Title: Synthesis, properties, and in vivo evaluation of sustained release albumin-mitoxantrone microsphere formulations for nonsystemic treatment of breast cancer and other high mortality cancers
CITATION PDF VIEWER THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00100827/00001
 Material Information
Title: Synthesis, properties, and in vivo evaluation of sustained release albumin-mitoxantrone microsphere formulations for nonsystemic treatment of breast cancer and other high mortality cancers
Physical Description: Book
Language: English
Creator: Hadba, Ahmad Robert, 1971-
Publisher: University of Florida
Place of Publication: Gainesville Fla
Gainesville, Fla
Publication Date: 2001
Copyright Date: 2001
 Subjects
Subject: Materials Science and Engineering thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Materials Science and Engineering -- UF   ( lcsh )
Genre: government publication (state, provincial, terriorial, dependent)   ( marcgt )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )
 Notes
Summary: ABSTRACT: Methods for preparing mitoxantrone (MXN)-loaded albumin microspheres for the treatment of breast cancer were developed. The effect of processing conditions on the particle size of unloaded and MXN-loaded microspheres was evaluated using multivariate analyses. The data suggested that the particle size of unloaded microspheres increased as protein concentration increased or the steric stabilizer concentration decreased. In addition, synergy between these two variables was observed. In situ-loading of MXN achieved loading efficiencies in excess of 80%. Comparable efficiencies were achieved with post-synthesis loading when the microsphere were prepared from albumin-poly(glutamic acid) blends. In vitro release of MXN in phosphate buffered saline under infinite sink conditions showed that the total amount of drug released increased as the glutaraldehyde concentration decreased. This trend was reversed when the microspheres were incubated in plasma. Nanoparticles were also prepared using ethanol desolvation. These particles were dispersible in saline and easily modified with amino acids. In addition, particle size could be varied by use of different non-ionic surfactants in the preparation. The effect of intratumoral (IT) versus intravenous (IV) drug administration on tumor response and systemic toxicity was investigated in vivo using the 16/C murine mammary adenocarcinoma tumor model. The data suggested that IT-treated animals had significantly smaller tumors and lower weight loss when compared to IV-treated animals. Furthermore, the addition of surgery to the chemotherapy further improved the survival of the animals.
Summary: ABSTRACT (cont.): Pilot studies using MXN-albumin microspheres suggested that microspheres could be safely administered IT in doses up to 48 mg/kg. However, there was no evidence that this higher dose resulted in improved long term survival when compared to the 32 mg/kg dose. The maximum tolerated dose of MXN given IT was approximately 12 mg/kg. The animal studies suggested that IT chemotherapy could result in significant reduction in tumor burden and systemic toxicity when compared to conventional IV chemotherapy. Microsphere-bound drug could be safely administered IT at very high doses. In addition, microspheres exerted an antitumoral effect over an extended period of time; however, complete tumor regression was achieved only by complete tumor perfusion.
Summary: KEYWORDS: microspheres, albumin, albumin microspheres, microparticles, nanoparticles, zeta-potential, murine, mouse, tumor model, murine adenocarcinoma, mitoxantrone, Novantrone, non-systemic treatment, intratumoral injections, local injections
Thesis: Thesis (Ph. D.)--University of Florida, 2001.
Bibliography: Includes bibliographical references (p. 179-196).
System Details: System requirements: World Wide Web browser and PDF reader.
System Details: Mode of access: World Wide Web.
Statement of Responsibility: by Ahmad Robert Hadba.
General Note: Title from first page of PDF file.
General Note: Document formatted into pages; contains xxii, 198 p.; also contains graphics.
General Note: Vita.
 Record Information
Bibliographic ID: UF00100827
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 49241232
alephbibnum - 002763574
notis - ANP1596

Downloads

This item has the following downloads:

master ( PDF )


Full Text











SYNTHESIS, PROPERTIES, AND IN VIVO EVALUATION OF SUSTAINED
RELEASE ALBUMIN-MITOXANTRONE MICROSPHERE FORMULATIONS FOR
NONSYSTEMIC TREATMENT OF BREAST CANCER AND OTHER HIGH
MORTALITY CANCERS

















By

AHMAD ROBERT HADBA


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001




























Copyright 2001

by

Ahmad Robert Hadba


























To my loving parents Khodr and Connie Hadba for their unconditional support and
infinite patience throughout my college experience.















ACKNOWLEDGMENTS

I would like to express my gratitude to my advisor and committee chairman, Dr.

Eugene P. Goldberg, for his guidance, patience, and support on this project. I would also

like to thank my committee members, Dr. Anthony Brennan, Dr. Christopher Batich, Dr.

Stanley R. Bates, and Dr. John Reynolds. In addition, I would like to recognize Dr.

Dietmar Siemann and his research group for generously providing the 16/C murine

mammary adenocarcinoma cell line and training in handling and transplanting this tumor.

Special thanks are extended to Dr. Guenther Hochhaus and his research group, especially

Intira Coowanitwong, for providing access to high performance liquid chromatography

analysis; Dr. Wolfgang Streit (Jake) for his help with tissue processing, imbedding and

sectioning; and Dr. Carol Detrisac for her help in analyzing the histology slides and

suggestions for future experiments.

I wish to acknowledge the following colleagues: Dr. James Marotta for his

support, enthusiasm, input, and experience-based advice inside and outside the

laboratory; Dr. Drew Amery for his help in learning new instrumental techniques and

reviving "dead" instrumentation; Dr. James Kirk for his experience-based input and

advice on microsphere synthesis and mitoxantrone properties; Dr. Christopher

Widenhouse for his scientific skepticism and unique perspective; Dr. Kaustubh Rau for

his help, advice on various research projects, and friendship; Dr. Michelle Fluegge for her

friendship, support, and help with animals; and Jennifer Russo (soon to be Mrs. Jennifer

Braunschweig) for her unconditional friendship, love, and support.









Special appreciation is extended to my colleagues for their assistance, advice, and

encouragement. These colleagues include Brett Almond, Joshua Stopek, Brian Cuevas,

Daniel Urbaniak, Amy Gibson, Patrick Leamy, Dr. Jeremy Mehlem, Jeanne McDonald,

Payam Chini, Dr. Lynn Peck, and Paul Martin.

I would also like to thank my family and friends, both in the United States and in

Lebanon, for their love and support. Special thanks go to my grandparents, Robert and

Louise Pavey, for their unconditional love and support; Carl and Doreen Pavey for

helping me get on my feet when I first moved to the United States; and to my brother,

Wissam Hadba, for remaining close to my parents and grandparents in Beirut during my

absence. A special thank you is extended to my uncle, Dr. Imad Hadba, for his financial

and moral support during my early years in the United States.
















TABLE OF CONTENTS

page

A C K N O W L E D G M E N T S ........................................................................ .....................iv

LIST OF TABLES ................................... .. ... .... ................. ix

L IS T O F F IG U R E S ............................................................................... ................. .. x ii

KEY TO SYMBOLS AND ABBREVIATIONS ...........................................................xvii

A B S T R A C T ...................................... .................................................. . x x i

1 INTRODUCTION ...................... .... ...................................... .....1.

2 BA CK GROUND ................................................... ....... ................ .5

2.1 Introduction................................................................ ... ..... ........ 5
2 .2 B rea st C an cer ................... ................... ............................... ............... 6
2.2.1 B reast C ancer Treatm ent .......................................... ........................... 7
2.2.1.1 Com m on breast cancer types..................................................................... 8
2.2.1.2 B reast cancer staging......................................... .............................. 9
2.2.1.3 Breast cancer treatm ent ................. ..... ........ .......... ................ ... ........... 9
2.2.2 Evolution of Conventional Breast Cancer Treatment............... .............. 10
2.2.3 Intratum oral C hem otherapy................................... ...................................... 13
2.3 N ovantrone (M itoxantrone) ...................... ..... ............................ .............. 16
2.3.1 M echanism of Action...................... ...................... ... ................. 16
2.3.2 Pharmacokinetic and Toxicological Properties .......................................... 19
2.4 Album in ........................ ............... ............. .......... 24
2.4.1 Albumin Composition and Structure.................... ............................ .... 24
2.4.2 A lbum in M icrospheres .............. ......................................................... 26
2 .4 .2 .1 M icrosphere synthesis..................................................... ... ................. 26
2.4.2.2 N anosphere synthesis ......... ................. ............................ .............. 30
2.4.2.3 Particle size control................................................. .......................... 31
2.5 16/C Murine Mammary Adenocarcinoma .................................................... 34
2.6 G oals of this R esearch........... .................. ....... ................... .............. 36

3 M ATERIALS AND M ETHODS ........................................ .......................... 38

3.1 M materials .......................... .............. ................. ......... 38
3.1.1 B suffered Solutions .................... ................. ................................... 40









3.1.2. Protein and Reagent Solutions ..... .............................................. 42
3.2 M methods ..................................................................................................... 44
3.2 .1 M icrosphere Synthesis......................................................... ... ................. 44
3.2.1.1 General m ethod ............. .. ......................... ..... .... . ...... .. 44
3.2.1.2 In situ-loaded albumin microspheres................................................. 46
3.2.1.3 M miscellaneous m icrosphere syntheses............................... .................... 47
3.2 .2 N anosphere Synthesis.......................................................... ... ... .............. 49
3.2.2.1 Suspension cross-linking...................... .... .......................... 49
3.2.2.2 D esolvation m ethod .................................................................. ... 50
3.2.3 Microsphere and Nanosphere Characterization........................................... 52
3.2.3.1 Electron m icroscopy........................................................ .............. 52
3.2.3.2 Particle sizing ................................. ......... ................... .. 54
3.2.3.3 M itoxantrone post-loading procedure............................... .................... 55
3.2.3.4 Microsphere drug content determination................... ................ 56
3.2.3.5 Zeta potential measurements.................. ........ .............. 58
3.2.4 In Vitro Release of Mitoxantrone from Albumin Microspheres ............... 58
3.2.5 A nim al Studies........................................ ... ................................ 60
3.2.5.1 M urine mammary adenocarcinoma................................... .... ............... 61
3.2.5.2 Tumor response to mitoxantrone..................... .................. 63
3.2.5.3 Tumor response to albumin encapsulated mitoxantrone....................... 66

4 SYNTHESIS AND RELEASE RESULTS AND DISCUSSION .............................68

4 .1 M icrosphere Synthesis ................................. ....................................................... 68
4.1.1 Effect of Processing Parameters on Particle Size .......................................... 68
4.1.1.1 Pilot study................................... 68
4.1.1.2 Full microsphere synthesis study........... ...................................... 70
4.1.1.3 Effect of in situ-loading on particle size ............................................ 80
4.1.2 Effect of Processing Parameters on Particle Morphology............................. 82
4.1.2.1 Full study...................................................................... ........... 82
4.1.2.2 In situ-loaded microspheres .............. ...... .......................... .......... 86
4.1.3 Additional Microsphere Syntheses ....................................................... 94
4.1.3.1 M ixed matrix microspheres.......................... ..................... 94
4.1.3.2 Cross-linking agents other than glutaraldehyde.................................... 99
4.2 M icrosphere L oading ....................................................................... 102
4.2.1 Post-Loaded M icrospheres .............. ...... ........................................ 102
4.2.2 In Situ-Loaded Microspheres............................ ............. 104
4.3 In Vitro Release of Mitoxantrone............................................................... 106
4.4 N anosphere Synthesis ................................................................................. 114
4.4.1 Nanosphere Particle Size Characterization............................................. 116
4.4.2 Preliminary Zeta Potential Measurements ............. ....... .......... .... 119
4.4.3 Synthesis of Magnetic Nanoparticles .......... ...................................... 122
4.4.4 Sum m ary of R esults............................ .......................... .............. 124

5 IN VIVO RESULTS AND DISCUSSION....... .......................................126

5 .1 Intro du action ..................................................... ............. 12 6









5.2 B baseline Study .......................................................... .. .......... 129
5.2.1 Tum or G row th .............................................. .................... .. 129
5.2.2 B ody W eight L oss .............................................. ............................. 131
5.2.3 Survival D ata ..................................... ......... .... .... .. .......... .. 134
5.3 M icrosphere Pilot Studies ................... .............. ............................. 141
5.3.1 Effect of Dose and Dosage Form on Tumor and Survival Pilot Study....... 141
5.3.2 Pilot Study of the Effect of Microsphere-Bound Mitoxantrone Dose on
Tum or and Survival ......................................................... ........ .............. 145
5.4 Summary and Discussion of the Results of the Animal Studies.................... 149

6 SUMMARY AND CONCLUSIONS ........................................... ............... 159

6.1 Introduction ........................................................................... ........ ........ ...... ........ 159
6.2 M icrosphere Synthesis ............................... ... .................................... 159
6.3 M icrosphere L oading ...................... .. .. ......... .. ....................... .............. 162
6.4 In Vitro Release......... ................................................................................. 163
6.5 Animal Studies ............................ ....... .............. 165
6.5.1 Baseline Study .......... .............................. .. .... .... ......... 165
6.5.2 Drug-M icrosphere Therapy Pilot Studies................................................... 165
6.5.3 General Conclusions ........... ..... ........................... 166
6.6 N anosphere Synthesis ................................................ ............................. 167

7 FU T U R E W O R K ......... ................. .......................................... ...........................168

G L O S SA R Y ...................................................... 174

A P P E N D IX ................................................................................................................ 1 7 6

L IST O F R E FE R E N C E S ........................................................................ ...................179

BIOGRAPHICAL SKETCH ...................................... ...................... ... ......... 197
















LIST OF TABLES


Table Page

2.1 Listing of the amino acid residues in human serum albumin and bovine serum
album in. ............................................................................25

2.2 Passive drug targeting based on the particle size of drug carrier and its route of
adm inistration. .......................................................................32

2.3 Biological characteristics of the 16/C murine mammary adenocarcinoma ..........35

3.1 The factors and factor levels investigated in the pilot study investigating the
effect of processing conditions on particle size. ....... ...... ............................. 46

3.2 The factors and factor levels investigated in the full study of the effect of
processing conditions on particle size. .............................. ............. ... ........... 46

4.1 Synthe sis conditions for pilot study microsphere batches. ................................69

4.2 Particle size and size distribution of batches synthesize in pilot study as
measured using digitized SEM images and the computer software package
S cio nTM ................................................................................ 7 0

4.3 Full study process conditions. Batches were synthesized in triplicate for this
study ..............................................................................7 1

4.4 Particle size and size distributions of Full Study microspheres. The following
data are averages of three batches (Run #6 was the average of two batches.).....72

4.5 ANOVA table for the calculated model describing the effect of processing
conditions on particle size ............................................................................... ...73

4.6 Model term coefficients for the second order model. ........................................73

4.7 Experimental conditions for the synthesis of in situ-loaded microspheres..........80

4.8 ANOVA table for the calculated linear model describing effect of process
parameters on particle size of in situ-loaded microspheres. .............................81









4.9 Model term coefficient estimates for the linear model describing effect of
processing conditions on particle size of in situ-loaded microspheres ..............81

4.10 Surface roughness rankings for in situ-loaded albumin microspheres. A rank
of 1 indicates a very smooth surface while a rank of 5 indicates a very rough
su rface ......................................................... ................ . 8 6

4.11 Loading efficiency of post-loaded albumin microspheres. The mitoxantrone
concentration data are averages of three samples for each condition quantified
by the depletion assay and the enzymatic degradation assay............................102

4.12 Effect of polypeptide and polypeptide concentration on MXN loading
efficiency in albumin microspheres. All samples were synthesized using 2 %
CAB and 10 % GTA concentration. The loading was quantified by depletion
assay only because the blend microspheres did not degrade enzymatically........104

4.13 Loading efficiency of in situ-loaded BSA microspheres. The MXN
concentration data are averages of three samples for each condition quantified
by the depletion assay and the enzymatic degradation assay............................105

4.14 Loading efficiency of in situ-loaded BSA microspheres synthesized in the
optimization study. The MXN concentration data are averages of three
samples for each condition quantified the enzymatic degradation assay. The
depletion data were calculated for an n= 1................................... ............... 105

4.15 Effect of capping amino acid and pH on albumin nanosphere particle size.
There was no evidence of a statistically significant difference between groups
having the same superscript at the 95 % level of confidence using a one-way
ANOVA followed by a Tuckey M CT ............................................................. 117

4.16 Effect of capping amino acid and pH on albumin nanosphere particle size.
There was no evidence of a statistically significant difference between groups
having the same superscript at the 95 % level of confidence using a one-way
ANOVA followed by a Tuckey M CT ............................................................. 118

4.17 Ionizable group pKa values for used capping materials. .....................................120

5.1 Normalized tumor weight (NTWT) data for all groups in the baseline study. ....131

5.2 Normalized tumor weight-adjusted animal weight (NAWT) data for all groups
in the baseline study. ............................... .... .......... ................ ............. 132

5.3 Contingency table describing effect of treatment on animal weight. Animals
in the large weight loss category experienced weight losses in excess of 10 %
their body weight on day of sacrifice or on day 10 after treatment, which ever
cam e first........................................................................................... . 134









5.4 Percent increase in life-span for the treatment groups over untreated control.
The ILS for all treatment groups was calculated using the MDD data for the
untreated control "C" group ..................................................... .....................137

A.1 Established factors that increase the relative risk of developing breast cancer
in w om en. .......................................................................... 176

A.2 TNM classification of breast cancer...................... ............... 177

A .3 Stage grouping ....................................................... ................. 178















LIST OF FIGURES


Figure Page

2.1 Chemical structure of mitoxantrone..... .................... ..............16

2.2 Mechanism of superoxide generation for doxorubicin. The reduction takes
place in the presence of reducing enzymes such as NADPH cytochrome P450
redu ctase ............................ .................... . ............................... . 17

2.3 Oxidative activation of mitoxantrone in the presence of horseradish
peroxidase and hydrogen peroxide................................ ......................... 18

2.4 Reactions of glutaraldehyde in aqueous media. (a) Reaction of a primary
amine with glutaraldehyde to form a Schiff base. (b) Formation of a and 3
unsaturated glutaraldehyde polymers and their subsequent reaction with
primary amines via Michael addition. (c) Formation of hemiacetals and their
subsequent reaction with primary amines. ................... ............... .......... 29

3.1 Image of a standard hemacytometer chamber with arrows pointing towards the
areas to be counted in the calculation the number of viable tumor cells. ............62

3.2 Treatment groups in the baseline response of the 16/C adenocarcinoma to
mitoxantrone delivered as free drug. ........................................ ............... 64

4.1 Effect of processing conditions on albumin particle size in micrometers. (a)
Effect of BSA and CAB concentration on particle diameter when the GTA
concentration is fixed at 10 %. (b) Effect of BSA and GTA concentration on
particle diameter when CAB concentration is fixed at 2.75 %. (c). Effect of
CAB and GTA concentration on particle diameter when the BSA
concentration is fixed at 20 % ............................................................ ............... 74

