiation and/or systemic chemotherapy.1 Even with advances in can cer treatment over the past 30 years, there have been only modest improvements in survival rates for high mortality cancers including colorectal and breast cancers.2, 3 For breast cancer in particular , a 0.4% increase in annual U.S. mortality was observed for the years betw een 1975 and 1990, followed by a 2.3% annual mortality reduction between 1990 and 2001. It is difficult to know how much of this improvement is due specifical ly to enhanced treatment regimens since it has been attributed to both improved detection and treatments.1, 4 Despite improvements in the treatment of breast cance r, mortality for lung cancer has continued to rise significantly over the last 30 years and is the leading cause of cancer death in both men & women illustrating a need for improved cancer treatments.5 The major limiting factor for the efficacy of intravenous chemotherapy is systemic toxicity. Since chemotherapy affects rapi dly dividing cells but cannot discriminate between cancerous (neoplastic) and healthy cells , a variety of toxic side effects occur. Rapidly dividing healthy cells such as thos e found in the gastroin testinal lining, hair follicles, and white blood cells are therefor e affected by chemotherapy and result in complications including nausea, vomiting, alop ecia, and leucopenia when these drugs are delivered systemically. Toxi c side effects of chemothera py not only reduce the patientÂ’s

PAGE 21

2 quality of life, but also can be life threatening in severe cases. For these reasons, improvements in chemotherapy are needed. A promising alternative to systemic treatments may be the use of intratumoral (IT) delivery of chemotherapy, which allows the administration of higher local doses of drugs directly to the tumor lesion with reduced systemic toxicity and would be particularly useful for cancers with well-defined primary tumors (including lung, colorectal, brain, and breast cancers). Additionally, IT chemotherapy may provide a basis for more effective tissue conserving surgeries for preoperative (neoadjuvant) applications. Furthermore, IT chemotherapy compositions with sustained drug release properties may provide improved efficacy through prolonged tumor exposure to drugs and may reduce cost s and increase patient quality of life by reducing the need for multiple treatments. Several studies have been conducted that illustrate the superior efficacy of IT treatment s over IV treatments in animal models. A recent review documents a number of such investigations.1 Additional recent human clinical studies with IT in jections free drugs for lung cancer treatment show marked improvement in airway obstruction and patient survival.6, 7 Previous research has shown the dependen ce of the drug release characteristics of MS on particle size, crosslink density, and in teractions between the MS matrix material and the drug. Both increased particle si ze and increased cross link density slow drug release, likely due to reduced surface area: volume ratio and decrease d particle swelling. Incorporation of anionic substances into the microsphere matrix can also increase cationic drug payload and/or prolong release from protein microsphere s. In particular, the use of polyglutamic acid (PGA) increased the loading of adriamycin in bovine serum

PAGE 22

3 albumin microspheres (BSA) up to ~40%.8, 9 In a similar study, PGA also increased drug loading in casein microspheres from 8% to 25%.10 The use of gelatin (G or Gel) as a MS matr ix material has also been investigated. The Gel MS showed improved cumulative drug release efficiency compared to past protein MS formulations, achieving in vitro release of nearly 80% of loaded drug within 12 hours in phosphate buffered saline.11 Despite the promise of Gel MS, research by Freeman, et al. in Dr. Goldberg’s laboratory indicated that the high degree of swelling of gelatin microspheres is likely to cause n eedle obstruction duri ng IT injections as determined in an in vivo study.12 In comparison, other recent studies have indicated that BSA MS for IT delivery of MXN, could be injected with litt le to no difficulty. BSA may therefore be a better clinical matrix material for IT injections unless swelling of Gel MS can be reduced using additional matrix materials.13, 14 In the research reported here, novel combin ations of MS matrix components were investigated for use in intratumoral chemot herapy applications. Based on the results of previous research, synthesis parameters were chosen to produce MS with mean particle diameters of 1-10 m (mesospheres). Particles in th is diameter range appear most promising to facilitate IT injection and also provide increased tumor perfusion. Gelatin – poly(glutamic) acid (PGA) compositions were investigated to determine if the incorporation of an anionic species would impr ove cationic drug loading and release properties, as well as reduce microsphere swelli ng through the introduction of ionic bonding characteristics. Gelatin has the advantages of being a natural biopolymer and can therefore be degraded in vivo by naturally occurring enzy mes. Gelatin is also readily available and inexpensive, and is readily crosslinked us ing pendant functional

PAGE 23

4 groups. PGA is an anionic polypeptide and shou ld facilitate increased loading of cationic drugs such as MXN through ionic interactions. To capitalize on the benefits of both bovi ne serum albumin and gelatin as MS matrix materials, blends of these two proteins were also investigated with the rationale that the incorporation of BSA may decrease particle swelling. Several formulations were synthesized and the resulting microspheres were characterized fo r particle morphology, size, and swelling. The novel use of DNA as a MS matrix biopolymer for chemotherapy applications was also investigated. The overall anioni c charge of DNA and its relative abundance made it an ideal candidate for such applica tions. Ionic interactions with cationic MXN may help prolong the release of therapeutic leve ls of drug. The use of DNA as an anionic constituent in a drug delivery matrix is nove l. In the late 197 0Â’s and early 1980Â’s, intravenous injection of fr ee DNA-daunorubicin complexes we re evaluated and showed marked reduction in drug cardiotoxicity and higher chemotherapeutic efficiency.15, 16 Though DNA has been incorporated into micropa rticles and liposomes to be released in gene therapy applications, it has not been used as a MS matrix material for purposes of prolonging drug release for chemotherapy. A dditionally, MS for gene therapy have typically been made from chitosan or poly(la ctide-co-glycolide). Th e study of MS with an albumin matrix that contains DNA has not been reported for either chemotherapy or gene therapy. Furthermore, although not the focus of this research, incorporation and delivery of viable DNA or plasmids from BS A mixtures as nanospheres or mesospheres may lead to useful gene therapy applications in future studies.

PAGE 24

5 1.1 Specific Aims A summary of the specific aims for this re search is outlined briefly below. The primary goal of these studies was to synthe size and evaluate novel protein-biopolymer blend microspheres for use in intratumoral ch emotherapy. Formulations were sought to increase MXN loading and/or prolong MXN release in vivo . Major overall objectives of IT research was increased safety and e fficacy of chemotherapy and improved patient quality of life by reducing the number of required treatments and reducing or eliminating the side effects commonly associated with tr aditional systemic chemotherapy. Each aim is described in more detail below. 1.1.1. Specific Aim 1: Develop Synthesi s for Mesospheres with Novel Protein Compositions. Gelatin-poly(glutamic acid) mesospheres (G-PGA), gelatin-bovine serum albumin mesospheres (G-BSA), and bovine serum al bumin-deoxyribonucleic acid mesospheres (BSA-DNA) were synthsized usi ng the steric stabilization process to form particles in the 1-10 m diameter range. Each of these prepara tions was crosslinked with glutaraldehyde using previously established BSA meso/microsphere parameters. The effect of aqueous phase component ratios on particle size, morphology, and swelling was studied for all mesosphere formulations. For some formulati ons the effect of changing crosslink density on these same parameters was also evaluated. Average dry particle size was measured quantitatively using a Coulter LS particle sizer and qualitatively using scanning electron microscopy. Swollen particle size was determined using an inverted microscope equipped with digital capture and measurem ent capabilities. Based on these results, compositions were chosen for further investigation in later aims.

PAGE 25

6 1.1.2. Specific Aim 2: Develop Synthesis for In Situ Mitoxantrone (MXN) Loaded Mesospheres with Novel Protein Compositions. In situ MXN loaded G-PGA (G-PGA-M) and BSA-DNA (BSA-DNA-M) mesospheres were synthesized with a ta rget mean particle diameter of 1-10 m. The effects of aqueous phase component ratios, drug loading, and crossl ink density (where appropriate) on particle size, morphology, swel ling, and MXN content were studied. All characterization techniques from Aim 1 were employed. Additionally, enzymatic digestion was used in conjunction with UV-vi sible spectroscopy (UV-Vis) to determine drug content of in situ MXN-loaded microspheres as a function of the above listed parameters. 1.1.3. Specific Aim 3: Determine the In Vitro MXN Release Properties of MXNLoaded Mesospheres. In situ MXN loaded mesospheres synthe sized in Aim 2 (G-PGA-M and BSADNA-M) were incubated in phospha te buffered saline (PBS) at 37 C for at least ten days. At predetermined timepoints, aliquots we re taken and analyzed using UV-Vis to determine the drug concentrations at each timepo int. These data were used to generate a release profile curve for MXN and were anal yzed to determine the effects of changing synthesis parameters in Aim 2 on release profile. 1.1.4. Specific Aim 4: Determine the In Vitro Cytotoxic Properties of In Situ MXNLoaded Meso/microspheres. In situ MXN loaded mesospheres (G-PGA-M) were incubated in rat glioma 2 (RG2) cell culture in complete media at 37 C and humid 8% CO2 atmosphere for a period of 4 days. On each day, an MTT assay was used to determine cellular viability and optical microscopy was used to assess cel lular morphology. MTT data and cellular

PAGE 26

7 morphology were compared to non-treatment controls to determine efficacy of the microsphere treatments in limiting or eliminating cellular viability. 1.1.5. Specific Aim 5: In Vivo Evaluations of Neoadjuvant IT Chemotherapy and Scheduled Multiple IT Injections. IT injection of BSA-M followed by surg ical removal of the tumor mass was evaluated in a 16/C murine mammary adeno carcinoma (MMAC) model to determine the efficacy of this treatment on animal survival . Surgical excision was performed on 1, 7, or 14 days post-treatment, except nontreatment c ontrols which underwent surgical excision at day 1 or 7 only. Tumors were examined hi stologically to evaluate the effect of IT treatment on initiating tumor necrosis as well as the distribution a nd degradation of MS within the tumor. Animal body weight ( normalized for tumor weight) was used as a measure of MXN-related toxicity. Add itionally, complete blood counts (CBC) were performed on designated days to explore th e mechanism of toxicity of MXN in this model. Additionally, in vivo examination of animal survival following multiple IT injections of MXN-loaded BSA MS (B SA-M; mean particle diameter 1-10 m) at 1 week intervals was carried out in a 16/C MMAC m odel. Animal body weight (normalized for tumor weight) was used as a meas ure of MXN-related toxicity. 1.2 Related Intratumoral Chemotherapy Work Patients with inoperable lung cancer ofte n suffer from life-threatening airway obstructions. Two studies were conducte d by Dr. Celikoglu us ing intratumoral chemotherapy as a means of relieving airway obstruction and prolonging patient survival. Analysis of intratumoral chemotherapy data from these human clinical trials utilizing

PAGE 27

8 intratumoral chemotherapy for lung cancer trea tment was also performed and the results of that analysis are presented here.

PAGE 28

9 CHAPTER 2 BACKGROUND 2.1 Cancer Cancer results from the uncontrolled pr oliferation of abnormal cells in the body and may be a result of heredity, environmenta l factors, or a combin ation of both. Cancer follows only heart disease as the leading cause of death in the Unites States, and accounts for 25% of all deaths. In 2005, the National Institutes of Health (NIH) estimated that cancer had an associated cost of over $74 billi on in direct medical co sts, and an overall cost of $209 billion.5 According to the 2005 Survei llance, Epidemiology, and End Results (SEER) Cancer Statistics Review (C SR), it is estimated that there will be over 570,000 cancer deaths this year. Lung, colore ctal, and breast cancer will account for a combined 46% of these deaths and are the top three leading causes of cancer in the US for both sexes combined. Lung cancer is the leading cause of cancer death among both men and women, with over 172,000 new cases and over 163,000 deaths estimated for 2005. Breast cancer is the second leading cause of cancer death among women behind only lung cancer and will account for 32% of new cancer cases in 2005.3 These statistics illustrate the ongoing need for improved cancer treatments, including those that may have an effect not only on patient surviv al, but also on quality of life. 2.2 Chemotherapy Success of cancer treatment re lies on the ability of the chosen treatment to remove or destroy cancerous cells without producing se rious systemic toxic effects. Traditionally chemotherapy is administered intravenously (IV), with the schedule and dose ranges

PAGE 29

10 varying depending on the drug(s) given, patient ch aracteristics, and type of cancer. The problem with this type of therapy is that chemotherapy targets rapi dly dividing cells and do not discriminate between rapidly dividing healthy and cancerous (neoplastic) cells, resulting in significant systemic toxicity when delivered intravenously. This toxicity is the limiting factor which restricts drug dos es that may be delivered, resulting in a marginal therapeutic index. In an effort to reduce systemic toxic ity, local chemotherapy techniques have been investigated. These techniques aim to incr ease local concentrations of drugs while limiting systemic exposure, thereby improving efficacy and reducing systemic toxicity. In general, there are four ways in which ch emotherapeutic efficiency might be optimized using local or regional administration: 1. Increase tumor exposure to cytot oxic drugs in terms of peak drug levels and area under the curve ( AUC) of concentration vs. time 2. Prolong tumor exposure to cytotoxic agents 3. Decrease systemic toxicity compared to IV administration 4. Enhance opportunity for synergy between multiple antineoplastic drugs17 To date, local chemotherapy modalities that have been used either experimentally or clinically include intraperitoneal (IP), intr a-arterial (IA), intrat hecal, and intrapleural injections. In a recent review of a regional IA therapy known as hepatic arterial infusion (HAI), results from 8 studies reported comp lete or partial response for patients receiving HAI of 5-fluorodeoxyuridase (FUDR) compared to IV delivery.18 Another review of studies on IP cisplatin for ovarian cancer reported that women receiving IP cisplatin showed significantly improved surviv al and reduced toxic side effects.19 These are a few

PAGE 30

11 examples of the promising clinical results for loco-regional deli very of chemotherapy agents reported to date. IT injections, as another a pproach to local therapy, invol ve the direct injection of chemotherapy into the tumor itself and therefore provide the highest possible concentration of drug reaching the tumor. C onsequently, this form of local chemotherapy may prove to be the simplest, most eff ective, and most clinically practical. Mitoxantrone. Mitoxantrone (1,4-d ihydroxy-5, 8-bis-((2-((2hydroxyethyl)amino)ethyl) amino)-9, 10-anthra cenedione dihydrochloride; MXN) is a synthetic anthracenedione with a significan tly broad spectrum of antineoplastic activity (Figure 2.1). MXN (trade name Novantrone®) is available commercially for clinical use and is sometimes used as a substitute for the more common Doxorubicin (DOX, an anthracycline) due to its somewhat reduced cardiotoxicity when compared to DOX.20 It has been recently investigated for the treatmen t of metastatic and hi gh-risk primary breast cancer by high-dose chemotherapy (HDCT) due to this reduced cardiotoxicity and other side effect considerations.21, 22 The dose-limiting toxicity of MXN for IV administration is myelosuppression, with other acute side eff ects occurring relatively infrequently at doses given clinically.22 MXN was chosen as the drug of choice for this research based on these reduced toxic effects and its broad app licability to a number of cancers. It also has a blue chromophore, making it easily de tectable in tissue hi stology and photometric analysis.

PAGE 31

12 OH OH O O N H N H N H OH N H OH Cl H 2 Figure 2.1: Mitoxantrone structure (MW=517.4 g/mol) MXN is active against solid tumors including breast and ovarian cancers, as well as many leukemias. In addition to standard IV tr eatments, intraperitoneal and intra-arterial delivery of MXN have also been evaluated.23-25 Though the exact mechanism is not thoroughly understood, it is believed that M XN works as an antineoplastic through the inhibition of DNA and RNA synthesis by intercal ating between base pa irs (Figure 2.2). The basic amino groups on the side chains of th e structure also allow electrostatic binding to phosphate groups in the DNA backbone, el ectrostatically cross-linking DNA strands.2527 It may also inhibit topoisomerase II, reducing DNA repair capabilities. MXN has been shown to kill more than 90% of cells in two hour exposure tests at a concentration of 0.5 g/mL with most cells killed within 1 hour in vitro .26, 27 The mechanism of action is not cell-cycle specific, but appear s to be more active during G2 phase.28 This phase of the cell cycle occurs between DNA synthesis (S phase) and mitosis (M phase). For illustrative purposes, a diagram of the cell cycle is presented in Figure 2.3. Mitoxantrone, like most cancer drugs, is typi cally delivered intravenously. In such cases, MXN is approximately 78% bound to plas ma proteins (primarily serum albumin), illustrating its albumin affinity. When delivered via IV infusion, MXN shows three phases of clearance, with half -lives of approximately 10 minutes, 1.1 to 1.6 hours, and 23 to 42 hours. Concentrations during the fina l phase are marginally above the cytotoxic level, indicating that the greatest cell kill bene fit is achieved within the first few hours of

PAGE 32

13 administration.25, 27 If MXN can be delivered locally over longer periods of time using a drug delivery device, it is clear that a si gnificant improvement in cell kill, and thus efficacy and potentially improved patien t quality of life, could be achieved. G C T A TA GC G C GC T A G C Planar ring structure of dru g Figure 2.2: Illustration on inter calation of a planar structure between base pairs in DNA. The exact mechanism through which MXN intercalates DNA is not understood, but there is an apparent preference for G-C base pairs. S G2 G1 M G0

PAGE 33

14 Figure 2.3: The cell cycle and phases. G1 i nvolves RNA and protein sythesis and is a period of normal metabolism. S phase involves DNA synthesis, G2 involves growth and further protein synthesis pr ior to cell division, and M phase is the mitotic phase, where the cell splits into 2 new cells. G0 denotes a prolonged or permanent G1 phase. (Figure adapted from Wilkes, et al.29) 2.2 Bovine Serum Albumin Albumin is the most abundant blood pl asma protein in the body, accounting for more than 60% of plasma proteins. The blood concentration is ~5%.30, 31 Albumin functions to bind, transport, and distri bute many different compounds throughout the body. These compounds are both endogenous an d exogenous compounds such as metal ions, fatty acids, amino acids, vitamins, and vi rtually all drugs. This feature, along with its inherent biocompatibility and abundance, make it also an ideal biopolymer for drug delivery systems. On average across species, albumin is comp rised of 585 peptide residues. Bovine serum albumin (BSA), which differs from hu man serum albumin (HSA) slightly in the number and proportion of amino acid residues, contains 583 residues and has a molecular weight of approximately 66,000 g/mol. Of th ese, 59 residues are lysine residues which contain primary amine functionality.32 This functionality enables BSA to be readily crosslinked using aldehydes.33, 34 Albumin microspheres have been investig ated for controlled delivery of various drugs and hormones, including insulin, anti-in flammatory drugs, antifungal agents, and drugs for chemotherapy. Though it has many advantages as a microsphere matrix material and one of the most widely studied biopolymers for such purposes, albumin is reported to have relatively low drug-loading capacities.33 Incorporation of anionic biopolymers can increase the loading efficien cy of BSA microsphere s and with cationic

PAGE 34

15 drugs may prolong drug release due to electr ostatic interactions between the drug and microsphere matrix.8, 9, 11, 14, 35 2.4 Gelatin Gelatin has been used in past and present research for controlled release systems in forms including microand nanosphere s for a variety of applications.36-38 Gelatin is attractive as a matrix for meso/micros pheres due to its minimal antigenity, biodegradation, and ready chemical cro sslinking through any nu mber of pendant functionalities.39 Gelatin is a natural biopolymer derived through denaturation of type I collagen. Type I collagen is found mainly in connec tive tissues such as skin, tendon, and bone.40 The primary structure of colla gen is that of a triple he lix, of which approximately 33% of amino acid residues are glycine a nd 25% are proline or hydroxyproline. It is this helical structure that is disrupted during denatura tion via the disruption of non-covalent bonds.39 Two of the three polypeptide ch ains in type I collagen ar e identical in composition. These two chains contain 1056 residues, wh ile the third contains 1038 residues. Due to its abundance in the sources desc ribed, gelatin is readily available and inexpensive. Additionally, its inherent biocompatibility a nd degradability make it very useful in drug delivery systems. One lim itation of gelatin is its relatively rapid dissolution in aqueous environments. This can, however, be overcome through chemical crosslinking using one or many of its pendant functionalities. Gl utaraldehyde, a common crosslinking agent and tissue fixative, reacts via a Schiff base reaction with primary amino groups of the lysine residues. Since gelatin contains ~ 30 lysine groups per 1000 amino acid residues it is readily cros slinked using glutaraldehyde.

