Citation
A combined reverse thermal gel-polymeric micelle system for sustained delivery of opthalmic drugs

Material Information

Title:
A combined reverse thermal gel-polymeric micelle system for sustained delivery of opthalmic drugs
Creator:
Famili, Amin ( author )
Place of Publication:
Denver, CO
Publisher:
University of Colorado Denver
Publication Date:
Language:
English
Physical Description:
1 electronic file (178 pages). : ;

Subjects

Subjects / Keywords:
Ophthalmic drugs ( lcsh )
Drugs -- Administration ( lcsh )
Ocular pharmacology ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Delivery of drugs to the eye is a challenging endeavor due to the myriad anatomical and physiological barriers preventing intraocular absorption of molecules that either contact the anterior surfaces of the eye or are present in the bloodstream. Topical administration{7f2014}the workhorse of ophthalmic drug delivery{7f2014}suffers from numerous drawbacks including poor intraocular bioavailability, excessive systemic absorption, need for frequent re-administration and a reliance on patient adherence for therapeutic efficacy, among others. To overcome these limitations, researchers have explored various controlled release drug delivery systems with the ultimate goal of sustaining consistent, localized drug levels at the target tissue thereby maximizing therapeutic efficacy and minimizing unintended adverse effects. Among the options explored, in situ-gelling polymeric systems may hold the most promise due to their ability to be administered directly at the target site by a minimally-invasive injection and form a stable physical gel while conforming to the specific anatomy of that space. To date, clinical application of such systems has been hindered by their limited ability to sustain long-term delivery of drugs. To overcome this limitation, we sought to develop a system comprising a reverse thermal gel (RTG) encapsulating drug-loaded polymeric micelles as a combined system for the sustained local delivery of poorly soluble drugs. The polymers comprising the RTG and the micelles{7f2014}both novel {7f2014}were first independently characterized as free-standing systems. The combined system was then evaluated and was found to retain the injectability and in situ gelling characteristics of the thermal gel, but additionally benefited from the superior ability of the entrapped micelles to encapsulate and sustain release of the poorly soluble corticosteroid triamcinolone acetonide (TA). Release of TA from the combined system was completely free of a burst release and was projected to continue at a steady rate for approximately twelve months. To our knowledge, this system is the first in the literature to achieve delivery time frames from an in situ-gelling polymer beyond a few months and has the potential to significantly improve delivery of TA and, more broadly, clinical treatment of posterior ophthalmic diseases.
System Details:
System requirements; Adobe Reader.
Thesis:
Bioengineering
General Note:
Department of Bioengineering
Statement of Responsibility:
by Amin Famili.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
899247464 ( OCLC )
ocn899247464

Downloads

This item has the following downloads:


Full Text
A COMBINED REVERSE THERMAL GEL-POLYMERIC MICELLE SYSTEM FOR
SUSTAINED DELIVERY OF OPHTHALMIC DRUGS
by
AMIN FAMILI
B.S., Lehigh University, 2009
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Bioengineering Program
2014


This thesis for the Doctor of Philosophy degree by
Amin Famili
has been approved for the
Bioengineering Program
by
Richard K. Benninger, Chair
Malik Y. Kahook, Advisor
Daewon Park, Advisor
Tom Anchordoquy
Jeff Stansbury
February 11, 2014
11


Famili, Amin (Ph.D., Bioengineering)
A Combined Reverse Thermal Gel-Polymeric Micelle System for Sustained Delivery of
Ophthalmic Drugs
Thesis directed by Professor Malik Y. Kahook and Assistant Professor Daewon Park
ABSTRACT
Delivery of drugs to the eye is a challenging endeavor due to the myriad anatomical and
physiological barriers preventing intraocular absorption of molecules that either contact
the anterior surfaces of the eye or are present in the bloodstream. Topical administration
the workhorse of ophthalmic drug deliverysuffers from numerous drawbacks including
poor intraocular bioavailability, excessive systemic absorption, need for frequent re-
administration and a reliance on patient adherence for therapeutic efficacy, among others.
To overcome these limitations, researchers have explored various controlled release drug
delivery systems with the ultimate goal of sustaining consistent, localized drug levels at
the target tissue thereby maximizing therapeutic efficacy and minimizing unintended
adverse effects. Among the options explored, in situ-gelling polymeric systems may hold
the most promise due to their ability to be administered directly at the target site by a
minimally-invasive injection and form a stable physical gel while conforming to the
specific anatomy of that space. To date, clinical application of such systems has been
hindered by their limited ability to sustain long-term delivery of drugs. To overcome this
limitation, we sought to develop a system comprising a reverse thermal gel (RTG)
encapsulating drug-loaded polymeric micelles as a combined system for the sustained local
delivery of poorly soluble drugs. The polymers comprising the RTG and the micelles
both novelwere first independently characterized as free-standing systems. The


combined system was then evaluated and was found to retain the injectability and in situ
gelling characteristics of the thermal gel, but additionally benefited from the superior
ability of the entrapped micelles to encapsulate and sustain release of the poorly soluble
corticosteroid triamcinolone acetonide (TA). Release of TA from the combined system was
completely free of a burst release and was projected to continue at a steady rate for
approximately twelve months. To our knowledge, this system is the first in the literature to
achieve delivery time frames from an in situ-gelling polymer beyond a few months and has
the potential to significantly improve delivery of TA and, more broadly, clinical treatment
of posterior ophthalmic diseases.
The form and content of this abstract are approved. I recommend its publication.
Approved: Malik Y. Kahook and Daewon Park
IV


ACKNOWLEDGEMENTS
I am grateful to all of the individuals that have helped medirectly and indirectlyin
reaching this milestone. It is a product of the countless teachers, professors, mentors,
advisors, friends and family members that have contributed to my development and success
and I am nothing without their contributions.
My advisors, Drs. Malik Kahook and Daewon Park, have been invaluable in their support,
advisement and guidance of my efforts. None of this would be possible without them, and
I am eternally grateful for their patience and unwavering backing. I also thank the members
of the Department of Ophthalmology for their expertise and constant willingness to help,
especially Dr. David Ammar who always took the time to help me accomplish and
understand any task at hand. My lab mates in the Translational Biomaterials Research Lab
were always there when I needed them and helped me in more ways than they may realize.
Two of my earliest career mentors and scientific role models, Saurabh Palkar and Bill
Baldy, put me on and pushed me along the path that got me here today. None of what I
have achieved would be possible without them and I am forever indebted to them for that.
The friends I have made along the way have been a constant source of happiness, help and
advice without which I wouldnt be who I am today. For all of the ways they have improved
my life and kept me going despite trials and tribulations, I express my deepest gratitude.
Finally, I thank my parents, about whom enough cannot be said. It was their immeasurable
efforts that instilled in me their curiosity, discipline, commitment and perseverance that
have, in turn, allowed me to progress in life. They have been fully supportive every step of
the way and have sacrificed enormously for me. Thank you, from the bottom of my heart.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.................................... 1
II. POLYMERIC MICELLES FOR SUSTAINED RELEASE OF A
POORLY SOLUBLE DRUG............................. 49
III. A REVERSE THERMAL GEL SYSTEM FOR LOCALIZED
DELIVERY OF TRIAMCINOLONE AC LION IDE............... 69
IV. POLYMERIC MICELLES ENCAPSULATED IN A REVERSE
THERMAL GEL AS AN INJECTABLE IN6Y///-GELLING
OCULAR DRUG DELIVERY SYSTEM..................... 91
V. SYNTHESIS AND CHARACTERIZATION OF A
BIODEGRADABLE POLY(A-ISOPROPYLACRYLAMIDE)-
BASED COPOLYMER.................................102
VI. DISCUSSION, LIMITATIONS AND FUTURE DIRECTIONS...119
REFERENCES..............................................136
APPENDIX
A. SYNTHETIC ROUTE OF PEG-PHS-PEG..................156
B. UV SPECTROMETRIC STANDARD CURVE FOR
TRIAMCINOLONE ACETONIDE CONCENTRATION
DETERMINATION...................................157
C. SYNTHETIC ROUTE OF PSHU-NIPAAM..................158
D. PSHU GPC ANALYSIS...............................159
E. PSHU MONOMER RATIO DETERMINATION................160
F. PSHU-NIPAAM MINIMUM GELLING CONCENTRATION
DETERMINATION...................................161
G. MODELING OF TA RELEASE FROM PSHU-NIPAAM GELS....162
H. RELEASE OF RANIBIZUMAB FROM PSHU-NIPAAM GELS....163
vi


LIST OF TABLES
TABLE
1.1 INTRAOCULAR CONCENTRATIONS OF TOPICALLY-
ADMINISTERED TIMOLOL MAI.LATE.................... 8
1.2 BENEFITS OF CONTROLLED-RELEASE DRUG DELIVERY
SYSTEMS OVER TOPICAL OR SYSTEMIC ADMINISTRATION
TECHNIQUES...................................... 17
2.1 PROPERTIES OF HIGH AND LOW MOLECULAR WEIGHT PEG-
PIIS-PEG BY GPC ANALYSIS........................ 59
2.2 POTENTIAL AVENUES FOR PEG-PHS-PEG MODIFICATION.. 59
3.1 ELEMENTAL ANALYSIS RESULTS OF PSHU, DPSHU,
PM PA AM AND PSIIl-MPAAVl....................... 80
E. 1 'H NMR PEAKS AND INTEGRATED AREAS USED IN PSHU
MONOMER RATIO CALCULATIONS......................160
F. 1 CONCENTRATION DEPENDENCE OF PSHU-NIPAAM GELLING.161
G. 1 FIT-TO-FIRST-ORDER PARAMETERS FOR VARIOUS TA
CONCENTRATION RELEASE PROFILES..................162
vii


LIST OF FIGURES
FIGURE
1.1 DIAGRAMMATIC REPRESENTATION OF THE EYE........... 2
1.2 SIMULATED KINETICS OF PULSATILE AND CONTROLLED-
RELEASE DRUG DELIVERY PARADIGMS.................. 9
1.3 ROUTES OF SUB-TENONS CAPSULE, SUB-CONJUNCTIVAL
AND INTRAVITREAL INJECTIONS..................... 11
1.4 DISTRIBUTION OF MARKETED AND INVESTIGATIONAL
DRUGS WITHIN THE BCS CLASSIFICATION SYSTEM...... 14
1.5 MECHANISM OF LIQUID-TO-PHYSICAL GEL PHASE
TRANSITION FOR BLOCK COPOLYMER-BASED REVERSE
THERMAL GELS.................................... 29
1.6 INFLUENCE OF DRUG HYDROPHOBIC IT Y ON
ENCAPSULATION EFFICIENCY IN PNIPAAM-BASED
REVERSE THERMAL GELS............................ 34
1.7 NANO-CARRIER DRUG DELIVERY SYSTEMS AND THEIR
MECHANISMS OF DRUG ENCAPSULATION................ 43
2.1 NMR ANALYSIS OF THE PEG-PHS-PEG COPOLYMER..... 58
2.2 PEG-PHS-PEG IS NON-CYTOTOXIC AGAINST RETINAL
PIGMENTED EPITHELIAL CELLS IN CULTURE........... 60
2.3 MICELLES PRODUCED BY A FILTER EXTRUSION METHOD
ARE SIGNIFICANTLY MORE MONO-DISPERSE THAN THOSE
PRODUCED BY A SONICATION TECHNIQUE.............. 61
2.4 MICELLES PRODUCED BY BOTH PROCESSES WERE
SPHERICAL AND POSSESSED A CORE-SHELL MORPHOLOGY 63
2.5 EFFICACY OF THE MICELLE PURIFICATION PROCESS IN
REMOVING DMSO FROM THE FINAL PRODUCT............ 64
2.6 RELEASE OF TRIAMCINOLONE ACETONIDE FROM PEG-PHS-
PEG MICELLES WAS NEAR ZERO-ORDER AND HIGHLY
SUSTAINED....................................... 66
viii


3.1 'h nmr spectrum of pshu copolymer and
CONFIRMATION OF BOC DEPROTECTION................ 79
3.2 FT-IR ANALYSIS CONFIRMED SUCCESSFUL CONJUGATION
OF PNIPAAM-COOH TO PSHU TO PRODUCE PSHU-NIPAAM.. 81
3.3 TEMPERATURE-DEPENDENT GELLING KINETICS OF PSHU-
NIPAAM BY OPTICAL AND RHEOMETRIC ANALYSES....... 82
3.4 PSHU-NIPAAM IS NON-CYTOTOXIC AGAINST RETINAL
PIGMENTED EPITHELIAL CELLS IN CULTURE................ 82
3.5 INTRAVITREAL INJECTIONS OF PSHU-NIPAAM IN RATS
DEMONSTRATED ITS IN VIVO BIOCOMPATIBILITY....... 84
3.6 RELEASE OF TRIAMCINOLONE ACETONIDE FROM PSHU-
NIPAAM GELS WAS DEPENDENT ON LOADING FRACTION
BUT INDEPENDENT OF PSHU-NIPAAM CONCENTRATION.... 85
3.7 DE-SWELLING KINETICS OF PSHU-NIPAAM GELS........ 87
4.1 PEG-PHS-PEG MICELLES CAUSED NO ADVERSE REACTIONS
AFTER INTRAVITREAL INJECTIONS IN RATS............... 96
4.2 RHEOLOGICAL ANALYSIS OF RTG ALONE AND THE
COMBINED RTG-MICELLE SYSTEM......................... 97
4.3 RELEASE OF TRIAMCINOLONE ACETONIDE FROM THE
COMBINED RTG-MICELLE SYSTEM WAS SUPERIOR TO THE
SAME FROM RTG OR MICELLES ALONE................ 99
5.1 PROPOSED DEGRADATION ROUTES FOR PSHU-NIPAAM
COPOLYMERS.....................................109
5.2 MASS LOSS OF PSHU-NIPAAM GELS IN PBS AND
CHOLESTEROL ESTERASE SOLUTIONS......................110
5.3 FT-IR SPECTRAL ANALYSIS OF PSHU-NIPAAM GELS IN PBS
AND CHOLESTEROL ESTERASE SOLUTIONS.............Ill
5.4 QUANTIFICATION OF FT-IR SPECTRA CONFIRM THAT
CARBOXYLIC ACIDS WERE NOT GENERATED AND
SIGNIFICANT AMIDE CLEAVAGE DID NOT OCCUR DURING
INCUBATION IN CHOLESTEROL ESTERASE SOLUTIONS...112
IX