4.2 SEM micrographs of albumin microspheres synthesized in the optimization
study. (a) SEM of sample BSAF1 synthesized using condition #1 in Table
4.3. (b) SEM of sample BSAF33 synthesized using condition # 3 in Table 4.3..84

4.3 SEM micrographs of albumin microspheres synthesized in the optimization
study. (a) SEM of sample BSAF36 synthesized using condition #6 in Table
4.3. (b) SEM of sample BSAF6 synthesized using condition # 6 in Table 4.3....85









4.4 Scanning electron micrographs of in situ-loaded albumin microspheres. The
images illustrate the effect of processing conditions on surface roughness and
porosity. The processing conditions are listed in Table 4.7. (a) SEM
micrograph of sample MXN3. (b) SEM micrograph of sample MXN5. (c)
SEM micrograph of sample M XN7 .................. ...............................87

4.5 Effect of processing conditions on surface roughness of in situ-loaded albumin
m icrospheres.................................... ................................ ........90

4.6 TEM micrographs of in situ-loaded microsphere sections showing the core
structure of the synthesized microspheres. (a) TEM of section from sample
MXN3 at a magnification of 5600 X. (b) TEM of section from sample MXN5
at a magnification of 11,880 X. (c) TEM of section from sample MXN7 at a
m agnification of 12,960 X ................................ ............... ............... 91

4.7 TEM micrographs of blank albumin microsphere sections showing the core
structure of the synthesized microspheres. (a) TEM of section from sample
BSF18 at a magnification of 4560 X. (b) TEM of section from sample BSF25
at a magnification of 11,880 X. (c) TEM of section from sample BSF27 at a
m agnification of 38,200 X .. ..................................................................... ......92

4.8 SEM micrographs of microspheres prepared from albumin and polypeptide
blends. (a) Microsphere matrix contains 5 % (w/w) poly(glutamic acid). (b)
Microsphere matrix contains 5 % (w/w) polylysine. ........................................ 96

4.9 SEM of microspheres prepared from albumin and CMC blends. (a)
Microsphere matrix contains 5 % (w/w) CMC. (b) Microsphere matrix
contains 10 % (w/w) CMC. (c) Microsphere matrix contains 15 % (w/w)
CMC. (d) Higher magnification of (c) showing the rough surface morphology
of the BCM C15 m icrospheres................................. ........................ ......... 97

4.10 Albumin microspheres cross-linked with ferric nitrate. (a) Blank albumin
microspheres. (b) In situ-loaded albumin microspheres.............. ................101

4.11 Normalized cumulative release of mitoxantrone from post-loaded albumin
microspheres in phosphate buffered saline pH= 7.4 at 37 C. The loading and
cross-linking agent concentration information was given in Table 4.11. It
should be noted that the number after the "BSAPL" denotes the GTA
concentration used in the synthesis of these microspheres..............................107

4.12 Normalized cumulative release of mitoxantrone from in situ-loaded albumin
microspheres in phosphate buffered saline pH = 7.4 at 37 C. The loading and
cross-linking agent concentration information was given in Table 4.13.............108

4.13 Normalized cumulative release of mitoxantrone from in situ-loaded albumin
microspheres in heat-treated human plasma at 37 C. The loading and
cross-linking agent concentration information was given in Table 4.13 ...........109









4.14 Normalized cumulative release of mitoxantrone from in situ-loaded albumin
microspheres in phosphate buffered saline pH = 7.4 at 37 C. The GTA
concentration was 8 % (w/w). The loading information was given in Table
4.14.................................................. ......... 111

4.15 SEM of albumin nanospheres synthesized using the suspension cross-linking
m ethod.............................................................................................. . 115

4.16 SEM of albumin nanospheres synthesized using the ethanol desolvation
m ethod.............................................................................................. . 115

4.17 The effect of capping group on nanosphere particle size at pH 4.0, 7.0 and 9.0.
The groups shown are uncapped nanospheres synthesized using 80 % ethanol
(B80), uncapped nanospheres synthesized using 75 % ethanol (B75),
nanospheres capped with glycine (Gly), glutamic acid (Glu), lysine (Lys), and
ethanol am ine (EA ). .................................................. .. .. .. .. ........ .. .. 116

4.18 The effect of surfactant on particle size at pH 4.0, 7.0 and 9.0. The groups
shown are uncapped nanospheres synthesized using 80 % ethanol and 2 %
Tween 80 (B80), 2 % Pluronic F108, 2 % Pluronic F127, 2 % Pluronic F68,
and Tetronic 908. The numbers above the bars are the particle diameter
averages (n = 9 m easurem ents). ................................................. .......... ..... 118

4.19 Effect of capping amino acid on C-potential of albumin nanospheres. The
groups shown are uncapped nanospheres synthesized using 80 % ethanol
(B80), uncapped nanospheres synthesized using 75 % ethanol (B75),
nanospheres capped with glycine (Gly), glutamic acid (Glu), lysine (Lys), and
ethanol am ine (E A ). ...................... .. .... ................ .................. ...........120

4.20 Effect of capping amino acid on C-potential of albumin nanospheres. The
groups shown are uncapped nanospheres synthesized using nanospheres
synthesized using 80 % ethanol and 2 % Tween 80 (B80), 2 % Pluronic F108,
2 % Pluronic F127, 2 % Pluronic F68, and Tetronic 908. ..................................121

4.21 Magnetite loaded albumin particles. Magnification 50,000 X. The circled
area shows the magnetite particles in an albumin nanosphere...........................123

4.22 Unprocessed magnetite particles. Magnification 50,000 X. ............................124

5.1 The effect of dose and route of administration on the average normalized
tumor weight. The tumor weights were normalized to the tumor weight at the
start of treatment. The treatment groups are intravenous injection at the
4 mg/kg dose (IV4, n= 24), intratumoral injection at the 4 mg/kg dose (IT4,
n = 24), intravenous injection at the 8 mg/kg dose (IV8, n= 24), intratumoral
injection at the 4 mg/kg dose (IT8, n= 28), and untreated control (Con,
n = 19). ........................................................................... 130









5.2 The effect of dose and route of administration on animal weight. The animal
weights were adjusted to the tumor weights then normalized to the animal
weight at start of treatment. The treatment groups are intravenous injection at
the 4 mg/kg dose (IV4, n = 24), intratumoral injection at the 4 mg/kg dose
(IT4, n= 24), intravenous injection at the 8 mg/kg dose (IV8, n= 24),
intratumoral injection at the 4 mg/kg dose (IT8, n = 28), and untreated control
(C on, n= 19). ..................................................................... 133

5.3 Percent survival for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. The groups received 8 mg/kg intratumoral MXN (IT8);
intravenous MXN (IV8); intratumoral MXN and surgery on day 10 (IT8S10);
intravenous MXN and surgery on day 10 (IV8S10); no treatment (C); or no
treatment and surgery on day 5 (CS5). Animals randomized to the IT8S10
that died prior to undergoing surgical resection of their tumor were dropped
fro m th e stu dy .................................................................. ............... 13 5

5.4 Survival plots for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. Animals randomized to surgery groups that died prior to
receiving surgery were dropped from the study. The groups received 4 mg/kg
intratumoral MXN (IT4); intravenous MXN (IV4); intratumoral MXN and
surgery on day 10 (IT4S10); intravenous MXN and surgery on day 10
(IV4S10); no treatment (C); or no treatment and surgery on day 5 (CS5). .........136

5.5 Survival plots for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. (a) Animals were treated with either intratumoral 8 mg/kg
or 4 mg/kg MXN. (b) Animals were treated with intravenous either 8 mg/kg
or 4 m g/kg M XN ................. .............................. .... .... .. .......... ... 139

5.6 Survival plots for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. (a) Animals were treated with either intratumoral 8 mg/kg
or 4 mg/kg MXN then had surgery on day 10. (b) Animals were treated with
intravenous either 8 mg/kg or 4 mg/kg MXN then had surgery on day 10..........140

5.7 Weight loss plots for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. The groups received 8 mg/kg FMXN (IT8); 16 mg/kg
FMXN (IT16); 24 mg/kg FMXN (IT24); 8 mg/kg BMXN (M8); 16 mg/kg
BMXN (M16); 24 mg/kg BMXN (M24); or unloaded microspheres (MO)
injections intratumorally. In addition, a no treatment control (C) was included. 142

5.8 Effect of dose and dosage form on C3H/HeJ mice bearing the 16/C murine
mammary adenocarcinoma. The groups received 8 mg/kg FMXN (IT8);
16 mg/kg FMXN (IT16); 24 mg/kg FMXN (IT24); 8 mg/kg BMXN (M8);
16 mg/kg BMXN (M16); 24 mg/kg BMXN (M24); or unloaded microspheres
(MO) injections intratumorally. In addition, a no treatment control (C) was
included. .......................................................................... 143









5.9 Survival plots for C3H/HeJ mice bearing the 16/C murine mammary
adenocarcinoma. The groups received 8 mg/kg FMXN (IT8); 16 mg/kg
FMXN (IT16); 24 mg/kg FMXN (IT24); 8 mg/kg BMXN (M8); 16 mg/kg
BMXN (M16); 24 mg/kg BMXN (M24); or unloaded microspheres (MO)
injections intratumorally. In addition, a no treatment control (C) was included. 144

5.10 Tumor growth plots for mice bearing the 16/C mammary adenocarcinoma.
The groups received 48 mg/kg BMXN (M48); 32 mg/kg BMXN (M32);
24 mg/kg BMXN (M24); Tsaline (ITO+); or unloaded microspheres (MO+)
injections intratumorally. A no treatment control (C) was included.................146

5.11 The effect of dose on tumor weight-adjusted animal weight. The groups
received 48 mg/kg BMXN (M48); 32 mg/kg BMXN (M32); 24 mg/kg BMXN
(M24); Tsaline (ITO+); or unloaded microspheres (MO+) injections
intratumorally. A no treatment control (C) was included. .................................147

5.12 Survival plots for mice bearing the 16/C mammary adenocarcinoma. The
groups received 48 mg/kg BMXN (M48); 32 mg/kg BMXN (M32); 24 mg/kg
BMXN (M24); Tsaline (ITO+); or unloaded microspheres (MO+) injections
intratumorally. A no treatment control (C) was included. .................................148















KEY TO SYMBOLS AND ABBREVIATIONS


ACN Acetonitrile.

AM Ametantrone.

BMXN Bovine serum albumin microsphere-encapsulated mitoxantrone.

BPG Name of microspheres prepared from blends of bovine serum albumin and
poly(l-glutamic acid).

BPL Name of microspheres prepared from blends of bovine serum albumin and
poly(l-lysine).

BSA Bovine serum albumin.

BSAx Bovine serum albumin microspheres in situ-loaded with mitoxantrone. The
"x" denotes the cross-linking agent concentration used in the synthesis.

BSAxPL Bovine serum albumin microspheres post-loaded with mitoxantrone. The "x"

denotes the cross-linking agent concentration used in the synthesis.


BSAFx Bovine serum albumin microspheres synthesized in the optimization study.
The "x" denotes the run number.

C Denotes animals in control group.

CSy Denotes animals in surgical control group. The "y" represents the day, after
inclusion in the study, the surgery was performed.

CAB Cellulose acetate butyrate.

CMC Sodium carboxymethyl cellulose.

D/C ratio Dispersed to continuous phase volume ratio.

DCC N,N'-Dicyclohexycarbodiimide










DCE 1,2-Dichloroethane.

DMPL Detoxified monophosphoryl lipid A.

DNA Deoxyribonucleac acid.

DNase Name of enzyme cocktail used to dissociate the 16/C adenocarcinoma tumor.

EA Ethanol amine.

EDC 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride.

EDTA Ethylenediamine tetraacetic acid disodium salt: dihydrate

ER Estrogen receptor protein.

FMXN Free drug mitoxantrone.

GTA Glutaraldehyde.

GTAD Glutaraldehyde solution in 1,2-cichloroethane.

GTN Gentamycin sulfate.

Glu Glutamic acid.

Gly Glycine.

HA Hyaluronic acid

HPLC High performance liquid chromatography.

HSA Human serum albumin.

I.A. or i.a. Intraarterial.

IACUC Institutional animal care and use committee.

I.P. or i.p. Intraperitoneal administration.

I.T. or i.t. Intratumoral administration.

ITx Intratumorally-treated animal. The "x" represents the dose in units of mg/kg.


xviii









ITxSy Inratumorally-treated animal. The "x" represents the dose in units of mg/kg.
The "S" denotes that the animal underwent surgical resection of its tumor on
day "y".

I.V. or i.v. Intravenous administration.

IVx Intravenously-treated animal. The "x" represents the dose injected in units of
mg/kg.

IVxSy Intravenously-treated animal. The "x" represents the dose in units of mg/kg.
The "S" denotes that the animal underwent surgical resection of its tumor on
day "y".

KDa Kilo daltons.

KPBS Potassium-containing phosphate buffered saline.

Lys Lysine.

Mx Animal treated with an intratumoral injection of microspheres. The "x"
represents the dose injected in units of mg/kg.

MAC Murine mammary adenocrcinoma.

MCT Multiple comparison test.

mOsm Milliosmoles.

MOT Mouse ovarian tetratocarcinoma.

MPC Methacryloyloxyethelene phosphoryl choline

MXN Mitoxantrone hydrochloride or mitoxantrone.

MXNx Bovine serum albumin microspheres in situ-loaded with mitoxantrone
synthesized for the optimization study. The "x" denotes the condition
number.

NADPH Reduced form of nicotinamide adenine dinucleotide phosphate.

NAWT Normalized tumor weight-adjusted animal weight.

NHS N-hydroxysuccinimide.

PBS Phosphate buffered saline.









PDI Polydispersity index.

PGA Poly(l-glutamic acid) sodium salt.

PLy Poly(l-lysine) hydrobromide.

RES Reticuloendothelial system.

SEM Scanning electron microscopy.

SPE Solid phase extraction.

SSA 5-sulfosalicylic acid.

tl/20y Initial drug distribution half-life.

tl/24 Secondary drug redistribution half-life.

t1/2Y Drug elimination half-life.

TCA Trichloroacetic acid.

TEM Transmission electron microscopy.

TFA Trifluoroacetic acid.

TNM Tumor-nodes-metastases system.

Tsaline Isotonic saline containing 0.5 % (v/v) Tween 80.















Abstract of Dissertation Presented 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 SUSTAINED
RELEASE ALBUMIN-MITOXANTRONE MICROSPHERE FORMULATIONS FOR
NONSYSTEMIC TREATMENT OF BREAST CANCER AND OTHER HIGH
MORTALITY CANCERS

By

Ahmad Robert Hadba

August 2001



Chairman: Dr. Eugene P. Goldberg
Major Department: Materials Science and Engineering

Methods for preparing mitoxantrone (MXN)-loaded albumin microspheres for the

treatment of breast cancer were developed. The effect of processing conditions on the

particle size of unloaded and MXN-loaded microspheres was evaluated using

multivariate analyses. The data suggested that the particle size of unloaded microspheres

increased as protein concentration increased or the steric stabilizer concentration

decreased. In addition, synergy between these two variables was observed. In

situ-loading of MXN achieved loading efficiencies in excess of 80 %. Comparable

efficiencies were achieved with post-synthesis loading when the microsphere were

prepared from albumin-poly(glutamic acid) blends. In vitro release of MXN in

phosphate buffered saline under infinite sink conditions showed that the total amount of

drug released increased as the glutaraldehyde concentration decreased. This trend was









reversed when the microspheres were incubated in plasma. Nanoparticles were also

prepared using ethanol desolvation. These particles were dispersible in saline and easily

modified with amino acids. In addition, particle size could be varied by use of different

non-ionic surfactants in the preparation.

The effect of intratumoral (IT) versus intravenous (IV) drug administration on

tumor response and systemic toxicity was investigated in vivo using the 16/C murine

mammary adenocarcinoma tumor model. The data suggested that IT-treated animals had

significantly smaller tumors and lower weight loss when compared to IV-treated animals.

Furthermore, the addition of surgery to the chemotherapy further improved the survival

of the animals. Pilot studies using MXN-albumin microspheres suggested that

microspheres could be safely administered IT in doses up to 48 mg/kg. However, there

was no evidence that this higher dose resulted in improved long term survival when

compared to the 32 mg/kg dose. The maximum tolerated dose of MXN given IT was

approximately 12 mg/kg. The animal studies suggested that IT chemotherapy could

result in significant reduction in tumor burden and systemic toxicity when compared to

conventional IV chemotherapy. Microsphere-bound drug could be safely administered

IT at very high doses. In addition, microspheres exerted an antitumoral effect over an

extended period of time; however, complete tumor regression was achieved only by

complete tumor perfusion.














CHAPTER 1
INTRODUCTION

The incidence (rate of diagnosis) of breast cancer in the United States is more

than 180,000 cases per year. This trend has remained relatively constant over the past

decade. Over the same period of time, breast cancer-related death (mortality) has

declined a statistically significant, yet modest 2.2 % per year. 1,2 The decline in mortality

has been attributed to improvements in treatment methods and the evolution of breast

cancer paradigms.3 It has become clear that chemotherapy in the adjuvant setting is

paramount for the success of a breast cancer treatment modality.

The accepted primary treatment of breast cancer is mastectomy or breast

conservation surgery, if indicated, followed by radiation and systemic chemotherapy.

Systemic chemotherapy is designed to eradicate metastases that may have spread from

the primary lesion. A growing body of evidence suggests that preoperative systemic

chemotherapy can not only down stage the disease allowing breast conserving procedures

to be performed, but can also reduce the incidence of tumor-positive axillary nodes.4-11

However, systemic chemotherapy remains associated with severe dose-limiting

side-effects, such as bone marrow suppression, stomatitis, nausea, and vomiting. 12 In

addition, alkylating agent and topoisomerase II inhibitor therapy-related leukemia has

been increasing in frequency over the past 20 years. 13 The toxicity of systemic

chemotherapy has encouraged oncologists to look for less toxic methods of treating

breast cancer.









Local delivery of chemotherapeutic drugs has long been recognized as one

potential method of delivering high drug doses at the target site(s) while minimizing

systemic exposure. Of these, intravesicular irrigation of bladder cancer with bacillus

Calmette Guerin (BCG),14 percutaneous ethanol injection or intraarterial (IA)

chemoembolization for the treatment heptaocellular carcinoma,15 and intratumoral (IT)

or intracavitary chemotherapy for the treatment of gliomasl2,16 have gained clinical

acceptance. Of these treatment modalities, the use of IT chemotherapy for the treatment

of breast cancer is of particular interest because of its potential ability to localize high

drug doses at the tumor site with reduced systemic exposure. In addition, drug leakage

from the tumor bed will more than likely be through the draining lymph nodes, the

primary site by which metastases spread. Another advantage of IT injections is that it is

easily adaptable to deliver particulate drug compositions directly to the tumor site.