PAGE 35

16 2.6 Deoxyribonucleic Acid Nucleic acids are biological polymers, w ith nucleotide monomers composed of a pentose sugar, phosphate group, and either a purine or pyrimidine nitrogen-containing base unit. Deoxyribonucleic acid (DNA) is specifically composed of 2Â’-deoxyribose sugars, phosphate groups, and combinations of two purine (adenine and guanine) and two pyrimidine (cytosine and thymine) bases. The phosphate group connects monomer units through a phosphodiester linkage between the phosphate group of one monomer unit and the sugar unit of the next, creating the regular backbone of the polynucleotide. The secondary structure of DNA is that of a doubl e helix, comprised of two strands of DNA. Purine-pyrimidine base pairs on the interior of the double helical st ructure of DNA serve to hold the double helix together via hydr ogen bonding (Figure 2.4). This helical structure results in a hydrophilic phosphate-s ugar backbone of each strand that lies along the outside of the helix in contact with the aqueous environment of the body and an overall negative charge at physiological pH. DNA has recently been studied as a biom aterial scaffold for tissue repair.41 Many of the same advantages that DNA offers for scaffold applications also make DNA potentially useful as a drug delivery biopolym er. These advantages include enzymatic degradation, hydrophilicity, and relatively abundance from animal and vegetable sources. Enzymatic degradation is particularly important as it allows biomaterial scaffolds and/or drug delivery devices to degrade in vivo , releasing drugs or ot her loaded compounds. When complexed with some drugs, DNA has been shown to reduce the inherent toxicity of these compounds. Trouet et al have s hown that when DNA-Adriamycin complexes are injected intravenously, they exhibit reduced toxicity, part icularly cardiotoxicity, and an increase in survival compared to IV administration of Adriamycin alone.15,16

PAGE 36

17 Figure 2.4: Schematic representation of DNA structure. Shading highlights the phosphate backbone of the structure. 2.6 Microsphere Synthesis Protein microspheres, such as bovine and human serum albumin, casein, and gelatin micro/mesospheres, have been synthe sized using a steric st abilization technique developed in this laboratory.11, 13, 14, 35 In this technique, steric stabilizers in the organic phase prevent particle coalescence through the duration of the crosslinking reaction. This technique is described in mo re detail in Chapter 3. Drug loading protein microspheres can be achieved through one of two methods: in situ loading or post loading. Post loading pro cedures involve the complete synthesis of microspheres in separate steps from drug load ing. Once particles are dried and collected, they are re-suspended in a drug solution for a given period of time. This not only lengthens the total synthesis time for MS, but also achieves relatively low drug loading. N C H N N C N N O O CH2 OH H O N C H N N C N H N H H O O CH2 O P H O O N C H N N C N N O O O CH2 O P H O O N C H N N C N H N H H O O P O O O O P O O O P O O O CH2 O P H O H H H H H H O P O O O P O O O N N O H O C H2 O P O O O H N H H H N N O H O C H2 O P O H C O H H H H N N O H H N O H O C H2 O P O H H N N O H OH C H2 O P O O O H C O H H H H O O O

PAGE 37

18 In contrast, in situ loading of drugs into microsphe res is done coincident with MS synthesis. Aqueous drugs such as MXN are included in the aqueous phase during synthesis and dried MS can be used without further modification. Th is type of loading has been shown to increase drug loading by at least 40% when compared to similarly synthesized post-loaded MS.14 2.7 Experimental Approach 2.7.1 Previous Meso/Microsphere Research Bovine Serum Albumin. Past in vivo animal studies involving BSA microspheres in Dr. GoldbergÂ’s research group have show n that IT injections are effective in prolonging survival and slowing or elimin ating tumor growth in both Lewis lung carcinoma and murine 16/C mammary ade nocarcinoma models. In the Lewis lung carcinoma model, up to 75% survival was obs erved in mice that were treated with microsphere formulations suspended in an MXN solution, with 92% survival when the same treatment was followed by surgery 10 days after IT MS drug delivery.42 In the mammary adenocarcinoma model, the LD50 for IT injection of in situ MXN loaded BSA microspheres (30-40 m diameter) was found to be roughl y three times higher than that of IT injected MXN. Additionally, cure rate s up to 80% were achieved with IT injections of some microsphere formulas in this model, compared to a 50% cure rate for free MXN delivered IT, and 0% for controls.13 All animals that receiv ed the highest BSA-M dose (48 mg/kg) survived at least 31 days following treatment, compared to just 16 days for IT MXN alone (8 mg/kg). These trends appear promising, though small group size prevented detection of a significant difference between free and BSA microsphere treatments. 13, 14

PAGE 38

19 Further in vivo animal research using BSA microsphe res of smaller particle size (510 m diameter) revealed that BSA micros pheres at a dose of 40 mg/kg MXN were statistically superior to the highest free-MXN dose (12 mg/kg; p=0.033) at prolonging survival. Nearly 80% of animals treated at the 40mg/kg BSA mi crosphere dose were cured at 60 days.13 Each of these formulations e xhibited release of MXN for no more than 24 hours, with the exception of one low crosslink density formulation which released MXN for approximately 120 hours at high doses.14, 43 Increased MXN loading and/or longer release times may impr ove survival outcomes further. Gelatin. Meso/microspheres synthesized usi ng a gelatin matrix have also been evaluated in vivo in a murine mammary adenocarcinoma model. Despite complete in vitro release of MXN from gelatin microspheres within 10 hours, in vivo results were promising. Mice treated with in situ MXN-loaded gelatin MS experienced prolonged survival compared to non-treatment controls . Several cases of toxicity were noted, however, and may be attributable to injection difficulties. On seve ral occasions gelatin MS swelled sufficiently to obstruct the need le during injection and consequently some animals received larger doses than intended as more pressure was applied to clear the obstruction. It is unclear whether toxicity in these animals is attributable to increased dosage or the extremely rapid release rate of the gelatin microspheres.12 This has limited the ability to achieve accurate survival results. 2.7.2 Research Goals The goal of the research proposed here wa s to synthesize mesospheres of various protein/polypeptide matrices that may achieve prolonged release and/or increased drug loading characteristics. Brie fly, Aims 1 and 2 involve the synthesis and ch aracterization

PAGE 39

20 of protein mesospheres (1-10 m mean diameter) of various matrix components and concentrations without and with in situ loaded MXN, respectively. Since it is known that crosslink density and partic le size affect drug release11, 13, 14, 33, 42, 44 these will be the primary parameters of interest evaluated in Aims 1 and 2. Aim 3 involves quantification of MXN loading and release from these microspheres, and Aim 4 evaluates the in vitro antiproliferative effect s of MXN-loaded microspheres on rat glioma-2 cells. Aim 5 investigates the ability of these form ulations to prolong survival in an in vivo murine mammary adenocarcinoma model. Additionally, statistical ev aluation of the survival and degree of airway obstruction results of 2 human clinical st udies examining IT chemotherapy treatment for patients with lung cancer was desired.

PAGE 40

21 CHAPTER 3 PROTEIN MESO/MICROSPHERE SYNTHESIS AND PROPERTIES 3.1 Introduction Protein microspheres have been prepared us ing a steric stabili zation developed in this lab.8, 11, 14, 34, 35, 42, 43, 45 Compositions have typically involved a single protein such as bovine serum albumin or protein-polypept ide blends as the microsphere matrix. Though these compositions have shown promise for chemotherapy in drug delivery applications, improved systems have been sought. For this reason, microspheres composed of gelatin-poly(glutamic) acid (G -PGA), gelatin-bovine serum albumin (GBSA), and bovine serum albumin-deoxyri bnucleic acid (BSA-DNA) blends were synthesized in this research; especially to optimize selected processing parameters for each of the compositions to prepare 1-10 m diameter mesospheres. 3.2 Materials All proteins and polypeptides were pu rchased from Sigma Chemical Company unless otherwise specified. Ge latin derived from calf skin was used in all gelatin containing microsphere compositions. B ovine serum albumin (BSA) fraction V lyophilized powder was used in all compositions utilizing BSA as a matrix component. Deoxyribonucleic acid from herring or sa lmon testes was used for DNA containing microsphere compositions. Due to inherent lot variation, obtaini ng uniform molecular weight DNA from herring testes was difficult. As a result, DNA from salmon testes was used as a replacement where noted. Glutar aldehdye (Grade II, 25% w/w aqueous) was obtained from Sigma Chemical Company and was used as the covalent crosslinking agent

PAGE 41

22 for all compositions. Cellulose acetate but yrate (CAB), 7% butyryl content obtained from Acros Organics was used as the st eric stabilizer duri ng synthesis in all compositions. Ultrapure water was used for all aqueous solutions and was prepared in the laboratory using a Barnstead/Thermolyne NANOpure Ultrapure Water system. A minimum resistivity of 17.0 M -cm was required for use in all experiments. HPLC grade methanol obtained from Fisher Scientific was used for particle size determination. All other solvents and salts were obtained from Fisher Scientific and were Certified A.C.S. grade unless otherwise specified. 3.3 Methods 3.3.1 Solution Preparation Phosphate Buffered Saline. Phosphate buffered saline (PBS) was prepared inhouse by mixing 0.1M sodium phosphate monobasic solution with 0.1M sodium phosphate dibasic solution until a pH of 7.4 is reached. The resulting PBS solution was filtered using 1 L vacuum filtration units with 0.22 m filter. Large stocks of PBS solution (4 L) were made in advance to avoi d daily variations in solution properties and were stored at 4 C until used. Bovine serum albumin. Due to the hygroscopic pr operty of albumin powder, BSA solutions were initially made in appr oximate concentrations and adjusted after solution characterization. Typically, BSA pow der absorbs ~10% of its weight in water from the atmosphere and also foams considerably upon mixing. As a result, BSA solutions were prepared by first weighing out ~10% excess of the desired amount of BSA. Ultrapure water was added in an appropr iately sized poly(propylene) conical in an amount slightly less than what was required for the final total volume. Using a vortex

PAGE 42

23 and bench-top rotator, solutions were allowed to mix completely at room temperature. Once solution was completely mixed centrifugation was used to collapse foam and the solutions were adjusted to the total final vol ume. Solution concentration (weight percent protein/solid) was quantified using a Me ttler LJ16 Moisture Analyzer at 130 C for 60 minutes. Any final adjustments to increase/de crease concentration were made and this process repeated until the solution was within a 5% difference tolerance level of the desired concentration. Solution density was de termined gravimetrically and was used in combination with moisture analysis results (w eight percent solid) to determine the weight per volume concentration of the solution. Gelatin. Like BSA, gelatin foams considerable during mixing. As a result gelatin solutions were prepared in much th e same manner as BSA solutions. However, since gelatin solubility in water at room temp erature is low, ultrapur e water was heated to approximately 30-35 C prior to adding to the appropriate quantity of gelatin (~110% of calculated quantity necessary). At room te mperature aqueous gelatin solutions will gel, inhibiting further mixing. As a result, periodic heating of the solution in a 40 C water bath was used as necessary to keep solution fluid during mixing using a vortex and bench-top rotator. Aqueous solutions for any microsphere composition using gelatin as a matrix material were prepared in this way, with addition of any other proteins/polypeptides once gelatin was nearly dissolved and additional heat as necessary to maintain fluidity. Deoxyribonucleic acid. DNA solutions were prepared using the same methodology as gelatin solutions outlined above . Due to its high molecular weight, DNA solutions above ~ 2% were very viscous and also required heating during the mixing

PAGE 43

24 process to maintain fluidity. Poly(glutamic) acid. PGA is very hygroscopic a nd picks up moisture from the atmosphere almost immediately when exposed to air in standard lab conditions. As a result, PGA was measured in a dry box unde r an Argon atmosphere with maximum 20% humidity directly into a 50 mL poly(propylene) conical. On ce the desired amount of PGA was weighed out, the conical was backfilled with argon and capped tightly for transfer from the dry box. Any additional solution components were weighed out separately on weigh paper and added to the conical cont aining PGA in ambient conditions. The remainder of the solution preparation remained the same as that for BSA and/or gelatin above. Glutaraldehyde. Glutaraldehyde was obtained as a 25% w/w aqueous solution but was required to be in or ganic solution for crosslinking du ring synthesis. To achieve this, the aqueous solution was vacuum distille d and the resulting dist illate was dissolved in 1,2-dichloroethane (DCE) to a concentration of 40 mg/mL. Cellulose acetate butyrate. Cellulose acetate butyrate solutions were prepared by dissolving solid CAB in DCE overnight with constant agitation in a volumetric flask. Concentrations of 3.0% w/v and 4.0% w/v were used in these studies. 3.3.2. Microsphere Synthesis Briefly, this method involves suspension of an aqueous solution (containing protein and other water soluble components, known as the aqueous phase) in an organic phase using a high speed paddle mixer. The organic phase contained steric stabilizers needed to prevent agglomeration or coalescence of par ticles until crosslinking is complete. After a given amount of time the crosslinking ag ent was added and the suspension stirred continuously while the crossli nking reaction takes place. The resulting crosslinked

PAGE 44

25 particles were washed with acetone and cen trifuged at 3000 rpm to force particles to settle out of suspension. This washing pro cess was repeated at least three times to remove organic and aqueous unreacted species. Once washing was completed, the particles were allowed to air dry for at least 48 hours. Th e resulting dried particles were then used for any necessary characterization methods. All above procedures were carried out at room temperature unless otherwise not ed. Figure 3.5 schematically illustrates this procedure. Aqueous Phase Organic Phase Crosslinking Suspension Aqueous Phase Organic Phase Aqueous Phase Aqueous Phase Organic Phase Crosslinking Suspension Crosslinking Crosslinking Suspension Figure 3.5: Illustration of susp ension crosslinking techniqu e for microsphere synthesis (diagram adapted from Cuevas 11). For all studies presented here, a 3% or 4% cellulose acetate bu tyrate (CAB) in 1,2dichloroethane (DCE) was used as the steric st abilizer in the organic phase. The aqueous phase contained various combinations of gelatin, PGA, BSA, DNA, and/or MXN as described in each study. A 40 mg/mL glut araldehyde (GTA)/DCE stock solution was used to crosslink particles in each case us ing the desired crossli nk density (w/w GTA to protein/polypeptide). Calculations of cross link density were based on the assumption that each GTA will primarily form crosslinks between lysine groups in two or more BSA or gelatin molecules, though it is underst ood that some intermolecular bonding may additionally occur.33, 34

PAGE 45

26 3.3.2.1 Gelatin-poly(glutamic) acid mesosphere synthesis Gelatin-poly(glutamic) acid mesospheres (G-PGA) were synthesized with minor changes to the previously desc ribed steric stabilization proc ess. Three mL of a solution containing the desired concentrations of ge latin and PGA was added to 47 mL of 3% CAB and mixed at 1250 rpm for 20 minutes. Du e to the gel nature of these solutions, synthesis was carried out at 40 C for initial portions of the synthesis procedure After 20 minutes, GTA was added to the mixing flas k to crosslink these compositions at a concentration of 4.5% (w/w) based on the exact concentrat ion of protein-polypeptide from moisture analysis and the mixing speed reduced to 600 rpm. The reaction was allowed to continue for 2 hr and 40 minutes longer and the resulting mesospheres were collected, washed, and dried as previous ly described in section 2.8. Table 3.1 summarizes the synthesis conditi ons for these compositions. Table 3.1: Synthesis condition summa ry for G-PGA mesospheres. Mixing Conditions RPM Time Temp (oC) 47 mL 3% CAB 1500 5 min 40 + 3 mL Gel/PGA soln 1500 5 min 40 heat 1500 10 min ambient (cooling) + x mL GTA 600 2 hr 40 min ambient TOTAL TIME3 hr The goal of this study was to determine th e effects of changing concentrations of gelatin and PGA on particle morphology, size, and swelling. Crosslink density, mixing speeds, and steric stabilizer concentration were chosen based on previous optimization studies using gelatin as the microsphere matrix material and were he ld constant for these studies.11 To attempt to keep the particle size in the 1-10 m range without changing other synthesis parameters, a total of 10% protein/biopolymer was used for all solution

PAGE 46

27 conditions. Gelatin mesosphere controls (G -C) were synthesized and characterized by Brian Cuevas in a parallel study using the same conditions. Table 3.2: Selected gelatin and PGA aqueous solution concentrations for synthesis of GPGA MS. Concentration (w/v) Condition Gelatin PGA Total G-C* 10% 0% 10% G-PGA 1 8.5% 1.5% 10% G-PGA 2 9% 1% 10% G-PGA 3 9.5% 0.5% 10% G-PGA 4 5% 5% 10% * synthesized by Cuevas in a parallel study11 3.3.2.2. Gelatin-bovine serum albumin mesosphere synthesis Using the same synthesis conditions desc ribed in Table 3.1, gelatin – bovine serum albumin MS (G-BSA) were synthesized. A so lution containing the desired concentration of gelatin and BSA was added in the second st ep instead of a gelatin-PGA solution. No other changes to this procedure were require d. Table 3.3 lists the conditions examined in this study. As with G-PGA a total of 10% pr otein was used for al l solution conditions. Gelatin mesosphere controls (G-C) were s ynthesized and characterized by Brian Cuevas in a parallel study using the same conditions. Table 3.3: Aqueous phase concentrations for G-BSA compositions Concentration (w/v) Condition Gelatin BSA Total G-C* 10% 0% 10% G-BSA 1 7.5% 2.5% 10% G-BSA 2 5% 5% 10% G-BSA 3 2.5% 7.5% 10% BSA-C 0% 10% 10% * synthesized by Cuevas in a parallel study11 3.3.2.3 Bovine serum albumin-deoxyribo nucleic acid mesosphere synthesis Bovine serum albumin-deoxyribonuclei c acid mesospheres (BSA-DNA) were synthesized using the previously describe d steric stabilizatio n method. Based on a

PAGE 47

28 previous research, 3 mL of BSA and DNA in designated proportions were added to 47 mL of 4% (w/v) CAB/DCE. Solutions containing DNA required heating to 30 C for 15 minutes prior to synthesis to reduce visc osity. A high speed paddle mixer set at 1250 rpm for 20 minutes was used to form a meso sphere suspension. A calculated amount of GTA was added to the suspension based on th e precise amount of protein and DNA in the aqueous phase and the desired crosslink dens ity. The stir rate was reduced to 600 rpm and the reaction was allowed to continue fo r an additional 2 hours and 40 minutes. The resulting MS were washed, dried, and collect ed as described prev iously (section 2.6). Table 3.4 summarizes this procedure. Table 3.4: Summary of steric stabiliz ation synthesis parameters for BDMS. Mixing Conditions RPM Time 47 mL CAB + 3 mL BSA / DNA1250 20 min + x mL GTA 600 1 hr 40 min + 50 mL acetone 600 1 hr TOTAL TIME3 hr An initial study was performed for feasibili ty and to help identify parameters of interest for a larger study. This study investigated 10% BSA – 1.5% DNA (BSA-DNA 1) as the aqueous phase using the synthesis conditions in Table 3.4. This condition was chosen specifically for comparison to prev ious mesospheres synthesized using this technique, namely those incorporatin g PGA and CMC into gelatin matrices.11 Following the pilot study and in order to determine the optimum parameter values for synthesizing mesospheres in the 1-10 m range, two parameters of interest were varied in a factorial design. These include the BSA:DNA ratio in the aqueous phase and crosslink density. Table 3.5 below lists all synt hesis conditions for the study. Note that ratios of BSA:DNA concentration that involv e a greater proportion of DNA compared to

PAGE 48

29 BSA were not analyzed, since DNA in its native helical form is not likely to crosslink sufficiently on an intermolecular level usi ng GTA due to a lack of primary amine functionality on the backbone. Previous unpub lished work by another researcher in Dr. Goldberg’s lab has shown that using DNA as th e sole mesosphere matrix material with GTA as a crosslinking agent resulted in aggl omeration of particles after the washing and drying processes, possibly i ndicating insufficient crossli nking under these synthesis conditions.46 For this reason it was believed that mesospheres formed with a greater proportion of DNA than BSA will likely not surv ive the washing and drying processes. Table 3.5: Aqueous phase concentrations fo r BSA-DNA compositions. X-Y: X refers to the BSA:DNA composition and Y refers to the crosslink density. Condition BSA Conc (w/v) DNA Conc (w/v) GTA Conc (w/w) BSA-DNA 110%1.5%4.33% BSA-DNA 2-10.67% BSA-DNA 2-24.33% BSA-DNA 2-3 7.5% 2.5% 8.00% BSA-DNA 3-10.67% BSA-DNA 3-24.33% BSA-DNA 3-3 5% 5% 8.00% BSA-C1 0.67% BSA-C2 4.33% BSA-C3 10% 0% 8.00% The effect of BSA:DNA ratio and cross link density on particle size, morphology, and swelling was studied. A BSA:DNA ratio of 10:0 was used as controls for all conditions, as these MS have been well ch aracterized in past experiments. 3.3.3 Particle Characterization 3.3.3.1 Particle morphology To qualitatively examine particle morphol ogy and particle size, scanning electron microscopy (SEM) was used. Dry particles were mounted on aluminum SEM stubs using double-sided tape. These stubs were coated with gold/palladium a lloy for 2.5 – 5 minutes

PAGE 49

30 (~ 300 Ã… coating) just prior to SEM imaging using a Technix Hummer V sputter coater. Parameter values that were used with a Je ol 6400 scanning electron microscope include an accelerating voltage of 5 KeV, a condens er lens setting of 8-10, and a working distance of 15mm. These parameters have been used for past studies with the assistance of Paul Martin.11, 13, 14, 47 Images were taken of ea ch mesosphere batch at 1000x, 2000x, and 3000x magnifications. In cases where addi tional characteristics or features needed further exploration, images at higher or lo wer magnifications were taken as deemed necessary. 3.3.3.2 Particle size analysis Average dry particle size was measured quantitatively using either a Coulter LS 230 or Coulter LS 13 320 particle sizer w ith small volume module, as previously described.11, 13, 14 Briefly, a 1% mesosphere suspension in 5 mL HPLC grade methanol was prepared and sonicated for at least 30 sec onds to facilitate complete dispersion just prior to analysis. The resu lting suspension was added dropw ise until obscuration reached 9-15%. Data were exported to Microsoft Excel for analysis, including mean, standard deviation, and particle size distri bution. It is important to not e that minimal to no particle swelling has been observed under these c onditions in past experiments. 3.3.3.3. Particle swelling A small amount (approximately 1-5 mg) of MS was placed on a standard glass slide and covered with a glass coverslip. Using an inverted optical microscope equipped with a high resolution digital cam era, at least three represen tative optical images were taken of dry mesospheres under 40x magnificati on. The slide was then removed from the microscope and 1 drop of PBS was adde d on the slide and a timer was started immediately. Once the coverslip was replaced, the slide was returned to the microscope

PAGE 50

31 stage and a representative area on the slide wa s selected. For initial examination of new compositions, optical images were taken at 1 minute following the addition of PBS, then at 30 second intervals for a total of 5 minutes , then at the 10 minute mark to investigate when full swelling was reached. If it appeared that full swelling was achieved within the first minute of adding PBS, images were taken only at 1 minute and 5 minutes posthydration for the remainder of similar compositions. The accompanying software has been calibra ted for use with the objectives on the microscope for measurement capabilities. Using this software, a minimum of 30 representative mesospheres on one representative dry image were measured. At least 30 mesospheres were measured in the same fa shion for the one-minute image. For new compositions, the same particles measured in the 1 minute image were measured on the 5 minute and 10 minute images. If it was discovered that all compositions appear to reach fully swollen size within 1 minute, only the dry image and 1 minute image were measured for the remainder of similar compositions. The mean particle diameter was calculate d for each optical image measured. Dry and hydrated particle size means were used to calculate the percent increase in particle diameter after full swelling was achieved. All particle size measurements from each image were used for statistical analysis to determine if differences in particle swelling existed between compositions. 3.3.4. Statistical Analysis Depending on the number of parameters of interest in a give n study, either a oneway or two-way analysis of variance (ANOVA) test was used to evaluate the effects of the desired parameters particle swelling. A signi ficance level of =0.05 was used for all ANOVA tests. TukeyÂ’s multiple comparis on test (MCT) was used to evaluate

PAGE 51

32 differences when applicable ( =0.05). Particle morphology a nd particle size distribution curves were evaluated only qualitatively and were not statistically analyzed. 3.4 Results and Discussion 3.4.1 Gelatin – Poly(glutamic) Acid Mesospheres Gelatin – PGA mesospheres (G-PGA) we re synthesized using the steric stabilization process described in sec tion 3.3.2.1. These studies were designed to determine the appropriate ratios of gelatin to PGA to achieve spherical microparticles in the 1-10 m average mean diameter. Based on th ese results, processi ng conditions were chosen for in situ MXN-loaded G-PGA (G-PGA-M). A total of four different concentration combinations of gelatin and PGA were examined to determine the effects of ch anging this parameter on particle size, morphology, and particle swelling. An initi al pilot study using an aqueous phase containing 5% gelatin and 5% PGA (G-PGA 4) wa s used to first determine the feasibility of a larger study. Processing parameters were set based on previous work by Cuevas for gelatin mesospheres and are listed in Ta ble 3.1. A concentrati on of 4.5% (w/w) of glutaraldehyde was used for crosslinking all compositions. Scanning electron microscopy results from the pilot study indicated that a number of particles were non-spherica l and appeared qualitatively mu ch larger than the desired mean particle size, with some pa rticles appearing as large as 30-40 m (Figure 3.6). Mean particle size was using the Coulter LS 230 particle size analyzer was measured to be 5.83 6.67 m. Though this appears to be with in the mean particle size range desired, the particle size distribution curve s howed there to be a la rge volume fraction of particles in the 20 – 40 m range (Figure 3.7). Particles of this dry size would likely be too large for injection upon swelling.