5.5 MPA CONCENTRATION DICTATES HO-PNIPAAM-COOH
MOLECUALR WEIGHT AND INCREASING MOLECULAR
WEIGHTS DECREASE THE I.CST...........................114
5.6 CONJUGATION OF HO-PNIP AAM-COOH TO PSHU YIELDED A
COPOLYMER WITH AN LCST BELOW BODY TEMPERATURE 115
5.7 ACCELERATED DEGRADATION TESTING OF PSHU-NIPAAM
AND PSHU-NIPAAM-OH GELS IN HCL AND PAPAIN
SOLUTIONS.......................................116
6.1 HYDROLYTIC AND ENZYMATIC DEGRADATION PATHWAYS
OF CARBOXYL-DERIVED AMIDES......................126
6.2 HISTOLOGICAL SECTIONS OF A CRUSHED OPTIC NERVE 7
DAYS AFTER RGD-FUNCTIONALIZED PSHU-NIPAAM
INJECTION SHOWED OPTIC NERVE DAMAGE BUT NO
FUNCTIONAL RECOVERY..................................134
A. 1 PEG-PHS-PEG FULL SYNTHETIC ROUTE...............156
B. 1 STANDARD CURVE FOR DETERMINATION OF TA
CONCENTRATION VIA UV SPECTROSCOPY...............157
C. l PSHU-NIPAAM FULL SYNTHETIC ROUTE...............158
D. 1 RESULTS OF GPC ANALYSIS OF PSHU................159
E. 1 'H NMR SPECTRUM OF PSHU WITH INTEGRAL VALUES FOR
PEAKS USED TO CALCULATE MONOMER RATIOS..........160
G. 1 FIRST-ORDER KINETIC MODELING OF TA RELEASE FROM
PSHU-NIPAAM GELS................................162
H. 1 SELECTIVE BOC REMOVAL FROM PSHU WAS ACHIEVED BY
MILD ACID-CATALYZED HYDROLYSIS..................163
H.2 PSHU-NIPAAM GELS RELEASED THEIR ENTIRE
RAMBIZl V1AB LOAD WITHIN 24 HOURS...............164
x


ABBREVIATIONS
ACA 4,4'-AZOBIS(4-CYANOVALERIC ACID)
AMHP 2,2'-AZOBIS[2-METHYL-N-(2- HYDROXYETHYL)PROPIONAMIDE]
ANOVA ANALYSIS OF VARIANCE
AUC AREA UNDER THE CURVE
BCS BIOPHARMACEUTICS CLASSIFICATION SYSTEM
BSA BOVINE SERUM ALBUMIN
CMC CRITICAL MICELLE CONCENTRATION
CMT CRITICAL MICELLIZATION TEMPERATURE
DCM DICHLOROMETHANE
DLS DYNAMIC LIGHT SCATTERING
DMEM DULBECCO'S MODIFIED EAGLE'S MEDIUM
DMF DIMETHYLFORMAMIDE
DMSO DIMETHYL SULFOXIDE
EDC N-(3 -DIMETHYLAMINOPROP YL)-A'- ETHYLCARBODIIMIDE HYDROCHLORIDE
xi


FBSF ETAL BOVINE SERUM
FESEM FIELD EMISSION SCANNING ELECTRON MICROSCOPY
FT-IT FOURIER TRANSFORM INFRARED
GFAP GLIAL FIBRILLARY ACIDIC PROTEIN
GPC GEL PERMEATION CHROMATOGRAPHY
HDI HEXAMETHYLENE DIISOCYANATE
HPMC HYDROXYPROPYL-METHYLCELLULOSE
IOP INTRAOCULAR PRESSURE
ISO INTERNATIONAL ORGANIZATION FOR STANDARDIZATION
LOST LOWER CRITICAL SOLUTION TEMPERATURE
MC METHYLCELLULOSE
MPA 3 -MERCAPTOPROPIONIC ACID
MPEG METHOXYPOLYETHYLENE GLYCOL
MTT 3 -(4,5 -DIMETHYLTHIAZOL-2-YL)-2,5 DIPHENYLTETRAZOLIUM BROMIDE
MW MOLECULAR WEIGHT
Xll


MWCO MOLECULAR WEIGHT CUT-OFF
NHS Y-HYDROXYSUCCINIMIDE
NMP Y-METHYL-2-PYRROLIDONE
NMR NUCLEAR MAGNETIC RESONANCE
P/S PENICILLIN/STREPTOMYCIN
PAA POLY (ACRYLIC ACID)
PBS PHOSPHATE-BUFFERED SALINE
PCL POLY (e-C APROL ACTONE)
PEG POLYETHYLENE GLYCOL
PEO POLY (ETHYLENE OXIDE)
PHS POLY(HEXAMETHYLENE-HZr-SERINOL)
PI POLYDISPERSITY INDEX
PLA POLY (LACTIC ACID)
PLGA POLY(LACTIC-CO-GLYCOLIC ACID)
PNIPAAM POLY (7V-IS OPROP YLACRYLAMIDE)
PPO POLY (PROPYLENE OXIDE)
PSHU POLY(SERINOL HEXAMETHYLENE UREA)
Xlll


RI
RPE
RSD
RTG
TA
TEM
TFA
UV
VPT
REFRACTIVE INDEX
RETINAL PIGMENTED EPITHELIAL
RELATIVE STANDARD DEVIATION
REVERSE THERMAL GEL
TRIAMCINOLONE ACETONIDE
TRANSMISSION ELECTRON MICROSCOPY
TRIFLUOROACETIC ACID
ULTRAVIOLET
VOLUME PHASE TRANSITION
xiv


CHAPTERI
INTRODUCTION
The task of ophthalmic drug delivery has been described as one of the most interesting
and challenging endeavours facing the pharmaceutical scientist[l]. The challenge is
defined by a multitude of factors including: a) the unique physiological barriers present
within the eye; b) the sensitive nature of intraocular tissues, which are easily affected by
exposure to foreign molecules; and c) the unique anatomical structures of the eye which
make accessing intraocular targets relatively difficult. These challenges exist because of
and in spite of the apparent ease of external access to the eye. Because of this unique
feature, the use of eye drops (i.e. topical administration of ocular therapeutics) remains by
far the single most commonly-employed delivery modality. As will be explored herein,
this modality inadequately addresses the needs and wants of the patient.
To this challenge, researchers have responded with great innovation and creativity. Most
commonly, polymeric systemsboth established and novelhave been employed for
their relative ease of design, synthesis and characterization, their cost-effectiveness and
their superior flexibility. In this work, two such polymeric drug delivery systems have been
developed. The first, a reverse thermal gelling polymer, is an injectable, in situ-gelling
aqueous polymer that addresses the issues of access and target retention required of ocular
drug delivery systems. The second, a polymeric micelle system, is an advantageous nano-
carrier system that addresses the challenges of delivery of poorly-soluble drug molecules
and their long-term sustained release. The synergistic combination of these two systems
a reverse thermal gel containing drug-loaded polymeric micellesuniquely addresses the


design criteria required of a translatable ocular drug delivery system and has the potential
to significantly improve the treatment of several ocular diseases.
In this introduction, the challenges of ocular drug delivery will be presented followed by
the role of ocular drug delivery systems in addressing these needs. Finally, the principles
and mechanisms of reverse thermal gelling polymers and polymeric micelle systems will
be described in addition to the limitations of current systems. Within these contexts, the
solution proposed by the work undertaken herein will be explored.
Ocular Drug Delivery
Anatomical and physiological challenges
The eye can broadly be separated into the anterior and posterior segments. The anterior
segment is composed of the structures that lie anterior to the vitreous humor: the cornea,
iris, ciliary body and lens, while the posterior segment includes the vitreous humor, retina,
choroid and optic nerve. These structures are illustrated in Figure 1.1.
Pupil
Posterior chamber
Zonular
fibres
Cornea
Anterior chamber
(aqueous humour)
Optic nerve ------
ary muscle
Suspensory
ligament
Retinal
blood vessels
Figure 1.1: Diagrammatic representation of the eye with major structures labeled. [2]
2


Similarly, ocular diseases are generally classified as occurring primarily either in anterior
or posterior segment tissues. In practice, there is a major difference in how anterior and
posterior diseases are therapeutically addressed. These differences are based primarily on
the significant difference in bioavailability at anterior and posterior tissues, respectively,
of drugs administered topically. In order to appreciate this difference, an analysis of the
defining anatomical and physiological characteristics of anterior and posterior structures is
necessary.
The anterior segment, consisting of tissues that are directly exposed to environmental
conditions when the eye is open, has developed a multitude of barriers to ensure pollutants
and potential toxins are not able to penetrate the eye and reach intraocular tissues, where
significant damage could occur. At the foremost anterior surface of the eye is the tear film,
which is a mucin-containing liquid layer that forms a protective hydrophilic layer covering
the cornea. The tear film has evolved to rapidly trap and clear pollutants from the anterior
surface of the eye. As a result, its minimal volume of 7-10 pL is turned over at a rate of
approximately 10.5% per minute[3], thereby reducing the retention time of molecules on
the ocular surfaces to only several minutes. Further, irritation of the ocular surface (e.g. by
foreign objects or molecules that disrupt the tear film) serves as an impetus to further
increase lacrimation and thereby the tear film turnover rate.
Beyond the tear film lies the cornea, which serves as the first mechanochemical barrier to
ocular penetration of exogenous substances. It is primarily divided into three layers: the
epithelium, stroma and endothelium. The anterior-most layerthe epitheliumis a highly
cellularized hydrophobic membrane which contains 90% of the total cells within the
cornea. At the superficial-most layer, these cells are connected by desmosomes and
3


surrounded by tight-junction complexes. As a result particularly of these tight junctions,
pericellular penetration of exogenous molecules through the epithelium is significantly
retarded. Posterior to the corneal epithelium lies the stroma, which is a highly-hydrated
matrix of collagen fibrils. Comprising approximately 90% of the corneal thickness, the
hydrophilic stroma presents a major challenge for hydrophobic drugs to partition across.
The posterior-most layer of the cornea is the endothelium, which provides minimal
retardation of the passaging of substances. Because of its location at the interface of the
cornea and the aqueous humor, the corneal endothelium is deliberately leaky to allow
diffusion of macromolecules from the aqueous humor into the corneal stroma and vice
versa. Based on the alternating hydrophobic-hydrophilic-hydrophobic structure of the
cornea, it can be understood that a drug molecule must be amphipathic in order to be able
to partition across the entirety of the cornea.
During their transit across the cornea, drug molecules are also susceptible to degradation
and adsorption via enzymes and proteins present in that tissue. The cornea is especially
rich in esterases, which can hydrolyze susceptible drug molecules as they passage across
the tissue[4]. Specific esterases (identified via study of the rabbit cornea) include
cholinesterase, acetylcholinesterase, carboxylesterase, acetylesterase, arylesterase and a
non-specific esterase[5]. Many topical administration strategies have used the presence of
these esterases to their advantage, designing pro-drugs that are metabolized into their active
form upon hydrolysis. One example is latanoprost, which is administered as an isopropyl
ester that is converted to its biologically-active free acid form by endogenous enzymes.
Since the isopropyl ester form also has better corneal permeability than the free acid, this
pro-drug strategy improves intraocular bioavailability.
4


Molecules that can successfully partition across the multitude of corneal barriers will reach
the aqueous humor, from which they can access and be distributed to surrounding tissues
such as the iris, ciliary body, lens, vitreous humor and choroid/retina. However, several
mechanisms also exist to prevent uninhibited access to these tissues. Aqueous humor
turnover and blood circulation in the anterior uvea are both major factors that physically
remove drug molecules from the anterior chamber. Of a total aqueous humor volume of
300 pL in humans, half of this volume is turned over approximately every 0.77 hours[6],
In addition, it has been suggested that metabolic pathways contribute to drug clearance
from the aqueous humor, thereby accelerating clearance[6]. However, counter to these
mechanisms, drugs in the aqueous humor have the ability to reversibly bind to tissues,
which can help prolong their residence times. In particular, the lens, vitreous humor and
especially the pigmented uvea have been found to participate in this binding process.
Non-comeal drug absorption routes after topical administration can also be considered, e.g.
through the conjunctiva or sclera. The conjunctiva in humans has 17 times greater surface
area than the comea[7], theoretically presenting a major potential route for drug absorption.
However, drug penetration across the conjunctiva is severely limited. The primary
mechanism for molecule exclusion is the presence of tight junctions in the superficial
conjunctival epithelium, which act similarly to those in the comeal epithelium, but may be
considered slightly more leaky[8]. In addition, the presence of conjunctival capillaries and
lymphatic vessels efficiently clear molecules that do achieve some level of absorption.
The sclera, on the other hand, represents a major absorption pathway. Due to its importance
in maintaining the structural integrity of the globe, the sclera is a dense, disordered network
of collagen fibrils. Significant amounts of proteoglycans and polysaccharides are also
5


present, making the sclera a highly hydrated tissue similar to the corneal stroma. As a result,
scleral permeability of small molecule drugs is generally significantly higher than corneal
permeability, by a factor of 1.2-11 for several common molecules[9,10].
Drugs administered by the other major administration routessystemic or oral
administrationalso face a daunting challenge in reaching intraocular targets. Drugs
administered by these techniques must reach ocular tissues from the bloodstream in order
to affect treatment. In the anterior segment, the blood-aqueous barrier limits passage of
drugs, while the blood-retinal barrier accomplishes this in the posterior segment. The
blood-aqueous barrier consists of two cell layersthe endothelium of the blood vessels in
the iris and ciliary body and the non-pigmented ciliary epitheliumexpressing tight
junctions that efficiently prevent the entry of molecules into the aqueous humor[ll]. The
blood-retinal barrier is maintained by retinal capillary endothelial cells and retinal
pigmented epithelial (RPE) cells. The RPE cell layer is situated between the neural retina
and the choroid and is chiefly responsible for regulating selective transport of molecules
between photoreceptors in the retina and capillaries in the choroid. However, tight
junctions between RPE cells heavily restrict intercellular transport. The role of the RPE
cell layer cannot be understated in protecting the highly-sensitive underlying neural retina.
The highly-vascularized choroid contains fenestrated capillaries, which allow equilibration
of plasma drug concentrations with the choroidal extravascular space. The RPE cell layer
is then tasked with excluding molecules in this space from reaching the neural retina. The
previously-mentioned tight junctions efficiently serve this purpose and make this transport
mechanism almost entirely non-productive[12].
6