The use of albumin microparticulate drug delivery systems for the treatment of

various diseases including cancer has been the topic of research since Wagner et al. first

reported the use of albumin microspheres for the study of the phagocytic capacity of the

reticuloendothelial systeml7 (RES) and the peripheral circulation in humans. 18 Albumin

microspheres were suggested for use in drug delivery to the RES because of their

chemical and physical stability and their rapid clearance by phagocytic cells.19 The

synthesis and use of albumin microspheres to improve the efficacy of therapeutic drugs

have been the topics of many reviews.20-23

The objective of this work was to develop mitoxantrone-loaded albumin

microsphere compositions for testing in a murine mammary adenocarcinoma model.

Mitoxantrone is a synthetic anthracenedione with marked antineoplastic activity. It has









lesser cardiac toxicity than anthracyclines such as doxorubicin, but has a narrower

antitumor activity spectrum.24 Prior to achieving this goal, several issues needed to be

addressed. First, understanding how processing variables control particle size and

cross-link density was necessary since they control microsphere properties and

potentially drug release characteristics. Second, studies to maximize drug loading and

loading efficiency were needed to develop an efficient drug delivery system. Previous

research in this laboratory utilized post-synthesis loading of microspheres. Because pure

albumin or casein microspheres generally had poor post-loading capacities, blends of the

proteins and poly(glutamic acid) (anionic polypeptide) were used to improve loading of

basic drugs. Loading of mitoxantrone (MXN) during microsphere preparation (in situ

loading) did not generally produce usable particles.

In this research, synthetic parameters that control albumin microsphere size were

studied. In addition, methods to improve the MXN loading capacity, both post-synthesis

and in situ loading methods, of prepared albumin microsphere compositions were

investigated. In vitro drug release properties of the prepared microsphere-drug

compositions were examined to compare the rank order of release from the different

compositions. Finally, selected microsphere-MXN preparations were tested in vivo in the

16/C murine mammary adenocarcinoma (MAC) tumor model. This tumor line was

selected because it was a highly metastatic spontaneous mammary tumor in the C3H/He

female mouse. It was responsive to MXN and had similar responses to chemotherapy

drugs as human mammary carcinomas.25,26 The animal studies included a baseline

investigation of MXN dose and route of administration on both mouse and tumor, i.e.

toxicity and efficacy of free MXN. This was followed by pilot studies investigating the









toxicity and therapeutic value of MXN-albumin microspheres in the 16/C MAC tumor

model.

Intravenous (IV) injectable nanoparticulates are desirable for treatment of various

illnesses such as viral or bacterial infections. One of the challenges of designing IV

injectable particulate formulations has been reticuloendothelial system (RES) evasion.

Particles less than 200 nm in diameter and/or surface chemistry modified have been

shown to evade the RES.27-29 Experience with liposomes suggests that a

phosphatylcholine head on the lipid or a poly(ethylene glycol) (PEG) coating may help

evade the RES.30-36 Thus, preparation of nanoparticles less than 200 nm in diameter

and/or nanoparticles that are surface modified with PEG, lecithin, or other

phosphorylcholine phospholipids may result in long-term circulation of drug-loaded

nanospheres. This work investigated methods for preparing such albumin nanoparticles.

In addition, preliminary investigations of the particle size and physiochemical

characteristics were conducted. This was done to lay the groundwork for future

investigations.

This research has resulted thus far in the presentation of three papers and the

co-authorship of a comprehensive review of intratumoral chemo-immunotherapy which

is currently in press in the Journal ofPharmacy and Pharmacology.37 In addition, two

manuscripts are currently in preparation and several others will result from these studies.














CHAPTER 2
BACKGROUND


2.1 Introduction

Breast cancer is the second leading cause of cancer-related death among women.

Advances in detection techniques have led to a substantial increase in early detection

rates; however, these innovations have translated into modest improvement in outcomes

or long-term survival. Between the years 1990 and 1997, the incidence and mortality

were 109.7 and 25.6 cases per 100,000 respectively. This represented an annual change

of 0.4 % in incidence and -2.2 % in mortality.2 Conventional treatment for breast cancer

has been surgical removal of the breast and axillary lymph nodes followed by radiation

and/or chemotherapy. In recent years, a growing body of evidence suggests that

preoperative chemotherapy is as effective as postoperative chemotherapy in terms of

outcome and overall survival. Since this systemic therapy has been shown to reduce the

tumor burden and down stage the disease, it can allow more breast-conserving surgery to

be performed, thereby improving the cosmetic outcome of the procedure without

compromising the efficacy of the treatment.4,7,9,38,39

Systemic chemotherapy is associated with serious side effects and dose-limiting

toxicities, such as stomatitis, leukopenia, bone marrow suppression, nausea, hair loss, etc.

Loco-regional administration techniques have been attempted to overcome these

limitations with varying degrees of success. Intratumoral injection of chemotherapeutic

agents is of particular interest because of the ability to localize drug toxicity by direct









injection into tumor lesions. This method delivers drug to the tumor site at doses so high

that they could not be tolerated systemically. More importantly from the view of this

work, this method can be used to deliver microdispersed drug delivery systems to prolong

the drug action at the target site(s). In this research, methods to synthesize and load

albumin microspheres with mitoxantrone (MXN) have been investigated. Mitoxantrone

is an anthracenedione that is active against a broad spectrum of human tumors. It is

structurally related to doxorubicin, a drug commonly used in breast cancer treatment

regimens. It has similar activity but lower cardiotoxicity.24

In this chapter, a brief discussion of the evolution of breast cancer treatment

leading to the current recommendations and possible future directions will be presented.

This will be followed by a discussion of the pharmacological and pharmacokinetic

properties of mitoxantrone, the drug of choice for this research. The properties of

albumin and albumin microsphere synthesis methods will be presented. Finally, a

description of the animal model and biological properties of the murine mammary

adenocarcinoma selected for drug-loaded microsphere formulations will be presented.

Although this study has been focused on mammary cancer treatment, it is pertinent to

improved local therapies for other high mortality cancers such as lung, colorectal, bone,

and prostate cancers.


2.2 Breast Cancer

The actual cause of breast cancer is unknown. However, epidemiological studies

have identified several risk factors based on personal behavior and environmental

exposure as well as biological and genetic characteristics.40-42 These risk factors and

their associated relative risk are tabulated in Table A. 1 in the appendix. The risk factors









suggest that breast cancer is demographic and age related, since the highest relative risk

factors are age and place of birth. Caution should be used when interpreting relative risk

data. The data are useful for quantifying an exposed group's increased risk of developing

breast cancer over a similar unexposed group. They do not provide information about

absolute elevation in risk over the general population.

Breast cancer progression is characterized by its prolonged duration and

heterogeneity within and between patients. In general, it is one of the slower growing

human tumors. The labeling index, a measure of the percentage of cells undergoing

mitosis at any given time,43 for adenocarcinomas is estimated at 2 % compared to 29 %

for lymphomas. The doubling time is 83 days for breast cancer and 27 days for

embryonal tumors while the growth fraction is 6 % compared to 90 % for embryonal

tumors and lymphomas.43 That is not to say that breast cancer is not aggressive. Some

women, particularly those under the age of 40, develop an aggressive form of the disease

and have poor prognosis.44 In others, particularly postmenopausal women, the disease is

indolent and survival rates are measured in decades even without treatment.

2.2.1 Breast Cancer Treatment

Approximately 95 % of diagnosed breast cancers are adenocarcinomas.42 A

majority of these (73-88 %) are infiltrating ductal carcinomas. The incidence of

infiltrating ductal carcinomas decreases with increasing patient age.45 Treatment

regiments for women diagnosed with breast cancer depend on several factors such as

patient age, tumor histological type, stage of the disease, and estrogen receptor status. A

brief description of the most common breast cancer types43,46,47 and their staging48 is

presented here.









2.2.1.1 Common breast cancer types

Carcinoma in-situ. These carcinomas refer to tumors arising from the ductal or

lobular epithelium and remain confined to those regions. These tumors have been sub-

classified based on their origin as either ductal or lobular. Ductal carcinoma in situ

(DCIS), or intraductal carcinoma, is a noninvasive neoplasm of ductal origin. It is

usually confined to one breast. Lobular carcinoma in situ (LCIS), or lobular neoplasia,

is a noninvasive neoplasm of lobular origin. It is usually diffuse, multicentric, and

bilateral. It does not usually become invasive, but women diagnosed with LCIS are at a

higher risk of developing invasive breast cancer.

Infiltrating or invasive breast cancer. Most of these carcinomas are ductal.

Infiltrating ductal carcinomas are the most common of all breast cancers. They are of

ductal origin and are characterized by stony hardness to palpation. They commonly

metastasize to axillary lymph nodes and have poor prognosis. Infiltrating lobular

carcinomas are areas of poorly defined thickening in the breast. They have a tendency to

grow around ducts and lobules. They are usually multicentric and bilateral with lymph

node involvement. The sites of metastasis are different from those of infiltrating ductal

carcinoma. Medullary carcinomas are relatively well-differentiated tumors, which may

attain large sizes, but demonstrate poor infiltration properties. They are distinguished by

poorly differentiated nuclei with lymphocyte and plasma cell infiltration. Tubular

carcinomas are well-differentiated carcinomas in which tubule formation is evident.

Axillary metastases are uncommon and the prognoses are usually better than either

infiltrating ductal carcinomas or infiltrating lobular carcinomas. Inflammatory breast

cancer is a rare type of locally advanced breast cancer.









2.2.1.2 Breast cancer staging

Staging refers to the grouping of cancers according to the extent of disease. It is

used to help choose a treatment regimen for the patient, estimate her prognosis, and

compare the results of different treatment regimens. Particular attention is paid to the

size of the tumor, the status of regional lymph nodes, and the presence of distant

metastases. The most widely used clinical staging system is based on the tumor-nodes-

metastases (TNM) system. Table A. 1 in the appendix summarizes the classification of

breast cancer whereas Table A.2 lists the stage grouping.

2.2.1.3 Breast cancer treatment

Breast cancer treatment depends on the extent of the disease. If the disease is

localized, the primary treatment is mastectomy, or breast conserving surgery

(lumpectomy) and radiation. This may be followed by systemic adjuvant (postoperative)

chemotherapy to attempt to eradicate micrometastases. Until recently, no standard

guidelines for the use of adjuvant chemotherapy regimens had been established. On

November 3rd, 2000, the National Institutes of Health Consensus Development

Conference on Adjuvant Therapy for Breast Cancer released its consensus statement.3 A

report conclusion is that refinements in adjuvant therapy have contributed to the recent

decrease in invasive breast cancer mortality rates (age adjusted mortality rates for all

female breast cancer patients decreased from 27.4 per 100,000 in 1990 to 24.3 per

100,000 in 1996).1 The use of adjuvant polychemotherapy, chemotherapy with multiple

types of drugs, improves the overall survival of women with localized breast cancer

regardless of their menopausal, nodal, and estrogen receptor status. In addition, regimens

containing anthracyclines result in a small but significant improvement in overall









survival. Hence, adjuvant polychemotherapy is recommended for the majority of

localized breast cancer patients. Hormonal therapy, such as with Tamoxifen, is

recommended for five years post operatively for patients whose lesions are estrogen

receptor (ER) positive. However, hormonal therapy is not recommended for patients

with ER negative lesions. In addition, localized postoperative radiotherapy for patients at

high risk of local recurrences is recommended. Similar recommendations have been

adapted by the International Consensus Panel on the Treatment of Primary Breast

Cancer.49 The biggest difference in the international panel's stance is the stronger

statement in favor of ovarian ablation for premenopausal women. Unlike the 2000

Consensus Development Conference Panel, the 1998 International Panel praised the

benefits of neoadjuvant preoperativee) chemotherapy. However, it stopped short of

recommending the use of such therapy leaving the decision for the doctor and patient.

Both consensus panels focused on localized disease. Little consideration of late stage or

inoperable disease (Stage IIIB and Stage IV) was evident in either report. The National

Cancer Institute provides treatment recommendations for such patients.50 Neoadjuvant

chemotherapy is recommended in an effort to down stage the disease. Enrollment in

clinical trials is also highly recommended. It should be mentioned that in 1999 the

National Comprehensive Cancer Network (NCCN) published its latest guidelines for the

treatment of breast cancer.51 These guidelines are intended to inform the patient of

treatment options and are not necessarily intended to be recommendations for clinicians.

2.2.2 Evolution of Conventional Breast Cancer Treatment

Historically, radical (Halsted) mastectomy was the standard treatment for breast

cancer. Although this procedure was successful in eliminating the primary tumor,









patients continued to die from distant metastases.52 Systemic adjuvant therapy was

introduced to improve the outcome of surgically curable breast cancer patients in the

1950s. The intent was to eliminate tumor cells that were systemically dispersed during

surgery. A 10-year follow-up of patients who received short-term adjuvant

chemotherapy failed to demonstrate the efficacy of adjuvant chemotherapy. 53,54 In light

of these results, the rationale for using chemotherapy shifted from treatment of

circulating tumor cells because of the surgery to eliminating micrometastases present at

the time of surgery. The efficacy of prolonged adjuvant chemotherapy was first

demonstrated in a randomized clinical trial. 55,56 The authors suggested that the effect of

adjuvant chemotherapy was systemic. Thus, less radical procedures such as full

mastectomy or lateral mastectomy may result in comparable disease control to radical

mastectomy but have better cosmetic value.55 In 1985, the National Institutes of Health

Consensus Development Conference concluded that adjuvant chemotherapy should be

the standard of care for node positive premenopausal women. 57 The 1985 panel did not

recommend adjuvant therapy for node negative women or postmenopausal women

regardless of nodal status. However, it recommended hormonal therapy for ER positive

postmenopausal women. In 1990, the National Institutes of Health Consensus

Development Conference concluded breast conservation treatment of early stage disease

(Stage I and Stage II) coupled with local radiotherapy was preferred over total

mastectomy because it resulted in equivalent survival while preserving the breast.58 The

1990 panel did not find the evidence at the time compelling enough to recommend the

use of adjuvant therapy for the treatment of node negative early stage breast cancer. In

2000, the National Institutes of Health Consensus Development Conference concluded









that adjuvant therapy was beneficial to the majority of women regardless of nodal,

menopausal, and receptor status,3 thereby apparently settling the argument of who should

receive adjuvant chemotherapy.

Interest in preoperative, or neoadjuvant, chemotherapy emerged from the belief

that reducing the size of an operable primary tumor would increase indications for breast

conserving procedures such as lumpectomies.53 Moreover, preoperative chemotherapy

was demonstrated to not only reduce the size of most breast tumors and thereby decrease

the need for mastectomy, but also to reduce the incidence of tumor-positive axillary

lymph nodes. Two clinical trials investigating the effect of preoperative chemotherapy

on operable breast cancer showed that neoadjuvant chemotherapy allows breast

conserving surgery with relapse-free survival comparable to standard therapy.7

However, these studies were not large-scale (greater than 1000 patients), randomized

clinical trials. In 1998, the National Surgical Adjuvant Breast and Bowel Project

(NSABP) B-18 released the results of a large-scale, randomized clinical trial comparing

preoperative chemotherapy to postoperative chemotherapy in women with primary

operable breast cancer. The results suggested that

preoperative chemotherapy is as effective as postoperative chemotherapy,
permits more lumpectomies, is appropriate for the treatment of certain
patients with stages I and II disease, and can be used to study breast cancer
biology. Tumor response to preoperative chemotherapy correlates with
outcome and could be a surrogate for evaluating the effect of
chemotherapy on micrometastases; however, knowledge of such a
response provided little prognostic information beyond that which resulted
from postoperative therapy.9 (p. 2672)

For a more comprehensive perspective on the evolution of breast cancer

management paradigms, the interested reader is directed to the report by Dr. Fisher.59









In summary, the current accepted treatment for breast cancer is surgical resection

followed by local radiation and systemic chemotherapy with or without endocrine

therapy depending on the ER status of the tumor. Preoperative chemotherapy has been

used to treat locally advanced or inoperable tumors, but not localized breast cancer. The

results of NSABP B-18 suggest that preoperative chemotherapy is as effective as

postoperative chemotherapy in the treatment of localized operable breast cancer. The

benefit of preoperative chemotherapy is that it allows more breast conserving procedures

to be performed, and thereby has better cosmetic and quality of life outcomes. The use of

preoperative chemotherapy has not been generally adopted as a standard therapy because

of the lack of data from other randomized, large-scale clinical trials with long term follow

up to verify the results of NSABP-B18. It is expected that these trials will be

forthcoming, and the question of the efficacy of preoperative chemotherapy will become

clearer in the near future.

2.2.3 Intratumoral Chemotherapy

It is important to note that the intratumoral injection of chemotherapeutic drugs

has been a route of drug administration largely ignored by the medical community

despite mounting evidence for superior efficacy and lower systemic toxicity in

comparison to the conventional intravenous drug delivery. The following is a brief

discussion of the benefits of intratumoral delivery of chemotherapeutic agents. The

interested reader is referred to several more comprehensive discussions on the topic in the

cited reports.12,16,37,60

All chemotherapeutic agents have significant dose-limiting toxicities, such as

bone marrow suppression, stomatitis, nausea, and vomiting.12 In addition, alkylating









agent and topoisomerase II inhibitor therapy-related leukemia has been increasing in

frequency over the past 20 years. 13 Localized administration of chemotherapeutic agents

has been recognized as a means to improve the efficacy of the drug while reducing its

systemic toxicity.61 Localization of the effects of various anti-tumoral agents has been

demonstrated by intraprostatic injections in dogs62 and rats.63 The clearance rates of

radio-labeled Bleomycin delivered intravenously (IV), subcutaneously (SC), and

intratumorally (IT) in tumor-free and tumor-bearing mice have been shown to be

comparable; however, the IV group showed higher organ distribution and lower activity

ipsilateral to the injection sites than the SC and IT groups.64 Moreover, in tumor-bearing

mice, the activity in the draining lymph nodes of the injection site was as high as that

seen in the draining lymph nodes following IV injection, but it was significantly lower on

the contralateral side lymph nodes. More recently, the effect of route of administration of

chemotherapeutic agents,65 immunoconjugates,66,67 or essential fatty acids68 on the

growth of various human carcinoma xenografts has been studied in nude mouse or rat

models. These studies further demonstrate that IT-delivered agents have higher

localization in target tissue, lower systemic exposure, longer tumor growth inhibitory

effects, and produce longer overall survival times than either IP (intraperitoneally) or

IV-delivered agents.