PAGE 52

33 Figure 3.6: SEM image of G-PGA 4 at 1000x magnification. Note the large nonspherical particles and the rough surface texture. Based on these results, three gelatin:PGA ratios were chos en for the larger study in addition to gelatin mesosphere experimental controls (no PGA content). Each set of parameters was run with 4 replicates, resulti ng in 16 total mesosphere batches. Aqueous phase concentrations are found in Table 3.2. All other processing conditions from the pilot study remained the same. Figure 3.7: Particle size distribution for G-PGA 4.

PAGE 53

34 Two different mixers were used during al l mesosphere synthesi s studies. Based on the SEM and particle size data , there appears to have been mixer variability introduced in these experiments, as batches run on each of these mixers were consistently different in particle size. Figure 3.8 show s representative SEM images ill ustrating this observation. Particle size distribution curves (Figure 3.9) also confirm this observation. Causes for these differences were not determined, but differences could have resulted from inaccurate mixer speeds since it is well known that mixer speed will affect particle size.11, 14 Differences could also be attrib uted to inaccurate or inconsistent temperature control by the thermocouple devices attached to each mixer setup. The gel nature of gelatin solutions make them highly temperature dependent and solution viscosity is also correl ated with particle size.14, 48 As a result of these discrepancies, mesospheres synthesized only using mixer 2 we re included for analysis resulting in a total of 8 mesosphere batches. Results from this point forward refer only to those G-PGA mesospheres synthesized on mixer 2 unless otherwise noted. Scanning electron microscopy images of G-PGA synthesized using mixer 2 show that mesospheres in all compositions appear to have mean diameters in the desired range of 1-10 m with few mesospheres exceeding 10 – 15 m. All G-PGA also appear to have a smooth texture on the surface of the part icles, and nearly all particles maintained a spherical shape. Some degree of clumping or agglomeration was both grossly observed and captured via SEM. All G-PGA compositions appear qualitativel y to have a large distribution of particle sizes, rang ing from submicron to roughly 10 m though there qualitatively appears to differences in pa rticle size distribut ion (Figure 3.10).

PAGE 54

35 Figure 3.8: Comparison of SEM images by high-speed mixer for G-PGA 1 composition. All images at 2000x magnification. G-PGA shown in A) and C) were synthesized using mixer 1 and those s hown in B) and D) were synthesized using mixer 2. Figure 3.9: Particle size di stribution comparison by highspeed mixer for G-PGA 1 composition. (M#a #b): #a denotes mixer number and #b denotes the batch number on that mixer. A B C D Mixer 1 Mixer 2

PAGE 55

36 Particle size observations from SEM imag es were confirmed using Coulter LS 230 particle size analysis. For nearly all GPGA compositions analyzed, the distribution curve centered on the 2 – 4 m range. Distribution curves ap pear to vary widely between distribution curves for G-PGA MS, even betw een batches of the same composition. All curves appear at least somewhat bimodal in the submicron – 20 m range, typical of microspheres synthesized using th is technique. In general, all peaks were broad with the first peak in the 0.4-2 m range and the second peak in the 4-10 m range. However, there was a significant peak at > 100 m for both batches of the G-PGA 3 composition. Typically particle sizes in this range are a result of agglomer ation of particles rather than the actual size of individual particles. Furthermore, no pa rticles of this size observed during SEM but some agglomeration was eviden t. To account for these observations, a particle size cut-off of approximately 50 m was set in the analysis of particle size data. Particle size data are tabulated in Table 3.6 and are illustrated graphically in Figure 3.12. There does not qualitatively appear to be an effect of changing Gel:PGA ratio on mean particle size, especially given the wide partic le size distributions and resulting standard deviations. Based on particle morphology and part icle size distribut ion, G-PGA 1 was identified as the most ideal of these compositions for in situ MXN loading. This composition has a narrower size distribution than the other compositions, with fewer particles larger than 10 m. Additionally, this compos ition has the highest amount of PGA in the mesospheres matrix and therefor e the highest potentia l for increased drug loading of cationic drugs such as MXN. Particle swelling was done on this composition for comparison with G-C. Figure 3.13 show s optical microscopy images of dry and

PAGE 56

37 swollen G-PGA 1 mesospheres at 40x magni fication. At least 30 randomly selected particles were measured on each image. Full swelling was achieved within 1 minute, as no further swelling was observed after this point . Mean particle diameter measurements are found in Table 3.7, and are represente d graphically in Figure 3.14. G-PGA 1 exhibited a 194% increase in particle diam eter (% IPD) from dry to swollen state compared to a 494% increase in size for G-C.11 Both compositions shared a similar dry particle size, but the G-C me sospheres swollen size was over twice that of G-PGA 1. Figure 3.10: SEM comparison of G-PGA co mpositions. All images taken at 3000x magnfication. Batch 1 Batch 2 G P GA3 G P GA 2 G P GA 1 A B C D E F

PAGE 57

38 Table 3.6: Particle size results for G-P GA compositions synthesized on mixer 2. Only data from particles in the size range indicated in Figure 3.11 are included. Sam p le M ean ( m ) SD ( m ) G-PGA 1a 2.4 2.9 G-PGA 1b 1.8 3.9 G-PGA 2a 1.8 5.1 G-PGA 2b 2.5 4.6 G-PGA 3a 2.2 5.5 G-PGA 3b 2.0 5.1 Figure 3.11: Particle size distribution comp arison by batch and G-PGA composition. An upper limit was set to ~50 m to avoid inclusion of agglomerates in analysis. 3.4.2. Gelatin – Bovine Serum Albumin Mesospheres Previous in vivo animal studies with G-C conducted by Cuevas uncovered a problem with the degree of swelling exhibited by these me sospheres. Several times during injection into subcutaneous tumors the needle became clogged and injection was either difficult or impossible. As a result of this finding and the swelling study data on G-C, gelatin-bovine serum albumin mesosphere blends (G-BSA) were synthesized using the same steric stabilization paramete rs as G-PGA (see section 3.3.2.1). Three concentration combinations of gelatin and BSA were examined, along with BSA

PAGE 58

39 controls. The GMS C mesosphe res synthesized by Cuevas in a parallel study were again used as gelatin controls. Concentrations for these compositions are found in Table 3.3. The goal of this study to was to determine th e effects of changing c oncentrations of BSA and gelatin on particle si ze, morphology, and swelling. 2.4 1.8 2.2 1.8 2.5 2.00.0 0.5 1.0 1.5 2.0 2.5 3.0 G-PGA 1 G-PGA 2 G-PGA 3Particle Size (um ) Batch 1 Batch 2 Figure 3.12: Batch and composition comparison of mean particle size using Coulter LS 230 particle sizer. Table 3.7: Mean particle diameter for G-PGA 1 from swelling study. Time Mean Particle Diameter ( m ) 0 (dry) 3.1 1.5 1 minute 8.9 6.9 5 minutes 9.0 6.7

PAGE 59

40 Figure 3.13: G-PGA 1 composition optical microscopy images A) dry and B) swollen at 1 minute post-hydration. Particle measurements are indicated on each image. A B

PAGE 60

41 3.13.2 8.9 19.0 0.0 5.0 10.0 15.0 20.0 G-PGA 1 G-CParticle Size (um ) Dry Swollen Figure 3.14: Dry and swollen mean particle diameter for G-PGA 1 and G-C. G-PGA 1 exhibited a 194% increase in particle diameter upon hydration, compared to 494% increase in size for gelatin controls . Raw data were not available for GC MS; as a result no error bars are included on those corresponding bars. Particle morphology analysis via SEM ag ain revealed some variation between mixers in this study. Two batches, both s ynthesized using mixer 1, appeared to have a much higher average particle size. Fi gure 3.15 shows SEM images for G-BSA 1 and illustrates this observation. Similar observa tions were made for G-PGA 2 (images not shown). Particle size distri butions further confirmed thes e observations, with particle size distribution of one batch in each of these compositions skewed heavily toward particles in the 40 m range (Figure 3.16). As with G-PGA, these batches were removed from further analysis (unless otherwise noted) and were attributed to inconsistent mixer speed and/or temperature control. Representative images of all G-BSA compositions are found in Figure 3.17. Overall, particles appear to be smooth a nd spherical for all co mpositions, with some

PAGE 61

42 mesospheres in each composition with rough surface texture. In comparison with GPGA, more particles appear to be ~10 m in diameter. No differences in particle size between compositions were detect ible qualitatively. Results from particle sizing analysis using the Coulter LS are found in Table 3.8. All compositions produced mesospheres with mean diameter of ~ 2 m. Particle size distributions for all compositions are found in Figure 3.18. Nearly all batches and compositions exhibited the usual bimodal distribution. In comparison to G-PGA com positions, the peaks in these distributions appeared narrower and the second peak of these distributions is shifte d, centered in the 620 m range, indicating that synthesis of these MS resulted in a larger number of particles over 10 m in diameter, as was visualized using SE M. In general synthesis of the G-BSA composition MS was more reproducible, with distributions from each batch in a given composition more closely resembling one another. Figure 3.15: Batch comparison for G-BSA 1: A) batch 1, B) batch 2, C) batch 3, and D) batch 4. As with G-PGA, mixer 1 pr oduced mesospheres with much larger average particle size (C). This phe nomenon was also observed for G-BSA 2 (not shown). All images taken at 2000x. A B C D

PAGE 62

43 Figure 3.16: Batch comparison of particle si ze for A) G-BSA 1 and B) G-BSA 3. Both compositions show one batch with skewed particle size distributions curves relative to the remaining batches. Th ese batches were both synthesized on mixer 1, which produced G-PGA mesospheres with similar results. Particle swelling studies re vealed differences between particle sizes of G-BSA compositions. One-way ANOVA of dry part icle size indicated that significant differences existed between several GBSA compositions (p<0.001). Tukey MCT revealed that the dry particle size of GBSA 1 was different fr om both G-BSA 2 and GBSA 3, but not different from BSA-C contro ls. The only additional differences were between G-BSA 2 and BSA-C. Full swelli ng was achieved within 1 minute of hydration with PBS for all compositions. As a result , only images taken at 1 minute post hydration were used for further analysis. No difference in particle diameter at 1 minute were found A B

PAGE 63

44 between any groups (p=0.222). It is importa nt to note that G-C mesospheres were not included in statistical analyses since raw sw elling data were unavailable. Based on the statistical results, it is likely that G-C swollen particle size would have differed significantly from all other compositions, as it is over twice that of all the others. Figure 3.17: Representative SEM images of A) G-BSA1, B) G-BSA 2, C) G-BSA 3, and D) BSA-C. All images taken at 2000x magification. Table 3.8: Mean particle size for G-BSA co mpositions as determined by Coulter LS particle size analysis. Sam p le M ean ( m ) SD ( m ) Sam p le M ean ( m ) SD ( m ) G-BSA 1a 1.85 4.34 G-BSA 2a 2.54 3.66 G-BSA 1b 2.16 4.55 G-BSA 2b 2.85 2.77 G-BSA 1c 2.02 3.66 G-BSA 2c 2.13 4.15 G-BSA 3a 2.27 3.48 BSA-C a 1.98 3.77 G-BSA 3b 2.16 3.14 BSA-C b 1.71 4.46 G-BSA 3c 2.26 3.11 BSA-C c 1.62 4.76 G-BSA 3d 2.73 3.43 BSA-C d 1.40 4.54 A B C D

PAGE 64

45 Figure 3.18: Particle size distributi on curves for A) G-BSA 1, B) G-BSA 2, C) G-BSA 3, and D) BSA-C. D B C A

PAGE 65

46 Table 3.9: Percent increase in mean partic le diameter (% IPD) for G-BSA compositions based on particle swelling studies. Condition Dr y ( m ) Swollen ( m ) % IPD G-BSA 1 6.7 2.79.9 4.4 48% G-BSA 2 3.8 1.67.3 3.0 93% G-BSA 3 5.0 2.27.9 3.5 59% BSA-C 5.7 2.38.7 3.6 53% G-C 3.2 19.0 494% 3.4.3. Bovine Serum Albumin – Deoxyribonucleic Acid Mesospheres Bovine serum albumin – deoxyribonucle ic acid mesospheres (BSA-DNA) were synthesized using the steric stabilization pr ocess described in sect ion 3.3.2.3. This study was designed to determine the appropriate ra tios of BSA to DNA to achieve spherical microparticles in the 1-10 m average mean diameter. Based on these results, processing conditions were chosen for in situ MXN-loaded BSA-DN A MS (BSA-DNA -M). An initial pilot study was run to test the f easibility of synthesizing BSA-DNA MS and to identify the appropriate proces sing parameters for this synthe sis. For direct comparison to previous mesosphere blend studies, an aqueous phase with 10% BSA and 1.5% DNA was chosen (BSA-DNA 1). Since BSA mesosp heres have been synthesized in the past and their processing parameters characterized in detail, the processing conditions used by Almond to produce 1-10 m average diameter mesospheres were chosen for this study. Analysis of particle morphology using SE M validated the feas ibility of using DNA as a mesosphere matrix material. Spherical mesospheres with smooth surface texture were synthesized. Qualitatively all mesos pheres appeared to fall in the desired 1-10 m range, with few to no particles larger than ~ 10 m (Figure 3.21). These observations were confirmed via particle size analysis. Mean particle size for BSA-DNA 1 was 1.3 4.5 m. The particle size distribu tion is found in Figure 3.22.

PAGE 66

47 Figure 3.19: Representative optical microscopy images of G-BSA MS from swelling study. The image on left is of A) G-BSA 2 dry mesospheres; the image on the right is the corresponding image of swollen G-BSA 2 mesospheres at 1 minute posthydration. A B G

PAGE 67

48 6.7 3.8 5.0 5.7 3.2 9.9 7.3 7.9 8.7 19 0.0 5.0 10.0 15.0 20.0 G-BSA 1 G-BSA 2 G-BSA 3 BSA-CG-CParticle Diameter (um ) Dr y Swolle n # + + # Figure 3.20: Particle swelling comparison by G-BSA composition. Matching symbols on dry particle sizes indicate statistical ly significant differences between those groups. No statistical differences were detected between th e swollen sizes of any of the G-BSA compositions. All G-BSA compositions exhibited less than 100% IPD, compared to almost 500% for G-C. Raw data for G-C were not available and could not, therefore, be included in the statistical analysis. Particle swelling was also analyzed for these mesospheres before continuing with drug loading and a larger st udy since it was unknown if the incorporation of DNA would increase the swellability of BSA mesosphere s. Results of the pa rticle swelling study indicated a significant difference between dry and swollen BSA-DNA 1 particles (p<0.001), with mean diameters of 5.0 2.1 m and 8.6 3.8 m, respectively (Figure 3.23 and Figure 3.24), corresponding to an ove rall 73% increase in diameter upon swelling. As with previous compositi ons, BSA-DNA 1 mesospheres reached fully swollen size within 1 minute of the addition of PBS.

PAGE 68

49 Figure 3.21: SEM image of BS A-DNA 1 at 2000x magnification. Figure 3.22: Particle size distribution for BSA-DNA 1. Based on these promising results, a more complete study was designed to determine the effects of changing BSA a nd DNA concentration ratios and crosslink density on particle size, morphology, and sw elling. Table 3.5 lists the conditions examined in this study.

PAGE 69

50 8.6 5.0 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 BSA-DNA 1Particle Diameter (um ) Dry Swollen Figure 3.23: Dry and swollen mean particle diameters based on swelling measurements using optical microscopy. BSA-DNA 1 MS dry and swollen particle sizes were significantly different fr om one another (73% IPD). Representative SEM images of BSA-DNA are found in Figure 3.25. All compositions containing DNA (BSA-DNA 2 and BSA-DNA 3) exhib ited increasingly spherical morphology with increasing crossli nk density. Higher crosslink densities produced smooth spherical mesospheres fo r each of these compositions. BSA-C compositions were spherical for all crossl ink densities, but surface roughness increased with corresponding increases in crosslink de nsity. Additionally, the majority of BSA-C qualitatively appeared to fall in the 1-10 m range particle diameter while BSA-DNA 2 and BSA-DNA 3 appeared to have larger mean particle diameters than BSA-C.

PAGE 70

51 Figure 3.24: Dry (A) and swollen (B) BSA-DNA 1 mesospheres via optical microscopy at 40x magnification. Swollen image taken at 1 minute post-hydration with PBS. A B

PAGE 71

52 Representative particle size distribution cu rves from Coulter LS particle sizing are found in Figure 3.26. Particle size distri butions between batches of the same compositions were more similar to one another than those for G-PGA MS in terms of shape and peak location, though not perfectly aligned (not shown). All curves show bimodal distributions, with most particles found in the 4-20 m range regardless of composition, though DNA containing compositions a ppear to have a slight of the peak shift toward 20 m compared with BSA controls. A summary of mean particle size data are found in Table 3.10. Particle swelling was analyzed for differences between compositions and crosslink density using two-way ANOVA for bot h dry and swollen particle sizes. No significant interactions were detected for dry particle size (p= 0.073), though there were differences in both composition (p=0.009) and crosslink density (p=0.013). Tukey MCT revealed that BSA-DNA 3 mesospheres had a significantly larger mean dry particle diameter than BSA-DNA 2 (p=0.009) overall and that mesospheres with the highest crosslink density had significantly larger dr y particle size than those with the middle crosslink density (p=0.009). No other differences were detected for dr y particle sizes. A significant interaction was detected for swollen particle sizes, indica ting that the swollen particle size depends not only of compositi on but also on crosslink density (p<0.001). Tukey MCT indicated that within high crossl ink density group there were no differences detected in composition. Howe ver, BSA-DNA 3 had larger swollen particle diameter than BSA-DNA 2 for both middle and low crosslink densities (p=0.038 and p<0.001, respectively). Examination of differences due to crosslink de nsity within each composition revealed that the highest crosslink density resulted in larger swollen particle

PAGE 72

53 size than the middle and low crosslink dens ities for BSA-DNA 2 compositions (p<0.001 in both cases). Within the BSA-DNA 3 com position, the highest cro sslink density again resulted in larger swollen particle diameter than that of the middle crosslink density (p=0.014). Additionally, the middle crosslink de nsity was found to be significantly larger than the low crosslink density within the BSA-DNA 3 composition (p<0.001). These results are represented graphically in Figure 3.27 and Figure 3.28. Though all DNA containing formulations experienced less than 80% IPD, their large mean swollen particle size indicates that injection of these particles may be difficult in some cases. Representative optical microscopy images from this study are found in Figure 3.24 and Figure 3.29 – Figure 3.31.