Conventional drug delivery modalities and their limitations
The net effect of these anatomical and physiological barriers to reaching the eye is that
traditional delivery techniques suffer from less-than-desirable distribution kinetics. For
example, in the case of topical administrationwhich accounts for 88% of the ocular
therapeutics marketless than 5% of the administered dose is able to reach the aqueous
humor[13]. The result is that the dose administered in each eye drop must be at least 20-
fold higher than the concentration needed at the aqueous humor. Much of that excess drug
is cleared through systemic absorption, especially by conjunctival vessels and the highly
vascularized nasal mucosa, where drugs can end up after clearance through the
nasolacrimal ducts[14]. These mechanisms so rapidly and efficiently clear topically
administered drugs that peak plasma concentrations are typically reached as soon as 5
minutes after topical instillation[15]. Systemic absorption has been implicated as a factor
in many observed side effects of topically administered drugs, especially when the patient
has underlying cardiovascular or respiratory diseases or is on other concomitant
medications[16-18].
The estimated 5% of the instilled dose that is available for intraocular distribution is true
only for anterior segment targets- the picture becomes even bleaker when posterior tissues
are concerned. A study in 1990 by Kyyronen and Urtti[19] investigated the distribution of
timolol maleatea highly soluble, small molecule beta-blockerin various intraocular
tissues after topical administration. Their findings are summarized in Table 1.1. They found
that as the drug partitioned further to posterior targets, the concentrations at those tissues
decreased by approximately one order of magnitude per barrier. Of the 125 pg/g instilled
dose, the concentration was 3 orders of magnitude lower in the vitreous humor (and
7


therefore available for posterior efficacy). For this reason, treatment of posterior ocular
diseases has especially suffered from the under-performance of topical administration.
Table 1.1: Intraocular concentrations of topically-administered timolol maleate[19].
Location Concentration (tig/g)
Instilled dose 125
Cornea 17.38
Aqueous humor 2.16
Lens 0.121
Sclera 2.66
Vitreous humor 0.083
Another major drawback of topical or systemic administration regimens is the pulsatile
nature of the drug concentration at the target tissue. When a dose of drug is administered
by either method, the drug follows the classic pharmacokinetic paradigm of absorption-
distribution-metabolism-excretion. Because doses are administered at discrete time points
at prescribed intervals, tissue drug concentrations typically fluctuate with time. This is
shown schematically in Figure 1.2. The reason this kinetic is less than ideal is that for each
drug and target tissue, a therapeutic window exists based on a lower bound defined by the
minimum effective concentration and a maximum bound defined by the maximum
tolerated dose. For many drug-dosage regimes, periods of time will be spent below and
above these bounds, respectively. Tissue drug concentrations above the maximum tolerated
dose are toxic to the tissue and result in many of the side effects attributed to specific drugs.
Those below the minimum effective concentration are not therapeutic and result in loss of
efficacy. Both scenarios result in less-than-ideal patient outcomes.
8


Pulsatile ----Controlled Release
Figure 1.2: Simulated kinetics of pulsatile and controlled-release drug administration
paradigms demonstrate the inconsistent tissue drug concentrations in pulsatile (e.g. topical
or systemic administration) techniques due to discrete time point dosing. In contrast, an
ideal controlled release system should be able to maintain consistent tissue drug
concentrations over a long period of time.
Perhaps the greatest challenge of topical or systemic administration regimens is the reliance
on patient adherence in determining therapeutic outcomes. Patients that are not compliant
with their dosing regimen have the potential to: a) suffer continuing vision loss due to lost
therapeutic efficacy[20,21]; and b) cause an unnecessary change by their physician in their
medical treatment plan as a result of perceived therapeutic inefficacy[22,23]. As a result of
the magnitude of these findings, scores of studies have evaluated all aspects of patient
compliance to identify specific problematic behaviors. One such study[23] found a
multitude of common problems including patients compressing doses during the day and
spacing them out at night, thus entirely missing doses and causing short- or long-term
interruptions in the medication. These findings suggest a combination of motivations
including convenience (taking more doses during the day and fewer at night) and
9


forgetfulness (missing doses), but improvements were noted when education addressing
the impact of compliance was provided to patients[24], However, a major challenge in
many ocular diseases is that they are largely asymptomatic until the later stages, when
significant irreversible damage has already set in. This disconnect between medication and
perceived symptoms may significantly contribute to non-compliance[25].
Studies quantitating patient dosing behavior have further confirmed the magnitude of the
challenge of non-compliance. One study found that during a four- to six-week period,
nearly 25% of patients had at least one day per month without any administered doses and
15% of patients took less than one half of prescribed doses[26]. Another found that over 3
months of prostaglandin analog treatment, the overall mean adherence rate was 71%, with
nearly 45% of patients administering drops less than 75% of the time[27]. The primary
obstacles reported in patient surveys were situational/environmental factors (e.g. being
away from home or changes to daily routines) and complexities of the medication
regimen[25]. As a result, significant efforts have been invested in developing systems that
can remove this burden from the patient and, thereby, improve medication efficacy and
therapeutic outcomes, as will be detailed in later sections.
Alternative delivery modalities
One strategy that continues to receive significant attention is administration of the drug (by
itself or encapsulated within a drug delivery system) directly at an intra- or periocular
target. This method represents the simplest and most direct means of bypassing the myriad
barriers preventing intraocular penetration and bioavailability of drug molecules. For
example, if the drug can be administered directly to the aqueous humor (e.g. by means of
a microneedle inserted through the cornea[28,29]), the drug will avoid all of the
10


permeability and clearance concerns of the cornea and gain immediate access to intraocular
targets. These alternative administration techniques are especially promising for improving
posterior drug bioavailability for the reasons previously described. The techniques that are
currently receiving the most attention are subconjunctival delivery, sub-Tenons capsule
delivery and intravitreal delivery. Pictoral representations of these routes are shown in
Figure 1.3.
Intravitreal
Figure 1.3: A diagrammatic representation of the location of sub-Tenons capsule, sub-
conjunctival and intravitreal injection routes.
Subconjunctival delivery of drugs or drug delivery systems bypasses both the cornea and
the conjunctiva, allowing improved drug levels both in the anterior space and the vitreous
humor[30,31], Improved posterior drug levels are likely related to transport by the
uveoscleral outflow pathway[32]. While subconjunctival delivery has proved promising
for drug depot formation, only drug delivery implants with relatively high release rates
11


were able to achieve and maintain therapeutic choroidal and retinal drug levels[33,34],
Amrite, et al.[35] found that neither nano- nor micro-particle formulations were able to
sustain detectable levels of a fluorescent marker in the posterior tissues following
subconjunctival administration.
The sub-Tenons capsule space extends posteriorly from the subconjunctival region and
occupies the potential space above the sclera and beneath Tenons capsule. The major
advantage of this space is that drugs or implants administered here are placed in direct
contact with the sclera, allowing for transscleral access to the choroid and retina.
Additionally, since the vitreous humor is not penetrated by this route, adverse effects
associated with intravitreal injections (e.g. retinal detachment and endophthalmitis) are far
less likely. One drawback of sub-Tenons injections is that pan-retinal drug levels have not
been demonstrated, and, in fact, the distribution of drug across the retina likely depends on
the location of injection and variations in scleral thickness[36]. For this reason, sub-
Tenons injections may be better suited for localized lesion treatment. Fernandez, et al.[37]
implanted several biomaterials within this space and found that while hydrophilic materials
were well tolerated, hydrophobic ones triggered inflammation and fibrosis, resulting in
complete fibrotic encapsulation within 12 weeks.
Intravitreal administration is the most common approach for achieving therapeutic
posterior drug levels. The injection is performed by a minimally-invasive intrusion through
the pars plana to reduce trauma. Ideally, use of a needle gauge of 25 or smaller is desired
to provide a self-healing wound (i.e. one that does not require suturing)[36]. Nevertheless,
the rate of complications can be fairly high for intravitreal injectionsespecially those
requiring repeated administrationincluding endophthalmitis, hemorrhage and retinal
12


detachment[38]. Because of the large volume of the vitreous humor, it is often used as a
space to form depots of drug for long-term treatment courses. This strategy is particularly
useful for poorly soluble drugs and those that can be crystallized as these will slowly
dissolve upon depot formation in the vitreous[39,40]. However, drugs that are not fully
solubilized can cause obstructions in the visual field, making them less favorable to
patients. Intravitreal injections have also been commonly employed for delivery of drug
delivery systems, including nanoparticles[41], microparticles[42,43], bulk polymeric
implants[44-46]and refillable devices anchored in the sclera[47,48], A major consideration
for intravitreal devices, however, is that drug diffusion rates are significantly lower in the
viscous vitreous humor resulting in less even drug distribution throughout the posterior
space[49,50].
Drug molecule considerations
The Biopharmaceutics Classification System (BCS) has been successfully employed for
decades in the evaluation of new therapeutic molecules. It differentiates drug molecules
based on two critical parameters: solubility and permeability (in particular, human
intestinal permeability). These parameters fundamentally control the rate and extent of
drug absorption and are also important determinants in the complexity of formulation and
delivery of the molecule[51]. Drugs are placed in one of four categories: high solubility
and high permeability (Class 1), low solubility and high permeability (Class 2), high
solubility and low permeability (Class 3) or low solubility and low permeability (Class 4).
As drug formulation progresses, Class 1 drugs are cherry-picked due to their ease of
development, the result being that few Class 1 molecules remain for development. Many
of the molecules undergoing development today are less soluble or less permeable than
13


ever before, necessitating the development of more sophisticated drug delivery methods.
In 2007, a meta-analysis of drug molecules and their BCS classifications was undertaken
[52], In particular, this study looked at drugs that were currently marketed at that time
versus those that were under investigation for marketing (new chemical entities, NCEs).
These data are summarized in Figure 1.4. This analysis highlighted a major shift in the
industry from a landscape dominated by Class 1 molecules to one dominated by Class 2
and Class 4 molecules. In short, this represented a move from generally soluble drug
molecules to poorly soluble ones.
Marketed Drugs New chemical Entities
Figure 1.4: A major shift was noted in the distribution of drugs within the BCS
classification system in 2007, with many more poorly soluble NCEs than marketed drugs.
One major consequence of this shift, especially as it relates to ocular therapeutics, is that it
complicates the picture of drug formulation and delivery. Because the vast majority of
ocular drugs are delivered topically, they must be able to be readily formulated as such.
Since topical formulations must be aqueous systems, solubility plays a critical role in this
process. This limits the delivery options for poorly soluble drugs, which, as discussed
earlier, are rapidly becoming a large part of the industry.
14


Efforts to improve ocular bioavailability of poorly soluble drugs have included many
strategies, but only ophthalmic suspensions have had a major impact on marketed drugs.
However, even suspensions suffer from several drawbacks that limit more extensive
clinical use. As described previously, patient compliance is a major challenge in
therapeutic success. Due to the additional requirement that suspension eye drop bottles be
shaken before dosing, they have been found to further reduce patient compliance[53]. In
addition, the actual gain in bioavailability from suspensions does not correlate with the
additional amount of drug present in those systems. For instance, the drug hydrocortisone
administered as a suspension with a drug content 33-fold higher than its solubility limit
only increased intraocular bioavailability of the drug 5-fold, as compared to a saturated
solution[54].
An alternative strategy to improve the intraocular bioavailability of poorly soluble drugs is
to form a water-soluble derivative of the drug of interest. This is most often accomplished
by forming a salt of the parent molecule[55]. The salt, having increased aqueous solubility,
now allows greater concentrations to be instilled by a topical drop. However, it must be
noted that this effect is at least partially counter-acted by the now lower permeability of
the drug. As was described earlier, the efficient corneal epithelium dictates that moderately
lipophilic molecules have the best corneal permeability. Salt formation reduces the
lipophilic nature of the molecule making it less permeable across the cornea. However, the
increased driving concentration is generally sufficient to overcome this loss in
permeability[56]. Of course, not all drug molecules are amenable to salt formation so this
strategy is limited in its applicability.
15


Drug Delivery Systems
A fundamentally different approach to this problem involves the application of drug
delivery systems. These systems are designed to address the various shortcomings of
traditional drug administration techniques and accomplish this task in significantly
different ways, as will be described in this section.
Motivations and considerations
As discussed in previous sections, the major limitations of topical and systemic ophthalmic
drug delivery are poor intraocular bioavailability, need for frequent re-administration,
inconsistent tissue drug concentrations and a strong dependence on patient compliance for
therapeutic efficacy. The motivation for advanced drug delivery systems stems from the
need to overcome these barriers and thereby reduce patient and provider burdens and
improve patient outcomes. The principle of controlled-release drug delivery is that a system
can be designed such that a reservoir of drug can be deployed near the tissue of interest
andin a controlled fashionrelease drug from the reservoir over a period of time.
Theoretically, this system would overcome the previously mentioned limitations in the
ways described in Table 1.2.
In order to achieve these goals, however, the system must be designed with several key
criteria, many of which are not met by commercially available drug delivery solutions. For
example, in order to improve intraocular bioavailability and bypass the complex
anatomical and physiological barriers described earlier, the system must be packaged in a
delivery system amenable to ophthalmic deployment. Given the compact and sensitive
16


structures of the eye and the limited number and size of potential spaces, this task is more
challenging in the eye than potentially any other tissue.
Table 1.2: Controlled-release drug delivery systems are theoretically able to overcome
many of the limitations of topical or systemic administration techniques
Limitation of topical/systemic administration Benefit of controlled release system
Poor intraocular bioavailability System can be deployed at or near the target tissue, overcoming physiological barriers
Need for frequent re-administration System can be loaded with sufficient drug to maintain therapeutic levels for months or years
Inconsistent tissue drug concentrations System can be tailored to achieve drug kinetics such that tissue drug levels are more consistent
Dependence on patient compliance System can sustain drug release for long periods of time without patient/provider action
Given these challenges, the biodegradability of the system also becomes a key criteria.
Since potential spaces are few in number and small in size, if a controlled release system
is deployed into one of these spaces, the system must be cleared from that space by the
time another system needs to be delivered. Take for example a system that is deployed into
the subconjunctival space of a glaucomatous patient to deliver a drug that reduces
intraocular pressure (IOP). Since glaucoma is a chronic disease, therapeutic levels of this
drug need to be sustained for the remainder of the patients life. If a controlled release
system can provide such therapeutic levels for a maximum of 6 months, a replacement
system would need to be administered at or near the 6-month time point. Since the sub-
conjunctival space is limited, the first system would need to be largely cleared from that
space in order to allow deployment of the replacement system. This requirement then
17


necessitates that biodegradation of the system be on a time scale similar to its maximum
time frame for drug release.
A successful controlled release system must also be able to precisely and reproducibly
control the kinetics of drug release from the system, hi order to allow a long usable life
span for the system, a large reservoir of drug is typically loaded into the system. The larger
this reservoir can be designed, the longer the system can theoretically provide therapeutic
levels of drug release. In the ophthalmic field, this time period should ideally be 3-6 months
or greater. However, with increasing drug loading, the potential for drug dumping or
localized drug levels above the maximum tolerated dose also increases. The kinetics of
drug release from (and biodegradation of) the system must be carefully controlled to
prevent these potentially harmful possibilities.
In order to understand the importance of release kinetics, a brief overview of the models
and terminology associated with the topic will be helpful. As drug molecules migrate from
their position at the interior of the system to the surface and then into the surrounding
medium, various factors can be involved in this process. The purpose of drug release
kinetic modeling is to gain insight into which factors are dominant on the time scale
relevant to drug release.
The first system that can be modeled is one in which the process of drug release from the
system is independent of the drug concentration, a zero-order kinetic. In this case, the
concentration, C, at time, /, can be expressed as:
C = C0 K0t
18


whereto is the zero-order rate constant and Co is the initial drug concentration. This kinetic
is generally not achievable for matrix systems where the drug is loaded throughout a
porous matrix, but is more common for reservoir systems where a bulk of drug is
surrounded by a rate-limiting membrane.
Matrix systems can more commonly fit a model in which the rate of the drug release
process is directly proportional to the drug concentration, a first-order kinetic. In this case,
the concentration with time is expressed as:
C = C0e~Klt
where Ki is the first-order rate constant. Note that in this case the process occurs as a
proportion of the drug concentration existing at that time (and therefore the initial drug
concentration) and so the release rate will decrease with time (as the remaining drug
concentration decreases) and will theoretically never reach completion.
The Higuchi model was derived specifically to describe drug release from a planar matrix
system and is based on the hypotheses that a) the initial drug concentration in the matrix is
much greater than drug solubility in the release medium; b) drug diffusion takes place only
in one direction; c) drug particles are much smaller than the systems thickness; d) matrix
swelling and dissolution are negligible; e) drug diffusivity is constant; and f) sink
conditions are always maintained in the release medium. The simplified equation
describing the Higuchi model is:
q = kh t0-5
where Q is the amount of drug released in time t and Kh is the Higuchi rate constant.
19