Another interesting aspect of IT injections is the possibility of systemic

immunological effects.61 Intratumoral injections of killed Corynebacterium parvum into

a mammary adenocarcinoma in mice resulted in rapid tumor and micrometastases

suppression. Moreover, the mice demonstrated immunity against the tumor cell line even

if 104 times the lethal dose of cells was injected SC.69 Immunization against the tumor









cell line was not demonstrated with intravenous injections of killed C. parvum.70 It may

be argued that the tumor suppression was not due to the cytotoxic effects of the killed

bacteria, but due to its antigenic effects. However, tumor suppression and anti-tumor

immunity has also been demonstrated with IT injections of cytostatic drugs in a guinea

pig hepatoma model.71 The importance of locoregional delivery has increased because

of the development of immunotherapies such as adenovirus transduction and retrovirus

transfection. Intratumoral injections of genetically modified adenoviruses induced

regression of tumors in a murine mammary adenocarcinoma model.72,73 This was also

observed in immune deficient mice.74

Intratumoral injections are not commonly used in the clinical setting. However,

intratumoral chemotherapy has been used to deliver chemotherapeutic drugs in human

clinical situations to improve the survival of patients with malignant gliomas.75 The

evidence suggests that the drug concentrations at the lesion site are higher than those

observed by traditional systemic delivery, the systemic toxicity is reduced, and the

survival rates are improved. 12 Intratumoral injections have also been used, with varying

degrees of success, for the treatment of skin recurrences of breast cancer,76 gastric

cancer,77 pancreatic cancer,78 squamous cell carcinoma,79,80 and laryngeal

papollomatois.81









2.3 Novantrone (Mitoxantrone)

Novantrone, 1,4-dihydroxy-5,8-bis[[2-[(2-hydroxyethyl)amino]ethyl]amino]-

9,10-anthracenedione dihydrochloride, is a synthetic anthracenedione with marked

antineoplastic activity. It has lower cardiac toxicity than anthracyclines such as

doxorubicin, but has a narrower antitumor activity spectrum.24 Mitoxantrone

hydrochloride or simply mitoxantrone (MXN) is the generic name for Novantrone. The

molecular weight is 517.41 g/mol. It is commercially available for clinical use as a

sterile 2 mg/ml saline solution having a pH between 3.0 and 4.5.




NH
OH O NH OH


*2 HC1


OH 0 n
OH O
NH
\^ NH OH

Figure 2.1 Chemical structure of mitoxantrone.



2.3.1 Mechanism of Action

The mechanism of action for MXN is not fully elucidated. Initially, it was

believed that MXN with its planar electron-rich polycyclic aromatic backbone exerted its

cytotoxic effect by intercalating DNA. However, the in vitro inhibition of nucleic acid

synthesis did not correlate with the in vivo tumoricidal activity suggesting additional

mechanisms of action. 82













SOH + 02


OCH3 O OH
OOC3 0 OH CH3 (c) Regenerated
Anthracycline Quinone
(a) Anthracycline H2N+ Superoxide
Quinone
OH
+le


-1 e
O' OH O
OH

OH
H O /+ 02

OCH3 O OH
0 0 CH3


(b) Anthracycline H2N
Semi-Quinone OH

Figure 2.2 Mechanism of superoxide generation for doxorubicin. The reduction takes
place in the presence of reducing enzymes such as NADPH cytochrome P450
reductase.



Anthracycline cytotoxicity is attributed to the reduction of the quinone moiety to a

free radical producing semiquinone in the presence of enzymes such as NADPH-

cytochrome P-450 reductase (Fig. 2.2, adapted from Doroshow, 1996.24). Quinone

reduction mechanisms and cytotoxicity have been reviewed.83 These free radicals

mediate non-protein associated DNA strand breaks and cell membrane lipid peroxidation.

However, only one third of MXN-associated DNA strand breaks are single stranded and

non-protein associated.84 Mitoxantrone does not produce free radicals via this pathway.









In fact, initially, it was believed that it did not produce free radicals and superoxide.82

Mitoxantrone does produce free radicals via a peroxidative conversion to an unstable

cyclic diimino compound (Fig. 2.3, adapted from Faulds et al., 1991.84). The oxidative

activation of MXN results in extensive DNA damage and could explain the non-protein

associated strand breaks.84,85


OH 0


NH OH NH O
'-\ 'NH -- NH OH


Figure 2.3 Oxidative activation of mitoxantrone in the presence of horseradish
peroxidase and hydrogen peroxide.



Like anthracyclines, MXN stabilizes the topoisomerase II-DNA cleavable

complex leading to inhibition of DNA strand reattachment.84,85 This leads to









protein-associated single and double DNA strand breaks effectively inhibiting any further

DNA metabolism (transcription, replication, recombination, and repair) and hindering

DNA packing. The extent to which this contributes to MXN cytotoxicity is unknown. In

addition to intercalation, electrostatic binding between the amine groups on MXN side

groups and the phosphate groups on DNA result in aggregation and compaction of

DNA. 82,84-86

2.3.2 Pharmacokinetic and Toxicological Properties

Since MXN oral absorption is poor, it has been administered intravenously. The

elimination of MXN after IV administration is triexponential; that is the elimination is

described using a three-compartment pharmacokinetic model. Reviews of MXN

pharmacokinetics reveal that there is great disparity in MXN elimination patterns

between patients.24,84,85,87,88 This has been attributed to the sensitivity of the

methods of detection used in some of these reports. However, it may just as well be due

to the difference in patients' metabolism and elimination patterns. In humans, the mean

initial distribution half-life (tl/2c) is between 3 and 10 minutes. This phase is one of

rapid distribution of the drug to erythrocytes, leukocytes, and platelets. It is followed by

partitioning into tissues and organs such as the liver, spleen, thyroid, and heart. This

secondary redistribution half-life (ti/23) is between 0.3 and 3.1 hours. The elimination

half-life (ti/2Y) may take as long as 215 hours. Mitoxantrone is eliminated predominantly

by the bile. However, only 30 % of the administered drug can be accounted for in the

form of metabolized and unaltered drug in the feces (20 %) and urine (10 %) five days

after administration. This suggests that MXN can penetrate deep tissues and has potential

for subsequent slow release from these compartments. Mitoxantrone and its metabolites









have been found in the organs of patients up to 272 days after discontinuing therapy84,87

further supporting the deep tissue penetration hypothesis. Mitoxantrone elimination and

metabolism are influenced by the patient's disease state. Liver dysfunction can prolong

MXN's terminal half-life, and a depot effect may occur in patients with abnormal third

spaces84,88 such as ascites, edema, fat, etc. Even with reduced toxicity, MXN has not

replaced doxorubicin because of lower activity against solid tumors when administered

intravenously at clinically tolerated levels.24

Recently, high-dose chemotherapy (HDCT) with hematopietic rescue for the

treatment of metastatic and high-risk primary breast cancer has received a great deal of

attention. The rationale behind HDCT is based on the observation that the doubling time

of human tumors increases as the tumor size increases.89 Thus, a relatively small

decrease in tumor size due to primary "induction" chemotherapy could be followed by

rapid regrowth if the tumor is left unchecked. Since the cytotoxic effect of chemotherapy

is exerted on rapidly proliferating cells, aggressive secondary ablativee" high-dose

chemotherapy regimens are warranted to impart complete responses.90 Mitoxantrone

emerged as a candidate for HDCT because of its demonstrated lower cardiotoxicity and

less severe side effects than other anthracyclines.91 This, coupled with a steep dose-

response curve, has prompted consideration of MXN as part of HDCT regimens. The

dose-limiting myelosuppression observed in conventional chemotherapy92 may be

overcome with peripheral blood progenitor or hematopietic cell support. Under stage I

and stage II clinical trial settings, MXN has been administered in concentrations up to six

times the standard clinical concentration of 14 mg/m2 with peripheral blood progenitor or

hematopietic cell support without any significant increase in hematological toxicity.93-96









Some oncologists have taken this approach a step further and forgone the conventional

primary "induction" chemotherapy and proceeded with multiple cycles of ablativee"

HDCT with hematopietic cell rescue. Significant improvement (30 %) in the outcome of

metastatic breast cancer patients receiving HDCT over conventional therapy have been

reported.90,97,98 The results of a study by Bezwoda98 became suspect after a review of

the findings revealed a disparity between the records and the reported results.99

Bezwoda's scientific misrepresentation has cast doubt on the effectiveness of the HDCT

since his report was the only one that showed significant benefit of HDCT in high risk

breast cancer. High dose chemotherapy has yet to be tested in large-scale randomized

clinical trial settings. Thus, the effects on overall survival have yet to be elucidated.

Localized delivery of MXN has been investigated as another approach to reducing

systemic burden while increasing local drug concentrations, thereby improving efficacy.

Intraperitoneal delivery of MXN has attracted much attention in the treatment of ovarian

cancer because of its high peak peritoneal cavity to plasma drug concentration ratio

(255:1). 100 Phase I and phase II clinical trials have demonstrated the feasibility of IP

administration in reducing MXN systemic burden such as leukopenia; however,

non-hematological toxicity became dose limiting at 23 mg/m2. The most common

side-effects were abdominal pain, chemical peritonitis, adhesion formation, and bowel

obstruction. 100-104 Attempts to reduce the local complications by reducing drug

solution dwell time or increasing the injection volume 100,104,105 have resulted in

reduced local complications. However, high local drug levels alone are not sufficient to

impart better patient outcomes. For the drug to be effective, it must penetrate the tumor

and diffuse from the periphery into the bulk exerting its cytotoxic effect on all of the









tumor cells. Mitoxantrone has a blue chromophore that is readily seen histologically. It

has been reported to penetrate only five to six cell layers into a tumor. 106 However, if

the drug is absorbed from the peritoneal cavity into the systemic circulation and is then

delivered back to the tumor via the vasculature, better outcomes may be expected. This

has been demonstrated in a stage I/II clinical trial whereby MXN peak plasma levels in

advanced ovarian cancer patients that showed clinical responses to IP administration

were significantly higher than those that did not. 102 Thus, IP MXN administration

seems insufficient to impart a significant clinical response without a tandem systemic

administration or alternative tumor perfusion.

Intraarterial (IA) delivery of drugs is another way to increase the local

concentrations at tumor sites. The idea is not only to increase the local drug

concentration but also to deliver it throughout the tumor by perfusing its vasculature.

Theoretically, this will circumvent the problem of limited diffusion into the tumor

encountered in IP administration. However, in the case of breast cancer, this approach

does not confine the drug to the tumor bed.107 To further localize the drug in the tumor,

angiographic placement of the catheter in the internal mammary artery and lateral

thoracic or subclavian arteries with or without embolisation of the distal internal

mammary artery have been attempted to eliminate the undesirable perfusion of the neck,

shoulder and arm. These treatments were successful in reducing systemic toxicity and

tumor size enabling surgical resection and local control. 107 However, in the absence of

well-controlled, large-scale randomized clinical trials, the effectiveness of the treatment

and its effect on the overall survival of patients is still unknown.









In summary, MXN has not replaced doxorubicin in traditional IV chemotherapy

regimens because of its lower effectiveness against solid tumors such as breast cancer or

ovarian cancer at systemically-tolerated doses. However, it has been utilized in HDCT

regimens with hematopietic support because of its lower cardiotoxicity. The rationale

behind HDCT may be sound, but the clinical data is insufficient to draw any meaningful

conclusions about efficacy and effect on the overall survival of breast cancer patients,

especially in the light of the Bezwoda controversy.

Attempts to localize drug concentrations at target sites have been investigated.

Intraperitoneal MXN administration for treatment of ovarian cancer patients aims at

capitalizing on the pharmacokinetic advantage of IP over IV administration. However,

hematological complications are replaced by local complications such as chemical

peritonitis, peritoneal adhesions, and bowel obstruction. In addition, the pharmacokinetic

advantage does not translate into a pharmacological advantage because of poor

diffusivity of MXN into tumors. Intraarterial administration of MXN seems to address

both the drug concentration localization and the tumor perfusion issues that limited the

use of IP administration. Angiographic catheter placement and selective blood vessel

occlusion can minimize the drug leakage and systemic exposure. This requires surgical

intervention and has the potential to induce further tissue damage due to the altered

arterial blood flow. This method of administration is not actively pursued in the clinical

setting, perhaps because of its complexity. As the benefits of localized delivery become

more appreciated by clinical oncologists, MXN will become more attractive in

chemotherapeutic regimens and may replace traditional anthracyclines such doxorubicin.









2.4 Albumin

Because serum albumin is the biopolymer of choice for the synthesis of

microsphere-drug compositions studied here, discussion of its properties is appropriate.

Albumin is the most abundant soluble protein in animals. It accounts for 60 % of the

total blood protein with an average concentration of 42.0 + 3.5 g/1 in adult humans. 108

(p. 256) By virtue of its high concentration and relatively low molar mass (66.5 kDa), it

is the principal protein controlling the colloid osmotic pressure of serum. It serves many

physiological functions including the binding, transport, and distribution of endogenous

and exogenous compounds and metal ions such as calcium, copper, and zinc. These

compounds include long chain fatty acids, steroids, thyroid hormones, vitamins D and

B12, and virtually all drugs. It also binds toxic byproducts and transports them to the liver

for metabolism and excretion. 108 (p. 234)

2.4.1 Albumin Composition and Structure

Recently, complimentary DNA techniques have been used to elucidate and

confirm the peptide sequences of serum albumin of many species. 108 (p. 162) The

peptide sequence of albumin differs between species; however, the average number of

peptides per molecule is 585 residues. Human serum albumin (HSA) has 585 peptide

residues while bovine serum albumin (BSA) has 583 (Table 2.1). The linearity of the

peptide sequences constituting the backbone of the protein is interrupted by disulfide

bonds formed between adjacent half-cysteine residues, resulting in eight and a half

double loops. The double loops form three globular homologous domains, each

containing two long loops separated by a short one. Each domain is further divided into

two subdomains. The domains, even though structurally and topographically similar, are









associated with different binding affinities and thus physiological functions. Albumin is

a highly helical molecule with 3 sheets, 3 turns and extended chains giving rise to a

triangular or heart-shaped appearance. High resolution x-ray data of carefully

crystallized human serum albumin predict that 67 % of the molecule is in ac helix

confirmation while the remainder is 3 sheets, 3 turns and extended chains. 109,110




Table 2.1 Listing of the amino acid residues in human serum albumin and bovine
serum albumin.
Amino Acid Human Bovine Amino Acid Human Bovine
Aspartic acid 36 40 Cysteine/2 35 35
Asparagine 17 14 Methionine 6 4
Threonine 28 34 Isoleucine 8 14
Serine 24 28 Leucine 61 61
Glutamic acid 62 59 Tyrosine 18 20
Glutamine 20 20 Phenylalanine 31 27
Proline 24 28 Lysine 59 59
Glycine 12 16 Histidine 16 17
Alanine 62 46 Tryptophan 1 2
Valine 41 36 Arginine 24 23
Adapted from Peters, 1996.108 (p.16)



In solution, albumin appears to be elliptical with an axial ratio of 3.5:1. Early

attempts to visualize the albumin molecule considered a linear arrangement of three

spheres of unequal size. However, it was recognized that the shape depends on the pH

and ionic strength of the solution. The molecule appears to be elongated at low pHs,

globular at neutral, and oblique spheroid at mildly basic pHs.108 The pH induced

conformational transitions of albumin, and its individual homologous domains have been

studied in detail 111 to elucidate the ligand-binding properties of each of the homologous

domains. Subdomains IIA and IIIA have been identified as the locations of the main









ligand-binding sites on albumin. 109,110 Binding site I, located in subdomain IIA, binds

long chain fatty acids, bilirubin, warfarin, salicylates, and other heterocyclic anions

whose charge is centrally located. Binding site II, located in subdomain IIIA, binds drugs

such as diazepam, ibuprofen, digitoxin, clofibrate, and 3'-azido-3'-deoxythymidine

(AZT). Anthracyclinesl 12,113 and anthracenediones such as mitoxantrone84,87 have

high binding affinities to albumin. However, the binding mechanism and binding site

locations) have not been identified.

2.4.2 Albumin Microspheres

Microspheres may be defined as homogeneous particles or monolithic

microcapsules ranging in size between 0.1 [im to 1000 rim. Some authors have further

sub-classified microspheres with diameters less than 1 [im as "nanospheres".

Consideration of albumin microspheres as carriers for targeting tissues and organs was

introduced in the early 1960's. Radio-labeled albumin microspheres were used to study

the phagocytic capacity of the reticuloendothelial system17 (RES) and the peripheral

circulation in humans. 18 Albumin microspheres were suggested for use in drug delivery

to the RES because of their chemical and physical stability and their rapid clearance by

phagocytic cells.19 The synthesis and use of albumin microspheres to improve the

efficacy of therapeutic drugs has been the topic of many reviews.20-23

2.4.2.1 Microsphere synthesis

The most common method of preparing albumin microspheres is suspension

cross-linking. The process involves the suspension of a small aqueous albumin solution

(dispersed phase) in an immiscible liquid (continuous phase) by some external physical

means such as a paddle mixer, vortex mixer, ultrasonicator, or homogenizer. The









resulting suspension is then chemically stabilized by either thermal denaturation or

chemical cross-linking, and the hardened particles are collected.

Dispersing agents. The critical step in the process is the formation of a stable

suspension prior to the cross-linking step. Investigators have used several dispersing

agents to accomplish this task. Polymers such as poly(methyl methacrylate), poloxamer

188,114,115 polycarbonate, 116 cellulose acetate butyrate, 116,117 and hydroxypropyl

cellulose 18 have been used as steric hindrance (stabilization) agents to stabilize the

initial suspension. Surfactants such as Span 85119-121 have also been reported. Some

investigators have taken advantage of albumin's long chain fatty acid binding ability to

stabilize the suspension, i.e. self-stabilize the suspension, when the continuous phase is

cottonseed oil22,122-124 or corn oil. 125

Heat denaturization. Albumin undergoes intermolecular aggregation and

irreversible structural changes as the temperature is increased above 80 C. This

denaturation is dependent on the albumin solution concentration, temperature, exposure

time, pH, and salt concentration. The aggregation is due to the formation of

intermolecular S-S bonds whose availability increases from 5 % at 60 OC to 47 % at

80 oC.108 (p. 66) Typical denaturation temperatures used to stabilize albumin

microspheres range from 80 to 160 OC with an exposure time between 10 and 30 minutes.