PAGE 73

54 Figure 3.25: Representative SEM images of BSA-DNA compositions. Each row repr esents one composition, with increasing crosslink density from left to right; A) – C) BSA-DNA 2, D) – F) BS A-DNA 3, G) – I) BSA-C. Increasing crosslink density A B C D E F G H I

PAGE 74

Figure 3.26: Comparison of representative particle size distribution curves by MS composition and crosslink density: A) BSA-DNA 2, B) BSA-DNA 3, and C) BSA-C. Each of these contains curves for low (X-1), medium (X-2), and high (X-3) crosslink densities. Note the ch ange in scale on part A compared with parts B) and C). A B C

PAGE 75

56 Table 3.10: Mean particle size data for BSADNA MS determined by Coulter LS particle sizing. Sam p le M ean ( m ) SD ( m ) BSA-DNA 2-1a 6.04 5.62 BSA-DNA 2-1b 10.23 9.31 BSA-DNA 2-1c 10.56 8.53 BSA-DNA 2-2a 18.93 13.85 BSA-DNA 2-2b 9.51 8.83 BSA-DNA 2-2c 11.86 11.78 BSA-DNA 2-3a 5.18 3.94 BSA-DNA 2-3b 14.66 12.49 BSA-DNA 2-3c 9.20 8.89 BSA-DNA 3-1a 10.25 9.77 BSA-DNA 3-1b 10.97 12.67 BSA-DNA 3-1c 10.96 15.00 BSA-DNA 3-2a 6.99 7.79 BSA-DNA 3-2b 10.48 13.19 BSA-DNA 3-2c 7.57 7.05 BSA-DNA 3-3a 9.46 8.92 BSA-DNA 3-3b 6.59 6.85 BSA-DNA 3-3c 6.91 8.43 BSA-C 1-1a 5.71 5.35 BSA-C 1-1b 6.97 5.39 BSA-C 1-1c 5.25 3.10 BSA-C 1-2a 3.63 3.48 BSA-C 1-2b 4.19 4.06 BSA-C 1-2c 4.15 3.79 BSA-C 1-3a 4.11 4.06 BSA-C 1-3b 4.27 4.30 BSA-C 1-3c 4.34 4.26 Table 3.11: Mean dry and swollen particle size measurement from optical microscopy imaging. All BSA-DNA compositions swell to less than 80% IPD, much less than that observed for G-C (494% IPD). Condition Dr y ( m ) Swollen ( m ) % IPD BSA-DNA 2-1 7.86 3.4010.48 4.14 33% BSA-DNA 2-2 8.39 5.059.99 4.76 19% BSA-DNA 2-3 9.92 4.3017.82 6.86 80% BSA-DNA 3-1 10.72 5.6818.97 7.92 77% BSA-DNA 3-2 8.67 4.2312.85 5.76 48% BSA-DNA 3-3 10.55 5.2516.22 6.86 54% BSA-C 1 6.25 2.4111.16 4.29 78% BSA-C 2 6.34 2.397.42 3.26 39% BSA-C 3 5.56 2.597.72 4.14 33%

PAGE 76

57 Figure 3.27: Particle swelling charts comp aring the effect of DNA content on dry and swollen sizes for each composition by crosslink density; A) low crosslink density, B) medium crosslink dens ity, and C) high crosslink density. 6.3 7.9 10.7 11.2 10.5 19.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 BSA C-1 BSA-DNA 2-1 BSA-DNA 3-1Particle Size (um ) Dry SwollenA 6.3 5.0 8.4 8.7 7.4 8.6 10.0 12.8 0.0 5.0 10.0 15.0 20.0BSA C-2BSA-DNA 1-2 BSA-DNA 2-2 BSA-DNA 3-2Particle Size (um ) Dry SwollenB 5.6 9.9 10.6 7.7 17.8 16.2 0.0 5.0 10.0 15.0 20.0 25.0 30.0 BSA C-3 BSA-DNA 2-3 BSA-DNA 3-3Particle Size (um ) Dry Swollen C

PAGE 77

58 Figure 3.28: Particle swelling ch arts comparing the effect of crosslink density on dry and swollen sizes for each composition by DNA content; A) 0% DNA, B) 2.5% DNA, and C) 5% DNA content. 6.36.3 5.6 11.2 7.4 7.7 0.0 5.0 10.0 15.0 20.0 BSA C-1 BSA C-2 BSA C-3Particle Size (um ) Dry SwollenA 7.9 8.4 9.9 10.5 10.0 17.8 0.0 5.0 10.0 15.0 20.0 25.0 30.0 BSA-DNA 2-1 BSA-DNA 2-2 BSA-DNA 2-3Particle Size (um ) Dry SwollenB 10.7 8.7 10.6 19.0 12.8 16.2 0.0 5.0 10.0 15.0 20.0 25.0 30.0 BSA-DNA 3-1 BSA-DNA 3-2 BSA-DNA 3-3Particle Size (um ) Dry Swollen C

PAGE 78

59 Dry Swollen A B C D E F Figure 3.29: BSA-DNA 2 swelling study optical microscopy images (40 x). Image on the left are of dry BSA-DNA 2 mesosphere s. Corresponding swollen images are on the right.

PAGE 79

60 Dr y Swollen A B C D E F Figure 3.30: BSA-DNA 3 swelling study images. Im ages on the left are of dry particles; those on the right are swollen, taken at 1 minute post-hydration. A

PAGE 80

61 Dr y Swollen A B C D E F Figure 3.31: Optical microscopy images of BSA-C

PAGE 81

62 CHAPTER 4 EVALUATION OF IN SITU DRUG LOADED PROTEIN MESO/MICROSPHERE COMPOSITIONS 4.1 Materials All proteins and biopolymers were purch ased from Sigma Chemical Company unless otherwise specified and were the same as those indicated in section 3.2 unless otherwise noted. HPLC grade methanol obtai ned from Fisher Scientific was used for particle size determination. All other solven ts and salts were obtained from Fisher Scientific and were Certified A.C.S. grade unless otherwise specif ied. Ultrapure water was obtained in the lab using the Barnstead NANOpure water system. Mitoxantrone (MXN) was obtained from Sigma Chemical Company (purity 97%) and was used for in situ microsphere loaded MS w ithout further modification. Enzymes and salts used for enzymatic digestion of MXN-loaded microsphere compositions were obtained from Sigma Chemical Company and used as supplied. Constituents of the digestion buffer included papain (from papaya latex source), bacterial protease Type VIII, ethylenediaminetetraac etic acid disodium salt:dehydrate (EDTA), and L-cysteine hydrochloride hydrate. Fo r compositions including DNA as a matrix material deoxyribonuclease (DNase) Type I fr om bovine pancreas was also used in the enzyme cocktail. Trichloroacetic acid was also during analysis of microsphere digests. In vitro RG-2 cell culture medium was made using DulbeccoÂ’s Modified Eagle Medium (DMEM), fetal bovine serum (FBS), a nd penicillin streptomycin (P-S) obtained

PAGE 82

63 from Gibco. Trypsin for cell detachment wa s obtained from Cellgro. RG-2 cells were obtained from the lab of Dr. Wolfgang Streit. 4.2. Methods 4.2.1. Solution Preparation Gelatin-poly(glutamic) acid-mitoxantrone. These solutions were prepared in the same manner as those presented in section 3.3.1. Prior to synthesis, 3 mL of this solution was isolated after characte rization and heated to 30 C for 30 minutes. To this, the appropriate quantity of MXN was added and allowed to di ssolve at room temperature, with additional periodic heati ng provided as necessary to maintain fluidity. Gel-PGAMXN solutions were heated a final time to 30 C just prior to dispersion to facilitate addition into the synthesis flask. Bovine serum albumin-deoxyribonucleic acid-mitoxantrone. Based on the known mechanism of action for MXN and on preliminary studies used to identify appropriate synthesis conditions, DNA a nd MXN cannot be combined in a single aqueous phase as the solution gels due to the interaction between the two. To account for this, aqueous solutions containing both BS A and MXN were prepared independently from DNA aqueous solutions. The BSA-M XN and DNA solutions were prepared as indicated in 3.3.1. Each of these solutions were made at a concentration twice that of the desired final concentration and were added in two separate steps to the synthesis flask in equal volumes to result in the final desire d concentration of e ach constituent. Enzymatic digestion buffer. An enzyme digestion buffer (EDB) was used for determination of drug loading in in situ MXN-loaded compositions. The buffer solution was prepared based on previous work by Habda14 and was prepared by dissolving 720 mg

PAGE 83

64 of EDTA, 80 mg L-cysteine, 50 mg papain, and 50 mg bacterial protease type VIII in 100mL of previously prepared 0.1M PBS (pH 7.0 ) at room temperature. When used with microspheres containing DNA, 50 mg of DNAse type I was also added to the buffer. EDB was prepared fresh on the day the diges tion study was to be in itiated to ensure maximum enzyme activity. 4.2.2. In Situ Mitoxantrone-Loaded Microsphere Synthesis 4.2.2.1. In situ MXN-loaded gelatin – poly(g lutamic) acid microspheres The same methodology used for synthesis of unloaded G-PGA MS was used for in situ loaded MXN gelatin – poly(glutamic) aci d microspheres (G-PGA-M), with 15% w/w MXN contained in the aqueous solutions. As with the unloaded GPGA, G-PGA-M were synthesized utilizing various ratios of gela tin and PGA in the aqueous dispersed phase (Table 4.12). Based on the results of th e unloaded G-PGA studies an aqueous phase containing 8.5% gelatin and 1.5% PGA (w/v ) was chosen. This composition had a narrower size distribution than other ratios investigated. Additionally, this composition contained the highest amount of PGA of those successfully synthesized and thus had the highest potential for increased MXN loading. For direct comparison to previous gelatinCMC blend microspheres, a second aqueous phase containing 10% gelatin and 1.5% PGA was also investigated. For these studi es a single crosslinke r concentration (4.5% w/w) was utilized to determine differences in MS properties based solely on the microsphere matrix components. Table 4.12: Solution parameters for G-PGA-M compositions Condition Gel Conc (w/v) PGA Conc (w/v) MXN Conc (w/w) G-PGA M-1 8.5% 1.5% 15% G-PGA M-2 10% 1.5% 15%

PAGE 84

65 Briefly, the steric stabiliza tion process was used to s ynthesize G-PGA-M using the solution and heat parameters listed in Table 4.13. Prior to MS synthesis, Gel-PGA-MXN solutions were heated at 30 C for 30 minutes or until fluid enough to be transferred into the mixing flask. G-PGA-M were washed a nd collected via centrif ugation at 2000 rpm in a series of four acetone wash es and were subsequently airdried, as described in section 2.8. Table 4.13: Synthesis conditions for G-PGA-M compositions Mixing Conditions RPM Time Temp (oC) 47 mL 3% CAB 1500 5 min 40 + 3 mL Gel-PGA-MXN soln1500 5 min 40 heat 1500 10 min ambient (cooling) + x mL GTA 600 2 hr 40 min ambient TOTAL TIME3 hr 4.2.2.2. In situ MXN-loaded bovine serum albumin – deoxyribonucleic acid microspheres BSA-DNA-M microspheres were synthesized via a modification of the steric stabilization method used for BSA-DNA micros pheres. Mixer speeds were adapted from previous studies on BSA microsphere synthesis.13, 14 One and one-half mL of the appropriate BSA-MXN aqueous solution (twice the desired final concentration) was dispersed in 47 mL of 4% CAB and allowe d to stir on the high speed mixer for 10 minutes. An equal volume (1.5 mL) of aque ous DNA solution of twice the desired final concentration was then added and allowed to mix for an additional 10 minutes. The addition of DNA in a second step while stirring eliminated the gelation that occurs when DNA and MXN are added into the same solution prior to mixing. All remaining steps in the synthesis remain the same as those prev iously described for unloaded BSA-DNA MS. Synthesis was carried out at ambient temperat ure. Table 4.14 summarizes this synthesis.

PAGE 85

66 Table 4.14: Summary of synthesi s conditions for BSA-DNA-M. Mixing Conditions RPM Time 47 mL CAB + 1.5 mL BSA / MXN 1250 10 min + 1.5 mL DNA 1250 10 min + x mL GTA 600 1 hr 40 min + 50 mL acetone 600 1 hr TOTAL TIME3 hr For direct comparison to past experiments us ing in situ loaded MXN with gelatin or BSA microspheres, a 15% (w/w) concentration of MXN was used in these experiments. In order to keep the number of samples to a minimum, MXN concentration was not introduced as a variable. Based on nonspherical particle morphology results from unloaded BSA-DNA MS, the lowest GTA con centration was not evaluated. Table 4.15 summarizes the solution parameters investigated in this study. Table 4.15: Solution and synthesis para meters for the BSA-DNA-M compositions. Condition BSA Conc (w/v) DNA Conc (w/v) MXN Conc (w/w) GTA Conc (w/w) BSA-DNA MC-2 4.33% BSA-DNA MC-3 10% 0% 15% 8.00% BSA-DNA M2-2 4.33% BSA-DNA M2-3 7.5% 2.5% 15% 8.00% BSA-DNA M3-2 4.33% BSA-DNA M3-3 5% 5% 15% 8.00% 4.2.3. Particle Characterization MXN-loaded microspheres were charact erized using the same methodologies outlined in section 3.3.3. In brief, particle morphology was qualitatively analyzed using SEM at 1000x, 2000x, and 3000x magnifications. Average dry particle size was qualitatively determined using a Coulter LS pa rticle sizer. This technique was also used to qualitatively analyze the particle size distribution of each batch and composition. Average particle swelling was determined us ing optical microscopy images of dry and

PAGE 86

67 swollen (in PBS) microspheres, with at least 30 microspheres measured using accompanying software. 4.2.4. Drug Content The concentration of MXN contained in in situ loaded microsphere compositions was evaluated using enzymatic digestion, followed by photometric analysis of the resulting digests. The enzymatic digestion bu ffer was prepared just prior to drug loading experiments as described in section 4.2.1. Approximately 5 mg of MXN-loaded microspheres were incubated in 10 mL of fres hly prepared EDB for at least 4 days at 37 C under constant rotation. Each microsphere composition was examined in triplicate along with non-microsphere contro ls. Controls each contained 200 L of a previously prepared 1000 g/mL aqueous drug solution in 10 mL of EDB and were used to evaluate drug degradation during the incubation peri od. After the initia l incubation period, samples were examined visually and using optical microscopy to determine if microspheres were fully digested. If any sa mples were not fully digested, these were returned to the incubator and checked ever y 48 hours until the microspheres were fully digested or at least 14 days had passed. At the end of the final incubation period samples were allowed to return to room temperatur e on a bench-top rotator. Two mL of each sample was added to 2 mL of 10% trichloroa cetic acid (TCA) and was allowed to react at room temperature for 30 minut es to precipitate dissolved protein. Samples were then centrifuged at 2000 rpm for 10 minutes. One mL supernatant aliquots were taken from each sample for UV-Vis analysis at max=610 nm using a Shimadzu UV-2401PC UV-Vis spectrophotometer. Mitoxantrone standards prepared in a 5% TCA solution ranging in

PAGE 87

68 concentration from 1 to 50 g/mL were used to prepare a linear calibration curve for quantification of unknown M XN concentrations. 4.2.5. In Vitro Mitoxantrone Release Approximately 2 mg of each composition of BSA-DNA-M were transferred into a 2 mL centrifugal filter tube (Ultrafree-CL centrifugal f ilter device; low binding 0.1 m membrane). These devices allow the microsphere s to be incubated in up to 2 mL of fluid, then separated from the supernatant using centrifugation. Samples were incubated at 37 C in 1 mL of phosphate buffer saline (PBS) under constant rotation. At predetermined timepoints, filter tubes were removed from the incubator and centrifuged for 10 minutes at 2000 rpm for aliquot collectio n. Aliquots were removed from tubes and replaced with 1 mL PBS, then samples return ed to the incubator until the next timepoint. UV-Vis analysis of aliquots was con ducted using a Shimadzu UV-2401PC UV-Vis spectrophotometer at max=610 nm against a matched matrix background. Absorption results were compared to previously anal yzed MXN concentration standards ranging from 1 to 50 g/mL to determine concentration of dr ug released. These data were then used to generate a release curve illustrating MXN release over time. This type of release conditions was used instead of the infinite si nk conditions used in th e past in order to more closely replicate the low fluid volume turnover tumor. 4.2.6. RG-2 Cell Culture Rat glioma 2 (RG2) cells were cultured fr om frozen stock in complete media and incubated at 37 C with humid 8% CO2 atmosphere. To begin the study, cells were harvested using standard cell culture techniques and counted using a hemocytometer. Cell concentration was adjusted to 3x104 cells/mL using complete media and 75 L of

PAGE 88

69 the resulting cell suspension was added to each well needed in five 96-well plates. These plates correspond to treatment days 0-4 for M TT. Plates were then incubated overnight to allow cells to attach to well bot toms. The following day (Day 0), 75 L of the appropriate treatment was added to each well in plates designated for days 1-4. Optical microscopy images were taken of Day 0 plate as baseline images for the study, then 10 L of MTT labeling reagent was added to each well and the plate returned to the incubator. After at least 4 hours, 100 L of solubization buffer was added to each well and the plate returned to the incubator overnight. The following day the plate was removed and the absorbance read using a micr oplate reader. The op tical microscopy and MTT steps above were repeated on each of th e remaining plates on their designated days. At the conclusions of the study, absorption data were then plotted over time to visualize the antiproliferative effects of MXN release as a function of the para meters of interest. An increase in absorption va lues corresponds to increas ed cellular proliferation. Treatment groups for these studies incl uded a non-treatment control (C), free MXN (F-MXN), and in situ MXN-loaded microspheres, with MXN doses of 0.5 g/mL, 12.5 g/mL, and 25 g/mL for all drug containing groups . These doses refer to total MXN dose based on loading and release stud ies, not microsphere dosage and were chosen for direct comparison to previous studies.11, 47 4.3 Results and Discussion 4.3.1. In situ MXN-loaded Gelatin – Poly(glutamic) Acid Microspheres Gelatin – PGA microspheres in situ loaded with MXN (G-PGA-M) were synthesized using the steric stabilization pr ocess described in sect ion 4.2.2.1. The goal of this study was to determine the mitxonatrone loading and release ch aracteristics of G-

PAGE 89

70 PGA-M for comparison to previously s ynthesized MXN containing microsphere compositions, namely gelatin (GC-M) and gelatin-carboxymethyl cel lulose (G-CMC-M) synthesized by Brain Cuevas in parallel studies. Ideally the incorporation of an anionic constituent such as PGA into the microsphe re matrix should increase MXN loading due to ionic interactions. Table 4.12 lists the two compositions examined. The concentrations of gelatin and PGA were chos en specifically for comparison to previous studies. Particle morphology was examined using SEM (Figure 4.32). Though some smooth spherical particles were synthesize d, there were also a large portion of microspheres that lacked spherical shape and had some degree of surface roughness. This phenomenon was observed for both G-PGA-M compositions. Qualitatively it appeared that the mean particle diameter wa s larger than unloaded G-PGA 1, with some particles exceeding 20-30 m diameter. Particle size distribution anal ysis using the Coulter LS pa rticle sizer resulted in mean particle diameters of 2.5 2.6 m and 3.4 3.1 m for G-PGA M-1 and G-PGA M-2, respectively. Though these mean sizes are lower than e xpected based on SEM analysis, the particle size di stributions extend to over 40 m diameter (Figure 4.33). The upper range of these distribut ions corresponds well with th e SEM qualitative analysis. The particle size distributi on curve for G-CMC-M is also included in Figure 4.33 for direct comparison to previous compositions. Overall, both G-PGA-M compositions app ear to have a simi lar distribution, though the G-CMC-M composition mean was somewhat larger (3.6 3.1 m) and contained a larger proportion of microspheres between 10 and 20 m than both G-PGA-M

PAGE 90

71 compositions and also contained micros pheres with diameters as large as 100 m. These distribution curves are different from those of unloaded G-PGA MS, lacking the pronounced bimodal peaks seen for protein MS synthesized using the steric stabilization method. Additionally, the curves for G-PGAM appear to be shifted toward smaller particle sizes compared to unloaded G-PGA, despite the number of larger particles observed via SEM. Particle morphology for G-CMC-M resembled that seen for G-PGA M-2 (image not available). Figure 4.32: SEM images of A) G-PGA M-1 a nd B) G-PGA M-2. All images taken at 1000x magnification. Figure 4.33: Particle size distribution for G-PGA M compositions. G-CMC-M is also included here for direct comp arison to previous studies. A B

PAGE 91

72 Analysis of particle swelling using optical microscopy revealed that swelling of GPGA-M compositions exhibited larger swollen particles sizes compared with their unloaded counterparts (Figure 4.35). These images also revealed that MXN appears to be unevenly distributed among microspheres a nd is compartmentalized in both G-PGA-M compositions rather than distributed throughout th e matrix. Some particles also appear to contain little to no MXN at al l. Dry and swollen particle sizes from this study are found in Table 4.16 and are graphically repres ented in Figure 4.35. One-way ANOVA of dry particle sizes indicated stat istical difference existed be tween compositions (p<0.001). Tukey MCT revealed that the mean dry particle sizes for G-PGA M-1 and G-PGA M-2 were larger than that of GPGA 1 (p<0.001 in both cases). Comparison of mean swollen particle size using one-way ANOVA indicated differen ces again exist between compositions (p<0.001). Subsequent Tukey MCT results indicated that G-PGA M-1 swollen particle size was significantly larg er than G-PGA 1 (p<0.001). Additionally, GPGA M-1 was found to be significantly la rger than G-PGA M-2 (p<0.001). Though swollen particle size appears much larger for G-PGA M-1 compared with G-PGA M-2, there were two very large particles in the G-PGA M-1 data set that likely skewed these data. Due to larger dry pa rticle sizes, % IPD for G-P GA-M was lower than that of unloaded G-PGA MS despite the larger sw ollen particles size of G-PGA-M MS. Table 4.16: Percent increase in mean particle diameter for G-PGA-M compositions based on particle swelling studies. Result s from G-PGA 1 are included here for reference. Condition Dry (m) Swollen (m) % IPD G-PGA 1 3.1 1.5 8.9 6.9 194% G-PGA M-1 8.3 2.9 17.1 11.3 106% G-PGA M-2 8.3 3.9 11.3 4.4 35%

PAGE 92

73 Dry Swollen A B C D E F Figure 4.34: Optical microscopy images of unloaded and MXN-loaded G-PGA compositions. All images at 40x magnifi cation; those on the left are of dry microspheres and those on the right ar e of fully swollen microspheres, 1 minute post-hydration with PBS. A) – B) are of unloaded G-PGA 1, C) – D) are of G-PGA M-1, and E) – F) are of G-PGA M-2.