The final relevant model is the Korsmeyer-Peppas model, which is described by the
equation:

= Kkp tn
where Mt/Mco is the fraction of drug released at time t, Kkp is the Korsmeyer-Peppas rate
constant and n is the release exponent. This release exponent can be used to characterize
the release mechanism, where n = 0.5 indicates Fickian diffusion, 0.45 < n < 0.89 indicates
non-Fickian transport, n = 0.89 indicates Case II transport and n > 0.89 indicates Super
Case II transport. These relationships hold true for the first 60% of drug release.
Polymeric drug delivery systems
Within the field of drug delivery, polymer-based systems are by far the most commonly
investigated. Polymers are long-chain molecules that consist of a large number of small
repeating units[57]. By selecting appropriate repeating units and thereby controlling the
final chemistry of the molecule, polymers can be designed with a wide variety of
properties, as required by the specific application. This immense flexibility is the primary
reason polymers are so commonly employed as biomaterials in general and drug delivery
systems in particular.
For a polymer system to be amenable to application as a drug delivery system, it must meet
several critical design criteria. First and chief among these is biocompatibility, which
defines the interactions between the host tissue and the implanted material. This interaction
is two-way and involves both the effect of the material on the host tissue and that of the
host tissue on the material. Most importantly, the implanted material must not cause
20


significant toxicity, inflammation, infection, tumorigenesis or any other adverse effect in
the patient[58]. In addition, the drug delivery system must be biodegradable. This entails
both that the device should be cleared by biological mechanisms within a pre-defined
timeframe, but also that its degradation products meet all of the biocompatibility criteria
set forth earlier[59]. In a drug delivery system, the ideal situation is that biodegradation
occurs on a similar timescale to complete drug release. While a drug delivery system will
need to meet all of the additional criteria common to biomaterials in general[60], one
additional criterion is that any interactions between the drug delivery vehicle and the
encapsulated drug molecules do not alter the therapeutic efficacy of the drug. This is
especially crucial for structure-sensitive drugs such as antibodies and proteins, which can
be degraded rapidly in the presence of certain stimuli [61-64],
In sita-gelling injectable systems
One specific class of polymeric drug delivery systems that are especially useful as
ophthalmic drug delivery systems is the in situ-gelling injectable system. This type of
system is characterized by a transition from a low-viscosity liquid to a physically or
chemically cross-linked gel upon introduction of a specific stimulus. From a practical
standpoint, this system is advantageous in that its initial liquid state allows it to be deployed
by a minimally-invasive injection through a syringe immediately at the desired anatomical
location. Upon injection, the system is designed to deploy to its gel state allowing it to be
retained at the injection site, resisting dilution or washing away by local fluid flows.
Transition to a gel state also allows entrapment of incorporated drug molecules and control
over their release from the system. In this way, release kinetics similar to bulk polymer
systems can be attained but with the added benefit of injectability. In situ-gelling systems
21


commonly employed in drug delivery systems can be categorized into two major modes of
sol-gel behavior: in situ-gelling systems and in situ polymer precipitation.
In situ-gelling systems are generally achieved by injecting either: a) a pre-mixed monomer
and initiator mixture or solution that polymerizes upon injection; or b) polymers that have
reactive functionalities amenable to crosslinking with or without a cross-linking agent. The
impetus driving the polymerization or cross-linking reaction upon injection can be
provided by thermal energy (e.g. using body temperature or an external heat source), the
application of light (e.g. using an ultraviolet (UV) probe) or the presence of ions. An
advantage of these types of systems is that the cross-linking density can be relatively
precisely controlled by modifying the ratio of components at injection. Since highly-
crosslinked systems are theoretically possible, the kinetics of drug release can be favorably
controlled. In addition, the pre-crosslinked system can be composed of sufficiently low
molecular weight (MW) polymers to make a low viscosity solution amenable to injection.
An example of a temperature-driven system was developed by Moore, et al. [65] in which
an acrylate-terminated copolymer of lactide and e-caprolactone is mixed with a thermal
initiator (benzoyl peroxide or N, Af-di m ethyl -/Mol ui di ne) immediately prior to
administration. Upon heating to body temperature, a redox reaction drives polymerization
of the system, which entraps pre-mixed drug molecules. While release kinetics of
flurbiprofen from the system were favorable, the system suffers from several distinct
disadvantages. First is the toxicity concerns of direct injection of free radical-producing
species (i.e. the initiators), which have commonly been shown to be tumorigenic[66]. As a
result, initiator concentrations must be kept low, which consequently results in long
gelation times. In the previously mentioned system, full gelation was reported to take
22


between 5 and 30 minutes. During this time, encapsulated drug molecules can readily
diffuse out of the lightly crosslinked system resulting in a large burst of drug release within
this time. Finally, many of the polymerization and/or cross-linking reactions used in these
systems are highly exothermic in nature resulting in significant heating of surrounding
tissues. One such system reported local tissue temperatures as high as 94C[67], which can
cause significant tissue necrosis around the injection site[68].
Photo-polymerizable systems may overcome some of the disadvantages of thermal-
initiated systems[69,70]. The mechanism of injection and crosslinking is very similar
between the two systems, but the additional control provided by photo-polymerizable
systems makes them more practical for controlled release systems. The primary advantage
is that curing times can be much shorter than in thermal-initiated systems, allowing for full
gel formation within several seconds when lasers are employed for curing[71]. While this
eliminates the issue of high initial burst release, the rapid curing and extent of
polymerization has been reported to result in significant shrinkage and brittleness in the
material. In one application, a photo-polymerized polyethylene glycol-co-poly(lactic acid)
(PEG-PLA) hydrogel was investigated for controlled release of various therapeutic
proteins[72]. While drug release was favorable with minimal burst and diffusion-controlled
kinetics, complete drug release was realized within only 5 days for the molecules studied.
In addition, the concerns surrounding direct injection of reactive species remain in these
systems.
One final class of in ,v/7//-crossl inked systems are those driven by ion-mediated gelation.
These systems almost exclusively use the natural class of polymers called alginates, which
form a gel upon contact with divalent cations (e.g. calcium ions)[73]. Alginates are natural
23


polysaccharide polymers isolated from brown seaweed that can undergo selective ion
binding to form inotropic hydrogels. The ratio of mannuronate to guluronic acid determines
the physicochemical and swelling properties of the resulting gel, which also dictate drug
release kinetics from the system[74]. While most physiological tissues do not contain
sufficient calcium content to drive gelation of alginate systems, the higher calcium content
of lacrimal fluid permits gelation[73], A study by Cohen, et al. [75] found that an alginate
system could efficiently entrap pilocarpine and form a gel within the conjunctival cul de
sac. This system administered in rabbits was able to sustain lower IOP for up to 24 hours
after instillation as opposed to 6 hours for free pilocarpine. As such, the system could
reduce the frequency of administration from 4 times to twice a day. Despite these findings,
two major limitations have prevented more widespread application of alginate-based
systems. The first is the potential immunogenicity of the gels[76] and the second is the
long timeframe for in vivo biodegradation (no noticeable degradation was observed after 3
months when implanted subcutaneously in rats)[77].
Drug delivery systems based on in .v/7//-preci pi taring polymers have also received
significant attention in recent investigations[78]. These systems are founded on the ability
of certain polymers to be injected in solution form but then precipitate from that solution
upon a secondary mechanism such as: a) solvent removal; b) change in pH; or c) change
in temperature.
In ,v/7//-precipi taring polymers based on solvent removal employ a water insoluble polymer
that is dissolved in a water-miscible organic solvent. Upon injection, the solvent diffuses
into the surrounding medium and water diffuses into the system. Because the polymer is
insoluble in water, it precipitates once enough water has entered the system and forms a
24


physically-entangled polymer matrix. Because the organic solvent phase diffuses into
surrounding tissues, physiologically-compatible solvents must be chosen to prevent
cytotoxic responses. Examples of such solvents include Af-methyl-2-pyrrolidone (NMP),
propylene glycol and dimethyl sulfoxide (DMSO), among others[73], Because of the
solvent exchange process responsible for gelation, drug release kinetics and burst release
behavior are strongly dependent on the partition coefficient of the drug[79]. While long-
term drug release with minimal burst behavior can be achieved, careful selection of solvent,
polymer, polymer molecular weight, polymer concentration and excipients is required for
each drug molecule of interest[80]. Since high polymer starting concentrations are more
favorable[81], starting viscosities of these systems tend to be higher and require larger
needles for injection (e.g. 22G)[80], Additionally, solvent toxicity remains a major concern
in these systems. While solvents such as NMP and DMSO are generally considered non-
cytotoxic at lower concentrations[82,83], some studies have indicated mytotoxicity[84]
and ocular-specific low dose cytotoxicity[85], As a result, use of these solvents in in situ-
forming polymer implants will remain controversial.
Reverse Thermal Gels
Principles and mechanism of operation
As a way to circumvent solvent-related concerns, polymers that precipitate in situ by
temperature change alone have been heavily investigated[86]. In such systems, an
entropically-driven phase separation process is triggered as the system reaches a certain
temperature. The thermodynamic driving force for this phenomenon is the release of
structured, bound water from the polymer resulting from interactions between groups
present in the backbone. This process results in phase separation between the water and
25


polymer and formation of a physically-crosslinked polymer matrix. The temperature at
which this phase separation occurs is termed the lower critical solution temperature
(LCST). Since the process is driven only by a thermodynamic competition between
hydration of the polymer backbone (at T < LCST) and interactions between polymer
molecules (at T > LCST), the process is fully reversible and the solution state can be
recovered by cooling. For this reason, these systems are termed reverse thermal gels
(RTGs).
From a drug delivery perspective, the utility of an RTG as a controlled release system will
depend on the physicochemical properties of the drug molecule of interest. For hydrophilic
small molecules, the expulsion of water associated with gelation of RTG systems will result
in a concomitant expulsion of the water-soluble drug. Applications in which immediate
release of a bolus of drug are desired would benefit from this process. However, most
controlled release systems are intended for applications where an extended period of drug
release is desirable. In these cases, application of an RTG would favor delivery of
hydrophobic drugs. At the LCST, development of hydrophobic interactions between
polymer molecules drives collapse and dehydration of the system. Since two phases are
formedthe polymer phase, which is now overwhelmingly hydrophobic, and the aqueous
phasethe drug will be forced to partition between these phases based on its partition
coefficient. The more hydrophobic the drug, the larger the amount of drug that will
partition into the polymer matrix as it is formed. In addition, more hydrophobic drugs will
partition out of the hydrophobic polymer matrix at a slower rate, thereby increasing the
delivery time frame.
26


Gelation of RTG systems is a complex phenomenon that relies on the dynamics of several
processes including heat transfer into and through the RTG from the surrounding medium,
the ability of water to diffuse out of the system and into the surrounding medium and the
associated build-up of hydrostatic pressure within the gel and polymer diffusion and
relaxation processes associated with collapse of the system[86]. To date, no models exist
that comprehensively take into account all of the factors.
The major advantage of an RTG system over thermal- or photo-crosslinked systems is that
no reactive species are present, which significantly reduces toxicity concerns. In addition,
for many RTG systems, the temperature-induced phase separation process is sharp and
rapid, allowing complete gelation within several seconds after injection. This rapid gelation
prevents much of the burst release behavior observed in in .v/7//-crossl inked system.
RTG chemistries
Various polymers exhibit thermally-induced phase separation in aqueous solutions making
them suitable for application as RTGs. These can be categorized as: a) polymers with ether
groups including poly(ethylene oxide) (PEO), random copolymers of PEO and
polypropylene oxide) (PPO), PEO-PPO-PEO block copolymers (known as poloxamers or
Pluronics), poly(lactic-co-glycolic acid) (PLGA)-PEO-PLGA block copolymers, alkyl-
PEO block copolymers and poly(vinyl methyl ether); b) polymers with hydroxyl
functionalities including poly(hydropropyl acrylate), hydroxypropyl cellulose,
methylcellulose, hydroxypropyl methylcellulose and poly(vinyl alcohol) derivatives; c)
polymers with substituted amide groups including poly(V-substituted acrylamides) [such
as poly(Af-isopropyl acrylamide) (PNIPAAm)], polyfV-acryloyl pyrrolidine), poly(A-
27