Glutaraldehyde cross-linking. Glutaraldehyde (GTA), 1,5-pentanedial, is one of

the most common dialdehydes used to cross-link bioprostheses and proteins. The

aldehyde group reacts with a primary amine to form a Schiff base. However, this is not

the only possible product. In aqueous media, GTA exists as a mixture of free aldehyde,

mono and dihydrated monomeric glutaraldehyde, monomeric and polymeric cyclic









hemiacetals, and unsaturated polymers in equilibrium.126 Reactions of glutaraldehyde in

aqueous media are shown in Figure 2.4, adapted from Hermanson. 127. Cyclic

hemiacetals, the products of aldol condensation, can form quaternary pyridinium

cross-links leaving aldehyde groups available for further reaction. Unsaturated

glutaraldehyde polymers may undergo vinyl addition (Michael addition) also leaving

aldehyde groups available for further reaction. 127 In comparison to other aldehydes,

GTA reacts relatively quicker and is able to span various distances between amino groups

both intra- and intermolecularly. In addition, it yields chemically, biologically, and

thermally more stable cross-links. 126 The cytotoxicity of GTA is a concern.

Glutaraldehyde cytotoxicity increases with the concentration of GTA used to fix

bioprostheses. 128 Initially, it was believed that unreacted aldehyde or free aldehyde as

the result of the hydrolysis of the Schiff base was the cause of the aldehyde toxicity

observed in glutaraldehyde-treated bioprostheses. However, the depolymerization of

glutaraldehyde polymers has been identified as the more likely cause of GTA-fixed

bioprostheses cytotoxicity. 129 It should be noted that glutaraldehyde cytotoxicity is not

necessarily disadvantageous for this research since the purpose of this work is the

localized delivery of cytostatic agents to adenocarcinomas. Of more concern is the

presence of unsaturated glutaraldehyde polymers because of their adverse impact on

cross-link density control. In this work, glutaraldehyde solutions in 1,2-dichloroethane

were prepared either by extraction from aqueous glutaraldehyde solutions or by dilution

of freshly distilled glutaraldehyde. These preparations minimize the presence of

unsaturated glutaraldehyde polymers in the glutaraldehyde solutions used for

microsphere synthesis.












(a) 2 R-NH2


0 0

H H


R--N N-R + 2H20


Primary amine


Glutaraldehyde


0 0

H-'- H


Glutaraldehy de


0 0
alkaline
HO H H
OH0 H' 0


a,p-unsaturated aldehyde polymer


R
O NH O

H H

H 0O Ha 0O

Michael addition product


(c) H
(c) H 'H


Glutaraldehyde


H2N-R Primary amine


HO 0 OH


Water


Aldol Condensation Product
(Hemiacetal)


+ 2 H3N-R

Primary amine


Quanternary pyridinium cross-link


Figure 2.4 Reactions of glutaraldehyde in aqueous media. (a) Reaction of a primary
amine with glutaraldehyde to form a Schiff base. (b) Formation of ac and 3 unsaturated
glutaraldehyde polymers and their subsequent reaction with primary amines via Michael
addition. (c) Formation of hemiacetals and their subsequent reaction with primary
amines.


Schiff base


H20

Water









Multivalent metal ions. The ability of anionic polysaccharides to chelate

multivalent metal ions is the basis of many industrial and medical applications.130

Multivalent metal ions have been used to thicken anionic polysaccharide

solutions 131,132 and to prepare stable microspheres. 133 Albumin microspheres have

been successfully prepared using Fe3+ ions to chelate pendant carboxylic acid

moieties. 117 The use of iron ions to prepare microspheres is of interest because they

mediate the formation of hydroxyl radicals thereby enhancing the cytotoxic effect of

mitoxantrone.

Other cross-linking agents. Several other cross-linking agents for albumin may

be used. Carbodiimides or so called zero-length cross-linkers are commonly used

because they mediate the conjugation of two molecules (e.g. a carboxylate and an amine)

without the addition of a spacer that may have biological cross-reactivity. Of most

interest is 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) because

of its water solubility. Albumin has an abundance of primary amine and carboxylate

groups, thus the potential for the formation of intra- and intermolecular cross-links is

high.

2.4.2.2 Nanosphere synthesis

So far in this discussion, the difference in the synthesis between nanospheres and

microspheres has been ignored. As with microspheres, suspension cross-linking is a

common method to synthesize nanospheres. The biggest difference in the synthesis of

nanospheres is the higher input power needed to generate sub-micron particles. Paddle

mixers and vortex mixers do not provide enough energy to form stable nano-dispersions,









so ultrasonicators or homogenizers are usually used to provide the necessary dispersion

energy.

Alternatives to suspension cross-linking have been devised to synthesize

nanoparticles. Investigators have reported using desolvation or coacervation methods to

prepare gelatin,134-136 chitosan,137,138 and albumin nanospheres.27,135,139,140

Briefly, a desolvating agent is slowly added to a macromolecule solution, with or without

stirring, until the system begins to coacervate. Energy (from an ultrasonicator,

homogenizer, vortex or paddle mixer) is then applied to form the nano-dispersions. The

nano-dispersions are then chemically stabilized by either glutaraldehyde cross-linking or

heat denaturization. The use of dispersing agents to form stable nano-dispersions is still

necessary. The advantage of this method over suspension cross-linking is the ability to

use less toxic solvents. In the case of chitosanl37 or gelatin nanoparticles,135,136

aqueous solutions of sodium sulfate have been used to desolvate the aqueous

macromolecule solutions. In the case of albumin, either alcohol desolvationl35 or

pH-coacervation with acetone27,139,140 have been reported.

2.4.2.3 Particle size control

Early biodistribution studies demonstrated that the particle size and route of

administration affect the site of localization of injected microspheres (see Table 2.5).

Particle size determines the first-order distribution,20, and the degree of stabilization

controls the degree of swelling, drug release kinetics and biodegradation of injected

particles.141 However, factors that affect circulatory lifetimes are less clear.









Table 2.2 Passive drug targeting based on the particle size of drug carrier and its
route of administration.
Particle Size Route of Site of Localization
(jm) Administration
< 0.1 IA/IV/IP Spleen, bone marrow, tumors
0.2-2.0 IA/IV/IP Liver
3.0-12.0 IV Lungs, liver, and spleen
7.0-12.0 IV Lungs
12.0 IA Microvasculature of gut, kidneys
IA = intra-arterial IV = intravenous, IP = intraperitoneal.
Adapted from Gupta and Hung, 1989.22



The effect of particle size on the tissue distribution of injected particles

necessitates particle size control in the development of particulate drug delivery systems.

The parameters controlling the particle size at steady state are the apparatus design, the

viscosity of the two immiscible phases, and the speed of mixing. The particle size

distribution is determined by the uniformity of the shear forces throughout the suspension

mixture at steady state. In general, the more uniform the mixing process, the narrower

the size distribution of the droplets.23 The time for the formed particles to reach steady

state at each stirring speed is usually between 10 and 30 minutes. Once the steady state

particle size is reached, care must be taken to maintain individuality of the initially

formed droplets during the course of the hardening stage. This is usually accomplished

by carrying out the stabilizing reaction under constant stirring in the presence of a

dispersing agent. It has been shown that for a given power input, the particle size and

size distribution decrease as the dispersion time increases from 1 to 8 minutes. The

dispersion time effect on particle size is dependant upon the dispersing agent composition

and its concentration. 114









The critical factors controlling the particle size and distribution have been widely

reported in the literature. Longo et al. reported that the size of the synthesized

microspheres is directly related to the power input and dispersion time. The effect of the

dispersing agent concentration was to stabilize the particles during the suspension and

cross-linking stage of the synthesis. 114 Gallo et al. studied the effect of several

processing parameters on the particle size of heat-stabilized bovine serum albumin (BSA)

nanoparticles. They studied the effect of BSA concentration, emulsification time, power

of emulsification, dispersed to continuous phase volume ratio, stirring rate during

stabilization, heat of stabilization, and continuous phase composition. Their results

suggest that the factors controlling particle size are the dispersing medium and heat of

stabilization temperature. The authors conclude that the emulsification power had a

relatively large effect on the particle size distribution. 122 El-Mahdy reported that

albumin particle size and size distribution are controlled by surfactant concentration,

albumin concentration, and glutaraldehyde (GTA) concentration. 121 The dispersed to

continuous phase volume ratio had an effect only after reaching a threshold. It should be

noted that the above experiments were one-factor-at-a-time experiments that lack

information about potential interaction of two or more variables. Miller et al. studied the

effect of several processing variables on the particle size and size distribution of BSA

nanoparticles using a central composite statistical design. They investigated the effect of

five process variables, namely BSA concentration, dispersed to continuous phase volume

ratio, emulsification time, glutaraldehyde (GTA) concentration, and drug

concentration. 118 Of the investigated factors, only the concentration of BSA and the

dispersed to continuous phase volume ratio had a significant effect on the particle size









and size distribution. Miller's work did not investigate the effect of dispersing agent or

cross-linking reaction time on the particle size and size distribution. The authors did not

offer an explanation why these seemingly important variables were neglected. The effect

of four factors, albumin concentration, dispersing agent concentration, pH, and dispersion

energy, on albumin particle size was recently reported. 115 The authors reported that all

the studied factors significantly affected particle size. The authors did not investigate the

effect of dispersed to continuous phase volume ratio, dispersion time, cross-linking agent

concentration, and cross-linking reaction time on particle size.

As can be deduced from the above discussion, there is much disparity in the

literature as to which factors are critical in affecting particle size of albumin

microspheres. Few multivariate experiments investigating the effect of process variables

on particle size have been conducted. Since the particle size of injected particles is

important for either IV or IT drug delivery, tissue permeation, drug release, and

determines their eventual organ distribution, the first aim of this research was to

investigate and clarify the effect of processing parameters on particle size.


2.5 16/C Murine Mammary Adenocarcinoma

A main objective of this research has been to formulate and evaluate the efficacy

of mitoxantrone microsphere compositions in a murine mammary tumor model. Most

mammary tumor models today utilize human tumor cell lines implanted in immune

compromised (nude) mice. Even though this approach provides information about the

effect of the chemotherapy on human tumor lines, information about immunological

effects is lost. To overcome this limitation, a spontaneous and weakly immunogenic

murine mammary carcinoma that is reportedly responsive to mitoxantrone was required.









The 16/C murine mammary adenocarcinoma (MAC) is a virus-associated tumor

that arises spontaneously in C3H/He female mice. It has been shown to be highly

metastatic after serial subcutaneous implantation if lung metastases are implanted. In

addition, the 16/C MAC response to chemotherapy appears to be similar to human

tumors. Because of the high metastatic rate and drug response, the 16/C MAC model has

been investigated for surgical adjuvant chemotherapy.25 Table 2.6 summarizes its

relevant biological characteristics. These characteristics make the 16/MAC in the

C3H/HeJ mouse an appropriate model to study the efficacy mitoxantrone microsphere

formulations.




Table 2.3 Biological characteristics of the 16/C murine mammary adenocarcinoma.
Origin Spontaneous virus associated tumor, 1974
Host C3H/He
Histological description Grade III mammary adenocarcinoma
Metastasis site 75-80 % to lungs
Time to grow to 500 mg 10-15 days (from a 30 mg subcutaneous implant)
Median day of death 29-46 days
Take rate 99%
Overall doubling time 1.6-2.4 days
200 mg chunk Approximately 2 x 10" cells
Adapted from Corbett et al., 197825 and Corbett et al., 1982.26



The activity of many of the clinically relevant cancer drugs against the 16/C MAC

have been reported.25,26,142-144 The 16/C MAC has been shown to be responsive to

anthracyclines and anthracenediones.25,26,142









2.6 Goals of this Research

The objective of this work was to develop mitoxantrone-loaded albumin

microsphere compositions for testing in a murine mammary adenocarcinoma model.

Prior to achieving this objective, several other issues were necessary to address. Since

particle size is a controlling factor for microsphere properties and potentially drug release

characteristics, evaluation of synthetic parameters that control particle size and cross-link

density was a necessary first step. Another issue of importance was microsphere drug

loading. Previous research in this laboratory utilized post-synthesis loading of

microspheres. Pure albumin or casein microspheres generally had poor post-loading

capacities. As a result, blends of proteins and poly(glutamic acid) were used to improve

loading. In situ loading of the microspheres, i.e. loading of drug during synthesis, did not

generally produce usable particles.

In this research, methods to synthesize in situ-loaded microspheres were

investigated. In addition, methods to improve the loading capacity of post-loaded

albumin microspheres were investigated. In vitro drug release properties of the resulting

microsphere-drug compositions were examined to compare different compositions.

Finally, selected microsphere-drug preparations were investigated in vivo in the 16/C

MAC tumor model. The animal studies included a baseline investigation of the effect of

dose level and route of administration of MXN as a free drug on both mouse and tumor,

i.e. toxicity and efficacy of free MXN. This was be followed by pilot studies

investigating the toxicity and therapeutic value of MXN-albumin in the 16/C MAC tumor

model.

Finally, because intravenous injectable nano-particulate microspheres appear

desirable for treatment of various illnesses such as viral or bacterial infections, one of the









challenges of designing IV injectable particulate formulations has been to evade the

reticuloendothelial system (RES). One way to evade the RES is to prepare particles that

are less than 200 nm in diameter. Experience with liposomes suggests that a

phosphatylcholine head on the lipid or a poly(ethylene glycol) (PEG) coating may help

evade the RES. Thus, preparation of nanoparticles less than 200 nm in diameter and/or

nanoparticles that are surface modified with PEG, lecithin, or other phosphorylcholine

phospholipids may result in long-term circulation of drug-loaded microspheres. This

work investigated methods for preparing such albumin nanoparticles. In addition,

preliminary investigations of the particle size and physiochemical characteristics were

conducted. This was done to lay the groundwork for future investigations.














CHAPTER 3
MATERIALS AND METHODS


3.1 Materials

Mitoxantrone (MXN) was generously donated by Lederle Laboratories. The

reported drug purity was 82.75 % and the balance was not disclosed. This drug was used

without any further purification. Ametantrone (AM) was obtained from Dr. Schultz at

the Drug Synthesis and Chemistry Branch of National Cancer Institute in Bethesda, MD.

The drug was used as an internal standard without any further purification. Gentamycin

sulfate (GTN) with a reported purity of 600 |tg gentamycin base/mg solid was purchased

from Sigma Chemical Company

All proteins, polypeptides, enzymes, and amino acids were purchased from Sigma

Chemical Company unless otherwise specified. Heparin sodium salt, Grade II

(151 Kunits/mg solid) was used for mouse blood collection and subsequent plasma

separation. Bovine serum albumin (BSA) fraction V and human serum albumin

fraction V (HSA) were used to synthesize microspheres or prepare blocking solutions.

Poly(l-lysine) hydrobromide (PLy) (Mw= 11,200 g/mol and PDI = 1.35) and

poly(l-glutamic acid) sodium salt (PGA) (Mw = 39,400 g/mol and PDI = 1.40) were used

to prepare blended microspheres. Papaya latex papain (22 U/mg solid), bacterial protease

Type VIII (11 U/mg solid), and pepsin (4550 U/mg solid) were used to prepare various

enzyme digestion buffers. Bacterial protease Type XIV (4.7 U/mg solid), collagenase

Type I (233 U/mg solid), and bovine pancreas deoxyribonuclease (650 K units U/mg









solid) were used to prepare the enzyme cocktail for tumor cell suspension injections.

L-glutamic acid (Glu) Sigma Grade and glycine (Gly) sodium salt were purchased from

Sigma Chemical Company. L-lysine hydrochloride 99+% (Lys) was purchased from

Acros Organics. Ethanolamine was purchased from Eastman Kodak Company.

Ethanolamine (EA) and the above mentioned amino acids were also used to cap pendant

aldehyde groups on some of the synthesized microspheres and nanospheres.

Sodium carboxymethylcellulose, Type 7LF (CMC) (reported molecular weight

90,000 g/mol) was donated by Aqualon. This was used to prepare microspheres from

either CMC alone or BSA/CMC blends. Cellulose acetate butyrate (CAB), 17 % butyryl

content, was purchased from Acros Organics. Pluronics (F68, F108, and F127) and

Tetronic 908 were purchased from BASF Corporation. Tween 80 was donated by ICI

Americas Inc. The CAB was used as a steric hindrance (stabilization) agent in the

preparation of microspheres and nanospheres. Pluronics, Tetronic, and Tween were

also used to prepare nanospheres.

Aqueous glutaraldehyde solutions (25 (w/w) % Type II and 50 (w/w) % Type I)

were purchased from Sigma Chemical Company. Formaldehyde (40 w/v % solution)

USP was purchased from Fisher Scientific. 1-Ethyl-3-(3-dimethylaminopropyl)

carbodiimide hydrochloride (EDC) and N,N'-dicyclohexyl-carbodiimide (DCC) were

purchased from Sigma Chemical Company.

Deionized Type I and Type II water were prepared in-house using the Barnstead

NANOpureTM Ultrapure Water System. The resistivity of the deionized water used in this

work was at least 17.4 MQ-cm. This water will be referred to as ultrapure water herein.

Ethanol USP, Absolute 200 proof, was distilled by AAPER Alcohol and Chemical









Company and purchased from the University of Florida Health Center Stores. Methanol,

Optima grade, and acetonitrile, HPLC grade, were purchased from Fisher Scientific. All

other solvents, reagents, and salts were purchased from Fisher Scientific or Acros

Organics and were Certified A. C. S. grade unless otherwise specified.

3.1.1 Buffered Solutions

Phosphate buffered saline. In vitro drug release and microsphere degradation

studies were carried out using in-house prepared buffers. Isotonic phosphate buffered

saline was prepared by first preparing a 50 mM phosphate buffer of appropriate pH then

adjusting the osmolarity of the resulting solution with sodium chloride (NaC1). The

50 mM phosphate buffer solution was prepared by mixing a 50 mM sodium phosphate

monobasic solution with a 50 mM sodium phosphate dibasic solution in a proportion that

maintained the pH at the desired level. The 50 mM sodium phosphate monobasic

solution was prepared by dissolving 10.36 g sodium phosphate monobasic monohydrate

in 1.5 L of ultrapure water. The 50 mM sodium phosphate dibasic solution was prepared

by dissolving 21.3 g sodium phosphate dibasic in 3.0 L ultrapure water. An approximate

proportion that maintained the pH at 7.4 at room temperature was 2.9 parts sodium

phosphate dibasic to 1 part sodium phosphate monobasic. Sodium chloride standards

with concentrations ranging from 0 to 10 mg/mL were prepared in the phosphate buffer.

The osmolarity of the standards was measured using a Precision Instruments [t OsmetteTM

osmometer. The calibration plot linear regression information, slope and y-intercept,

were used to calculate the amount of NaCl needed to prepare a 300 mOsm solution. The

appropriate amount of NaCl was then dissolved in the above prepared phosphate buffer









solution. The resulting phosphate buffered saline solution was then filtered through a

0.20 |tm filter and/or autoclaved.