PAGE 93

74 3.1 8.38.3 3.2 8.9 17.1 11.3 19.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 G-PGA 1G-PGA M-1G-PGA M-2G-CParticle Size (um ) Dry Swollen # # ^ ^ + + Figure 4.35: Dry and swollen particle size comparison for G-PGA-M compositions. GPGA (unloaded) and G-C (unloaded) ar e included for comparison. Raw data were not available, therefore G-C was not included in statistical analysis. Matching symbols indicate significant differences. Once particle morphology, size, a nd swelling were characterized in vitro MXN loading and release were char acterized. Microspheres were enzymatically digested to determine the total drug content in in situ MXN-loaded G-PGA MS according to the procedure outlined in section 4.2.3. Aliquots of the enzyme digests were analyzed using spectrophotometric techniques to determine dr ug loading. Results of this analysis are found in Table 4.17 and Figure 4.36. Results from CuevasÂ’s parallel research on Gel and G-CMC MS are included for comparison. Table 4.17: In situ MXN-loading results for G-PGA-M and related compositions. Condition MXN Loading % (w/w) G-C* 11.5 1.3 G-PGA M-1 6.7 0.7 G-PGA M-2 5.2 1.3 G-CMC M* 7.8 0.3 * synthesized by Cuevas in a parallel study

PAGE 94

75 0 5 10 15 G-C MG-PGA M-1G-PGA M-2G-CMC MMXN Loading % (w/w) Figure 4.36: MXN loading comparison for GPGA-M and related compositions. G-C-M and G-CMC-M were synthesized by Cuev as and are included for comparison only. Microspheres were incubated in PBS at 37 C as discussed in section 4.2.5. Release for both G-PGA M compositions was ongoing for over 100 days (Figure 4.37). Initial release was very slow, very atypical of past protein mi crosphere compositions. While past experience has shown a burst effect with all other compositions, with 30 – 50% release achieved within the first 20 hours, G-PGA M compositions did not reach 10% release for over 12 days. A summary of th ese results is found in Table 4.18. Results from the free MXN control indicate that ~8% of drug was lost to either binding to the filter device membrane during centrifugation or to drug degradation over the duration of the study. Release results pres ented here may, therefore, be a slight underestimate of actual drug release.

PAGE 95

76 Table 4.18: MXN release results for G-PGA M and related compositions. Condition Release % G-PGA M-1 32.0 3.1 G-PGA M-2 36.0 9.5 G-CMC M 78.1 7.6 G-C M 49.7 2.9 F-MXN 92.1 1.5 36.0% 32.0% 78.1% 49.7% 92.1% 0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0% 110100100010000 Time (hours)Percent MXN Release (%) G-PGA M-2 G-PGA M-1 G-CMC M G-C M F-MXN Figure 4.37: Release curve for G-PGA-M and re lated compositions. F-MXN is a free MXN control used to quantify and MXN degradation and the binding affinity of MXN to the filter membrane. In vitro RG-2 cell culture studies were used to determine the efficacy of G-PGA at inhibiting and/or preventing ne oplastic cellular viability. Pr evious studies were done by Cuevas showed that gelatin alone had no inhi bitory effects and as a result this control condition was not further examined here. Tr eatments included 1) non-treatment controls

PAGE 96

77 (C), 2) MXN controls (F-MXN), and 3) GPGA M-2 microspheres, with n=4 for all treatments, except n=8 for controls. Dosage s for each of these treatments are found in (Table 4.19). Cellular viability and morphology were mon itored daily for the duration of the 4day experiment as described in section 4.2.6. Cellular viability was measured using an MTT assay. This assay contained a yellow tetrazolium salt that was metabolized by active cells to form water insoluble purple formazan crystals. These crystals were made soluble by addition of a buffer in a secondary step. The concentration of solubized formazan in the resulting solution was quan tified by reading the absorbance at 550 nm. This concentration is directly proportional to the number of metabolically active cells; a greater the number of active cells results in a greater number of formazan crystals, and therefore a larger absorbance value. On da y 4 post-treatment, absorbance values were normalized to the average absorbance for non-treatment controls. Normalized mean cellular viability values were used for statistical anal ysis and are listed in Table 4.19. Table 4.19: In vitro RG-2 cell culture treatment groups and MXN doses and resulting normalized MTT absorbance values on day 4. Condition MXN Dose (g/mL) Absorbance C 0 1.000 0.046 F-MXN 1 25 0.000 0.000 F-MXN 2 12.5 0.000 0.000 F-MXN 3 0.5 0.166 0.020 G-PGA M-2 125 0.038 0.021 G-PGA M-2 212.5 0.161 0.039 G-PGA M-2 30.5 0.518 0.381 In addition to MTT analysis, optical mi croscopy images were taken daily to evaluate cellular morphology. Extended pro cesses and cells spread ing along the floor of the culture dish are signs of healthy cells. Cells exhibiting these characteristics were

PAGE 97

78 considered to be experiencing little to no affect of treatment. Cells th at appeared to have an ameboid morphology with few to no pro cesses extending outward, and especially those that had evidence of cell wall rupture/cellu lar debris were considered to have been killed by treatment. Cellular morphology was used in conjunction with MTT away results to make conclusions on the effectiv eness of G-PGA MS in inhibiting and/or preventing proliferation of RG-2 cells. 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0.800 0.900 01234 DayAbsorption F-MXN 1 F-MXN 2 F-MXN 3 G-PGA M-2 1 G-PGA M-2 2 G-PGA M-2 3 Control Figure 4.38: Mean absorbance value comparison for G-PGA M in vitro RG-2 cell proliferation study. A two-way ANOVA followed by TukeyÂ’s MCT wa s used to compare the effects of treatment condition and MXN dose on cellula r proliferation on Day 4 (normalized absorbance values). Results from this test indicated that significant differences existed for both treatment group and dose level (p=0.005 and p<0.001, respectively). Non-

PAGE 98

79 treatment controls were found to have signi ficantly higher absorbance values than all other dose levels (p<0.001 in al l cases). The highest dose, 25 g/mL MXN, was shown to significantly reduce cellular proliferation wh en compared to both the lowest dose, 0.5 g/mL (p<0.001) and the middle dose of 12.5 g/mL (p=0.005). Treatment condition was also found to be significantly different (p=0.005), with F-MXN treatments showing significantly lower absorbance (more cellular inhibition) than G-PGA M-2 treatments. This is likely attributable to the slow re lease of MXN from G-P GA MS (Figure 4.37). Longer studies may show that G-PGA-M MS in hibit cellular prolifer ation equivalent to or better than free MXN cont rols. However, studies shoul d also be done to evaluate MXN activity during longer studies to en sure the activity is not compromised. Representative optical microscopy images from day 2 post treatment are found in Figure 4.39. Non-treatment controls app eared healthy, with processes extending outward. Cells the received either low dose treatment appear somewhat healthy, though there is some apparent effects of receiving tr eatment. While there are some cells with extended processes, there are also some cells with retraced processes and ameboid morphologies. No cells appeared viable in either the middle or high dose F-MXN treatments, with all cells exhi biting ameboid morphologies and most cells appear to have a blue color, indicating internalization of MXN within the cell. Some cellular debris is also visible in each of these groups. Though no MXN internalization is evident in the high dose G-PGA M-2 treated cells, there are no cells with extended processes in these wells. Middle dose G-PGA M-2 treated cells show evidence of toxic effects of MXN, though some cells appear to have extended proce sses. There appear to be much fewer of these healthy cells however, in comparison to low dose G-PGA M-2 treated cells.

PAGE 99

80 A BC DE FG 0.5 g/mL 12.5 g/mL 25 g/mL F-MXN GPMS M-2 Figure 4.39: Day 2 optical microscopy images of RG-2 cells by treatment type and MXN dose. Image A) is of non-treatment controls. All images taken at 40x.

PAGE 100

81 4.3.2. In situ MXN-loaded Bovine Serum Albumin – Deoxyribonucleic Acid Microspheres Bovne serum albumin – DNA microspheres in situ loaded with MXN (BSA-DNAM) were synthesized using the steric stabi lization process descri bed in section 4.2.2.2. Synthesis conditions, including aqueous phase concentratio ns and crosslink concentrations, were chosen based on th e results from unloaded BSA-DNA MS. An initial study was completed as a tria l to verify the feasibility of in situ loading BSA-DNA MS with MXN. The aqueous phase for this study consisted of 10% (w/v) BSA, 1.5% (w/v) DNA, and 15% (w/w) MXN (BDMS M-1). These parameters were chosen for direct comparison to GPMS M-2 and GCMS M compositions. The BSA-MXN solution was added separately from an additional DNA solution, with each so lution prepared at twice the desired concentrati on and added in equal volumes to arrive at the desired concentration. These modified synthesis parameters are found in Table 4.14. BSA-DNA M-1 were synthesized using these parameters to verify the corrected synthesis parameters would produced spherical microparticles in th e desired particle size range. Particle morphology analysis with SEM verified that this modified synthe sis technique would produce particle with the desired character istics (Figure 4.40). BSA-DNA M-1 appeared to fall in the 1 – 10 m size range with spherical mor phology. The resulting particles, however, exhibited a rough surface textur e compared with unloaded BSA-DNA 1 microspheres. The particle size distribution for BSA-DNA M-1 microspheres was similar to that of unloaded BSA-DNA 1, with somewhat larg er quantity of microspheres in the 5-10 m range (Figure 4.41). Mean particle diameter for BSA-DNA M-1 was 3.12 2.65 m.

PAGE 101

82 Figure 4.40: SEM image of BSA-DNA M1 microspheres at 1000x magnification. Particle swelling analysis revealed th at BSA-DNA M-1 microspheres experienced only a 28% increase in particle diameter a nd achieved full swelling within 1 minute of the addition of PBS. Optical microsc opy images of dry and swollen BSA-DNA M-1 microspheres are found in Figure 4.42. MXN containing BS A-DNA compositions swelled less than their unload ed counterparts (Figure 4.43), possibly attributable to the interaction between MXN and DNA acting as an additional source of crosslinking. Oneway ANOVA analysis of swelling data indica ted that significant differences existed between dry particle sizes of BSA-DNA 1 and BSA-DNA M1 microspheres (p<0.05). There were no significant differe nces detected however, between dry and swollen particle sizes for BSA-DNA M-1 microspheres (p>0.05). Furthermore, there were no significant differences detected between swollen sizes of MXN-loaded and unloaded BSA-DNA 1 microspheres (p>0.05).

PAGE 102

83 Figure 4.41: Particle size di stribution comparison for MXN-loaded and unloaded BSADNA 1 compositions. Since in situ MXN-loaded BSA-DNA appear to be feasible, a larger study was designed to examine varying concentrations of BSA and DNA, as well as 2 crosslink densities. These parameters were chosen based on the unloaded BSA-DNA study. Based on non-spherical partic le morphology results from unloa ded BSA-DNA, the lowest GTA concentration was not evaluated (Table 4.15). Modified synthesis parameters used for BSA-DNA M-1 were used to synthesize these microspheres as well (Table 4.14). Particle morphology was evaluated using SE M; images are found in (Figure 4.45). BSA-DNA M-2, both middle and hi gh crosslink densities had smooth surface texture and maintained a spherical morphology. Both BSA-DNA M-3 compositions lacked spherical morphology for many of the microspheres but maintained a smooth surface texture. Mean particle size for these microspheres ap peared qualitatively larger than BSA-DNA M-2. In comparison, many microspheres of both BSA-C M compositions lacked both smooth surface texture and spherical morphology. In situ MXN-loaded BSA

PAGE 103

84 AB Figure 4.42: Optical microscopy images of BSA-DNA M-1 A) dr y and B) 1 minute post-hydrat ion. Both images at 40x magnification.

PAGE 104

85 5.0 7.0 8.6 8.9 0 5 10 15 BSA-DNA 1BSA-DNA M-1Particle Size (um ) Dry Swollen Figure 4.43: Comparison of partic le swelling for unloaded and in situ MXN-loaded BSADNA 1 microspheres. Matching symbols indicate significa nt differences. microspheres have been well characterized in past experiments by other researchers and were thus used as controls for these experiment s. At this time the cause of this abnormal behavior is unknown, as micros phere synthesis parameters were essentially unchanged from these past experiments. All BSA-DNA-M compositions showed some degree of bimodality, with the larger second peak centered over the 6-20 m range. Both crosslink densities of the BSA-DNA M3 composition showed peaks out at 40 m and larger, indicating the presence of agglomerates at this concentration of DNA. Representative distri bution curves are found in Figure 4.44 and a summary of mean particle size data from Coulter LS particle sizing is found in Table 4.20. The unusually large st andard deviation found in BSA-DNA M2-2

PAGE 105

86 and BSA-DNA M3-3 is likely caused by agglom eration, evident for th e latter in Figure 4.44. Figure 4.44: Representative particle size dist ributions curves for MXN-loaded BSA-DNA MS compositions. Table 4.20: Summary of mean pa rticle size from Coulter LS particle sizing. Note the large SD of BSA-DNA M2-2 and BSA-DNA M3-3, likely caused by agglomeration. Condition Mean (m) SD (m) BSA-DNA M2-2 12.44 37.37 BSA-DNA M2-3 4.91 4.85 BSA-DNA M3-2 8.40 9.78 BSA-DNA M3-3 18.94 43.21 BSA-C M2 5.70 4.57 BSA-C M3 7.61 18.67 Particle swelling was agai n analyzed using optical microscopy imaging. A summary of the results from this study is found in Ta ble 4.21. Two-way ANOVA was used to analyze dry and swollen particle si zes as a function of microsphere composition and crosslink density. No significant differences were detected in dr y particle size based on composition (p=0.098) or crosslink dens ity (p=0.591). In contrast, significant differences were detected in swollen particle size as functions of both crosslink density

PAGE 106

87 (p=0.003) and composition (p<0.001). BSA-C M compositions were found to have significantly larger swollen size than BSA-DNA M3 within both crosslink densities (p<0.05 in both cases). Similarly, BS A-C M microspheres were found to have significantly larger swollen particle sizes than BSA-DNA M2 within the middle crosslink density (p=0.007). However, within the hi gh crosslink density BSA-DNA M2 swollen particle size was found to be significantly larger than th at of BSA-C M (p=0.027). BSADNA M2 was also found to have significantl y larger swollen particle size than BSADNA M3 within the high cro sslink density (p<0.001). Th ese results are represented graphically in Figure 4.46 a nd Figure 4.47. Representative optical microscopy images from this study are found in Figure 4.48 – Figure 4.50. BSA-DNA M compositions were enzymatically digested and the resulting digests were analyzed using spectrophotometric tech niques to quantify th e concentration of MXN loaded into the microspheres as outlined in section 4.2.4. BSA-DNA M1-2 loading and release were char acterized prior to synthesis of the remaining BSA-DNA M compositions. Results of this analysis are found in Table 4.17. Note that the percent loading for the BSA-DNA M1-2 composition is much lower than that of the other compositions. This study was completed prior to synthesis of subs equent conditions and complete digestion was never fully achieve d. Optical microscopy out to day 20 showed little to no evidence of any digestions, with many blue microspheres still visible (Figure 4.51). This study was repeated using fresh enzy me with the same results. As a result, release data for this composition will not be reported as percent MXN release. MXN doses for i n vitro studies will be based off of the maximum total release of MXN rather than percent loading data.

PAGE 107

88 A B C D E F Middle GTA High GTA Figure 4.45: SEM image comparison of BSADNA M-2 (A & B), BSA-DNA M-3 (C & D), and BSA-C M (E & F). All imag es were taken at 2000x magnification.

PAGE 108

89 Table 4.21: Percent increase in mean partic le diameter for BSA-DNA M compositions from particle swelling studies. Condition Dry (m) Swollen (m) % IPD BSA-C M-2 8.7 2.9 12.2 3.4 40% BDMS MC-3 8.4 2.5 9.5 2.5 13% BDMS M1-2 7.0 2.3 8.9 3.1 28% BDMS M2-2 8.2 3.6 9.6 3.7 18% BDMS M2-3 8.3 3.4 11.1 3.4 33% BDMS M3-2 7.8 4.0 9.9 3.8 26% BDMS M3-3 7.5 3.5 7.8 3.7 5% Figure 4.46: Particle swelling charts co mparing dry and swollen sizes for each composition by crosslink density; A) middle crosslink density, B) high crosslink density. 8.7 7.0 8.2 7.8 12.2 8.9 9.6 9.9 0.0 5.0 10.0 15.0 20.0 BSA-C M2 BSA-DNA M1-2 BSA-DNA M2-2 BSA-DNA M3-2Particle Size (um ) Dry SwollenA 8.48.3 7.5 9.5 11.1 7.8 0.0 5.0 10.0 15.0 20.0 BSA-C M3BSA-DNA M2-3BSA-DNA M3-3Particle Size (um ) Dry SwollenB

PAGE 109

90 Figure 4.47: Comparison of dry and swollen pa rticle sizes by composition; A) BDMS MC, B) BDMS M2, and C) BDMS M3. 8.7 8.4 12.2 9.5 0.0 5.0 10.0 15.0 20.0 BSA-C M2 BSA-C M3Particle Size (um ) A 8.2 8.3 9.6 11.1 0.0 5.0 10.0 15.0 20.0 BSA-DNA M2-2 BSA-DNA M2-3Particle Size (um ) Dry SwollenB 7.8 7.5 9.9 7.8 0.0 5.0 10.0 15.0 20.0 BSA-DNA M3-2 BSA-DNA M3-3Particle Size (um ) Dry SwollenC

PAGE 110

91 Dry Swollen A B C D Figure 4.48: Particle swelling images for BS A-DNA M-2; A & B middle crosslink dens ity; C & D high crosslink density. All images at 40x.

PAGE 111

92 92 92 Dry Swollen A B C D Figure 4.49: Particle swelling images for BS A-DNA M-3; A & B middle crosslink dens ity; C & D high crosslink density. All images at 40x.

PAGE 112

93 93 93 Dry Swollen Figure 4.50: Particle swelling images for BS A-C M; A & B middle crosslink density; C & D high crosslink density. All image s at 40x.

PAGE 113

94 Table 4.22: In situ MXN-loading results for BSA-DNA M compositions. Condition MXN Loading % (w/w) BSA-C M-2 9.9 0.6 BSA-C M-3 11.2 0.6 BSA-DNA M1-2 3.8 0.2* BSA-DNA M2-2 8.3 0.4 BSA-DNA M2-3 8.4 0.2 BSA-DNA M3-2 8.6 0.8 BSA-DNA M3-3 9.5 0.2 Figure 4.51: Optical microscopy image of BS A-DNA M1 microspheres after 20 days in EDB at 37 C (40x magnification). Full digestion was never achieved for at th e conclusion of the enzymatic digestion study involving all other BSA-DNA M com positions either, though these compositions appeared to have undergone some microsphere degradation. Lack of full degradation was evidenced by the color of the protein pellet following the TCA protein precipitation step of the procedure. BSA-C Mcompositions appeared to fully degrade, as this pellet was white in color indicating that little to no MXN remained bound to microspheres and/or protein. In contrast, all compositio ns containing DNA had dark blue pellets, indicating some MXN remained in microsphe res and/or bound to protein/DNA. As a result, MXN loading results for BSA-DNA M2 and BSA-DNA M3 compositions are

PAGE 114

95 95 likely underestimated here. It is unknown w hy BSA-DNA M1 MS did not degrade at all while BSA-DNA M2 and M3 showed signs of at least some degradation. 0 4 8 12 BSA-C M BSA-DNA M2 BSA-DNA M3 BSA-DNA M1*MXN Loading (% ) Middle GTA High GTA Figure 4.52: MXN loading comparison by BS A-DNA composition and crosslink density. * BSA-DNA M1 loading data are lik ely grossly underestimated (see explanation above). Mitoxantrone release extended beyond 1000 hours for BSA-DNA M1 microspheres. This prolonged release was at tributed to the strong ionic interaction between MXN and the DNA incorporated into th e microsphere matrix. This release was substantially longer than G-C M and GCMC M compositions, and total cumulative MXN release was exceeded compared to these compositions as well. Release of MXN from G-PGA M lasted longer than BSA-D NA M1, but cumulative MXN released was lower for G-PGA M than BSA-DNA M1 (Figure 4.53).