acryloyl piperidine) and poly(acryl-L-amino acid amides); and d) other polymers including
poly(methacrylic acid).
Block copolymer-based chemistries are perhaps the most commonly employed for drug
delivery applications due to their ease of synthesis and flexibility of the block chemistry.
The general principle behind these polymers is to achieve a balance between hydrophobic
and hydrophilic segments within the polymer. At lower temperatures (e.g. room
temperature), the polymers are sufficiently hydrophilic to allow solvation. With increasing
temperature (e.g. near body temperature), hydrophobic domains aggregate to minimize the
hydrophobic surface area contacting the aqueous medium, which reduces the amount of
structured water surrounding the hydrophobic domains and maximizes the solvent entropy.
As a result, a stable physical gel can be formed. This process is illustrated diagrammatically
in Figure 1.5.
One of the most common classes of block copolymer-based RTG chemistries is the
poloxamer, which is constructed from a hydrophobic PPO block grafted at both ends with
hydrophilic PEO blocks (PEO-PPO-PEO)[87-90]. Because of their commercial
availability and favorable thermal gelling properties, poloxamer-based systems have been
investigated for myriad tissue engineering[91,92] and drug delivery[93-95] applications.
Most poloxamer systems actually exhibit cloud point temperatures well above body
temperature, but can be made to form physical gels by injecting high concentrations of the
polymers (e.g. 20 wt%). The dynamic underlying this principles is that poloxamer-based
solutions at lower concentrations form micelle nano-structures, and there is not sufficient
polymer present for these nano-structures to coalesce into a cohesive physical gel [96], Only
at higher polymer concentrations is the polymer content per unit volume sufficient to form
28


long-scale polymer structures amenable to complete gelation. As a result, injections require
larger gauge needles and may be hyperosmolar to the surrounding tissue at
administration[73], both of which consequences are detrimental to their clinical utility.
Figure 1.5: Block copolymers containing hydrophobic (red) and hydrophilic (blue)
segments can undergo a reversible solution-to-gel phase transition at elevated temperature.
The gel state is characterized by formation of hydrophobic domains (dashed circles) within
the physical matrix, which can entrap hydrophobic drug molecules.
Nevertheless, several studies have shown the potential for poloxamer-based drug delivery
systems. Veyries, et al.[97] demonstrated that a 25 wt% poloxamer solution could prolong
the residence time of vancomycin for administration in high infection risk areas (e.g. post-
surgical). In vivo, high concentrations were maintained near the site of injection for 24
hours and therapeutic concentrations were maintained for 8 days. Further, the poloxamer
system did not alter vancomycin activity and was well tolerated after injection. Another
study by Miyazaki, et al.[98] evaluated a poloxamer system for sustained anti-turnoral
activity using mitomycin C. Again using a 25 wt% system, they found that tumor-bearing
mice had longer life spans after treatment with the drug-loaded poloxamer system as
compared to free drug. Most importantly, controlled release properties of the gel system
allowed administration of drug concentrations that would be toxic on their own. This
property is of great value in anti-cancer therapies.
29


Despite these positive findings, further application of poloxamer-based systems has been
hindered by several critical drawbacks, most of which are driven by the need to employ
high concentration solutions to achieve body-temperature gelation. One physiologically-
observed effect is that intraperitoneal injections of poloxamer systems resulted in sustained
hypertriglyceridemia and hypercholesterolemia[99]. Both effects were found to be related
to stimulation of 3-hydroxy-3-methylglutaryl-co-enzyme A reductase activity in the liver,
presumably during metabolic clearance of the polymer[100]. Another troublesome clinical
finding for poloxamer-based systems is that certain formulations inhibit P-glycoprotein,
resulting in increased cellular uptake of certain drugs and thereby increasing their
cytotoxicity[101-103]. While the exact mechanism of this effect has not been discovered,
it has been proposed that it may be related to the ability of poloxamer molecules to
incorporate into cell membranes and induce changes in plasma membrane order,
morphology and mechanical stability[ 104,105],
In order to circumvent these clinical challenges, other block copolymer chemistries have
also been investigated. Another relatively common formulation is that comprised of the
common building blocks PLGA and PEG, synthesized as PEG-PLGA-PEG or PLGA-
PEG-PLGA. Jeong, et al. [106] investigated a PEG-PLGA-PEG system for in vitro release
kinetics of both a hydrophilic (ketoprofen) and hydrophobic (spironolactone) model drug.
Because of the formation of hydrophobic domains within the physical gel and the ability
of these domains to control release of hydrophobic molecules, release kinetics were more
sustained for the hydrophobic drug at 55 days to complete drug release compared to 6-10
days for the hydrophilic drug. While the PLGA block permits complete biodegradation of
these systems within 3-4 months[107], like the poloxamer-based systems, high polymer
30


concentrations (greater than 20 wt%) are required for adequate gelation and drug loading
levels beyond 0.25 wt% are generally not reported.
PLGA-PEG-PLGA is a commercially-available thermal gelling polymer sold under the
name ReGel. As would be expected, its physical properties are very similar to those of
PEG-PLGA-PEG systems, but its increased overall hydrophobicity makes it better suited
for controlled release of hydrophobic drugs. Zentner, et al.[108] demonstrated that the
system could sustain in vitro release of paclitaxel for 50 days with complete in vitro
degradation of the system within approximately 10 weeks. The amphiphilic nature of the
copolymer also helped solubilize the hydrophobic drugs cyclosporine A and paclitaxel by
400 to 2000-fold. However, the ReGel system was not an advantageous carrier of
hydrophilic molecules. Complete in vitro release of several model proteins was observed
within 10-15 days for all of the molecules studied. Additionally, as with previously
described systems, high polymer concentrations (greater than 20 wt%) were still needed
for gelation, limiting the delivery options of the system.
PNIPAAm-based RTGs
In order to circumvent the challenges associated with block copolymer systemsas
described in the previous paragraphsmany groups have turned to specific thermo-
sensitive homopolymers. While many examples were given at the beginning of this section,
none have received more attention than PNIPAAm[109]. In its simplest homopolymer
form, PNIPAAm is a water-soluble polymer that exhibits a sharp, reversible LCST near
32C[110], positioning its sol-gel transition point comfortably above room temperature and
sufficiently below body temperature. While many polymers possess LCSTseven in this
temperature rangePNIPAAm is unique in its sharp, almost discontinuous,
31


transition[ 111]. This property makes it particularly well-suited for in situ gelling
applications, where rapid gel formation is desirable.
The LCST behavior of PNIPAAm is generally attributed to an entropically-driven coil-to-
globule transition[l 12-115], At temperatures below the LCST, PNIPAAm molecules exist
in a solvated state in a flexible and expanded coil conformation due to an abundance of
hydrogen bonding between amide groups and surrounding water molecules. With
increasing temperature, these coils collapse to a globule state due to increasing
intramolecular hydrogen bonding between amide groups and hydrophobic interactions
between isopropyl groups. From an enthalpic point of view, this transition process is
endothermic, indicating the polar amide groups prefer hydrogen bonding with water to
existing at the non-polar interior of the collapsed globule. From an entropic point of view,
this transition decreases the conformational degrees of freedom possessed by the polymer
molecules, but increases the degrees of freedom for the water molecules that were
previously hydrogen-bonded to the amide groups. Additionally, the space occupied by the
polymer molecules is decreased significantly in the globule state, which decreases the
excluded volume and further increases the overall configurational energy of water
molecules in the system. Therefore, the coil-to-globule transition is driven by: a)
dominance of the enthalpic term at low temperatures, which stabilizes the expanded coil
state of PNIPAAm molecules through hydrogen bonding between amide groups and water
molecules; and b) a transition to dominance by entropy terms at higher temperature, driven
by the increased configurational entropy of water molecules, which stabilizes the compact
globule state.
32


Practically, this process results in several critical changes in an aqueous PNIPAAm
solution that permit facile characterization and application. The first is that in dilute
aqueous solutions, phase separation at the LCST causes significant turbidity, allowing
LCST measurement by monitoring the change in optical density of a PNIPAAm solution
with increasing temperature. The second is that once in the collapsed globule state, globules
can aggregate to further increase the entropy of free water, thereby forming a physical gel.
Thermodynamically, this aggregation process will occur either slowly at low temperatures
or in dilute solutions, or quickly at high temperatures or in concentrated solutions. The
temperature dependence is driven by the increased driving force to maximize the
configurational entropy of water molecules at higher temperatures and the concentration
dependence is driven by the mean distance between globules, which must be overcome in
order to permit aggregation.
Because the coil-to-globule transition is driven by a loss of structured water previously
bound to PNIPAAm molecules, which allows formation of a more compact globule state,
significant water is liberated from the PNIPAAm phase during this transition. This volume-
based transition from expanded coil to compacted globule is termed the volume-phase
transition and has been thoroughly characterized in previous studies[l 11], Of particular
note is the estimation by Shibayama, et al.[l 16] that 13 water molecules are liberated from
the PNIPAAm phase per monomer unit. The net effect of this phenomenon is that during
gelation of PNIPAAm-based systems, a significant amount of water is expelled from the
system resulting in a largely dehydrated physical gel that is predominantly hydrophobic in
nature[112].
33


Figure 1.6: During the sol-gel transition of PNIPAAm-based RTGs, a hydrophobic
physical gel is formed concomitant with the expulsion of significant amounts of water. For
a hydrophobic drug, most of the drug in the system will partition into the physical gel
during this transition. However, for a hydrophilic molecule, most of the drug will be
expelled along with the water resulting in low loading and a large burst of drug release.
From a drug delivery perspective, the combined effect of the expulsion of water content
and the formation of a hydrophobic gel core presents a favorable environment for
entrapment and controlled release of hydrophobic drugs and an unfavorable one for the
same of hydrophilic drugs, as illustrated in Figure 1.6. For a hydrophobic drug molecule,
the expulsion of water and hydrophobic interactions driving physical gel formation will
encourage drug molecules to be entrapped within the gel. In addition, after gel formation,
the drug will be released slowly into the surrounding aqueous medium as it partitions out
34


of the hydrophobic gel core. However, for a hydrophilic molecule, the drug will follow
water molecules out of the polymer phase of the system during the sol-gel transition,
resulting in poor entrapment efficiencies and a large burst of drug release concomitant with
the transition.
This theoretical difference in drug release capabilities has also been borne out in the
literature. For example, Zhang, et al.[117] explored a PNIPAAm-based system for
controlled release of the water-soluble protein bovine serum albumin (BSA). Even though
BSA is a large molecule (MW: 69 kDa) model compound, the group found that 20-50% of
the drug load was released within the first several hours, presumably due to the volume
phase transition associated with PNIPAAm systems. Additionally, Han, et al.[118]
investigated release of 5-fluorouracil from PNIPAAm-chitosan hydrogels and also noted a
large 40-60% burst within the first hour and complete release of the drug payload within
24 hours. In general, release kinetics of hydrophilic drugs from PNIPAAm-based systems
exhibit a large burst release and complete drug release within 24-48 hours. Systems
intended for controlled release of hydrophobic drugs have generally fared better. Wilson,
et al.[l 19] studied release of the hydrophobic small molecule drug colchicine from a
PNIPAAm-based thermal gel and found the system could sustain therapeutic drug levels
for 14 days. Hoare, et al.[120] studied uptake and release behaviors of several different
drug molecules and found that more hydrophobic drugs more rapidly and more completely
partitioned into the hydrophobic cores of PNIPAAm-based microgels and that release rates
for these hydrophobic drugs were slower in vitro. These results indicate that PNIPAAm-
based drug delivery systems are well suited for encapsulation and controlled delivery of
hydrophobic drugs.
35


PNIPAAm-based RTGs also have the distinct advantage over block copolymer-based
systems of de-coupling polymer hydrophobicity and thermal gelling characteristics.
Because block copolymer systems rely on a specific hydrophobic-hydrophilic balance to
achieve thermal gelling, introduction of particularly hydrophilic or hydrophobic
moleculesespecially covalently tethered to the polymer itselfstand to compromise
thermal gelling characteristics. For example, Xun, et al. [ 121 ] synthesized a thermo-
sensitive block copolymer hydrogel using a poly[(s-caprolactone)-co-lactide]-PEG-
pol y [(s-caprol acton e)-co-l acti de] (PCLA-PEG-PCLA) triblock copolymer.
Functionalization with the peptide KRGDKK was found to shift the rheological sol-gel
transition temperature down by 5C and also significantly altered drug release kinetics
from the system. While adjustments could be made to the polymer chemistry to take these
changes into account, the magnitude of this effect will be different depending on a large
number of parameters including: functionalized molecule, functionalization density,
loaded drug molecule physicochemical properties, drug loading fraction, etc. This would
result in a resource-consuming optimization process and an inflexible system. Since
PNIPAAm-based RTG chemistries are largely unaffected by these parameters, they are
considered more flexible systems that can be employed as platform technologies.
In AitM-Gelling Systems for Ophthalmic Drug Delivery
Motivations
In situ-gelling systems are uniquely capable of overcoming challenging delivery scenarios
to provide a minimally-invasive deployment strategy for some of the most complex
anatomical challenges. Perhaps nowhere is this of greater importance than in the field of
ophthalmology, where minimally-invasive deployment, harmonization with the host-tissue
36


anatomical space and non-reactivity with highly sensitive cells and tissues are of critical
concern. For these reasons, RTGs present an especially advantageous option for
ophthalmic drug delivery.
First, the ability of RTGs to be injected through small-gauge needles make them more
acceptable to patients and providers as it minimizes patient discomfort and reduces the
complexity of the procedure, leading to lower risk of complications and better patient
outcomes. This will be especially true for RTG chemistries that are low viscosity solutions
at room temperature and can be injected through 30 or 32G needles, as are commonly used
in ophthalmic procedures (e.g. intravitreal injections). This requirement precludes most
alginate and block copolymer systems as they require larger needles for injection. While
larger needles (up to 22G) can be used for intravitreal injections without the need for
suturing, such systems have not been as well-accepted clinically. For example, the Ozurdex
system was an extruded PLGA matrix containing dexamethasone that was injected into the
vitreous humor for long-term treatment of macular edema (secondary to branch retinal vein
occlusion or central vein occlusion) and uveitis. Because the device was fashioned as a
bulk polymer rod, it required an applicator with a 22G diameter. Many papers associate the
large needle size to reported side effects including endophthalmitis, intraocular
inflammation, increased IOP, retinal detachment and conjunctival hemorrhage[122], In
addition, reports of complications were found to be higher in patients that needed repeated
implantations[123]. In order to avoid the potential for such complications, an RTGthat can
be delivered using a much smaller gauge needle would be of clinical interest.
Another advantageous characteristic of RTGsand of many in situ-gelling polymersis
their ability to conform to the anatomical confines of the delivery space. As described in
37