Phosphate buffered saline was also used to suspend adenocarcinoma cells or

pieces. The buffer recipe obtained from Dr. Seimann's laboratory (University of Florida

Department of Radiation Oncology) required the presence of potassium salts. This buffer

was prepared by dissolving 800 mg potassium phosphate monobasic, 4.60 g sodium

phosphate monobasic, 800 mg potassium chloride, and 32.0 g sodium chloride in 4.0 L

ultrapure water. The pH of the final solution was adjusted to 7.4 with 1.0 N hydrochloric

acid (Fisher Scientific). The buffer was then sterile filtered as needed using sterile

disposable Coming filter apparatus. The exact filter apparatus used depended on the

volume needed at the time of filtering. This buffer will be referred to as KPBS herein.

Digestion enzyme buffers. The drug payload of loaded albumin microspheres

was measured after enzymatic digestion of the microsphere matrix proteins. Two

digestion buffers were used. The first digestion cocktail buffer was prepared by

dissolving 720 mg ethylenediamine tetraacetic acid (EDTA) disodium salt: dihydrate

(Sigma Grade from Sigma Chemical Company), 80 mg L-cysteine hydrochloride hydrate

(Sigma Chemical Company), 50 mg papaya latex papain, and 50 mg bacterial protease

Type VIII in a 100 mL 0.1 M phosphate buffer solution (pH= 7.0). The phosphate buffer

solution was prepared using the same methods used to prepare the PBS with the

appropriate salt concentration modifications. This enzyme cocktail will be referred to as

the Protease Digestion Buffer herein. The second digestion buffer used was a

pepsin-containing solution reported by Luftensteiner et al.145. In brief, a 0.1 M

potassium chloride solution was prepared. The pH of the solution was adjusted to 2.0









using a 0.1 M HC1 solution. Enough pepsin was dissolved in the above buffer to prepare

a solution with 9000 U/mL activity. This enzyme buffer will be referred to as the Pepsin

Digestion Buffer herein.

Enzyme cocktail. An enzyme cocktail buffer was used to dissociate the 16/C

tumor cells prior to freezing or transfer to animals. This recipe was obtained from Dr.

Seimann's laboratory. In 100 mL sterile PBS, 25 mg collagenase Type I, 50 mg bacterial

protease Type XIV, and 40 mg bovine pancreas deoxyribonuclease were dissolved. The

enzyme cocktail solution was then sterile filtered through a 0.22 |tm corning disposable

filter unit. This enzyme cocktail will be referred to as the Dnase Enzyme Cocktail herein.

Neutral Buffered Formalin. All harvested animal tissue samples were fixed

using a 10 % formalin buffer solution. The buffer was prepared by dissolving 4.0 g

sodium phosphate monobasic and 6.5 g sodium phosphate dibasic in 900 mL ultrapure

water. The buffer volume was made up to 1000 mL using a 40 w/v % formaldehyde

solution. The pH of the buffer was adjusted to 7.0 using either 1 N HC1 or 1 M NaOH.

The resulting solution was then sterile filtered through a 0.22 |tm DuraporeTM

hydrophilicc poly(vinylidine fluoride) membrane filter; Millipore Corp.) filter.

3.1.2. Protein and Reagent Solutions

Albumin solutions. Bovine serum albumin is hygroscopic-it absorbs up to 10 %

of its weight in moisture. In addition, BSA solutions foam significantly when agitated

making their preparation using volumetric flasks cumbersome. Thus, BSA solutions

were prepared at a concentration close to the desired concentration, then the true

concentration was measured gravimetrically. The density of each prepared BSA solution

was measured. Then, samples of these solutions were dried at 130 OC on a Mettler LJ16









Moisture Analyzer. The density of the prepared BSA solution and the weight of the dried

albumin were used to calculate the concentration of the solution in weight per volume

percent. It should be noted that sufficient volume of each sample was dried so that the

mass of dry protein was at least 1 g.

Albumin and polypeptide solutions. Albumin and polypeptide solutions were

prepared to synthesize microspheres with different matrices and functionality

concentrations. Albumin and PLy blend solutions were prepared by first dissolving the

PLy in water then adding the BSA. Albumin and PGA blend solutions were prepared by

first dissolving the BSA then adding the PGA to the albumin solution. The PGA was

weighed in a dry box because it is very hygroscopic. It was necessary to add the PGA

powder to the BSA solution because adding the albumin as a powder to a PGA solution

did not result in a homogeneous solution. This was also observed previously in our

laboratory.

Plasma. Volunteer human plasma was obtained from the Shands hospital blood

bank or Civitan blood bank. Human plasma was heat treated to minimize daily changes

in protein concentration due to protein precipitation. The heat treatment consisted of

maintaining the plasma at 60 OC for 5 minutes, then centrifuging the plasma at 5000 rpm

for 10 minutes using the Beckman J2-21 centrifuge (Beckman Coulter, Inc., Fullerton,

CA). The plasma was pooled, preserved using 0.02 % sodium azide, then stored at

-20 C until used.

Glutaraldehyde solutions. Glutaraldehyde solutions (GTA) in

1,2-dichloroethane (DCE) were prepared according to the methods reported by KirkI 17

Approximately 30 mL of 1,2-dichloroethane were transferred to a weighed 50 mL









centrifuge tube. The weight was recorded. Approximately 10 mL of a 25 % aqueous

glutaraldehyde solution were added. The weight was recorded. The mixture was

vortexed for one minute on a Genie 2 vortex mixer (Scientific Industries Inc., Bohemia,

NY), then placed on a rotary tumbler (Built in-house) to mix overnight. The mixture was

recovered and centrifuged at 2500 rpm for 10 minutes using a DynacTM II bench top

centrifuge (Clay AdamsTM BD Biosciences, Franklin Lakes, NJ) to separate the organic

layer from the aqueous layer. The aqueous layer was carefully removed and discarded.

The weight of the tube was then recorded. The difference in the weight tube after

removal of the aqueous layer and the weight of tube with DCE was the weight of GTA

extracted from the aqueous solution.

Glutaraldehyde solutions in DCE were also prepared by vacuum distilling the

25 % aqueous GTA solution and recovering the glutaraldehyde distillate. The distillate

was then dissolved in DCE to a final concentration of 40 mg/mL. This is the preferred

method of preparing GTA solutions in DCE.


3.2 Methods

3.2.1 Microsphere Synthesis

3.2.1.1 General method

Microspheres were synthesized using the suspension cross-linking method. In

brief, a BSA solution was suspended in a CAB solution in DCE using a high-speed

mechanical stirrer equipped with a 2-inch, three-blade propeller. The stirring rate was

maintained at 1250 rpm for 20 minutes after which the appropriate amount of GTA in

DCE was added to cross-link the BSA. The stirring rate was reduced to 600 rpm and the

reaction was allowed to continue for two hours. At the 2-hour mark, 50 mL acetone were









added and the reaction allowed to continue for another hour. The resulting particles were

collected by centrifugation at 3000 rpm. The particles were washed by resuspension in

acetone and centrifugation. This was repeated four times, then the particles were dried

overnight in a fume hood. In a typical preparation, the particle weight per preparation

was 600 mg. The volume of acetone used per wash cycle was approximately 30 mL.

The particles were then stored at 4 OC until further use. The reaction vessels used in these

syntheses were 300 mL LabconcoTM lyophilization flasks and that the total volume of

liquid in the flasks was fixed to 50 mL. Three different mechanical stirrers were used in

these syntheses. The mechanical stirrers were LightninTM Lab Mixer Model LIU08

(General Signal; Dublin, Ireland), CaframoTM Type RZRI (Caframo; Wiarton, Ontario,

Canada), and RW20DZM (Jenke & Kunkel GMBH & Co., Germany). It should be noted

that even though the term "suspension cross-linking" is commonly used in the literature

to describe this method of microsphere preparation, it is not descriptive of the process.

This term is used to describe methods that introduce the cross-linking agent through the

aqueous phase and those that introduce it through the organic phase. The method

described here introduces the glutaraldehyde through the organic phase. This method is

expected to produce particles that have a higher cross-link density on the surface of the

microsphere than within the core 1l4,116,146

Initially, a pilot study was conducted to investigate the effect of the following

factors on particle size: CAB concentration, BSA concentration, GTA concentration,

dispersed to continuous phase volume ratio (D/C), cross-linking reaction time, and

stirring rate during cross-linking. The results of this pilot study were used to statistically

design a three-level, four-factor experiment investigating the effect of processing









conditions on particle size. The conditions for the pilot study are given in Table 3.1. The

conditions for this "full" study are given in Table 3.2.


Table 3.1 The factors and factor levels investigated in the pilot study investigating
the effect of processing conditions on particle size.
Processing Condition Low Level High Level
Cross-Linking Reaction Time (hrs) 1.5 3.0
GTA Concentration (% w/w) 6.4 12.7
BSA Concentration (% w/v) 15 30
Stirring Rate During Cross-linking (rpm) 300 600
CAB Concentration (% w/v) 0.5 2.0
D/C Ratio (% v/v) 5 10


Table 3.2 The factors and factor levels investigated in the full study of the effect of
processing conditions on particle size.
Processing Condition Low Level Middle Level High Level
BSA Concentration (% w/v) 10 20 30
D/C Ratio (% v/v) 5 10 15
CAB Concentration (% w/v) 0.5 2.0 5.0
GTA Concentration (% w/w) 5.0 10.0 15.0


A note about concentration of glutaraldehyde: Each percent glutaraldehyde is

equivalent to 0.1 mmol per gram albumin. In addition, there are 59 lysine units per

albumin molecule. On a per gram BSA basis, there are 0.901 mmol of lysine. Aldehydes

are believed to preferentially react with the lysine residues in proteins. Thus, for each

percent GTA used, the GTA to lysine molar ratio is 1:9.


3.2.1.2 In situ-loaded albumin microspheres

Microspheres were loaded with drugs such as mitoxantrone either after synthesis

(post-loaded) or during synthesis (in situ-loaded) microspheres. In situ-loaded

microspheres were synthesized using the same techniques described above except that the









starting BSA solution had the appropriate amount of drug dissolved in it. It should be

noted that all the drugs used in this study were readily soluble in water and insoluble in

DCE and acetone. Post-loading will be described in section 3.2.3.3.

3.2.1.3 Miscellaneous microsphere syntheses

Microsphere matrix blends. Microspheres prepared from blends of BSA and

polypeptides (PGA or PLy) or carboxymethylcellulose (CMC) were prepared using the

same techniques described above.

Carboxymethylcellulose microspheres. Carboxymethylcellulose microspheres

were prepared using a modification of the methods reported by Kirk 17. In brief, 2 mL

of a 3.5 % (w/v) aqueous CMC solution were suspended in 16 mL of the CAB solution

using a vortex mixer set to the highest setting. The vortexing commenced for 2 minutes

after which the cross-linking agent was added. The suspension was vortexed for another

minute then placed on roto-tumbler for at least 2 hours. The collection steps used were

the same as those described in section 3.2.1.1. The cross-linking was achieved by

microwave heating or ferric nitrate ionic cross-linking instead of glutaraldehyde

cross-linking. Ferric nitrate was substituted for GTA in this case because of the lack or

amine groups on the CMC backbone. In the case of microwave heating, the cross-linking

step was as follows. The suspension was placed in the microwave (1200 W Samsung

household microwave oven). The sample was heated on high for 30 seconds. At this

point, the sample was warm to the touch. This was followed by another 15 second

heating step. In the case of ionic cross-linking, the ferric ion was used to chelate at least

two carboxyl side groups on the CMC backbone. The Aqualon 7LF CMC has 0.7

carboxyl group and 4.6 hydroxyl groups per repeat unit 147 resulting in a repeat unit









formula weight of 218 g/mol. Assuming that each ferric ion would chelate with three

carboxyl groups, then 0.432 g of ferric nitrate nonahydrate per gram CMC are needed to

saturate all the carboxyl groups. A 0.5 M solution of Fe(N03)3.9H20 in water was

prepared and used to cross-link the CMC suspension. Enough ferric ion solution was

added to saturated all carboxylic acid groups on the CMC backbone; typically, 4 mL of

the 0.5 M ferric ion solution per g CMC was required. The reaction was allowed to take

place overnight at room temperature.

Microspheres with different cross-linking agents. Albumin microspheres were

also prepared using Fe(N03)3.9H20 as an ionic cross-linker. Briefly, 2 mL of a

20 % (w/v) BSA solution were suspended in 16 mL of a 2 % CAB solution using a vortex

mixer set to the highest setting. The vortexing commenced for 2 minutes after which the

Fe(N03)3.9H20 was added. The suspension was vortexed for another minute then placed

on roto-tumbler for at least 2 hours. Bovine serum albumin has 99 peptides (aspartic acid

and glutamic acid) with available pendant carboxyl groups. Assuming each ferric ion

would chelate three carboxyl groups, then 33 moles of ferric ion per mole of BSA are

needed to saturate all carboxyl groups. The molecular weight of BSA is 65.5 kDa. Thus,

0.204 g Fe(N03)3.9H20 are needed per gram of BSA. Approximately 1.00 mL of the

0.5 M Fe(N03)3.9H20 per gram BSA was added to saturate all the available carboxyl

groups.

Attempts to prepare both CMC and BSA microspheres using carbodiimide as a

cross-linker failed to result in stable microspheres. The microspheres were collectable,

but dissolved when suspended in aqueous media. The techniques used to prepare these

microspheres were similar to those described above, except that the cross-linking agent









was N,N'-dicyclohexyl carbodiimide (DCC) dissolved in DCE. The DCC was added in

at least 3-fold molar excess to account for any losses in the reaction of DCC with water.

Microspheres loaded with various drugs. Microspheres loaded with

gentamycin sulfate (GTN) were prepared using the methods reported in section 3.2.1.2.

These microspheres were intended as a degradable alternative to the controlled release

devices used for the treatment of osteomyelitis. They were sent to Dr. James Marotta at

Alfred University for in vitro testing. Microspheres loaded with lipopolysaccharide

(LPS) were prepared using the same techniques described in section 3.2.1.2. LPS is a

potent bacterial endotoxin used to stimulate the immune system. These microspheres

were sent to Dr. Streit at the University of Florida Brian Institute for evaluation in

microglia cell culture.

3.2.2 Nanosphere Synthesis

Nanospheres were synthesized using suspension cross-linking and desolvation

methods.

3.2.2.1 Suspension cross-linking

The following method is a modification of the techniques developed in our

laboratory and reported by McCluskyl48. In brief, 2 mL of a 30 % (w/v) BSA solution

were added to 50 mL of a 2 % (w/v) CAB solution in a 300 mL LabconcoTM

lyophilization flask. The mixture was cooled to 0 OC using an ice bath. The cooled

mixture was then dispersed using a SonicatorTM Cell Disruptor Model W-375 (Heat

Systems-Ultrasonics Inc., Plainview, NY) equipped with a micro-tip probe. The

ultrasonic energy (125 W) was applied continuously for 10 minutes after which, the

appropriate amount of GTA was added to cross-link the BSA. The ultrasonic energy was









applied for another 10 minutes. The suspension was transferred to 50 mL centrifuge

tubes, placed on the roto-tumbler, and allowed to react at room temperature for another

three hours. The centrifuge tubes were then collected and at least 20 mL of acetone were

added to each. The particles were collected at 4 OC by centrifugation at 16,000 rpm

(35,000 x g) for 10 minutes using the J-17 rotor on the Beckman centrifuge. The

particles were washed with acetone three times, then incubated in a 1 % albumin solution

(approximately 20 mL) for two hours at room temperature. This step was added to react

any residual aldehyde groups on the surface or in the bulk of the nanospheres. The

particles were washed with ultrapure water three times then with acetone three times.

The resulting pellet was dried overnight in a fume hood.

3.2.2.2 Desolvation method

Albumin nanospheres were also prepared using the desolvation or coacervation

method. Reports of preparation of stable gelatin21,134,135,149 and chitosanl37

nanoparticles without the use of halogenated organic solvents was the motivation to

prepare albumin nanoparticles using these techniques. The cited reports utilized

concentrated salt solutions to "salt out" the protein from the aqueous solution. This

approach was not successful in desolvating albumin from solution. The commercial

purification of serum albumin is by fractionation using cold ethanol. This reversible

denaturization of albumin was applied to prepare albumin nanospheres. The process

involved the addition 20 mL of an 80 % (v/v) ethanol solution to 5 mL of a 5 % (w/v)

albumin solution containing 2 % (w/v) Tween 80 in a 50 mL conical centrifuge tube.

The contents of the centrifuge tube were mixed for 1 minute using the vortex mixer. The

solution turned milky white. The suspension was cooled in an ice bath and sonicated at









75 W for 10 minutes using the Heat Systems sonicator. To cross-link the albumin,

100 p.L of a 25 % (w/w) aqueous glutaraldehyde solution were added to the suspension

then the ultrasonic energy was applied for another 10 minutes. The suspension was

placed on the roto-tumbler, and the cross-linking reaction was allowed to continue at

room temperature for another three hours. The particles were collected at 4 OC by

centrifugation at 16,000 rpm for 10 minutes using the J-17 rotor on the Beckman

centrifuge. The particles were washed twice with 100 % ethanol then three times with

acetone. The resulting pellet was dried overnight in a fume hood.

Capping of albumin nanoparticles. In an attempt to improve the resuspension

of the synthesized albumin nanospheres, the pendant aldehyde groups were reacted with

several amino acids. It was hypothesized that capping the surface aldehyde groups with

amino acids would reduce the probability interparticle Schiff base cross-link formation

during the particle collection step. The use of amino acids to accomplish this had the

added advantage of altering the surface charge of the nanoparticles thereby increasing the

interparticle repulsion. L-glutamic acid (Glu), glycine sodium salt (Gly), 1-lysine

hydrochloride (Lys), and ethanolamine (EA) were used to cap the pendant aldehyde

groups. The capping step was carried out as follows. Nanoparticles were synthesized as

described above. However, two hours after the initiation of the cross-linking reaction, a

sufficient amount of ethanolamine or amino acid solution was added to the suspension to

quench all the glutaraldehyde added during the cross-linking step (approximately 0.45

mmol). The capping reaction was allowed to continue overnight. The particles were

collected as mentioned in section 3.2.2.2.









Other surfactants. Nanoparticles were also synthesized using the nonionic

surfactants such as Pluronics and Teteronic instead of Tween 80. The same

techniques described in section 3.2.2.2. were used to synthesize these particles.

However, the albumin solution was suspended in either a 2 % (w/v) solution of

Pluronic F68, Pluronic F108, Pluronic F127, or Tetronic 908 in 80 % (v/v) ethanol.