PAGE 115

96 96 0 10 20 30 40 50 60 70 80 110100100010000 Time (hrs)Concentration MXN (ug/mL) G-C M G-CMC M G-PGA M2 BSA-DNA M1 Figure 4.53: MXN release concentration pe r milligam of microspheres for BSA-DNA M1-2. Gelatin, gelatin-CMC, and gelati n-PGA microsphere release curves are included for comparison. Mitoxantrone release for other BSA-DNA M compositions was not as promising as it initially appeared base d on the release study for BSA-DNA M1. BSA-DNA M2 and BSA-DNA M3 compositions exhibited a sign ificant burst release, with MXN release nearly complete by 25 hours (Figure 4.54). It is unknown why these release profiles differ so greatly. One explanation is that the source of DNA used in BSA-DNA M2 and M3 was different from that used in BS A-DNA M1. BSA-DNA M1 microspheres were synthesized using DNA from herring teste sour ce. When this was reordered to have sufficient quantity for the remainder of the st udies, the molecular weight of the second lot was noticeably lower than that of the or iginal lot based on an obvious decrease in viscosity of DNA solutions made with the new lot. As DNA from the original lot was no longer available, the DNA used for subseque nt BSA-DNA syntheses was from salmon

PAGE 116

97 97 teste source. This DNA appeared qualitatively to be much cl oser to the original herring DNA. The molecular weight of any of th ese DNA samples has not been measured as access to the GPC/MALS system was not availabl e for the duration of these experiments. Figure 4.54: Release profile for BSA-DNA M co mpositions for A) the entire release study and B) the first 25 hours only. In vitro RG-2 cell culture studies were conduc ted to determine the effect of BSADNA M1 MS on the inhibition of cellular viability as outlined in section 4.2.6. Treatments for this study were similar to those used for the G-PGA M MS study, 0% 20% 40% 60% 80% 100% 120% 050100150200250300350 Time (hours)% MXN release A 0% 20% 40% 60% 80% 100% 120% 0510152025 Time (hours)% MXN release BSA-C 2 BSA-C 3 BSA-DNA M2-2 BSA-DNA M2-3 BSA-DNA M3-2 BSA-DNA M3-3 B

PAGE 117

98 98 including 1) non-treatment controls (C), 2) MXN controls (F-MXN), and 3) BSA-DNA M1 MS (n=4 for all groups, except n=6 for controls). Normalized absorbance results from th e MTT assay are found in Table 4.23 and Figure 4.55. Two-way ANOVA anal ysis of MTT absorption data on day 4 indicated that there were no significant di fferences detected between treatment groups (p=0.482). There were, however, differences between MXN dose (p< 0.001). Subsequent Tukey MCT analysis revealed that all treatment gr oups were superior at inhibiting viability compared to controls (p<0.001 in all cases). Additionally, middle and high doses were superior to the low dose at inhibiting viabi lity (<0.001), though there were no differences between these 2 doses. Table 4.23: In vitro RG-2 cell culture MTT assay trea tment groups and mean normalized absorbance values of day 4. Condition MXN Dose (g/mL) Absorbance C 0 1.000 0.077 F-MXN 1 25 0.000 0.000 F-MXN 2 12.5 0.000 0.000 F-MXN 3 0.5 0.161 0.016 BSA-DNA M1 1 25 0.000 0.000 BSA-DNA M1 2 12.5 0.000 0.000 BSA-DNA M1 3 0.5 0.217 0.020 Representative optical microscopy images , used to evaluate cellular morphology, are found in Figure 4.56. Nontreatment controls appear healthy, with processes extending outward and the number of cells in creasing indicating viab ility. Cells that received a low dose of either MXN treat ment (F-MXN or BSA-DNA M1) showed signs of toxicity evidenced by ameboid morphology, though some cells still appear to have healthy processes. Both middle and high dos es of both MXN treatments appeared very effective at inhibiting ce llular viability, as there did not a ppear to be many, if any, viable

PAGE 118

99 99 cells in any of these groups. Additionall y, uptake of MXN was evident with blue coloration visible within the cells. 0.000 0.250 0.500 0.750 1.000 00.511.522.533.54 DayAbsorptio n Control F-MXN 1 F-MXN 2 F-MXN 3 BSA-DNA M1 1 BSA-DNA M1 2 BSA-DNA M1 3 Figure 4.55: Normalized mean absorbance values for BSA-DNA M1 in vitro RG-2 cellular viability study. Dosage key: 1 = 0.5 g/mL, 2 = 12.5 g/mL, and 3 = 25 g/mL.

PAGE 119

100 100 F-MXN BSA-DNA M1 0.5 g/mL 12.5 g/mL 25 g/mL Control Figure 4.56: Optical microscopy images of RG-2 cells from in vitro RG-2 cell culture study. All images were taken at 40x.

PAGE 120

101 CHAPTER 5 IN VIVO EVALUATION OF BOVINE SE RUM ALBUMIN MICROSPHERES FOR INTRATUMORAL CHEMOTHERAPY 5.1 Introduction Though in vitro examinations of microsphere fo rmulations gives insight on the effectiveness of microsphere formulations in causing cell death, these studies do not closely replicate the physiologi cal environment and final clin ical use of microspheres as an intratumoral delivery vehicle. To more closely replicate these, in vivo experimentation is necessary. Previous in vivo animal studies have shown that intratumoral (IT) modalities are superior to traditional intrave nous therapies in terms of animal survival and tumor regression. In a Lewis lung carci noma model, 92% of mice survived to the endpoint of the study (35 days) when these an imals were treated with IT chemotherapy followed by surgical resection 20 days later.42 Further studies using a 16/C mammary adenocarcinoma model resulted in an increase in survival of over 300% for animals receiving IT MXN injections followed by surg ery 10 days after treatment compared to controls.14 These improvements can be attribut ed to several advantages offered by intratumoral therapies, including the in creased MXN drug dosage reaching the tumor mass and the reduction in toxicity associated with such therapies. In studies investiga ting the use of 20-40 m diameter in situ MXN-loaded bovine serum albumin (BSA-M20) MS for IT chemotherapy in a 16/C MMAC model, up to 50% survival was achieved compared to 0% for c ontrols. Subsequent studies using smaller diameter in situ MXN-loaded BSA MS (5-10 m; BSA-M) achieved survival of up to

PAGE 121

102 102 80% of mice compared to 0% for controls13. These studies provide d the basis for further studies evaluating BSA-M. In conjunction with Brett Almond and Shema Freeman, two separate in vivo studies were carried out in C3H/HeJ mice with the following goals: 1. histological evaluation of MMAC tumors to evaluate extent of perfusion of BSA-M MS and tumor necrosis after IT treatment, 2. to further assess toxicity mechanisms using complete blood count analysis, 3. to examine the effect of elapsed tim e between neoadjuvant IT chemotherapy and subsequent surgical resection. 4. to assess the advantage of administeri ng repeated IT treatments at scheduled intervals 5.2 Materials All proteins, enzymes, and solvents were obtained as described in sections 3.2 and 4.2 unless otherwise specified. In these studies, the MXN used for in situ loaded microspheres was donated by Lederle Laborator ies. General medical supplies were obtained through University of Florida Health Center Stores, Henry Schein, or Webster Veterinary Supply. Ketamine HCl (100 mg/mL), Flunixin Meglumine (50 mg/mL), and Xylazine (20 mg/mL) were used as anesthetics/analgesics for all animals studies. Methoxyfluorane was mixed 1:1 with mineral oil and was used as an inhaled anesthetic. 5.3 Methods 5.3.1 Microsphere Synthesis & Characterization In situ MXN-loaded BSA mesospheres (BSA -M) with 2% and 8% (w/w) GTA crosslink densities with a mean diameter of 5-10 m were synthesized by Brett Almond using the steric stabilization process descri bed in section 3.3.2. A brief summary of the synthesis and characterization of microsphere s used for these studies is found below.

PAGE 122

103 103 Table 5.24: Synthesis parameters for BSA-M microspheres. Mixing Conditions RPM Time 47 mL 4% CAB + 3 mL 20% BSA-MXN 1250 20 min + x mL GTA 600 2 hr + 50 mL acetone 600 1 hr TOTAL TIME3 hr 20 min A summary of the synthesis paramete rs is found in (Table 5.24). MXN concentration was kept at 15% (w MXN/w pr otein) for all studies. BSA-M MS were characterized using SEM (Figure 5.57) and C oulter LS particle si zing (Figure 5.58) as described previously. Particles with low crosslink density (2% GTA) show increased porosity, with obvious pitting or pores in the MS structure. Higher crosslink density (8% GTA) resulted in smooth, round microspheres. In vitro MXN loading and release were also quantified using previously described techniques (Figure 5.59). Figure 5.57: SEM images of BSA-M used in these in vivo studies: A) 5-10 m BSA-M MS with high crosslink density (8 % GTA; 500x magnification) and B) 5-10 m BSA-M MS with low crosslink de nsity (2% GTA; 750x magnification) B A

PAGE 123

104 104 Figure 5.58: Particle size distribut ion curves for unloaded BSA MS, in situ MXN-loaded BSA MS with low crosslink dens ity (2% GTA w/w; BSA-M L), and in situ MXN loaded BSA MS with high cro sslink density (8% GTA w/w; BSA-M H). Figure 5.59: In vitro release of MXN from BSA-M20 (20-30 m) and BSA-M (5-10 m) with low and high crosslink densi ties under infinite sink conditions. BSA MS BSA-M L BSA MH

PAGE 124

105 105 5.3.2 In Vivo Animal Study Protocols 5.3.2.1 Tumor passage The 16/C murine mammary adenocarcinoma cell line is not viable ex vivo. For this reason, tumors were maintained in C3H/HeJ mice. In order to harvest the tumor for passage, mice with tumor were euthanized via CO2 inhalation. Within 5 minutes of death, the tumor was excised using blunt di ssection through an openi ng in the skin and was finely minced under aseptic conditi ons. The tumor was resuspended at a concentration of 0.5 mg tumor/mL in calci um-free PBS. Each receiving mouse was anesthetized using met hoxyfluorane inhalation and was injected with 50 L of tumor suspension subcutaneously into the flank. Tumor inoculation using these methods resulted in growth of a 500 mg tumo r within 14 days of injection. 5.3.2.2 Standard protocol for in vivo 16/C MMAC studies All in vivo studies were conducted with approva l of the University of Florida Institutional Animal Care and Use Comm ittee (IACUC). Mice were housed in the University of Florida Animal Care Services (ACS) facilities. Ten to fourteen week old syngenic fema le C3H/HeJ mice were inoculated with 16/C MMAC tumor according to section 5.3.2.1 . Mice were weighed and monitored daily for signs of tumor growth. When the resulting tumor reached 10 cm in its largest dimension (~ 2.5% body weight based on an estimated weight of 0.5g for a 1x1 tumor), mice were randomly assigned to a treatment group, based on the spec ific study design. For animals receiving IT treatment(s), anim als were anesthetized using methoxyfluorane inhalation and 100 L the designated treatment was admi nistered in 5 injections, four around the tumor perimeter and one in the center of the tumor. Drug doses were based on an average mouse weight of 20 g and all animals received the same dose independent of

PAGE 125

106 106 their actual weight. The day of treatment was designated Day 0 a nd differed from mouse to mouse based on tumor growth rate. This method was chosen instead of treating all animals on the same day to ensure that tumo r size and condition at the time of injection was similar for all animals. Mice were observed daily for tumor gr owth, body weight, and general signs of malaise. Animal weight and tumor dimensions were recorded at le ast every 2 days for 40 days post-treatment. Tumor weight was calcu lated based on these measurements and the equation for the volume of an ellipse (tumor weight = a x b2/2000) where a is the largest measured tumor diameter and b is the diameter perpendicular to a , both in measured in mm. Mice that exhibited body wei ght loss of 20% or more (adjusted for tumor mass) were considered to be suffering from toxicity and were euthanized via CO2 inhalation. Additionally, mice displaying signs of discomfort, severe de hydration, or other general malaise were euthanized. Treatment failure was defined as tumor mass exceeding 10% of initial body weight; these animals were also euthanized. Any mice su rviving to at least 40 days were considered “cured”. 5.3.2.3 Blood collection Blood samples were collected for analys is of complete blood counts with differentials. Approximately 200 L of blood was collected from the ventral tail artery of each mouse. The tail was disinfected w ith povidone iodine solution, then nicked carefully with a scalpel. Blood was collected into pediatric EDTA blood collection tubes and given to the ACS clinical lab to perfor m CBCs with differentials. Following blood collection, pressure was applied to the nick in the tail artery until bleeding subsided.

PAGE 126

107 107 CBC and differential results we re analyzed for signs of leukopenia, indicative of doselimiting toxicity for MXN. 5.3.2.4 Tumor histology Any tumors removed during surgery were collected and fixed in 10% neutral buffered formalin for at least 24 hours for histological analysis. Prior to tissue processing, tumors were dehydrated in a graded ethanol series (3 washes at each of 75%, 90%, 95%, and 100% dry ethanol co ncentrations for 15 minutes per wash). Tissues were embedded in paraffin, sectioned, and Hematoxi lin & Eosin (H&E) stained using standard techniques by the University of Florida Research Histology Core Lab (Department of Pathology, Immunology, and Labor atory Medicine). All slides were evaluated and interpreted by Dr. Carol Detrisac from the College of Veterinary Medicine, Department of Pathobiology. 5.3.3 Study Design for In Vivo Animal Studies 5.3.3.1 Neoadjuvant intratumoral chemotherapy and tumor histology In order to examine the effect of timing of surgical resection of the tumor after IT injection, mice received one the following trea tment groups on Day 0: 1) IT injected MXN (MXN), 2) BSA-M MS with low crossl ink density (2% GTA; MS L), 3) BSA-M MS with high crosslink density (8%; MS H), or 4) non-treatm ent control (C). Surgical resection was scheduled for either Day 1, 7, or 14 for IT treated animals (n=3 per timepoint and treatment; n=33 total). Animals assigned to the non-treatment control group were scheduled for surger y on Day 1 or 7 only since th ey were not expected to survive past an average of 10 days. Blood sa mples were also collected for assessment of toxicity mechanisms following IT MXN treatments. Tumor histology was also completed for this study to identify ex tent of tumor perfusion and necrosis in vivo . Table

PAGE 127

108 108 5.25 summarizes these treatment groups. Prot ocols for tumor passage, IT treatment procedures, animal monitoring, histology, and blood collections described in sections 5.3.2.1 through 5.3.2.4 were followed for this study. Table 5.25: Design for neoadjuvant IT ch emotherapy study. CBC values were also analyzed from mice with no tumor (NT). Treatment (Day 0) Day 1 Day 7 Day 14 24 mg/mL BSA-M (2% GTA) + 4 mg/mL MXN H, CBC H H, CBC 24 mg/mL BSA-M (8% GTA) + 4 mg/mL MXN H, CBC H H, CBC 8 mg/mL MXN H, CBC H H, CBC Control H, CBC H H = histology, CBC = complete blood count + differential 5.3.3.2 Scheduled intratumoral chemotherapy injections Clinincal chemotherapy regimens ofte n involve a series of chemotherapy treatments rather than a single treatment. To more closely replicate th is clinical situation, a study was designed to evaluate the efficacy of scheduled mu ltiple IT injections of BSAM for increased animal survival. Ten animals were assigned to each of three IT treatment groups, using dosages similar to those admini stered in the neoadjuvant IT chemotherapy study: 1) BSA-M MS, 2) MXN, or 3) non-tr eatment control. Dosage and scheduling information is found in (Table 5.26). In a clinical setting, re peated chemotherapy treatments are often scheduled weeks apart to allow for recovery from any toxic effects of each treatment. However, the rapid growth rate of the 16/C MMAC tumor line necessitated these injections to be scheduled closer togeth er in this study. Tumor size typically reaches 10% of body mass within 10 days left untreated, thus any unperfused portions of the tumor would reach a level requi ring euthanasia long before several weeks passed between treatments. As a result, inje ctions were scheduled for days 0, 7, and 14. There is an inherent increased risk of toxicity with these in jections spaced more closely together, requiring close monitoring of an imal body weight following treatment.

PAGE 128

109 109 Standard procedures for tumor passage, IT injection, and animal care were used for this study (see sections 5.3.2.1 through 5.3.2.2). Table 5.26: Design for scheduled IT injection study. Dose Treatment Day 0 Day 7 Day 14 BSA-M MS 24 mg/kg BSA-M + 4 mg/mL MXN 24 mg/kg BSA-M + 4 mg/mL MXN 24 mg/kg BSA-M + 4 mg/mL MXN MXN 8 mg/kg 4 mg/kg 4 mg/kg Control 5.3.4 Statistical Analysis Animal survival was analyzed using the right-censored Kaplan-Meier method. The log-rank test was used on the resulting surv ival curves to determine any statistical differences between groups with a significance level of =0.05. When differences were detected, TukeyÂ’s multiple comparison test (MCT) was used to make pair-wise comparisons. One-way ANOVA was used for determinati on of differences between treatment groups for continuous numerical data (ani mal weight, CBC values, etc.), followed by TukeyÂ’s MCT when applicable ( =0.05 for all tests). Ordina l data were analyzed using the Kruskal-Wallis ANOVA with =0.10. 5.4 Results & Discussion 5.4.1 Neoadjuvant Intratumoral Chemotherapy and Tumor Histology This study was designed to achieve thr ee goals: assess the efficacy of IT chemotherapy on animal survival, evaluate tumor histology to determine the amount of tumor necrosis and MS distribution follo wing IT treatment, and to explore the mechanisms of toxicity for MXN following IT chemotherapy. In order to accomplish these goals, 33 mice were randomly assigned to one of 11 treatment groups, varying in IT

PAGE 129

110 110 treatment received and time following IT treatm ent to surgical resection of the residual tumor. Treatments included IT BSA-M of 2 crosslink densities delivered in a MXN solution, IT MXN, or non-treatment controls with surgery following treatment at day 1, 7, or 14 (Table 5.25). On designated surgical days, mice had blood samples and tumors collected for complete blood counts with differen tials and histological an alysis. Note that the non-treatment control group did not include su rgery at the 14 days due to the fact that these animals would experience tumor growth sufficient to require euthanasia prior to reaching this point. Four mice died during or immediately fo llowing surgery due to anesthesia or hypothermia. These animals were excluded from survival analysis. Some, but not all, of these animals were replaced in the study, resu lting in only two mice in the 7 day surgery MXN group (MXN-7) and one mouse in the 1 day surgery MXN group (MXN-1). One mouse if the MXN-14 group was replaced, as the mouse originally assigned to this group died prior to surgery. 5.4.1.1 Surgical observations Observations were recorded during surgic al excision of tumors, including tumor size and appearance. As expected, tumors removed on day 7 from non-treatment control mice were the largest tumors and the most difficult to excise cleanly. Tumor margins were not always identifiable, increasing the chance of tumor recurrence from an unresected portion. All tumors that were administered MXN via IT injection were typically much smaller at surgery and were stai ned dark blue. Little to no staining of the surrounding tissues indicates that most or all of the MXN remained inside the tumor mass. There were cases, however, of staining of the overlying skin. In cases where IT treated tumors exhibited tumor regrowth, the observations of small area of stained tumor

PAGE 130

111 111 surrounded by tumor tissue normal in appear ance supported the hypothesis that these cases of failure were due to incomplete perfusion of MXN into the tumor mass. 5.4.1.2 Histology results Tumors that were excised during surg ery were sectioned and stained for histological analysis. Microspheres were visible in tumor sections of all tumors injected with BSA-M. Degradation of low crossl ink density (2% GTA w/w) microspheres was evident at 14 days, with porous structur e and a lower degree of eosinophilic staining (Figure 5.60). 5.4.1.3 Blood cell counts Blood was collected from mice in each treatment group on days 1 and 14 (Table 5.25) to examine the mechanism of toxicity fo r IT injected MXN. Mice with no tumor (NT) were employed as a sec ondary control in th is study to compare C3H/HeJ mice with and without tumor. CBC data are found in Table 5.27, along with one way ANOVA results comparing treatment groups. These data are also presented graphically in Figure 5.61 through Figure 5.63. Leukopenia was hypothesized to be the primar y mechanism of toxicity. However, there was no evidence of leukopenia or an emia based on the lack of significant differences detected in WBC and RBC counts between treatment groups (p=0.501 and p=0.194, respectively). There was also no signi ficant difference evident between groups for hematocrit (p=0.06), though BSA-ML trea ted animals appear to have lower hematocrit than those in the remaining groups, particularly non-tumor bearing controls. Platelet count was found to be signifi cantly different between groups (p=0.008). Tukey MCT revealed that mice treated with BSA-ML-1 had lower significantly lower platelet counts than both NT and MXN-14. BSA-ML-14 mice had a hi gher platelet count

PAGE 131

112 112 than day 1 surgical controls (C) and M XN-1 treatments. It is not understood why chemotherapy would cause an increase in pl atelet count, and this difference may be attributed to an abnormally high plat elet count in a single mouse ( > 1,000,000 platelets/ L) in the BSA-ML-14 group. This mo use was also noted to have abnormal platelet morphology. Figure 5.60: Tumor sections from A) MS H (day 1) and B) MS L (day 14) treatment groups. Microspheres are visible in both images, with significant degradation of MS L evident (B) by increased poros ity and reduced eosinophilic staining. Both images at 100x. A B