previous sections, potential spaces within the eye are relatively small (e.g. 100 pL for the
subconjunctival space) and abnormally shaped. In addition, devices that are too large or
obtrusive to the ocular anatomy canat the leastcause an increase in IOP. As a result,
polymeric bulk implants can be problematic in that they are of a fixed geometry and cannot
adapt after implantation. In contrast, because RTGs are injected initially as a solution and
take several seconds or more to fully gel, they can conform to the anatomical space into
which they are injected before gelation, allowing a more anatomically-appropriate fit.
Examples
A number of previous works have investigated in situ-gelling polymer systems for
ophthalmic drug delivery. These systems have employed naturally-derived and synthetic
polymers and have included ion-, pH- and thermally-activated gelation mechanisms. A
brief review of some of the more notable systems is included to highlight their successes
and failures, and where room exists for improvement in the field.
Naturally-derived cellulose derivatives such as methylcellulose (MC) and hydroxypropyl-
methylcellulose (HPMC) exhibit phase transition temperatures above body temperature (>
40C); however, in the presence of various salts this can be reduced to below body
temperature[124], Bhowmik, et al.[124] employed this strategy to develop thermo-
reversible gels for ophthalmic delivery of ketorolac. They found that increasing salt
concentrations lowered the sol-gel phase transition temperature and also increased the
longevity of drug release from their system, from 180 to 240 minutes. This increased drug
release time frame was attributed to an increase in the viscosity of the system resulting
from the addition of salts. However, practical application of this systems would
undoubtedly be very challenging due to the requirement that 6-8 wt% salts be added to the
38


system to achieve the desired properties. This makes the system severely hyperosmolar,
which could pose a serious detriment to the tissues at the site of administration[78].
In order to overcome this problem, several groups investigated MC or HPMC in
combination with poloxamers to both improve the gelation kinetics of MC/HPMC and the
sustained release properties of poloxamer-based systems. El-Kamel[125] found that a
solution containing 15 wt% poloxamer (F127) and 3% MC gave the slowest release
kinetics of timolol maleate and, applied topically, increased the intraocular bioavailability
of the drug 2.4 fold in rabbits, presumably due to the increased retention time accorded by
the higher viscosity as compared to free drug solutions. Paavola, et al.[126] had similar
results with poloxamer-HPMC and poloxamer-MC composites for sustained release of
lidocaine HC1 and ibuprofen. However, release time frames beyond 2-3 days could not be
achieved with any of the systems, making them applicable only to short-term delivery
applications. Desai, et al.[127] found that an FI27 poloxamer mixed with MC, HPMC or
PEG could significantly reduce the release rates of pilocarpine as compared to poloxamer
alone. Gelation kinetics were also improved, which were attributed to an increase in the
mixed systems viscosity. However, release time frames for this system were also limited,
achieving only several hours of sustained release. Thrimawithana, et al.[128] investigated
thermal gelling dispersions of MC and carrageenan for transscleral delivery of
macromolecules. While the system could transition from a dispersion to a gel below body
temperature, rheological properties of the system were complex, exhibiting several
transitions at various temperature. Sub-Tenons injections of the system loaded with a
Cx43 antisense oligonucleotide led to a significant reduction in Cx43 levels in the choroid
39


of rats 24 hours after administration. The longevity of this effect was not evaluated,
however.
In the realm of synthetic polymers, poloxamers have received a great deal of attention for
ocular drug delivery, as in other areas, due primarily to their commercial availability and
well-characterized history. Gupta, et al.[129] found that a poloxamer-based temperature
and pH-sensitive (through addition of chitosan) gel could enhance delivery of timolol
maleate and improve intraocular bioavailability. Srividya, et al.[130] found that a pH-
sensitive thermal gel based on poly(acrylic acid) (PAA) and HMPC could sustain delivery
of ofloxacin for up to 8 hours. Ma, et al. [ 131 ] employed a poloxamer-PAA copolymer to
sustain release of gatifloxacin for up to 12 hours with favorable gelation kinetics. For
posterior applications, Lee, et al.[132] investigated the F127 poloxamer for transscleral
sustained delivery of dexamethasone. In vitro studies determined that the system could
provide up to 48 hours of delivery with good scleral permeation.
Other block copolymers, especially those of PLGA and PEG, have also received attention
in this field. Duvvuri, et al. [ 133] explored PLGA-PEG-PLGA thermal gels for delivery of
ganciclovir and found the system could maintain therapeutic release rates for
approximately 10 days. Similarly, after intravitreal injections in rabbits, maintenance of
vitreal drug concentrations was achieved for 14 days after injection[134], A different block
copolymer chemistry, composed of PEO-PCL-PEO was developed by Wang, et al. [ 135]
for intravitreal injection and delivery of the antibody bevacizumab. The polymer exhibited
excellent biocompatibility both in vitro and after intravitreal injection in rabbits and
sustained in vitro delivery of the large molecule for 11 days in vitro. Similarly, Park, et
al.[136] developed a PEGylated polyurethane block copolymer for intravitreal delivery of
40


bevacizumab and found excellent in vitro and in vivo biocompatibility of the system, which
could sustain release of the antibody for 17 weeks from a 20 wt% gel.
Shortfalls of existing systems
A major limitation of the above-described systems is that most were evaluated as a direct
replacement for eye drops. As such, delivery timeframes were severely limited (to several
hours) and the systems would be expected to suffer the same limitation of limited posterior
bioavailability. None of the afore-mentioned studies investigated the impact of their
systems on posterior drug levels. While some excellent systems were described that were
designed specifically for intra- or periocular administration and could provide therapeutic
posterior drug levels, release time frames of these systems were still fairly limited, with the
longest system achieving 17 weeks of release. This falls short of the currently-marketed
Ozurdex system, which can provide 4-6 months of dexamethasone delivery.
In order to achieve clinical relevance, the delivery time frame of RTGs intended for
posterior delivery of ophthalmic drugs must be extended. As has been elucidated in
previous sections, this problem has plagued nearly all RTGs that have been investigated,
even outside the ophthalmic realm. What is clear is that hydrogel-based systems (such as
poloxamers) have an especially poor ability to sustain drug release due to their highly
hydrated state even after gelation, which permits relatively free diffusion of drug molecules
out of the RTG matrix. PNIPAAm systems have generally fared better in terms of release
kinetics due to their formation of a dense, hydrophobic matrix, but their application to intra-
or peri-ocular drug delivery scenarios could not be found in the literature. Nevertheless,
even PNIPAAm-based RTGs are limited to 1-2 months of delivery in most reported
studies. As such, a novel approach is necessitated.
41


Micelles as a secondary delivery vehicle
One such approach is the incorporation of a nano-carrier system within the reverse thermal
gel. Nano-carrier systems as a class of drug delivery vehicles typically include
nanoparticles, microparticles, micelles and liposomes. These are illustrated in Figure 1.7.
Nanoparticles and microparticles are bulk particles composed of non-water soluble
polymers compacted into a particle form. Drugs can be encapsulated within the particle
during fabrication, which is typically realized through an oil-in-water emulsification
process[137]. Sustained drug release is achieved due to the time taken either: a) for the
drug to partition out of the hydrophobic particle into the surrounding medium; or b) for the
particles polymer matrix to degrade to a sufficient extent to allow the drug to diffuse out
of it. Their application in ophthalmic drug delivery is vast[138-140], but will not be
extensively covered herein. Similarly, liposomes have been applied for several
ophthalmology-specific applications[12,141,142]. Liposomes are artificially-fabricated
vesicles composed of a lipid bilayer encapsulating an aqueous solution. They can be loaded
with therapeutic molecules either in the interior of the particle (typically for hydrophilic
molecules) or within the hydrophobic segment of the lipid bilayer (for hydrophobic drug
molecules). Liposomes are especially well-suited for targeted drug delivery as they can be
designed to release their contents only after fusion with a cell membrane (which can be
mediated by cell-specific antibodies/receptors) or to permit diffusion of molecules out of
the liposome only after cellular internalization.
Micelles possess some character similar to both nanoparticles and liposomes. Like
liposomes, they are typically self-assembled from amphiphilic molecules. However, their
final morphology is composed of a hydrophobic core (much like a nanoparticle)
42


surrounded by hydrophilic segments, which make up the exterior of the particle (commonly
called the corona). Most micelle systems will self-aggregate into this structure in an
aqueous environment when the correct conditions (typically of temperature and
polymer/surf actant concentration) are met. The minimum concentration required for self-
aggregation is termed the critical micelle concentration (CMC) and the minimum
temperature required is called the critical micellization temperature (CMT)[143],
Encapsulated drug molecule
Figure 1.7: Nano-carrier systems typically employed for drug delivery applications
include nanoparticles composed of bulk polymer (left), liposomes composed of a lipid
bilayer surrounding an aqueous solution (center) and micelles with a hydrophilic shell and
hydrophobic interior (right). Typical modes of drug encapsulation are illustrated.
For a molecule to be amenable to this micellization process, it must possess both
hydrophilic and hydrophobic character. For this reason, the two main categories of
molecules that are used to form micelles are surfactants and amphiphilic copolymers, with
copolymers being the most common choice. Amphiphilic block copolymers are typically
synthesized either as A-B or A-B-A block copolymers, where A is a hydrophilic polymer
and B is a hydrophobic polymer. Upon micellization, the hydrophobic B segments
aggregate to form the core of the micelle and the hydrophilic A segments form the corona,
which stabilizes the core-shell structure. The thermodynamic driving force behind
micellization of block copolymers is the decrease of free energy in the overall system due
to removal of hydrophobic polymer segments from the surrounding aqueous medium,
allowing primarily hydrophilic segments to interface with water molecules[144]. Because
43


micellization is driven by aggregation of hydrophobic polymer segments and the final
micelle structure is composed of a highly hydrophobic core, micelles are particularly
advantageous carriers for poorly soluble drug molecules. This property gives micelles a
distinct advantage in: a) solubilization of poorly-soluble drug molecules thereby improving
their bioavailability; b) reducing toxicity and other adverse effects due to high drug loading
only within the micelle core; and c) substantially altering drug permeability across
physiological barriers[145].
Sustained drug release kinetics are also generally more favorable from a micelle system.
The rate of drug release from a micelle will depend on several factors including the rate of
diffusion of the drug out of the micelle core (this will also be influenced by the partition
coefficient of the encapsulated drug molecule), micelle stability and the rate of degradation
of the polymer comprising the micelle. If the micelles are stable and the rate of degradation
is slow, the release rate will be largely determined by the drugs affinity for the micelle
core, which can be driven by chemical, physical or electrostatic interactions. This suits
micelles well for sustained release of hydrophobic drug molecules as their hydrophobic
core will result in relatively slow drug partitioning into the surrounding aqueous medium.
However, these kinetics also imply thatbeyond micelle chemistry alonedrug release
behavior will also be dependent on the partition coefficient of the drug itself.
Micelle fabrication and concomitant or subsequent drug loading are realized by one of
several methods including dialysis, oil-in-water emulsification and film-based methods.
The dialysis method consists of dissolving the polymer and drug in a suitable organic
solvent (one that is water-miscible and solubilizes both the polymer and drug) and
dialyzing this solution against water. As the solvent diffuses out of the dialysis chamber
44


and is replaced by water, the polymer will aggregate to form micelles and a fraction of the
drug will be encapsulated. This method is perhaps the most simple of those available, but
is commercially impractical due to the length of time required to ensure complete removal
of the organic solvent (generally several days)[146]. The oil-in-water emulsification
technique is similar to that used to produce nanoparticles and comprises adding an organic
phase consisting of the drug dissolved in a volatile solvent to an aqueous phase consisting
of the polymer dissolved in water followed by sonication and evaporation of the volatile
organic solvent. This method is rapid and simple, but generally yields low drug
encapsulation efficiency, broad micelle polydispersity and significant levels of residual
solvent, which are often toxic[146]. Film-based methodswhile subject to various
modificationsgenerally begin with dissolution of the polymer and drug in a volatile
organic solvent followed by deposition of this solution onto a surface, evaporation of the
solvent to produce a polymer-drug film and rehydration of the film in water. Because
solvents that are generally considered to be less cytotoxic (e.g. DMSO or NMP) are non-
volatile, they cannot be used with this procedure and so residual solvent levels remain a
major concern.
Micelles have been fairly extensively investigated for purposes of ophthalmic drug
delivery. Due to the reasons discussed previously, they have most commonly been
employed for delivery of solubility-limited molecules and those with poor ocular
permeability. For example, cyclosporin A (CsA) is challenging to formulate for topical
ophthalmic administration due to poor solubility and permeability. Di Tommaso, et
al. [147] developed PEG-hexyl substituted poly(lactide) di-block copolymer-based micelles
for topical administration of CsA. These micelles showed excellent CsA encapsulation and
45


ocular compatibility and significantly increased intraocular bioavailability of the drug as
compared to topical eye drops. Li, et al.[148] investigated a micelle system for improved
solubilization and intraocular bioavailability of the poorly-soluble drug diclofenac. The
micelle chemistry employed was a di-block copolymer consisting of PEG and PCL. The
micelles were ocular-compatible and increased penetration of the drug across rabbit cornea
by 17-fold compared to free drug and increased the area under the curve (AUC) of the drug
by 2-fold, also compared to free drug. Gupta, et al. [149] developed a crosslinked micelle
based on a copolymer of NIPAAm, acrylic acid and A'-vinyl pyrrol i done crosslinked by
Af,Af-methylene /v'.s-acrylamide. The produced micelles were both pH- and temperature-
sensitive and could sustain release of the poorly-soluble drug ketorolac for 10 hours with
a 2-fold increase in trans-comeal permeation ex vivo. Civiale, et al.[150] developed
polyhydroxyethylaspartamide-based micelles and demonstrated a 2-fold increase in
intraocular bioavailability of dexamethasone in vitro. Finally, Lin, et al. [ 151 ] evaluated
poloxamer-coated chitosan micelles for improved topical penetration of metipranolol.
They demonstrated a prolonged reduction of IOP in rabbits after topical administration of
the micelles as compared to standard eye drops.
Because micelles are well-suited to solubilization and sustained release of poorly-soluble
drugs, they present an excellent system to supplement the release capabilities of an RTG.
By encapsulating pre-formed micelles (pre-loaded with drug) within an RTG, the micelles
can now provide the drug release-limiting dynamics in addition to the RTG, which should
result in longer term sustained release of the drug. One additional setback of using micelles
alone is that they are prone to dissociation upon dilution (e.g. after injection) and can
therefore face major stability challenges. By encapsulating the micelles within the RTG
46