These particles were not capped with albumin or amino acids.

Magnetic particles. Magnetite particles (reported size 30 nm) were donated by

(Nanophase, Inc.; Burr Ridge, IL). In addition, magnetite particles (reported size 30 nm)

were donated by (Materials Modification, Inc.; Fairfax, VA). The methods reported in

section 3.2.2.2 were used with a slight modification to synthesize magnetic particles.

Magnetite (50 mg) particles were suspended in 10 mL of a 5 % albumin solution over

night. The magnetite suspension was collected by centrifugation. The pellet was then

resuspended in 5 mL of a 5 % albumin solution containing 2 % Tween 80. This

suspension was then suspended in 20 mL of 80 % ethanol. The rest of the process is as

described in section 3.2.2.2.

3.2.3 Microsphere and Nanosphere Characterization

3.2.3.1 Electron microscopy

Scanning electron microscopy (SEM) was used to examine the morphology of all

synthesized particles. Dry particles were sprinkled on double-sided sticky tape mounted

on aluminum SEM stubs. Loose particles were blown off the stub using a pressurized air

duster. The samples were coated with a gold/palladium alloy using the Technix

Hummer V sputter coater (sputtering time 2.5 minutes). The samples were analyzed

using the JEOL SEM-6400 scanning electron microscope or JEOL JSM-6330F field









emission scanning electron microscope (JEOL, Ltd., Peabody, MA). The samples were

analyzed at a 5 KeV accelerating voltage and a condenser lens setting of 8 to 10. This

corresponds to a current ranging from 3 x 10-9 to 6 x 10-10 A. The working distance was

15 mm. These conditions were those that minimized charging while maximizing

resolution for the analysis of microspheres. Nanospheres were imaged using a 15 KeV

accelerating voltage and a condenser lens setting of 8 to 10.

Transmission electron microscopy (TEM) was used to examine the morphology

of synthesized magnetic nanoparticles. Samples were prepared by first suspending a

small amount of nanoparticles in acetone (concentration less than 0.5 %). Copper grids

with SiO backing were then dipped and allowed to air dry. These samples were then

examined on the JEOL JEM-200CX transmission electron microscope (JEOL, Ltd.

Peabody, MA). Diffraction patterns of magnetite powder and magnetic microspheres

were also obtained. This analysis was conducted by Dr. Viswanath Krishnamoothy at the

Major Analytical Instrumentation Center (MAIC) at the University of Florida.

Transmission electron microscopy was also used to examine sections of loaded

and unloaded albumin microspheres in an attempt to ascertain their internal structure.

Microsphere samples were processed and analyzed by Karen Kelley at the

Interdisciplinary Center for Biotech Research (ICBR) at the University of Florida. The

sample preparation involved the embedding of the microspheres in an acrylic resin. After

the resin was cured, the samples were thinly sectioned (200 to 500 nm thick sections)

using an ultra-microtome. The sections were then floated on a warm water bath and

collected on the copper grids. These samples were then analyzed on a Hitachi H-7000









transmission electron microscope (Nissei Sangyo America, Ltd. Rolling Meadows, IL a

division of Hitachi Instruments, Inc., Japan).

3.2.3.2 Particle sizing

Initially, microspheres were manually sized using digitized SEM images and the

ScionTM image analysis software. To get a fair estimate of the mean particle size, at least

200 microspheres per sample were measured.

The Coulter LSTM 230 particle size analyzer with the small volume module

(Beckman Coulter Inc., Fullerton, MA) was also used to measure the average particle size

of albumin microspheres. For this analysis, approximately a 1 % particle suspension in

water was prepared. The particles were first wet with acetone (equivalent concentration

after water addition was less than 5 %). Water (house de-ionized water at the particle

research center) was added to each sample just prior to analysis. The sample was mixed

on a vortex mixer for approximately 1 minute. The sample was then added drop-wise

until the obscuration read at least 12 but less than 15. The Fraunhofer optical model was

used to calculate the particle size. The data was processed using the Coulter LS 2.1 la

software. In situ-loaded microspheres were also sized using this technique. However,

since this is a light scattering technique, changes in the refractive index of the liquid

medium during measurement will confound the data. Thus, loaded particles were

exhaustively extracted in 100 % methanol prior to analysis. In addition, the sizing

measurements were also conducted in HPLC grade methanol.

The Coulter MultisizerTM II (Beckman Coulter Inc., Fullerton, MA) equipped

with the 100 orifice tube (100 |tm aperture diameter) was also used to analyze particle

size. In this case, the samples were suspended in the Coulter IsotonTM electrolyte solution









(isotonic saline). Measurements were collected until at least 100,000 particles were

counted. The data was processed using the Multisizer AccuComp 1.19 software.

Nanosphere particle size was analyzed using the ZetaPlusTM zeta potential

analyzer (Brookhaven Instruments Corporation, Holtsville, NY). This analysis was

conducted by Amy Gibson at the University of Florida Engineering Research Center.

Approximately, a 0.1 mg/mL particle suspension was prepared in a 50 mM phosphate

buffer (pH was 4.0, 7.0, or 9.0). The data was processed using the ZetaPlus sizing

software V2.27.

3.2.3.3 Mitoxantrone post-loading procedure

Microspheres were loaded with MXN using the following procedure. Into

10x75 mm disposable culture tubes, 100 mg of microspheres were weighed out. To each

tube, 4.00 mL of a 5.0 mg/mL MXN solution was added. The contents of the tube were

mixed on a vortex mixer for 1 minute, then placed on the roto-tumbler overnight. The

samples were recovered, then centrifuged for 10 minutes at 2000 rpm using the Dynac II

bench top centrifuge. The supernatant was carefully collected for analysis. The samples

were washed with water once and with acetone twice. The supernatants were also

carefully collected in each of the washing steps. The collected supernatants were diluted

to a volume of 50.00 mL then stored for spectrophotometric analysis. Microspheres

loaded using this technique included albumin microspheres, albumin/poly(l-glutamic

acid) blended microspheres (BPG), and albumin/poly(l-lysine) blended microspheres

(BPL).









3.2.3.4 Microsphere drug content determination

Drug content of microspheres was determined using both enzymatic digestion and

solution depletion assays.

Enzymatic digestion. Several enzyme recipes were used in these experiments.

A review of Dr. Jean Quigg's notes revealed that the best results were obtained using the

protease digestion buffer described in section 3.1.1. Freshly prepared protease digestion

buffer (10 mL) was added to loaded microspheres (5 mg). The samples were incubated

in triplicate at 37 C in the TBS (Triangle Biomedical Sciences, Durham, NC) incubator

for at least 48 hours. Longer times were needed for highly cross-linked formulations.

Control drug solutions were incubated to account for any drug degradation during the

digestion. Controls were either 200 ptL of a 1000 [tg/mL MXN solution incubated in

10.0 mL Protease Digestion Buffer or in 10.0 mL of the protease digestion buffer

containing 0.5 mg/mL BSA. After the prescribed time, the samples were recovered and

cooled to room temperature. The samples were examined microscopically to determine

whether the microspheres were fully digested. The dissolved proteins or protein

fragments in each sample were precipitated by incubating 2 mL of each sample in 2 mL

of 10% (w/v) trichloroacetic acid (TCA) at room temperature for 30 minutes. The

samples were then centrifuged at 3000 rpm for 10 minutes using the Dynac II bench top

centrifuge. The supernatants were collected for spectrophotometric analysis.

The techniques reported by Luftensteiner et al. 115 were also used to digest the

loaded microspheres. The Pepsin Digestion Buffer described in section 3.1.1 was used

for these experiments. The incubation and collection procedures used were identical to

those reported in the above paragraph.









Depletion assay. Neither digestion buffer resulted in the complete degradation of

microspheres synthesized with more than 4 % GTA (0.4 mmol GTA per gram BSA)

cross-linking agent concentration. In addition, the BPG microspheres did not degrade

regardless of GTA concentration. This necessitated a second technique to measure drug

payload. The underlying assumption of depletion assays is that the difference between

the concentration of the loading solution prior to incubation with microspheres and that

after incubation with microspheres is equal to the amount of drug taken up by the

microspheres. If proper controls are incubated along with the samples to account for any

losses due to drug oxidation or binding to the glassware, a reasonable estimate of the drug

payload can be calculated.

The drug payload of post loaded was determined using the depletion assay

method. The incubation procedure and sample collection was described in section

3.2.3.3. The controls in this case were 4.00 mL of the loading solution transferred to

10x75 mm disposable culture tubes and incubated as the microspheres were. The loading

solutions underwent all the centrifugation and washing steps described in the

post-loading section. The concentration of the loading solutions and controls were

analyzed using UV-Visible spectrophotometry.

The drug payload of in situ-loaded microspheres was also analyzed using the

depletion assay technique. In this case, the collected solution was the wash solution from

the microspheres collection steps described in section 3.2.1.1. It was difficult to match

the matrix of every sample collected using this technique, so the standard additions

methodl50 (p. 141) was used to quantify the amount of drug in solution.









3.2.3.5 Zeta potential measurements

The zeta potential of nanoparticles was measured using the ZetaPlusTM zeta

potential analyzer (Brookhaven Instruments Corporation, Holtsville, NY) at different pHs

to determine the effect of capping material on surface charge. This analysis was

conducted by Amy Gibson at the University of Florida Engineering Research Center.

The zeta potential measurements were correlated to the particle size at any given pH to

deduce the effect of surface charge on particle agglomeration and potentially swelling.

Samples were prepared in 50 mM phosphate buffer solutions (pH = 4.0, 7.0. or 9.0).

These were the same solutions used to measure particle size. This salt and particle

concentration was too high to measure the Zeta potential. The samples had to diluted ten-

fold prior to analysis. The data was analyzed using the Brookhaven Zeta Potential

Analyzer Ver. 2.18 software.

3.2.4 In Vitro Release of Mitoxantrone from Albumin Microspheres

The in vitro release of MXN from loaded microspheres formulations was tested in

phosphate buffered saline (PBS, 0.05 M, pH = 7.4). In addition, the release of MXN

from in situ-loaded MXN microspheres was tested in heat-treated human plasma.

Incubation procedure. Loaded microspheres (20 mg) were incubated in 100 mL

of 0.05 M PBS at 37 C in the TBS incubator under constant agitation. At predetermined

times, 1.00 mL aliquots were sampled with replacement. The samples were collected in

disposable acrylic cuvettes (Semi-UV semi-micro cuvettes; Fisher Scientific) and stored

at 4 C until analysis.

The release of MXN from in situ-loaded microspheres was also examined in

heat-treated human plasma. In this case, 20 mg of loaded microspheres were incubated in









10 mL of plasma at 37 C. At predetermined times, 9 mL aliquots were sampled and

then replaced with fresh plasma. The collected samples were stored at 4 OC until

analysis.

Unloaded (blank) microspheres were incubated using the same procedures

described above. These microspheres were used as negative controls for the loaded

microspheres. In addition, unloaded microspheres were used to measure the degradation

rate of the microspheres in PBS (pH= 7.4) at 37 C. The amount of protein or protein

hydrolysis products were quantified using the Bio-Rad Protein Assay (Bio-Rad

Laboratories; Richmond, CA). This assay utilizes the shift in absorbance maximum of

Coomassie Brilliant Blue from 465 nm to 595 nm upon binding to basic or aromatic

protein residues. The samples were analyzed using the Perkin-Elmer UV-Vis at 595 nm.

Spectrophotometric analysis. Samples collected from the in vitro release

studies were analyzed spectrophotometrically using either the Perkin-Elmer Lambda 3B

UV-Vis Spectrophotometer (Perkin-Elmer Instruments, Inc.; Norwalk, CT) or the

Shimadzu UV-2401PC UV-Vis Spectrophotometer (Shimadzu Scientific Instruments,

Inc.; Norcos, GA). The absorbance was measured at 610 nm against a matched blank.

Calibration plots were prepared in the range of 1 [tg/mL to 50 [tg/mL. In order for the

calibration data to be accepted, the coefficient of determination (R-squared) for linear

regression of the calibration data was at least 0.990. If the R-squared value was deemed

unacceptable, new standards were prepared and analyzed. Standards were also randomly

analyzed with the unknowns to account for any instrumental variation throughout the

analysis.









Depletion assay samples (section 3.2.3.4) were analyzed using the same

techniques discussed in the above paragraph. In the case of in situ-loaded samples, the

standard additions method was used to quantify the amount of MXN in the wash

solutions. This method was used because of the difficulty of matching the wash solution

compositions. In addition, quartz cuvettes were used for this analysis because the DCE

and acetone would have dissolved the acrylic disposable cuvettes.

Samples collected for microsphere degradation analysis were analyzed using the

microassay procedure described in the Bio-Rad Protein Assay Kit literature. Briefly, The

BSA standard provided in the kit was diluted five times to prepare standards in the range

of 1.2 [tg/mL to 10.0 [tg/mL (the linear range of the assay). The dye reagent was diluted

five fold with ultrapure water. Into disposable acrylic semi-micro cuvettes, 800 ptL of

each standard and 200 ptL of the dye reagent were quantitatively transferred. The

samples were incubated for 10 minutes at room temperature then their absorbance was

measured at 595 nm. Samples were treated in the same manner as the standards.

3.2.5 Animal Studies

All mice utilized in this work were 10 to 12 weeks old female C3H/HeJ purchased

from Jackson laboratories (The Jackson laboratory, Bar Harbor, ME). The animals were

housed at Animal Resources in the University of Florida Health Center. They were fed

ad libium and maintained on a 12-hour light schedule. All animal procedures were

approved by the University of Florida Institutional Animal Care and Use Committee

(IACUC).









3.2.5.1 Murine mammary adenocarcinoma

The 16/C MAC was obtained from Dr. Seimann's laboratory in the Department of

Radiation Oncology at the University of Florida. The 16/C adenocarcinoma was

maintained by serial passage in female C3H/HeJ mice. Two techniques were used to

passage this tumor.

Transplantation of tumor pieces. Tumor bearing mice were sacrificed using

CO2 asphyxiation. The tumor was excised and placed in a pre-weighed centrifuge tube

containing sterile-filtered, potassium-containing, pho sphate-buffered saline (KPBS). The

weight of the tumor was determined as the difference between the weight of centrifuge

tube with the tumor and its weight without the tumor. The tumor was then placed on a

sterile 60 x 15 mm disposable polystyrene petri dish (Fisher Scientific). The tumor was

diced into a fine paste using curved scissors. The appropriate amount of sterile KPBS

was then added to prepare a 500 mg/mL tumor suspension. The tumor was thoroughly

suspended by repeatedly aspirating then decanting the fluid using a 10 mL sterile

disposable polystyrene serological pipette. The tumor suspension was then injected

subcutaneously onto the backs of new female C3H/HeJ mice using a 23 gauge 1-inch

hypodermic needle. Each mouse received a 50 ptL injection, the equivalent of 25 mg of

suspended tumor pieces.

Transplantation of tumor cells. As in the case of transplantation of tumor

pieces transplantation, the tumor was diced into a fine paste on the sterile petri dish. The

tumor paste was then suspended in 10 mL of freshly prepared Dnase enzyme cocktail

(described in section 3.1.1). The tumor was then transferred to a sterile 50 mL centrifuge

tube containing 25 mL Dnase enzyme cocktail. It was incubated in the TBS incubator at









37 C for one hour. The tumor suspension was then passed through a sterile fine mesh

screen or a 70 jtm cell strainer into a new sterile 50 mL centrifuge tube. The strainer was

washed with 10 mL KPBS. The cell suspension was centrifuged at 2000 rpm at 4 OC for

10 minutes using the Beckman centrifuge. The supernatant was discarded, and the plug

suspended in 1.5 mL of fresh KPBS. This will be referred to as the Master (M). A

sample of the Master was diluted by 50 fold (M/50). A sample of the M/50 (250 ptl) was

mixed with trypan blue (50 ptl). A small sample of this mixture was transferred to a

hemacytometer. Trypan blue will bind to dead cells only. The live (uncolored) cells in

each of the four outer quadrants (see arrows on Figure 3.1) were counted. The number of

cells per milliliter suspension in the M/50 dilution can be calculated by the multiplying

the average number of cells per quadrant by 1.2 x 104. Thus, the number of cells per

milliliter in the master will be 50 times the above number. The concentration of the

Master was then adjusted to yield an injection of 5 x 106 cells in the 50 ptl injection.


\,,,,,,, /


-z-:::::::::::::::::::-
.77-

.I. . .. | . .


Figure 3.1 Image of a standard hemacytometer chamber with arrows pointing towards
the areas to be counted in the calculation the number of viable tumor cells.









Initially, the results of the tumor cell transplants were discouraging. The tumor

growth and take rate was not reproducible. The take rate is defined as the proportion of

animals that develop a palpable tumor within the 10 to 15 days after transplantation.

However, the transplantation of tumor pieces was reproducible with high take rates

(greater than 95 %). This tumor transplantation method was used for all animal studies

reported in this research. It should be noted that the tumor cell transplantation method

was not abandoned. However, recent attempts have been successful at transplanting

tumor cells that develop tumors with growth rates similar to the growth rate of the tumor

pieces transplants. These methods may be used in future animal studies reported by this

group.

3.2.5.2 Tumor response to mitoxantrone

This study investigated the effect of local preoperative MXN treatment on the

16/C adenocarcinoma. The local (intratumoral) administration of MXN was compared to

the standard intravenous administration. Two dose levels, 4 mg/kg and 8 mg/kg, were

chosen for this study. It should be noted that the intravenous LD5o for MXN in mice is

6.60 mg/kg. In order to assess the effect of preoperative chemotherapy, surgical resection

groups were added to compare preoperative chemotherapy to chemotherapy alone. A

total of 10 treatment groups of 12 animals each were needed to complete this study.

Figure 3.2 shows the breakdown of the treatment groups.

Little was known about the effect of the intratumoral delivery of free MXN on the

tumor and animal. Therefore, this study was conducted in two parts. The 8 mg/kg dose

level was investigated first. Tumor bearing animals (24 per group) were treated either

intravenously or intratumorally with the 8 mg/kg dose. The animals were then

randomized to either the surgery or the non-surgery subgroups (12 animals each).









64










0







o

-e-


'C













o











-e
C0
rA







t4
O+


















-Iu
.,-









Animals in the surgery groups underwent surgical resection of their tumor 10 days

after receiving treatment. Since none of the non-treatment controls lived to the 10 day

time point, surgical control animals underwent surgical resection of their tumors five

days after inclusion in the study. When the 8 mg/kg groups were all treated, the 4 mg/kg

groups were treated.