PAGE 132

113 113 Table 5.27: CBC data for mice enrolled in this study. MS L refers to low crosslink density BSA-M and MS H refers to high crosslink density BSA-M; NT refers to mice with no tumors. (n=3 per group) WBC RBC HCT (%) Platelets Neutrophils Lymphocytes Group AVE SD AVE SD AVE SD AVE SD AVE SD AVE SD NT 4867 681 8.8 0.2 49.4 1.1 806333 99771 1161 260 3577 421 C 4967 1457 8.9 0.5 47.3 2.2 688333 9074 1151 208 3584 1000 MXN-1 5000 436 8.4 0.2 44.8 2.4 683000 21656 2374 488 3561 2361 MXN-14 5231 2134 8.3 0.8 46.9 3.4 826000 10817 2135 326 2678 1287 MS L-1 3567 451 8.3 0.5 43.2 1.1 560667 53910 2193 371 1280 362 MS L-14 4200 1300 7.7 0.5 43.0 3.8 939667 97910 2467 919 1507 488 MS H-1 4967 1518 8.2 0.4 45.7 1.6 738000 87230 2136 198 2771 1450 MS H-14 3533 473 8.1 0.7 45.1 1.3 820333 210965 1723 601 1633 299 p value 0.501 0.194 0.06 0.008 0.018 0.102 Though there was no evidence of leukopeni a, MXN treated mice had increased leukocyte cell population (p=0.018). TukeyÂ’s MC T results indicated that both control groups had significantly lower neutrophil coun ts than all treated animals with the exception of MXN-14 and BSA-MH-1. Though th ese differences were statistically significant, it is unlikely that this slight eleva tion in cell count is enough to clinically indicate acute inflammation for these groups. 5.4.1.4 Survival Survival analysis was conducted to dete rmine if time between treatment and surgery had an impact on animal survival. These data are summarized in Table 5.28. The Kaplan-Meier survival curve is found in Figure 5.64. Due to the low number of animals per gr oup, no differences were detected except between groups with 100% surv ival and controls (0% surviv al). This was sufficient, however, the determine that IT treatment in either form (as MXN alone or in MS form) was an effective treatment for increasing animal survival in the 16/C MMAC model compared to controls. Though not statistically detectable, it appear s that surgery on Day

PAGE 133

114 114 1 was more effective than surgery on day 7, since these control animals survived 5 days longer than their day 7 counterparts. This difference was likely due to the more defined tumor margins and smaller size at the tim e of resection for the day 1 group. Additionally, all animals that received any IT MXN treatment and had surgery on Day 1 survived, whereas at least 2 animals in each of the MXN and MS L groups that received surgery at a longer interval af ter IT treatment died. Longe r times between treatment and surgery allow for unperfused regions of the tu mor to grow to a significant size, and increase the risk of small foci of tumor left after surgery. Figure 5.61: Comparison of WBC and RBC coun ts by treatment group. No significant differences were detected between any groups. 0 2000 4000 6000 8000 ControlsMXNMS LMS HWhite Blood Cell Count (cells/uL) Day 0 Day 1 Day 14 0.0 2.0 4.0 6.0 8.0 10.0 ControlsMXNMS LMS HRed Blood Cell Count (x 106 cells/uL) Day 0 Day 1 Day 14

PAGE 134

115 115 Figure 5.62: Comparison of hematocrit values and platelet counts by treatment group. No significant differences were detected for hematocrit, though it appears that microsphere treatments may result in lower hematocrit values compared to controls. Large variability in these data likely prevented detection of significance for this parameter. Signi ficant differences were detected for platelet counts. Matching symbols indi cate differences between these groups. 35.0 40.0 45.0 50.0 55.0 ControlsMXNMS LMS HHematocrit (%) Day 0 Day 1 Day 14 0 0.2 0.4 0.6 0.8 1 1.2 ControlsMXNMS LMS HPlatelet (x106 cells/mL) Day 0 Day 1 Day 14** *

PAGE 135

116 116 Figure 5.63: Comparison of neutrophil and lympohctye counts by treatment group. Control group neutrophil percentage wa s found to be significantly lower than all groups except MXN-14 and MS H-1. Table 5.28: Survival summary for neoadjuvant IT chemotherapy study. All groups with 100% survival showed significantly increas ed survival time versus controls. Control MXN MS H MS L Surgery Day 1 7 1 7 14 1 7 14 1 7 14 Median Survival a 21 16 * * 33 * * * * 34 * % ILS b >150% >150% 106% >150% >150% >150% >150 % 113% >150% % cures c 0% 0% 100% 100% 33% 100% 100% 100% 100% 50% 100% a Time to reach 50% survival post-treatment (in days) b ILS=increase in median life span compared to median for all controls (16 days) c Animals alive at 40 days 0 1000 2000 3000 4000 ControlsMXNMS LMS HNeutrophils (cells/uL) Day 0 Day 1 Day 14 0 1500 3000 4500 6000 7500 ControlsMXNMS LMS HLymphocytes (cells/uL) Day 0 Day 1 Day 14

PAGE 136

117 0% 20% 40% 60% 80% 100% 051015202530354045 Days after Tumor ResectionSurvival (%) C-1 C-7 MXN-1 MXN-7 MXN-14 MS H-1 MS H-7 MS H-14 MS L-1 MS L-7 MS L-14 Figure 5.64: Survival chart for neoadjuvant IT chemotherapy study. Control (C) mice received no chemotherapy treatment but underwent surgical resection on Day 1 or 7. Mice that received IT MXN treatmen ts on Day 0 received either MXN at 8 mg/kg dose or 24 mg/kg of BSA-M MS delivered in 4 mg/k g MXN, where BSA-M MS were of either low or high crosslink density (2% or 8% GTA w/w, respectively). Surgical resection was then performed on either Day 1, 7, or 14 following IT treatment.

PAGE 137

118 5.4.2 Scheduled Intratumoral Chemotherapy Injections This study was designed with the intent of more closely replicating current human clinical chemotherapy treatments and to al low for treatment of possible unperfused regions of tumor after a single IT MXN in jection. As such, 3 injections of IT chemotherapy were scheduled at one week intervals on days 0, 7, and 14. Treatment groups included non-treatment controls, IT MXN injections, and IT BSA-MS (8% GTA w/w) delivered on MXN solution. The study design including dosages is found in Table 5.26. A total of 30 mice were included in this study, 10 in each treatment group. All mice enrolled in the study received at least 2 treatments. Fifty percent of animals receiving IT MXN in either form di d not receive the third injection for one of three reasons. Three mice experienced > 15% loss in TABW and did not receive the third injection due to suspected toxicity. Four mice were c onsidered to have no viable tumor remaining, with any palpable mass consider ed to be scar tissue due to its dense feel and therefore did not receive the third in jection. The remaining three mice were euthanized prior to the third injection due to tumor regrowth or weight loss in excess of 20%. Also of note is one control animal e xperience spontaneous tumor regression. This animal was considered to be an outlier and was not included in survival analysis. 5.4.2.1 Animal weight Since there was an apparent risk of t oxicity with this study, animal body weight loss (adjusted for tumor weight; TABW) was closely monitored and was analyzed as a parameter of interest (normalized for animal weight at Day 0; NTABW) on day 5, 7, 14, 21 and 28. A summary of these data is f ound in Table 5.29 and displayed graphically in Figure 5.65.

PAGE 138

119 Significant weight loss differences were observed at day 5 (one way ANOVA, p=0.002). Anaylsis via TukeyÂ’s MCT indicated that controls had significantly higher body weight compared to both IT treatments, with no difference detected between these latter treatments. BSA-M treated mice ha d significantly lower NATBW compared to those in the MXN group at days 7, 14, and 21 da ys (p < 0.045 in all cases). However, by 28 days post treatment mice in the BSA-M group recovered their weight loss and no differences were detected (p=0.206). Table 5.29: Body weight summary for the sche duled IT injections study at selected timepoints (n=10 per group). Mean Normalized Tumor Adjusted Body Weight (S.D.) Treatment Day 5 Day 7 Day 14 Day 21 Day 28 Control 1.01 (0.03) MXN 0.99 (0.03) 1.00 (0.02) 1.00 (0.02) 1.03 (0.05) 1.04 (0.05) BSA-M 0.95 (0.02) 0.94 (0.04) 0.92 (0.07) 0.97 (0.06) 1.00 (0.07) 0.85 0.90 0.95 1.00 1.05 051015202530 Time after Treatment (days)Normalized TABW Control MXN BSA-M Figure 5.65: Normalized TABW over time by treatment group. Gray lines represent times selected for statistical analysis.

PAGE 139

120 This observed toxicity compared to contro ls was likely due to high levels of MXN delivered at closely spaced intervals. The additional weight loss experienced by animals in the BSA-M group can be attributed to the prolonged release of MXN from the microsphere matrix, further increasing MXN do se at later timepoint s. Weight loss was recovered for both IT treate d groups by 4 weeks after treatment, with animals in the MXN group experiencing minor weight gain by this time. 5.4.2.2 Survival Survival analysis revealed th at both IT treatments were effective, with both groups surviving longer than non-trea tment controls (p<0.001). Mice receiving IT treatments survived 400% longer than cont rol, with 70% of animals in each group surviving at least 40 days. Of those animals surviving less than 40 days in th e BSA-M group, two mice were euthanized due to seve re toxicity marked by excessi ve weight loss (> 20% TABW) and the third experienced tumor regrowth to over 10% body mass and was euthanized on day 38. All three mice in the MXN treatment group with less than 40 day survival experienced tumor regrowth. These result s are summarized in Table 5.30 and (). Table 5.30: Survival summary for scheduled IT injection study. Control MXN BSA-M Medial survivala 8 * * % ILSb >400% >400% % curesc 0% 70% 70% a Time to reach 50% survival post-treatment (in days) b ILS=increase in median life span compared to median for controls c Animals alive at 40 days

PAGE 140

121 0% 20% 40% 60% 80% 100% 120% 051015202530354045 Time after Initial Treatment (Days)Survival (%) Control MXN BSA-M Figure 5.66: Animal survival by treatmen t group for scheduled IT injection study.

PAGE 141

122 CHAPTER 6 RELATED INTRATUMORAL CHEMOTHERAPY WORK 6.1 Introduction Lung cancer is the overall leading cause of cancer death, with an estimated 28% of all cancer deaths attributed to lung cancer in 2006. Furthermore, nearly 60% of people diagnosed with lung cancer will not survive 1 year after diagnosis. The majority of lung cancers fall into one of two main classifi cations: small cell lung cancer (SCLC) or nonsmall cell lung cancer (NSCLC). Roughly 85% of all lung can cers fall into this second category. These are further subdivided in to 3 types: squamous cell carcinoma, adenocarcinoma, and large-cell undifferentiate d carcinoma. Only about 15% of people diagnosed with NCSLC will survive more than 5 years.5 Treatment for NSCLC patients typically involved surgery, chemotherapy, ra diation therapy or some combination of these modalities.49 Surgery remains the pr eferred treatment for patients diagnosed with this disease. However, for patients in poor general health or advanced stage disease, surgery may not be possible. In lung cancer patients with inoperable disease, endobronchi al obstruction is a common and life-threatening complication. Th ese patients are typically treated with combined irradiation and chemotherapy. In cas es of severe obstruction a more urgent approach to airway debulk ing is necessary. Direct intratumoral injection of chemotherapy has shown promise using a mixed regimen of 5-fluororuracil (5-FU), mitomycin, methotrexate, bleomycin, and mitoxantrone.49 Based on this study, two

PAGE 142

123 further clinical studies were performed to eval uate the efficacy of IT injection of 1) 5-FU alone and 2) cisplatin on incr ease in lumen diameter (reduc tion of bronchial obstruction) and patient survival. These studies were conducted by Dr. Firuz Celikoglu and Dr. Seyhan Celokoglu at Florence Nightingale Hospita l in Istanbul, Turkey. Each of these studies is described briefly be low, including statistical analys is of these data, performed as IT work related to this research. 6.2 Intratumoral Chemotherapy with 5 -fluorouracil for Palliation of Bronchial Cancer in Patients with Severe Airway Obstruction 6.2.1 Study Summary Patients with nearly complete obstruction (> 50%) of at least one major airway were enrolled in the study (n=65, age 28-82 years). All patients were cl assified according to health and tumor status. Complete medi cal history, physical examination, bronchoscopy, and computed tomography (CT) were among the preliminary studies performed on each patient. Injections of IT chemotherapy were perf ormed using a flexible bronchoscope with a flexible, retractable 23-gauge needle typica lly used for needle aspiration biopsy. The bronchoscope was introduced transnasally into the trachea with the needle retracted until within 2 cm of the area to be injected. The needle was inserted directly into the tumor mass and 0.5-1 g of 5-FU solution (50mg/mL ) was injected in a fanning manner to maximize tumor perfusion. Increase in lumen diameter was bronchos copically evaluated and assigned a semiqualitative value based on a 3 point scale (Tab le 6.31) assigned such that an increase in value corresponded to an increase in the degree of efficacy of IT treatment (i.e. increase in lumen diameter). Patient initial and final lumen opening (LO) measurements were

PAGE 143

124 used to determine if any difference lumen di ameter occurred following IT injection with 5-FU. Wilcoxon matched-pairs signed-rank analys is was used for these data based on the following: a) this test is designed to detect differences either between baseline and treatment values of the response variable, allowing each patient to serve as their own control and b) is non-parametric, requiri ng no assumption of data distribution. A significance level of =0.05 was used for all tests. Table 6.31: Semi-quantitative numerical valu e assignment for degree of increase in lumen diameter (ILD). Category Value > 50% ILD 3 25-50% ILD 2 < 25% ILD 1 6.2.2 Results Effects of IT injection with 5-FU were vi sible endoscopically. An average of three sessions in two-weeks relieved symptoms a nd restored airway patency partially or completely in 57 or 65 patients (Table 6.32) . Patients with tumor recurrence were received IT 5-FU treatments weekly until for up to 2 months until tumor tissue was completely cleared, with treatments repeated in this fashion for any further recurrences. Lung collapse was reversed in 14 of 20 patie nts who presented w ith this on initial examination. Minimal adverse reactions to IT 5-FU were observed. Table 6.32: Number of patients in each ILD category by tumor location Tumor Location >50% ILD Good response 25%-50% ILD Partial response < 25% ILD No response Total Trachea 5 3 0 8 Carina + Main Bronchus 1 3 1 5 Main bronchus or bronchus intermedius 18 11 3 32 Lobar bronchus 10 6 4 20 Total 34 23 8 65

PAGE 144

125 There were no indications of toxicit y, with no patients experiencing nausea, interstitial pneumonitis, bone-marrow suppressi on, inflammatory reaction, or hair loss. Additionally, no pain or discomfort was reported during IT injection. Results of statistical anal ysis using the Wilcoxon Matc hed-Pairs Signed-Rank test indicated that there were signi ficant differences in lumen diameter before and after IT injection with 5-FU (p<0.001). Upon examina tion of each tumor lo cation individually, data indicated that there was significant im provement in lumen diameter for all tumor locations except those involving the carina and main bronchus (p=0.225). This group had the lowest number of patients a nd also exhibited large variabil ity, likely the cause of this lack of significance. Four of five patients w ith tumors in this location experience at least partial response to IT 5-FU injections. Stat istical results are summarized in Table 6.33. Table 6.33: Results of Wilcoxon Matched-Pair s Signed-Rank tests using lumen opening values from individual patient examina tions before and after IT treatment. Tumor Location n p value Trachea 8 0.014 Carina + main bronchus 5 0.225 Main Bronchus or bronchus intermedius 32 < 0.001 Lobar bronchus 20 < 0.001 Total 65 < 0.001 6.3 Intratumoral Administra tion of Cisplatin through a Bronchoscope Followed by Irradiation for Treatment of Inoper able Non-Small Cell Lung Cancer 6.3.1 Study Summary Patients with inoperable NS CLC and greater than 50% symptomatic obstruction of the trachea or a main bronchus were enrolled in the study (n=23, age 43-78 years). All patients had stage III diseas e of squamous or adenocar cinoma cell types. Patients underwent complete medical history, physic al examination, chest x-rays, full blood count, bronchoscopy, and computed tomography (C T) prior to treatment in addition to

PAGE 145

126 other examinations. Tumor size and burden as determined via bronchoscopy were recorded before and after treatment. Four IT injections of up to 40mg of cisp latin (4mg/mL) were administered to all patients in the same manner as the 5-FU study described is section 6.2.1 over a three week period. Debulking efficacy was again graded using the semi-quantitative three point scale in Table 6.31. Within 7 days of the final session of IT chemotherapy, patients underwent irradiation therapy w ith a curative intent using a standard 60 Gy dose for NSCLC whenever possible. If residual tu mor was evident one month post-irradiation a second course of 6 weekly IT cisplatin injec tions was started. Survival was determined with a six month minimum follow-up period. Patient initial and final lumen opening (LO) measurements were used to determine if any difference lumen diameter occurred following IT cisplatin injection and irradi ation using a Wilcoxon Matched-Pairs SignedRank test ( =0.05). Survival data were analyzed using one-way ANOVA ( =0.05), followed by TukeyÂ’s MCT when applicable to determine if treatment efficacy or tumor location affected survival time. 6.3.2 Results Nineteen of 23 patients experi enced at least a moderate response to IT cisplatin injections, with 11 patients ach ieving greater than 50% increas e in lumen diameter (good response). Using Wilcoxon Matched-Pairs Sign ed-Rank analysis, an overall significant improvement in lumen opening after IT cispla tin injections (p<0.001). Examination of the response to IT injections of individual tumor locations revealed that lumen opening was significantly improved for tumors in a ll locations except thos e involving the carina and main bronchus (Table 6.34).

PAGE 146

127 Table 6.34: Summary of result s using lumen opening measurements from individual patients before and after IT treatment by tumor location. Tumor Location n p value Trachea 4 0.050 Carina + main bronchus 5 0.186 Main bronchus or bronchus intermedius 14 0.001 Lobar 5 0.030 Total 28 <0.001 Analysis of mean survival by response level to IT cisplatin injections indicated that patients with good response (>50% ILD) had si gnificantly longer mean survival (684 days) than patients with both moderate and small responses (369 and 213 days, respectively). Additionally, four patients were still alive at the time of this analysis with no local recurrence. Tumor location was also found to have an effect on mean survival time (p=0.003). Tukey MCT comparisons revealed that patients with tumors located in the trachea had significantly longer survival ( 852 days) compared with all other groups. There was no significant difference detecte d, however, between any of the remaining groups. These analyses are summar ized in Table 6.35 and Table 6.36. Table 6.35: Patient survival by efficacy of IT chemotherapy debulking using IT cisplatin injections. Response to IT treatment n Residual tumor Mean Survival (95% CI-days) Patients alive at study end (survival days) p value Good† + 11 0 684 (509, 858) 3 (485, 824, 751) < 0.001 Moderate+ 8 3 369 (26, 475) 1 (446) Small† 4 4 213 (70, 355) 0 † and + denote significant differences between categories with matching icon. Table 6.36: Patient survival following IT cisplatin treatment by tumor localization on bronchial tree. Tumor Location n Mean Survival (95% CI-days) Patients alive at study end (survival days) p value Trachea + † * # 3 852 (412, 1293) 0 0.003 Trachea + MB+ 1 237 0 Carina + MB† 4 256 (-10, 522) 0 MB or BI# 13 488 (357, 620) 4 (485, 824, 446, 751) Lobar bronchus* 2 335 (-910, 1580) 0 †,+,*, and # denote significant differences between categories with matching icon. MB = main bronchus; BI = bronchus intermedius

PAGE 147

128 6.4 Discussion Patients with inoperable lung cancer often suffer from potentially life-threatening airway obstructions. Several treatment m odalities are available for airway debulking including mechanical removal of tumor tissue, laser resec tion, or endobronchial radiation therapy. However, these interventional bronc hoscopic procedures are not available at all treatment centers and may be contraindicated. In such cases a new treatment method is desired. IT chemotherapy has been shown to be an effective a lternative and provides rapid relief from airway obstr uction in the present studies. IT injection of 5-FU was shown to be safe, achieving high local chemotherapy doses with no major adverse side effects or toxicity. Over 87% of patients receiving IT 5-FU injections experienced at least 25% ILD, with over half of patients attaining a minimum of 50% ILD. Endoscopic IT injection of cisplatin wa s found to be effective at debulking obstructed airways in patients with inoperable NSCLC with no serious procedure-related complications or systemic toxicity. Over 82% of patients receiving IT cisplatin treatment experienced at least 25% incr ease in lumen diameter. No residual tumor was found in any patients with over 50% ILD (48% of pa tients), and patients that underwent a second course of IT injections showed no residual tumor at the end of this course. Patients that experienced greater than 50% ILD survived to a median of 636 days, more than double what has been reported for irradiation alone.50 Additionally, 4 patients were alive at the conclusion of the study and none of the pati ents with >50% ILD experience local recurrence during the follow-up period. These studies illustrate the promise of IT chemotherapy as a safe, effective treatment of cancers with solid primary tumors. Further, more complete clinical studies

PAGE 148

129 should be done to further elucidate any di fferences between IT chemotherapy and other type of intervention.