system, they are protected from this dilution effect because they are physically entrapped
within the RTG matrix, whichas discussed previouslyis largely dehydrated upon
gelation. Finally, the RTG presents an advantageous secondary vehicle for micelle delivery
because of its in situ gelation, which will prevent micelles from being carried away by local
fluid flows after injection. This effect has been shown to severely limit micelle retention at
the inj ection site, especially after intravitreal inj ection where micelles were found to escape
into the aqueous humor and be cleared systemically within 14 days[152].
To assess this systems ability to deliver hydrophobic drugs, the drug triamcinolone
acetonide (TA) was chosen as a model poorly-soluble drug. TA has gained attention for
ophthalmic indications in recent years by use as an intravitreal injection for treatment of
exudative age-related macular degeneration, diabetic macular edema, proliferative diabetic
retinopathy, and branch and central retinal vein occlusion, among others. These ocular
diseases are characterized by a) a pathological proliferation of intraocular cells such as
RPE cells in proliferative vitreoretinopathy and vascular cells in proliferative diabetic
retinopathy, which are generally accompanied and stimulated by intraocular inflammatory
processes; or b) defects in the blood-retinal barrier due to capillary leakage, which results
in the accumulation of fluid in the intra- or sub-retinal spaces[153]. Corticosteroids have
been shown to be effective in reducing intraocular inflammation, tightening capillary
vessel walls andat sufficient concentrationssuppressing both neovascularization and
the proliferation of endothelial and fibroblast cells[ 154], In order to achieve sufficiently
high concentrations of corticosteroids in posterior tissues, direct intravitreal injections were
explored and found to be highly effective[155]. However, the use of soluble corticosteroids
allowed for their rapid clearance from the vitreous humor resulting in the need for frequent
47


re-injections. To overcome this problem, large-dose injections of the poorly-soluble,
crystalline corticosteroid TA (solubility at 37C: 24 pg/mL[156], log P: 2.53[157]) were
employed as a method to form an intravitreal depot of the corticosteroid, which could
sustain efficacious drug levels for a longer period of time[158]. This method continues to
suffer from the use of a large single dose of the drug, which has been attributed with many
side effects, but especially a spike in IOP which can persist for weeks to months if
untreated[159]. In addition, the current treatment paradigm requires intravitreal TA
injections to be repeated every 2-5 months, which can be burdensome for both the provider
and the patient.
The use of a controlled release system, such as the combined RTG-micelle system, may
improve the intraocular pharmacokinetics of TA administration by providing for a slow,
continuous release of the drug over a longer time frame than is currently achievable. By
minimizing the initial concentration spike of the drug (as is seen in intravitreal injections)
and improving the longevity of a single administration, this system may reduce side effects
such as IOP spikes and reduce the burden of treatment, potentially resulting in improved
clinical outcomes. The ability for such a combined polymeric micelle/RTG system to
encapsulate and sustain release of TA will be explored within this thesis.
48


CHAPTER II
POLYMERIC MICELLES FOR SUSTAINED RELEASE OF A POORLY
SOLUBLE CORTICOSTEROID
Poor aqueous solubility is a major limiting factor in the administration of existing and
newly-developed drugs. To overcome this challenge, a polymeric micelle system was
developed based on a novel PEG-capped poly(ester urethane) tri-block copolymer for
efficient encapsulation and long-term delivery of a model poorly-soluble corticosteroid.
Low and high molecular weight variants of the base polymerdesigned to modulate drug
release kineticswere characterized by GPC and NMR spectroscopy to confirm their
structural properties. The polymer was found to be highly biocompatible based on in vitro
cytotoxicity testing against human retinal pigmented epithelial cultures. Micelles were
fabricated from the low and high molecular weight variants using both an established
sonication technique and a newly-developed filter extrusion process. Micelles produced
by both techniques possessed the expected core-shell structure, as confirmed by electron
microscopy. However, those produced by the extrusion technique were significantly more
monodisperse in diameter with a relative standard deviation of 12.3% compared to 42.3%
for those produced by the sonication technique. Release kinetics of the corticosteroid
triamcinolone acetonide were dependent on the molecular weight of the base polymer,
but all conditions were capable of maintaining zero-order release kinetics for over one
year with minimal burst release. This novel micelle system is well-suited for extended
delivery of poorly-soluble small molecule compounds, and the filter extrusion process
developed herein provides a rapid fabrication technique that produces a highly
49


monodisperse micelle population and eliminates the need for additional sterilization
steps.
Introduction
Poor aqueous solubility is a major limiting factor in the administration of existing and
newly-developed drugs. Nearly 40% of new drug candidates exhibit poor solubility leading
to low bioavailability, large intra- and inter-subject variability and poor dose
proportionality[160]. While strategies can be employed to facilitate solubilization of these
compounds (such as micronization and salt formation)[55], these methods do not typically
improve the duration of therapeutic activity and, as such, require frequent re-administration
to maintain therapeutically effective levels of the drug in tissue.
To simultaneously improve delivery and duration of therapeutic activity of poorly soluble
drugs, advanced drug delivery systems have been developed that encapsulate the
compound in a nanocarrier system. Such systems that have garnered the most attention
include liposomes, nanoparticles and micelles; however, application of these systems
continues to be hampered by various limitations. Liposomal systems tend to have limited
drug loading capacity since higher drug loads can destabilize the membrane and affect
overall stability of the system[161]. Nanoparticulate systemsmost commonly formulated
from poly(lactic-co-glycolic acid) (PLGA) or similar biodegradable polymerstypically
produce acidic byproducts during biodegradation[162] leading to concerns of local
inflammation[163] and drug stability[61],
Micellar systems can overcome many of these limitations due to their unique core-shell
structure. Typically fabricated from di- or tri-block amphiphilic copolymers, micelles self-
50


aggregate into nanostructures possessing a hydrophobic core and hydrophilic shell. This
structure is highly advantageous for use as an advanced drug delivery system because
poorly soluble drugs can be efficiently entrapped in the hydrophobic core at high
concentrations while the hydrophilic shell ensures high biocompatibility and prevents
particle aggregation[164].
In this work, a novel micellar system has been developed using a functionalizable
amphiphilic copolymer, which lends significant flexibility both in the optimization of the
physico-chemical properties of the micelles and in its conjugation with targeting ligands or
other biomolecules. In addition, a novel filter extrusion process was developed and
characterized for fabrication of the micelles, which possesses several advantages over
current fabrication techniques such as sonication. While this process is commonly
employed in fabrication of liposomes, to our knowledge this is the first time such a process
has been reported for the facile fabrication of polymeric micelles.
In order to assess drug release kinetics, the corticosteroid triamcinolone acetonide (TA,
molar mass 434.5 g/mol) was employed as a model drug. TA has limited aqueous solubility
but has gained extensive popularity recently, especially for ocular indications such as
uveitis, macular edema, vitreoretinopathy and choroidal neovascularization secondary to
age-related macular degeneration[165]. Current clinical administration techniques (such as
intravitreal or periocular injection) rely on injecting large doses of a suspension of the drug
to avoid the need for frequent re-administration, but this method carries a high incidence
of highly damaging side effects including elevated intraocular pressure, endophthalmitis,
cataract and retinal detachment[166].
51


The micelle system developed herein could significantly improve the clinical use of TA by
acting as an injectable long-term sustained release system. The synthesis, characterization
and in vitro biocompatibility testing of the precursor polymer are presented in addition to
characterization and in vitro drug release testing with TA of micelles fabricated by both a
common sonication technique and an adapted filter extrusion technique.
Material and methods
Materials
Dimethylformamide (DMF), hexamethylene diisocyanate (HDI), methoxypolyethylene
glycol (mPEG-550), DMSO and serinol were obtained from Sigma-Aldrich (St. Louis,
MO). Di-tert butyl di-carbonate and ethyl acetate were obtained from Alfa Aesar (Ward
Hill, MA). TA was obtained from Fluka (Sigma-Aldrich, St. Louis, MO). Hexane was
obtained from EMD Chemicals (Philadelphia, PA) and diethyl ether from Fisher Scientific
International, Inc. (Waltham, MA). All organic solvents were anhydrous and chemicals
were used as received. Phosphate-buffered saline (PBS), fetal bovine serum (FBS),
penicillin/streptomycin (PS) and Dulbecco's Modified Eagle's Medium-nutrient mixture F-
12(1:1 DMEM:F-12) were obtained from Thermo Scientific (Logan, UT). Vybrant MTT
Cell Proliferation Assay Kit was obtained from Molecular Probes (Eugene, OR).
Synthesis of N-Boc serinol
Serinol (21.5 mmol) was dissolved in 20 mL of absolute ethanol and stirred at 4C. Di-tert-
butyl dicarbonate (26.0 mmol) was dissolved in 20 mL of absolute ethanol and added
dropwise to the serinol solution over a period of one hour, while maintaining 4C and
constant stirring. The solution was heated to 37C with vigorous stirring and reacted for
52


one hour. The ethanol was removed by rotary evaporation at 45C and 10 mbar vacuum
and the solid was re-dissolved in a 1:1 mixture of hexane and ethyl acetate by gentle
heating. Additional hexane was added until precipitation was observed and the resulting
suspension was stored at 4C overnight to allow recrystallization. Subsequent vacuum
filtration yielded a white flaky product.
Synthesis of poly(hexamethylene-alt-serinol) (PHS) block
A-Boc serinol (1.31 mmol) was weighed out and lyophilized for 12 hours at -45C and
0.045 mbar. In a reaction vessel, the A-Boc serinol was dissolved in 3 mL dry DMF and
heated to 80C under gentle stirring and a nitrogen atmosphere. Then HDI (1.31 mmol)
was added drop-wise and the reaction was carried out for either 3 or 5 days to produce low
or high molecular weight PHS, respectively. After the specified time, additional HDI (2.62
mmol) was added to cap each end of the polymer with an isocyanate group and terminate
the reaction. After 24 hours, the product was rapidly precipitated twice in diethyl ether.
Conjugation of mPEG
The isocyanate-terminated PHS from the previous reaction was immediately re-dissolved
in 3 mL dry DMF and heated to 80C under gentle stirring and a nitrogen atmosphere. An
excess of mPEG (5 mmol) was lyophilized and added to the reaction. The PEGylation
reaction was carried out for 12 hours and the resulting product, polyethylene glycol-block-
poly(hexamethylene-£///-serinol)-/}/o6Vf-polyethylene glycol (PEG-PHS-PEG), was
purified by three precipitations in diethyl ether and then dried completely by extended
rotary evaporation at 50C and 10 mbar vacuum. The full synthetic route of PEG-PHS-
PEG can be found in Appendix A.
53


Micelle fabrication by sonication
PEG-PHS-PEG and TA were dissolved in 2 mL DMSO at 2.5 and 0.25 wt% respectively
(polymer/DMSO and drug/DMSO). This solution was then added dropwise to a beaker
containing 40 mL of purified water (milliQ or equivalent) submerged in an ultrasonic bath
(VWR International, West Chester, PA; 48W RF power). The resulting emulsion was
sonicated for 10 min. Removal of DMSO was carried out by centrifugation at 4500 ref for
5 min, pouring off the supernatant and then re-suspending the micelles in purified water.
This DMSO extraction procedure was carried out 3 times. The resulting micelles were
either used immediately or lyophilized at -45C and 0.045 mbar to produce a dry product.
Micelle Fabrication by extrusion
PEG-PHS-PEG and TA were dissolved in DMSO at 2.5 and 0.25 wt% respectively
(polymer/DMSO and drug/DMSO). This solution was then added to purified water (milliQ
or equivalent) at a 1:20 organic:aqueous phase ratio and loaded into one end of the
extrusion apparatus (Avestin, Inc., Ottowa, ON). A silver membrane filter (100 nm average
pore size, Sterilitech Corp., Kent, WA) was used as the extrusion filter. The emulsion was
passed through the filter 11 times and then transferred to a centrifuge tube. Removal of
DMSO was carried out by centrifugation at 4500 ref for 5 min, pouring off the supernatant
and then re-suspending the micelles in purified water. This DMSO extraction procedure
was carried out 3 times. The resulting micelles were either used immediately or lyophilized
at -45C and 0.045 mbar to produce a dry product.
54


Residual DMSO quantification
After each iteration of the above-described DMSO extraction procedure, the supernatant
was analyzed by UV-spectroscopy to determine the DMSO concentration by measuring
the absorbance at 208 nm against a pre-constructed standard curve. This yielded the amount
of residual DMSO left in the system as a function of purification step. In addition, a
secondary method was employed to directly measure the amount of residual DMSO
contained in fabricated micelles. After three and five purification iterations, micelles were
suspended in dThO and sonicated for 1 hour at elevated temperature in order to lyse the
micelles and permit release of entrapped DMSO. After centrifugation, the supernatants in
these samples were also measured by UV spectroscopy to determine the residual DMSO
concentration. The measured DMSO amount was divided by the total micelle mass in each
sample to express these data as the percent DMSO content remaining in the system.
Polymer and micelle characterization
NMR spectra were collected on an Inova 500 MHz NMR spectrometer (Varian, Inc.,
Palo Alto, CA) at 25C with a Nalorac NTR probe using CDCb as the solvent. Spectra are
displayed in ppm using the solvent peak as an internal reference. Molecular weight data
were collected on an EcoSEC GPC (Tosoh Biosciences, King of Prussia, PA) with a
refractive index (RI) detector using DMF/LiBr as a mobile phase and poly(methyl
methacrylate) standards.
Particle size data was determined by dynamic light scattering (DLS) on aNicomp 380 ZLS
(PSS, Port Richey, FL) with micelles suspended in PBS at a concentration of 1.0 mg/mL.
All DLS data are plotted using an intensity-weighted distribution. For particle morphology
55


assessments, dry micelles were sputter coated with 5nm Au and examined by field emission
scanning electron microscopy (FESEM) on a JSM 7401F (JEOL EISA, Peabody, MA).
Cytotoxicity testing
Cytotoxicity tests were carried out per ISO 10993-5 using an extract exposure procedure
against human retinal pigmented epithelial (ARPE-19) cells and assessed by MTT assay.
Cells were maintained at 37C and 5% CO2 in 1:1 DMEM:F-12 supplemented with 5%
FBS and 1% PS. To prepare extractions, the polymer was added to complete media at 5
wt% and incubated at 37C for 48 hours. Dilutions of the extraction were prepared by
diluting the extraction with complete medium. For cytotoxicity tests, cells were seeded at
10,000 cells/well and incubated for 24 hours at which point the media was replaced with
the extraction medium. After 24 hours of further incubation, cellular metabolic activity was
assessed by MTT assay per the manufacturers instructions with absorbance read at 540
nm on a Synergy 4 Multi-Mode Microplate Reader (BioTek, Winooski, VT). Data were
normalized to the absorbance of control wells incubated in pure media. Statistical
significance was assessed by ANOVA (MATLAB, MathWorks, Natick, MA) at a
significance level of 0.05.
In vitro TA release testing
Immediately after fabrication, micelles were suspended in 1 mL sterile PBS (pH 7.4) and
inserted into a 3500 MWCO dialysis chamber (Float-A-Lyzer G2, Spectrum Labs, Rancho
Dominguez, CA) floating in a conical tube containing 14 mL of sterile PBS. The tubes
were incubated at 37C on a rocking platform and samples were pulled periodically by
removing 5 mL of the release medium outside of the dialysis chamber and replacing it with
56


fresh PBS. The concentration of TA was determined by UV/visible spectroscopy
(GENESYS 10S UV-Vis, Thermo Scientific, Logan, UT) at 237 nm against a pre-
constructed standard curve (see Appendix B).
Results and discussion
Polymeric micelles have been effectively employed in myriad drug delivery applications
and are particularly advantageous in their capacity for stabilized encapsulation of
hydrophobic compounds[167]. Micelles are characterized by a hydrophobic core and
hydrophilic shell structure, wherein hydrophobic drugs can be encapsulated at the core
while the hydrophilic shell provides solubility and protection against in vivo clearance
mechanisms. The block co-polymer structure most commonly employed in micelle design
to achieve the contrasting core and shell properties lends great flexibility in tuning the
micellar properties to the molecule of interest.
Polymer synthesis and characterization
The structure of A-Boc serinol was confirmed by 'H NMR spectroscopy (500 MHz,
CDCb) 5 5.25 (1H, s), 3.70 (4H, d), 1.45 (9H, m). The tri-block copolymer PEG-PHS-PEG
was synthesized in low and high molecular weights by the methods described earlier. Its
structure was confirmed by *H NMR as shown in Figure 2.1: (500 MHz, CDCb) 5 5.25
(2H, s), 4.17 (1H, t), 4.09 (4H, d), 3.64 (4H, t), 3.40 (3H, s), 3.18 (4H, t), 1.51 (4H, quin),
1.45 (9H, m), 1.35 (4H, quin). Peak assignments were corroborated by NMR modeling
(Advanced Chemistry Development). Results of GPC analyses of the low and high
molecular weight copolymers are shown in Table 2.1.
57