Treatment conditions. Animals were treated when their tumor measured

between 10 and 12 mm in the longest dimension. Tumor dimensions were measured

using Vernier calipers. Tumor weight was estimated using the following equation:

axb2
Weight(g) -
2000

The weight of the tumor was calculated in grams. The "a" dimension is the tumor's

longest dimension in mm. The "b" dimension is the tumor's dimension orthogonal to

"a". Animals that developed more than one tumor nodule prior to treatment were

excluded from the study. In addition, animals that showed signs of malaise prior to

treatment were excluded from the study.

Measured parameters. Tumor dimensions and animal weights were measured

every other day. Animals were monitored for signs of lethargy or malaise. Animals that

exhibited these signs were isolated. In addition, animals whose tumors ulcerated and

bled after treatment were also isolated. The animals were isolated to minimize the

chances of cannibalism by their cage mates. Animals whose tumor weight reached 10 %

of their body weight were sacrificed. Animals that exhibited body weight losses of 20 %

or more(after adjusting for tumor weight) were also sacrificed. The tumor and liver of

each sacrificed animal was harvested for histological examination. The tumors and livers

were fixed with 10 % normal buffered formalin. The fixed tissues were stored at room









temperature until processed for histological evaluation. From the above collected data,

tumor growth and/or regression were calculated. In addition, animal body weight

changes were calculated as one measure of drug toxicity. Histological evaluation of the

excised livers was another measure of drug toxicity.

Statistical evaluation. This study was designed to detect a 25 % difference in the

average tumor weight between the treatment groups with a 95 % level of confidence.

The differences in average tumor weight and animal body weights between treatment

groups were analyzed using 1-way ANOVA and 2-way ANOVA tests followed by

appropriate multiple comparisons tests. The proportions of animals that experienced

large weight losses to those that did not were compared using X2 tests or Fisher exact

tests.

3.2.5.3 Tumor response to albumin encapsulated mitoxantrone

Pilot studies investigating the toxicity and tumor response of encapsulated

mitoxantrone were conducted. In situ-loaded microspheres synthesized with 8 % GTA

concentration (0.8 mmol per gram BSA) were used for these studies. These microspheres

exhibited an intermediate release rate when incubated in plasma. Initially, intratumoral

administration of free MXN was compared to dispersions of MXN-loaded microspheres

in saline at comparable drug concentrations. Mitoxantrone-loaded microspheres were

administered using a 25 gauge needle at the 8 mg/kg, 16 mg/kg, and 24 mg/kg dose

levels. A total of eight treatment groups of four animals each were treated. The control

groups were untreated animals and drug-free microsphere treated animals. The treatment

conditions and measured parameters were the same as those measured in the baseline

study (section 3.2.5.3). This study was followed by another pilot study investigating the






67


toxicity and tumor response to escalating doses of microsphere-bound mitoxantrone. In

this case the doses were 24 mg/kg, 32 mg/kg, and 48 mg/kg. The microspheres were

suspended in isotonic saline containing 0.5 % Tween 80 (Tsaline) to improve

microsphere suspension stability and reduce sticking to the syringe. The controls were

untreated animals, Tsaline injected animals, and drug-free microsphere injected animals.














CHAPTER 4
SYNTHESIS AND RELEASE RESULTS AND DISCUSSION


4.1 Microsphere Synthesis

Albumin microspheres were synthesized using the suspension cross-linking

method described in section 3.2.1.1. The effect of processing parameters on particle size

was investigated in some detail because of the paucity of such data in the literature. Of

particular interest was the potential synergistic effect of processing parameters on particle

size. The effect of CAB (steric hindrance agent) concentration, BSA (matrix protein)

concentration, GTA (cross-linking agent) in DCE concentration, D/C (dispersed to

continuous phase) volume ratio, cross-linking reaction time, and cross-linking reaction

stirring rate were investigated in a statistically designed pilot study. The total volume of

liquid in the reaction vessel was fixed to 50 mL. The reaction vessel and propeller shapes

were kept constant. The initial stirring rate was maintained at 1250 rpm for 20 minutes.

The process variable conditions are tabulated in Table 3.1. The results of the pilot study

were then used to design a more comprehensive statistical study of the most important

processing parameters controlling particle size. The conditions of this study are tabulated

in Table 3.2.

4.1.1 Effect of Processing Parameters on Particle Size

4.1.1.1 Pilot study

The process conditions of synthesized batches are tabulated in Table 4.1. The

particle sizes were measured using digitized SEM images and the ScionM image analysis










software as described in section 3.2.3.2. The measured particle sizes are tabulated in

Table 4.2. The particle size distribution was described as the mean to median ratio. The

obtained particle size data were statistically analyzed using the CARDTM software

package. The analysis suggested that CAB concentration, BSA concentration, and D/C

ratio were the most important parameters controlling particle size. In addition, the

analysis suggested that interaction terms between the process variables were also

important.


Table 41


Synthesis conditions s


Sample Run Time GTA Cone. BSA Cone. CAB Cone. Stirring D/C
Name (hrs) (% w/w) (% w/v) (%o w/v) Rate Ratio (%)
(rpm)
BSA1 3.0 6 15 0.5 300 10.0
BSA2 1.5 6 30 0.5 300 5.0
BSA3 1.5 6 15 0.5 300 5.0
BSA4 1.5 6 15 0.5 300 10.0
BSA5 3.0 6 15 0.5 300 10.0
BSA6 1.5 12 30 2.0 300 5.0
BSA7 1.5 6 30 2.0 600 10.0
BSA8 3.0 6 15 2.0 300 5.0
BSA9 3.0 12 30 2.0 600 10.0
BSA10 1.5 6 15 2.0 300 5.0
BSA11 3.0 12 15 0.5 300 10.0
BSA12 1.5 12 30 2.0 600 10.0
BSA13 1.5 12 15 0.5 300 10.0
BSA14 3.0 6 15 0.5 600 5.0
BSA15 3.0 6 15 2.0 300 5.0
BSA16 3.0 12 15 2.0 300 10.0
BSA17 1.5 6 15 0.5 600 5.0
BSA18 3.0 12 30 0.5 300 5.0
BSA19 3.0 6 30 2.0 600 10.0
BSA20 3.0 12 15 0.5 300 5.0
BSA21 1.5 12 15 0.5 600 5.0
BSA22 1.5 12 15 0.5 300 5.0
BSA23 3.0 6 30 0.5 300 5.
BSA24 1.5 12 30 0.5 600 10.0
BSA25 3.0 12 15 0.5 600 5.0
BSA26 1.5 12 30 2.0 300 5.0









Table 4.2 Particle size and size distribution of batches synthesize in pilot study as
measured using digitized SEM images and the computer software package ScionTM.
Sample Name Average Diameter Standard Deviation Mean to Median
([im) ([tm) Ratio
BSA1 10.2 5.5 1.072
BSA2 21.7 13.9 1.070
BSA3 9.9 4.1 1.017
BSA4 8.0 4.3 1.096
BSA5 10.9 5.4 1.054
BSA6 11.7 6.2 1.171
BSA7 6.9 5.2 1.202
BSA8 11.0 7.9 1.167
BSA9 11.1 6.9 1.036
BSA10 4.5 3.8 1.205
BSA11 12.3 5.5 1.045
BSA12 8.0 5.5 1.248
BSA13 8.9 4.8 1.067
BSA14 12.1 6.6 1.008
BSA15 6.1 3.8 1.127
BSA16 4.6 3.1 1.232
BSA17 12.3 6.2 0.995
BSA18 20.3 12.0 1.028
BSA19 7.5 5.2 1.210
BSA20 18.2 6.4 0.987
BSA21 9.6 6.4 1.112
BSA22 10.9 5.8 1.141
BSA23 24.0 11.6 0.996
BSA24 16.8 10.7 1.120
BSA25 12.5 4.7 1.037
BSA26 11.7 6.4 1.182


4.1.1.2 Full microsphere synthesis study

The results of the Pilot Study were used to design a larger scale study with

expanded process variable ranges. This will be referred to as the "Full Study" herein.

The GTA concentration was included as a variable even though the pilot study suggested

that the GTA concentration effect was not significant. This was done for two reasons.

First, the interaction term between GTA and CAB concentrations in the pilot study was

significant suggesting synergy between the two variables and thereby both were

significant. Second, the GTA concentration controls microsphere cross-link density and










therefore swelling, drug entrapment, and release characteristics. These are all variables

of interest to the scope of this work. The experiment was designed using CARDTM's

Design of Experiment module. Each experimental condition was run in triplicate. Table

4.3 lists the run numbers and the associated process conditions.


Table 4.3
this study.


Full study process conditions. Batches were synthesized in triplicate for


Run # BSA Cone. CAB Cone. GTA Cone. D/C Ratio
(% w/v) (% w/v) (% w/v) (% v/v)
1 10 0.5 15 5.0
2 10 5.0 15 15.0
3 10 5.0 15 5.0
4 20 2.0 10 5.0
5 20 2.0 10 15.0
6 10 5.0 5 15.0
7 20 2.0 5 10.0
8 20 0.5 10 10.0
9 30 2.0 10 10.0
10 30 0.5 5 5.0
11 10 0.5 15 15.0
12 30 0.5 15 15.0
13 30 5.0 5 5.0
14 10 5.0 5 5.0
15 30 5.0 15 15.0
16 10 0.5 5 15.0
17 30 5.0 15 5.0
18 20 0.5 10 10.0
19 20 2.0 10 10.0
20 10 2.0 10 10.0
21 20 2.0 15 10.0
22 20 2.0 10 10.0
23 30 0.5 15 5.0
24 30 5.0 5 15.0
25 20 2.0 10 5.0
26 30 0.5 5 15.0
27 20 5.0 10 10.0
28 10 0.5 5 5.0
29 30 5.0 15 5.0
30 30 0.5 5 5.0









Table 4.4
following data


Particle size and size distributions o'Full Study microspheres. The
e ra averages of three batches (Run #6 wa )


The particle sizes of the various batches were analyzed using the CoulterTM LS

230 particle size analyzer. The obtained data were averaged then tabulated in Table 4.4.

The Span is a measure of particle size distribution. It was calculated by dividing the

difference between the 90th percentile diameter and 10th percentile diameter by the 50th


Run Average Average Standard Average Average Span
Number Diameter Deviation Mean to
([im) ([tm) Median Ratio
1 14.6 2.4 0.727 1.185
2 7.2 2.6 1.133 2.328
3 3.9 2.6 0.890 2.677
4 12.2 2.8 0.762 1.675
5 9.2 2.7 0.827 2.380
6 7.5 2.0 1.061 2.122
7 15.8 3.1 0.889 3.618
8 16.4 2.6 0.751 1.391
9 12.8 2.8 0.732 1.566
10 27.9 2.6 0.759 1.309
11 15.7 2.8 0.811 2.072
12 27.9 2.6 0.770 1.184
13 8.4 2.6 8.98 2.695
14 6.0 2.7 1.015 3.492
15 6.4 2.9 0.859 2.979
16 15.5 2.5 0.799 1.749
17 4.4 2.6 0.829 2.028
18 17.4 2.6 0.723 1.301
19 10.1 2.7 0.718 1.537
20 9.6 2.6 0.782 1.814
21 8.6 2.7 0.729 1.750
22 9.9 2.7 0.754 1.730
23 24.3 2.6 0.738 1.069
24 6.4 2.6 0.884 2.727
25 11.4 2.9 0.796 2.355
26 24.8 2.4 0.757 1.289
27 4.0 2.1 0.864 1.657
28 15.9 2.5 7.49 1.362
29 3.8 2.5 0.773 1.792
30 30.2 2.4 0.761 1.080









percentile diameterl51. The averaged data were analyzed using the statistical software

package JMPIN version 3.2.1. A second order least squares regression analysis was used

to model the effect of process conditions on particle size. A model that fit the data with a

coefficient of determination (R2) equal to 0.956 was calculated. The analysis of variance

(ANOVA) table for the whole model is presented in Table 4.5. The model term estimates

and their associated F-ratio and p-values are tabulated in Table 4.6.


ANOVA table for the calculated model
on particle size.


describing the effect of processing


Source Degrees of Sum of Mean F-Ratio p-value
Freedom Squares Square
Model 6 1606.282 267.714 83.158 < 0.001
Error 23 74.045 3.219
Total 29 1680.326






Table 4.6 Model term coefficients for the second order model.
Model Term Coefficient Standard Error F-Ratio p-value
Intercept 22.222 3.143 49.985 <0.001
BSA 0.604 0.065 87.429 <0.001
CAB -4.198 1.207 12.102 0.002
GTA -2.377 0.714 11.089 0.003
BSAxCAB -0.124 0.019 43.224 <0.001
CAB2 0.622 0.202 9.4334 0.005
GTA2 0.108 0.036 9.294 0.006


Table 4.5
conditions


















"-
E


.2

u


10 11 12 13 1415 1 7 1 2 2


BSA Conc 26 27 28 29 30


al1 *20-25
015-20
010-15
05-10
10-5














5
3.5
2 CAB Conc
0.5


014-16
S12-14
110-12
18-10
S6-8
S4-6
S2-4
*0-2


13

GTA Conc


1->>J% %Vlu R1. -D
CO
Figure 4.1 Effect of processing conditions on albumin particle size in micrometers.
(a) Effect of BSA and CAB concentration on particle diameter when the GTA
concentration is fixed at 10 %. (b) Effect of BSA and GTA concentration on particle
diameter when CAB concentration is fixed at 2.75 %.















S20-25


2 15-10
0 0-5






I- 0

U.







0.5 0.7
1.0 1.2 1.4 1.6 1.9 GTA Conc
CAB9 2.1 2.3 2.5 2.8 3.0 .2 3.4 3.7 3.9 4.1 4.3 4.6 4.8 5.0
CAB Conc

Figure 4.1 (c) Effect of CAB and GTA concentration on particle diameter when the
BSA concentration is fixed at 20 %.



Sample BSAF6, prepared using Run # 6 conditions, was dropped from the

analysis because it did not suspend well in water prior to analysis. Thus, the data

obtained from that analysis were erroneous. The SEM micrograph revealed that the

microspheres in this batch were highly agglomerated. Other samples made using

condition # 6 were also agglomerated, but they resuspended after vigorous sonication.

More about the morphology of the various batches will be presented in section 4.1.1.2.

Figures 4. la, b and c graphically show the effect of the processing conditions on particle

size.









The model predicted that albumin particle size increased as the concentration of

CAB decreased or as the concentration of BSA increased. However, the effect of GTA

was not as obvious. It exhibited a minimum at 11.0 % (1.1 mmol GTA/gBSA) and

therefore, the particle size decreased between 5 and 11 % and increased between 11 and

15 %. It should be noted that 1 % is equivalent to 0.10 mmol GTA per gram BSA. In

addition, the interaction between the BSA and CAB concentrations was significant.

There was no evidence to suggest that variation of the D/C ratio within the range

investigated had a significant effect on the particle size.

The increase in particle size as the stabilizing agent concentration

decreases 115,121 or as albumin concentration increases29,115,118,121,122 is

documented and generally accepted in the literature. In addition, empirical relationships

relating particle size of microspheres prepared by suspension cross-linking methods have

been proposed. These relationships predict that the particle size is directly related to the

viscosity of the dispersed phase and inversely proportional to the viscosity of the

continuous phase and the concentration of stabilizing agent.23

The increase in particle size as the concentration of CAB decreases can be

explained if one considers the mechanism of steric stabilization. Steric repulsion occurs

when the distance between approaching polymer-covered surfaces drops below a few

times the radius of gyration of stabilizing polymer. The exact distance will depend on the

interaction between polymer and surface, the surface coverage, and the solvent

properties. There are two driving forces for steric repulsion. When the separation

distance between the approaching surfaces is small enough that the stabilizing polymer

layers start to overlap, the local concentration of polymer in the overlapped region is









higher than the bulk solution giving rise to a repulsive osmotic pressure that drives the

surfaces apart. 152,153 In addition, the overlap of the stabilizing polymer layers would

reduce the entropy of the system making this situation energetically unfavorable. 152 As

the stabilizing polymer concentration drops to a point below full surface coverage,

flocculation (Bridging flocculation) can occur. 153,154 This occurs when the separation

distance between two particles is small enough that polymer molecules on the surface of

one particle can interact with the surface of the other thereby bridging the gap. In the

case of liquid-liquid interfaces, the surfaces are fluid and instead of flocculation, droplet

coalescence occurs. Thus, the particle size is expected to increase as the concentration of

CAB decreases. The albumin concentration controls the amount of particle-forming

material required to stabilize the suspension.

As the concentration of albumin increases, the viscosity of the aqueous phase

increases as does the resistance to deformation and disintegration under shear increases.

Thus, for a given shear rate, the droplet size is expected to increase as the viscosity of

dispersed phase increases. In addition, as the concentration of BSA increases, more

particle forming material is available. Thus, for a given CAB concentration, the particle

size is expected to increase as a function of BSA concentration due to the shortage of

stabilizing polymer. The interaction between the CAB and BSA concentration is not

surprising since the concentration of BSA controls number of droplets formed in

suspension and the CAB concentration controls the stabilization of these particles. Thus,

these two variables have a synergistic effect on particle size.

The above analysis did not provide evidence to suggest that the D/C ratio has a

significant effect on particle size. However, several published reports suggest that









increasing the D/C ratio increases particle size. Muller et al. 118 reported that the particle

size increased as the D/C ratio increased whereas El-Mahdy et al. 121 suggested that the

D/C ratio had an effect only after reaching a threshold value. The above studies do not

necessarily contradict this study. Although both reported using comparable D/C ratio

ranges, the emulsification methods and more importantly the particle size range of

interest were different. Muller et al. used a combination of ultrasonication and static

mixing to prepare sub-micron particles and El-Mahdy et al. used a membrane

emulsification method to prepare particles less than 2 |tm in diameter. The D/C ratio is

the volume fraction of the aqueous phase in the organic phase. For a given set of

conditions, it will control the number of stable droplets in suspension. As the droplet size

decreases, the percentage of material at the interface increases and thus the interfacial

properties of the dispersed and continuous phases increasingly dominate the stability of

the dispersion. In the case of larger particles, the bulk properties, e.g. density, dominate

the stability of the dispersion. Thus, it is expected that the small particle size regime

addressed in the above previous work was a major contributor to the D/C ratio effect.

Most of the research discussed thus far was conducted using the classical one

factor at a time methodology. Only the reports of Luftensteiner et al. 115 and Muller et

al. 118 were statistically-designed multivariate studies and thus more relevant to this

research. Muller et al. studied similar processing variables to those reported here.

However, they were interested in a different particle size range (sub-micron) and used

different emulsification methodologies. Lufetensteiner et al. prepared super-micron

particles. However, they investigated a different set of processing parameters than those

studied in this work. The only variables in common were protein concentration and




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - - mvs