PAGE 149

130 CHAPTER 7 CONCLUSIONS The focus of this research was the synthesis of mesospheres of various protein/biopolymer matrices in an effort to achieve pr olonged release and increased loading of mitoxantrone for intratumoral chemotherapy applications. The specific aims of this research included: 1. Synthesis unloaded novel protein/biopolymer meso/microsphere compositions. 2. Synthesis of in situ MXN-loaded novel protein meso/microsphere compositions. 3. Evaluation of in vitro MXN-loading and release from each in situ MXN-loaded composition 4. Investigation of the in vitro cytotoxicity of in situ MXN-loaded meso/microspheres. 5. In vivo evaluation of neoadjuvant IT chem otherapy and scheduled multiple IT injections of in situ MXN-loaded protein mesospheres. 7.1. Unloaded Meso/Microsphere Synthesis 7.1.1. Gelatin-PGA Mesosphere Synthesis Gelatin-poly(glutamic) acid mesospheres (G -PGA) were successfully synthesized with four different aqueous solution con centrations of gelatin and PGA. G-PGA microspheres containing 5% (w/v) PGA were larg er than the desired particle size range and most particles lacked spherical mor phology. In contrast, 1.5% (w/v) or less PGA content resulted in smooth, spherical mesos pheres with an overal l mean dry particle diameter of approximately 2 m. Based on particle size di stributions the majority of these microspheres fell within a range of 0.04 – 10 m. Some agglomeration of particles occurred upon drying, verified visually vi a SEM and also by the presence of large

PAGE 150

131 “particles” on the distribution curves in excess of 50 – 100 m. These compositions exhibited much lower increase in particle diameter when fully swollen in PBS (194%) compared to mesospheres synthesized from ge latin alone (G-C) using the same synthesis parameters (494%). 7.1.2. Gelatin-Bovine Serum Albumin Mesosphere Synthesis Smooth, spherical gelatin-bovine serum albumin mesospheres (G-BSA) were synthesized, though some mesospheres exhibite d a rough surface textur e within each of the four different aqueous solution concentratio ns of gelatin and BSA. Overall, the mean dry particle size for these co mpositions was approximately 2 m, with nearly all particles in the 0.04 – 20 m range. Particle distribution curves for each composition indicated very few particles in the 10 – 20 m range, regardless of composition. Swelling for these compositions was also found to be much lo wer than G-C and averaged less than 65% increase in swelling diameter compared to 494% for G-C. 7.1.3. Bovine Serum Albumin-Deoxyribo nucleic Acid Mesosphere Synthesis Bovine serum albumin-deoxyribonucleic aci d in varying concentrations produced spherical mesospheres (BSA-DNA) for middle and high crosslink densities with varying degrees of surface rougness. BSA-DNA MS with low crosslink densities lacked spherical morphology, likely due to insufficient crosslinking. Keeping crosslink density the same, dry and swollen particle size as measured in swelling studies appeared to increase with increasing DNA content. Th is was to be expected, as increasing DNA content greatly increases the overall viscosity of the aqueous solution used in synthesis as a result of its high molecular weight. Particle s with the lowest cro sslink density appeared to swell the most, also as expected, with the exception of one composition. The highest

PAGE 151

132 crosslink density composition for mesospheres made from 7.5% BSA and 2.5% DNA (w/v) content experienced a larger degree of swelling than all other compositions, with an 80% increase in particle diameter compared to less than 35% for the lowest crosslink density in this composition. At this time, th is cause of this discrepancy has not been uncovered. Mean swollen particle sizes fo r some BSA-DNA MS compositions neared 20 m, making these questionable for IT injections based on previous in vivo studies with G-C. 7.2. In situ Mitoxantrone-Loaded Microsphere Synthesis and In Vitro MXN Release 7.2.1. In situ MXN-loaded Gelatin-PGA Mesospheres Synthesis & Release In situ MXN-loaded G-PGA were successf ully synthesized, though particle morphology lacked sphericity for a large numbe r of particles. Rough surface texture was also observed for both compositions (8.5% gelatin – 1.5% PGA and 10% gelatin – 1.5% PGA, w/v). Particle swelling was lower than unloaded counterparts, with mean increase in particle diameter of 106% and 35%, resp ectively. Mean particle diameter was also somewhat larger at 2.5 and 3.4 m, respectively. Particle size distribution curves revealed that some particles had diameters in excess of 20 – 30 m. These results correlated well the drastically slower MXN release compared with previous BSA and gelatin compositions. MXN was released in excess of 1000 hours, compared with less than 10 hours for gelatin (Gel MS) an d less than 24 hours for most BSA meso/microsphere (BSA MS) compositions. Cumulative MXN release was lower than expected, at less than 40%, compared with 50 – 80% of previous Gel and BSA microspheres. MXN loading was also lower de spite the addition of an ionic species in the microsphere matrix. MXN loading may have been somewhat underestimated due to lack

PAGE 152

133 of degradation in vitro. The exceptionally slow release of these mesospheres may preclude them from being effective in an IT chemotherapy application. However, these MS may prove useful in combination with othe r faster releasing systems, or in another application altogether. 7.2.2. In situ MXN-loaded Gelatin-PGA Mesospheres Synthesis & Release Smooth, spherical BSA-DNA were synthesi zed with an average dry particle diameter of approximately 7 m. Swelling was low compared with other MXN containing compositions and with unloaded BSA-DNA MS, with an average increase in mean particle diameter of less than 25% and ranged from 5% to 40% depending on composition. The largest mean swollen particle size was 12.2 m, which should be easily injectable for IT treatment. MXN loading was underestimated due to lack of complete digestion during these studies, but averaged approximately 8.5% (w/w) compared to ~10% for BSA controls. A pilot composition synthesized from 10% BSA and 1.5% DNA showed promising in vitro release results, with release extending past 1000 hours and a higher cumulative total MXN dose ( g/mL) released compared to GPGA and G-C microspheres. However, al l additional BSA-DNA compositions were synthesized with DNA from a different sour ce and release results were disappointing. Release extended only to approximately 10 – 12 hours and was comparable to G-C alone with no DNA incorporation. 7.3. In Vitro Cytotoxic Properties of In Situ MXL-Loaded Mesospheres 7.3.1. Gelatin-PGA MS In Vitro Cell Culture Studies In vitro RG-2 cell culture studies indicated decreased cellular proliferation compared with non-treatment controls at a ll dose levels. Free MXN positive controls

PAGE 153

134 however outperformed G-PGA MS with complete cellular in hibition at higher doses. This was not entirely unexpected based on the slow release of MXN from G-PGA-M microspheres. 7.3.2. BSA-DNA MS In Vitro Cell Culture Studies In vitro RG-2 cell culture studi es revealed that BSADNA M1 MS (herring DNA source) were as effective as F-MXN alone at eliminating cellular viability. The middle and high doses resulted in complete inhibi ting of cellular viabili ty by day 1 for both treatment groups. All cells in both groups at these doses show ed no signs of health, with no processes extended outward, blue colorati on indicating uptake of MXN into the cells, and some cellular debris at day 4. Low doses of both groups reduced cellular viability significantly, though some viable cells were still present at day 4 of the study. Further studies should be conducte d using BSA-DNA-M MS that incorporated salmon DNA and had a more rapid release profile to assess the ability of thes e MS to eliminate viability. 7.4. In vivo Evaluation of Protein Mesospheres 7.4.1. Neoadjuvant IT Chemotherapy The use of IT chemotherapy in combination with surgery proved to be effective at improving survival of C3h/HeJ mice in a 16/ C murine mammary adenocarcinoma model. This combination of an initi al IT injection of 5-10 m BSA mesospheres followed by surgical excision of the tumor 1 to 15 days later resulted in cure rates up to 100%, compared to 0% for controls (surgery only). Surgery at time points clos er to the initial IT injection facilitate tumor excision by reduc ing tumor size compar ed with non treated controls. No significant toxicities were obs erved with this therapy, as evidenced by the lack of indication of leucopenia or acut e inflammation found in blood cell counts of control and IT chemotherapy treated animals.

PAGE 154

135 7.4.2. Scheduled Multiple IT Injections The use of multiple IT injections on a sc heduled regimen more closely resembles the current clinical practi ce with systemic chemotherapy. Animals receiving IT injections at 1 weeks intervals for 2 or 3 w eeks showed marked impr ovement in survival, with 70% survival rate for IT MXN and IT MXN combined with protein MS. Toxicity was a factor in this survival , as 20% of mice receiving co mbination therapy died as a result of toxic side effects ra ther than tumor regrowth. Spaci ng injections farther apart or reducing the initial and/or follow-up doses may further increase survival by avoiding or reducing these effects.

PAGE 155

136 CHAPTER 8 FUTURE WORK As studies progress through much of any re search, new avenues of interest present themselves. The following are research interests that should be investigated in the future. 1. Synthesis of in situ MXN-loaded Gel-BSA mesospheres. Unloaded G-BSA MS showed promise, as these swelled less th an MS of gelatin alone, and may combine the benefits of increased MXN release of gelatin MS with the decreased swelling and slower degradation of BSA MS. 2. Synthesis of MS with DNA as the sole matr ix material. Preliminary synthesis of these MS is promising, though synthesis c onditions need to be optimized. DNA is an excellent drug carrier for MXN base d on its natural interactions, and these MS may prove to have more optimal rel ease and loading profiles. Additionally, these MS may possess the unique ability to be crosslinked by MXN alone, requiring no additional crosslinki ng agent during synthesis. 3. Evaluation of DNA releas e from BSA-DNA MS and/or DNA MS. Release of DNA or plasmid may lead to gene therapy applications for th ese MS in addition to the currently investigated chemotherapy applications. 4. Evaluation of combinations of different sized MS to optimize or select release custom release profiles. As different part icle sizes release drugs differently based on changes in surface area:volume ratio, mixing various sizes may allow release profiles to be tailored.

PAGE 156

137 5. Application of IT chemotherapy using MS to other cancer types in vivo . None of the MS presented here have been examined for treatment of any other cancer line than 16/C MMAC. These MS may also be useful for other cancers with well defined primary tumors such as lu ng, colorectal, and brain cancers. 6. Investigation of other prot eins and/or biopolymers as matrix materials for MS. Different proteins and co mbinations of proteins and anionic biopolymers may prove to have improved release and loadi ng characteristics, and consequently may be more effective in vitro and in vivo . 7. Evaluation of in situ loading other cytotoxic drugs into protein MS. Such drugs may include cisplatin, doxor ubicin, and cyclophosphamide.

PAGE 157

138 LIST OF REFERENCES 1. Goldberg EP, Hadba AR, Almond BA , Marotta JS. Intratumoral cancer chemotherapy and immunotherapy: opport unities for nonsystemic preoperative drug delivery. Journal of Pharmacy and Pharmacology. 2002;54(2):159-180. 2. Ries LAG, Eisner MP, Kosary CL, et al. SEER Cancer Statistics Review, 19752001. http://seer.cancer.gov/csr/1975_2001/ . Accessed October 4, 2004. 3. Ries LAG, Eisner MP, Kosary CL, et al. SEER Cancer Statistics Review, 19752002. http://www.seer.cancer.gov/csr/1975_2002/ . Accessed April 12, 2006. 4. Cancer Facts & Figures 2004 . Atlanta, Ga: American Cancer Society; 2004. 5. Cancer Facts & Figures 2006 . Atlanta, GA: American Cancer Society; 2006. 6. Celikoglu F, Celikoglu S. Intratumoral chemotherapy with 5-fluorouracil for palliation of bronchial cancer in pati ents with severe airway obstruction. Journal of Pharmacy & Pharmacology. 2003;55:1441-1448. 7. Celikoglu F, Celikoglu S, York AM, Goldbe rg EP. Intratumoral admisitration of cisplatin through a bronchoscope follow ed by irradiation for treatment of inoperable non-small cell obstructive lung cancer. Lung Cancer. 2006;51:225236. 8. Longo WE, Iwata H, Lindheimer TA , Goldberg EP. Hydrophilic albuminpolyglutamic acid-adriamycin microsphe res for localized chemotherapy. Paper presented at: Society of Biomaterials, 1982. 9. Goldberg EP, Iwata H, Longo WE. Hydrophilic albumin and dextran ionexchange microspheres for localized chem otherapy. In: Davis SS, Illum L, McVie JG, Tomlinson E, eds. Microspheres and drug delivery . Amsterdam: Elsevier Science Publishers; 1984:309-325. 10. Knepp WA, Jayakrishnan A, Quigg JM, Sitren HS, Bagnall JJ, Goldberg EP. Synthesis, properties, and intratumoral evaluation of mitoxa ntrone-loaded casein microspheres in Lewis lung carcinoma. J Pharm Pharmacol. Oct 1993;45(10):887-891.

PAGE 158

139 11. Cuevas BJ. Synthesis and properties of pr otein micro/mesosphere-drug compositions designed for intratumoral cancer therapy [Ph.D. Dissertation]. Gainesville: Materials Science & Engin eering, University of Florida; 2003. 12. Freeman ST, York AM, Cuevas BJ. Evaluation of gelatin mesospheres for the treatment of 16/C murine mammary adenocarcinoma : University of Florida (unpublished); 2003. 13. Almond BA. Mitoxantrone-loaded albumin microspheres for localized intratumoral chemotherapy of breast cancer [Ph.D. Dissertation]. Gainesville: Materials Science & Engineering, University of Florida; 2002. 14. Hadba AR. Synthesis, properties, and in vivo evaluation of sustained release albumin-mitoxantrone microsphere formul ations for nonsyste mic treatment of breast cancer and other high mortality cancers . Gainesville: Materials Science & Engineering, University of Florida; 2001. 15. DePrez-DeCampeneere D, Jaenke R, Trouet A, Baudon H, Maldague P. DNAanthracycline complexes II. Comparativ e study of the acute lesions induced in mice after intravenous administrati on of free and DNA bound Adriamycin. European Journal of Cancer. 1980;16(8):987-998. 16. Trouet A, Baurain R, Deprez-De Campen eere D, Layton D, Masquelier M. DNA, liposomes, and proteins as carriers for antitumoral drugs. Recent Results in Cancer Research. 1980;75:229-235. 17. Markman M. Principles of regional antin eoplastic drug delivery. In: Markman M, ed. Regional chemotherapy: clinic al research and practice . Totowa: Humana Press, Inc.; 2000:1-4. 18. Kemeny NE, Atiq OT. Intrahepatic ch emotherapy for metastatic colorectal cancer. In: Markman M, ed. Regional chemotherapy: clinical research and practice . Totowa: Humana Press, Inc.; 2000:5-20. 19. Markman M. Intraperitoneal antineoplastic drug deliver y: rationale and results. Lancet Oncology. 2003;4(5):277-283. 20. Doroshow JH. Anthracyclines and Antr acenediones. In: Chabner BA, Longo DL, eds. Cancer chemotherapy and biotherapy . 2nd ed. Philadelphia: LippencottRaven; 1996:409-434. 21. Faulds D, Balfour JA, Chrisp P, Lang try HD. Mitoxantrone. A review of its pharmacodynamic and pharmacokinetic properties, and therapeutic potential in the chemotherapy of cancer. Drugs. 1991;41(3):400-449.

PAGE 159

140 22. Crossley RJ. Clinical safety and to lerance of mitoxantrone (Novantrone). Cancer Treatment Reviews. 1983;10(Suppl B):29-36. 23. McArdle CS. Regional chemotherapy for lo cally advances breast cancer. In: Kerr DJ, McArdle CS, eds. Regional chemotherapy: theory and practice . Amsterdam: Harwood Academic Publishers; 2000:51-63. 24. Nagel JD, Varossieau FJ, Dubbelman R, ten Bokkel Huinink WW, McVie JG. Clinical pharmacokinetics of mitoxantr one after intraperit oneal administration. Cancer Chemotherapy and Pharmacology. 1992;29(6):480-484. 25. Doroshow JH. Anthracyclines and Antr acenediones. In: Chabner BA, Longo DL, eds. Cancer chemotherapy and biotherapy . 2nd ed. Philadelphia: Lippencott Williams & Wilkins; 2001:409-434. 26. Durr FE, Wallace RE, Citarella RV. Mol ecular and biochemical pharmacology of mitoxantrone. Cancer Treatment Reviews. 1983;10(Suppl B):3-11. 27. Faulds D, Balfour JA, Chrisp P, Lang try HD. Mitoxantrone. A review of its pharmacodynamic and pharmakokinetic properties, and therapeutic potential in the chemotherapy of cancer. Drugs. 1991;41(3):400-449. 28. Reents S. Mitoxantrone. Clinical Phar macology, Gold Standard Multimedia Inc]. 09/14/2004; http://cpip.gsm.com/ . 29. Wilkes GM, Ades TB. Consumer's Guide to Cancer Drugs . Boston: Jones and Bartlett Publishers; 2000. 30. Sugio S, Kashima A, Mochizuki S, Noda M, Kobayashi K. Crystal structure of human serum albumin at 2.5 A resolution. Protein Engineering. 1999;12:439-446. 31. Marieb EN. Human Anatomy & Physiology . 4th ed. Menlo Park: Addison Wesley Longman, Inc.; 1998. 32. Peters T. All about albumin: biochemistry, genetics, and medical applications . San Diego: Academic Press, Inc.; 1996. 33. Chen Y, Burton MA, Gray BN. Pharm aceutical and methodological aspects of microparticles. In: Willmott N, Daly J, eds. Microspheres and regional cancer therapy . Boca Raton: CRC Press; 1994. 34. Jarakrishnan A, Jameela SR. Glutaralde hyde as a fixative in bioprostheses and drug delivery matrices. Biomaterials. 1996;17(5):471-484. 35. Longo WE, Goldberg EP. Hydrophilic albumin microspheres. Methods in Enzymology. 1985;112:18-26.

PAGE 160

141 36. Bhavsar MD, Tiwari SB, Amiji MM. Formulation optimization for the nanoparticles-in-microsphere hybrid oral delivery system using factorial design. Journal of Controlled Release. 2006;110:422-430. 37. Park H, Temenoff JS, Holland TA, Taba ta Y, Mikos AG. Delivery of TGF-beta1 and chondrocytes via injectible, biodegra dable hydrogels for cartilage tissue engineering applications. Biomaterials. 2005;26:7096-7103. 38. Morimoto K, Chono S, Kosai T, Seki T, Tabata Y. Design of novel injectable cationic microspheres based on aminated gelatin for prologned insulin delivery. Journal of Pharmacy & Pharmacology. 2005;57:839-844. 39. Bigi A, Bracci B, Cojazzi G, Panzavolta S, Roveri N. Drawn gelatin films with improved mechanical properties. Biomaterials. 1998;19:2335-2340. 40. Yannas IV. Natural Materials. In: Ratn er BD, Hoffman AS, Schoen FJ, Lemons JE, eds. Biomaterials Science: An Intr oduction to Materials in Medicine . San Diego: Academic Press; 1996:84. 41. Stopek JB. Biopolymer-microglia cell compos itions for neural tissue repair [PhD dissertation]. Gainesville: Materials Sc ience & Engineeri ng, University of Florida; 2003. 42. Kirk JF. Protein microspheres for controlled dr ug delivery and related analysis of biopolymers [Ph.D. Dissertation]. Gainesville: Materials Science & Engineering, University of Florida; 1997. 43. Almond BA, Hadba AR, Freeman ST, et al. Efficacy of mitoxantrone-loaded albumin microspheres for intratumoral chemotherapy of breast cancer. Journal of Controlled Release. 2003;91:147-155. 44. Yapel Jr AF. Albumin microspheres: heat and chemical stabilization. Methods in Enzymology. 1985;112(Part A):3-18. 45. Longo WE, Iwata H, Lindheimer TA, Go ldberg EP. Preparation of hydrophilic albumin microspheres using pol ymeric dispersing agents. J Pharm Sci. 1982;71(12):1323-1328. 46. Enriquez I. DNA microsphere synthesis: University of Fl orida (unpublished); 2004. 47. York AM, Cuevas BJ, Martin PJ, Goldbe rg EP. Synthesis and characterization of gelatin-polyglutamic aci d microspheres for local delivery of chemotherapy agents. Paper presented at: Surfaces in Biomaterials Annual Meeting & Exhibition, 2003; Savannah, GA.

PAGE 161

142 48. Arshady R. Albumin microspheres a nd microcapsules: methodolgy of manufacturing techniques. Journal of Controlled Release. 1990;14:111-131. 49. Celikoglu S, Karayel T, Demirci S, Celi koglu F, Cagatay T. Di rect injection of anti-cancer drugs into endobronchial tu mors for palliation of major airway obstruction. Postgraduate Medical Journal. 1997;73:159-162. 50. Petrovich Z, Stanley K, Cox JD, Paig C. Radiation in the management of locally advanced luncg cancer of all cell types. Cancer. 1981;48(1335-40).

PAGE 162

143 BIOGRAPHICAL SKETCH Amanda Michele York was born on May 22, 1978, in Indianapolis, Indiana, to Dennis and Paula York. She liv ed in Indiana until the age of eight, when her family moved to Homestead, Florida. Amanda gr aduated from South Dade High School in 1996 as Salutatorian of her class. Her interest in science and mathematics took her to Stetson University for undergraduate studies in physic s. She received a Stetson Undergraduate Research Experience grant to conduct research over the summer of 1998. It was here that she discovered her love for research. She began to work on what would become her senior research project en titled “Development of a TV Holography System for Modal Analysis of Musical Instruments” under the guidance of Dr. Kevin Riggs. While at Stetson, Amanda was a member of the varsity fastpitch softball team from her freshman through junior years. She graduated Cum Laude with a Bachelor of Science degree with a physics major and mathematics minor from Stetson University in May 2000. Amanda began graduate school at the Un iversity of Florida in August 2000 to pursue a Master of Science degree in biomed ical engineering with a concentration in biomaterials under the guidance of Dr. Eugene Goldberg. She received her MS in May 2003, but her love of learning and research kept her at the University of Florida to earn a Doctor of Philosophy degree in biomedical engineering as well. She completed her research on drug-loaded protei n microspheres for intratumoral chemotherapy in spring 2006 while working full-time as a research & development intern at Regeneration

PAGE 163

144 Technologies, Inc. in Alachua, Florida. Amanda received her Doctor of Philosophy degree in May 2006. Amanda enjoys being creative in her free time, including making hand made cards using rubber stamping, scrapbooking, woodworking, and jewelry making. She is an avid sports fan and enjoys watching nearly all sp orts, including football, basketball, and auto racing. She also loves being outdoors and playing with her dog, Peanut. She looks forward to a research career in the biom edical industry and to starting a family.