F
Figure 2.1: *H NMR (500 MHz, CDCI3) analysis of the PEG-PHS-PEG copolymer
confirmed the structure of the hydrophobic PHS block (N-Boc serinol: B, G, H and J and
HDI: A, C, D and I) and successful capping by mPEG (E and F). Peak assignments were
corroborated by NMR modeling (Advanced Chemistry Development).
The PEG-PHS-PEG copolymer was used as-is in this work, but we predict that it can also
be extensively modified to tailor it to the specific molecule being investigated. Specific
methods of modification and their predicted impact on micelle performance are
summarized in Table 2.2. The PHS blockitself a copolymerbenefits primarily from
the inclusion of Boc-protected serinol within the repeating unit. With all serinol primary
amines protected, the frequent Boc groups contribute significant hydrophobicity to the PHS
block and undoubtedly also influence micelle core density due to steric effects. As these
are selectively removed, a) the hydrophobicity and steric effects of the PHS block are
attenuated, allowing for a denser but less hydrophobic micelle core; and b) primary amines
are exposed.
58


Table 2.1: Properties of high and low molecular weight PEG-PHS-PEG by GPC analysis.
Polymer Mn [kDa] Mw [kDa] PI Hydrophobic: hydrophilic ratio3)
Low MW PEG-PHS-PEG 25.3 48.2 1.91 42.9
High MW PEG-PHS-PEG 32.5 92.9 2.86 83.4
This latter outcome opens up an array of possibilities in that the primary amines can be
used to conjugate antibodies or other targeting moieties for localized micelle
effect[164,168]. This procedure would involve a partial or complete deprotection of the
Boc groups of PEG-PHS-PEG after polymer synthesis is complete and then conjugation of
the biomolecule of interest after micelles are fabricated. Additionally, as the fraction of
Boc groups that are removed during the deprotection routine can be precisely controlled
(data not shown), a balance can be found between retaining sufficient hydrophobicity to
stabilize the micellar structure and exposing enough amines to allow conjugation.
Investigation of this phenomenon is currently underway.
Table 2.2: Potential avenues for PEG-PHS-PEG modification
Modification Predicted impact on micelle performance
PHS block molecular weight Modification of micelle core hydrophobicity and density
Selective Boc group deprotection in PHS block Reduced hydrophobicity of micelle core, increased potential for hydrogen bonding with drug, introduction of moderate positive charge to core
Selective Boc group replacement Modification of micelle core hydrophobicity, direct tethering of therapeutics or other biomolecules
Poly(methylene) chain length in diisocyanate monomer Modification of micelle core density
PEG block molecular weight Resistance to rapid in vivo clearance and protein adsorption
59


Cytotoxicity testing
Cytotoxicty of the copolymer was assessed in vitro per ISO 10993-5 against ARPE-19 cells
using the MTT assay for mitochondrial activity. As shown in Figure 2.2, mitochondrial
activity between negative control samples incubated in pure media and experimental
samples incubated in media with PEG-PHS-PEG extracts at all concentrations were
statistically indifferent (p=0.265). Positive control samples incubated in complete media
with 10% DMSO were the only samples that were statistically different (p=1.65 x 10"5).
p< 0.0001
"o
c
o
U
o
ox
O
c
-O
S
o
X)
<
Figure 2.2: The PEG-PHS-PEG copolymer was non-cytotoxic against ARPE-19 cells as
assessed by MTT mitochondrial activity assay. Positive control samples (10% DMSO in
culture media) were the only samples that were significantly different (p<0.0001); all
experimental groups were statistically insignificant from the negative control sample of
pure culture media (p=0.265). Data were normalized to the media only control sample;
means + standard deviations shown (n=5).
Micelle fabrication
p = 0.265
10% DMSO Media Only Extract,
Full Cone
Extract, Extract,
1:2 Dilution 1:4 Dilution
Drug-loaded micelles were fabricated by two methods: sonication and extrusion. The
sonication procedure is similar to a nanoparticle fabrication procedure in which an oil-in-
60


water emulsification is formed, followed by sonication to produce nano-scale particles. The
extrusion procedure also starts with an oil-in-water emulsification but uses serial passage
through a filter instead of sonication to produce the final particles. A comparison of the
size distribution of particles fabricated from high MW PEG-PHS-PEG by the two
techniques is shown in Figure 2.3.
50 100 200 450 900 1800
Diameter (nm)
Figure 2.3: The size distribution of micelles obtained by extrusion was significantly more
monodisperse than that obtained by sonication (12.3% relative standard deviation (%RSD,
a) versus 42.3%, respectively). The mean diameter (p) was also smaller for extruded
micelles at 217.5 nm compared to 357.4 nm for sonicated micelles. Both populations were
produced using high MW PEG-PHS-PEG. Plot shown for intensity-weighted distributions.
Micelles fabricated by the sonication technique had a mean diameter of 357.4 nm with a
42.3% relative standard deviation (RSD). In stark contrast, those obtained by extrusion
through a 100 nm pore size filter had a mean diameter of 217.5 nm with a 12.3% RSD,
indicating a significantly more monodisperse population. That the mean diameter of
extruded micelles is larger than the average pore size of the filter material used can be
explained by: a) the oil phase of the emulsification forming cylindrical micelles as they
61


pass through the filter pores, which then collapse to spheres larger in diameter than the
pore size after passing through the filter; and b) potential coalescence of micelles after
extrusion but before DMSO is removed.
Beyond producing a more monodisperse population of micelles, which will reduce batch-
to-batch variability during the fabrication process, the filter extrusion process also carries
other advantages over current techniques. First, fabrication time is greatly reduced; the
process takes under one hour, whereas a typical dialysis procedure takes over a day to
complete. Second, the size of micelles produced by this process can be readily controlled
by simply by changing the average pore size of the extrusion filter membrane. Third, unlike
the sonication procedure, the extrusion technique contains no steps that would be
potentially damaging to structure-sensitive molecules such as proteins or antibodies.
Finally, the greatest advantage of the extrusion process is thatso long as the filter
membrane pore size is less than 200 nmthe micelles are pre-sterilized and need not
undergo secondary sterilization procedures, as would be the case with other fabrication
techniques. These secondary sterilization techniques can be especially damaging to
polymeric systems as they can increase the toxicity of or alter the physico-chemical
characteristics of the final system[ 168-170],
In addition to characterization by DLS, micelle morphology was examined by FESEM and
TEM. TEM micrographs confirmed the core/shell structure that was expected, showing the
higher contrast hydrophobic core surrounded by the lower contrast PEG shell (Figure 2.4,
Panel D). Additionally, FESEM was used to investigate sonicated samples immediately
after fabrication (Figure 2.4, Panel A) and at 2 weeks after incubation in PBS (Figure 2.4,
Panel B) and extruded samples immediately after fabrication (Figure 2.4, Panel C). At all
62


time points and by both methods, micelles had high sphericity. The differences noted in
polydispersity by DLS were also evident in the micrographs; extruded samples were
significantly more monodisperse. There was also a large difference noted in micelle
diameters between the 0 and 2 week samples. As determined by post-hoc image analysis,
micelles immediately after fabrication had a mean diameter of 369 nm, correlating fairly
closely with DLS results. However, the mean diameter increased drastically within 2 weeks
to 962 nm, indicating significant swelling of the micelles when incubated in PBS.
Figure 2.4: Micelle size and morphology assessed by FESEM after fabrication by
sonication (A) and extrusion (C) corroborated DLS data and also demonstrated sphericity
of produced particles. Micelles fabricated by sonication and incubated in PBS for 2 weeks
(B) had significantly larger diameters than post-fabrication samples, indicating significant
swelling of the particles upon incubation. TEM imaging (D) confirmed the core/shell
structure expected of micelles with a poly(ester urethane) core and PEG shell.
63


Residual DMSO levels
Solvents often used in pharmaceutical formulations such as DMSO and NMP are generally
considered non-cytotoxic at lower concentrations[82,83]. However, some studies have
indicated mytotoxicity[84] and ocular-specific low dose cytotoxicity [8 5] of DMSO, which
raise concerns about its use in biomedical applications. In order to address concerns related
to the use of DMSO in our micelle fabrication technique, we evaluated the efficacy of our
DMSO removal process. The amount of DMSO removed at each iteration of the
purification process is shown in Figure 2.5. Of the 55 mg of DMSO used in the initial
process, 50.7 1.57 mg was removed in the first step, 1.71 0.184 mg was removed in the
second step and 0.035 0.002 mg was removed in the third step. No DMSO was detectable
in the supernatants at the fourth and fifth steps. These data confirm that the purification
process used in the micelle fabrication process efficiently removes residual DMSO within
three iterations.
£
03
"c3
£
a
zj
cft
O
GO
2
a
Purification Iteration Number
Figure 2.5: The amount of DMSO removed per purification iteration decreased with
each additional iteration. The vast majority was removed in the first iteration with
smaller amounts removed in the second and third iterations. The amount of DMSO in the
supernatant at the fourth and fifth steps was below detectable levels. Means and standard
deviations of n=3 samples are shown.
64


Samples of micelles were also evaluated at three and five iterations directly for the amount
of DMSO remaining within the product. After the third iteration DMSO accounted for
0.303 0.260% of the system and this value dropped to 0.137 0.030% after the fifth
iteration. These findings confirm that the current purification process efficiently lowers
DMSO to low levels within 3-5 iterations. Additionally, both of these residual DMSO
values are below the 1.0% minimum concentration found to cause ocular cytotoxicity by
Galvao, et al.[85]
Release of triamcinolone acetonide
TA was loaded into micelles during fabrication by co-dissolution with PEG-PHS-PEG in
the organic phase of the emulsion. The TA loading ratio was fixed at 10 wt% and the effect
of varying the PEG-PHS-PEG molecular weight was examined by in vitro release testing.
As shown in Figure 2.6, all samples showed varying levels of burst release within the first
several days of testing, but for all samples this effect involved less than 6.0% of total drug
load and was as low as 2.0% for samples produced by sonication. The release rate of TA
from extruded micelles was inversely related to molecular weight: the average release rate
for low and high molecular weight extruded micelles was 0.321% and 0.117%/day,
respectively, compared to 0.175%/day for those produced by sonication. These release
rates represented significantly slower release than free drug, which was released at
6.49%/day. Excluding the initial burst period, R2 fit to zero-order was 0.939 and 0.943 for
low and high molecular weight extruded micelles, respectively, and 0.986 for sonicated
samples. All of the micelle systems were capable of sustaining near zero-order release of
TA for extended periods of time. Assuming consistent release rates were to be maintained,
extrapolated zero-order kinetics indicates that low and high molecular weight extruded
65


micelles would continue releasing TA for 350 52.5 and 1220 267 days, respectively,
compared to 15 2.72 days for free drug. These projected release time frames from PEG-
PHS-PEG micelles represent a significant improvement over free drug and previous
systems intended for TA delivery[165,171,172] and could potentially significantly
improve the clinical administration of TA.
100 -
-o
S3 80

"3
o£
60 -
>
-D
-5 40
U
20 -
u
TT
ll
Free Drug Low MW, Extruded
High MW, Extruded High MW, Sonicated




0
20
40
60 80 100 120 140 160
Time (days)
Figure 2.6: Release of TA from extruded and sonicated PEG-PHS-PEG micelles was near
zero-order for all conditions. TA release rate was inversely related to the molecular weight
of PEG-PHS-PEG. Means standard deviations shown for n=3 samples.
To put this release behavior in perspective, Ozurdex is a currently-marketed PLGA
intravitreal implant for the delivery of dexamethasone. Dexamethasone is a corticosteroid
very similar in structure and indications to TA with a molar mass of 392.5 (as compared to
434.5 for TA). Even though Ozurdex is a large-scale polymer implant, its 6-month
sustained dexamethasone delivery timeframe[173] is far shorter than what PEG-PHS-PEG
micelles were able to achieve with TA. Further, since Ozurdex is a bulk polymer implant,
66


the PEG-PHS-PEG micelles will enable a much more facile injectable delivery, thereby
reducing the trauma and potentially the side effects associated with the procedure.
Specifically compared to current administration techniques for TA, micelle systems have
a further advantage over suspensions in that they are less obstructive to vision. Suspensions
of TAwhich are opaque by naturehave been reported to cause transient obstructions
in the visual fields of patients for several days or longer after administration[174], until the
drug settles at the inferior inner limiting membrane of the retina. Nanoparticulate systems
such as the micelles developed in this work avoid this clinical drawback due to their small
size and hydrated state (i.e. PEG shell), which would be expected to render them invisible
to the patient upon injection.
Mechanistic insights from release kinetics
To better understand the capabilities of the PEG-PHS-PEG micelle system, an analysis of
the time-dependent release kinetics and FESEM micrographs can elucidate the
mechanisms driving TA release from the micelle core. Based on these data, the mechanism
of drug release appears to be a combination of swelling of the micelle core and simple
diffusion. Based on the release kinetics in the first several days and the significant diameter
increase observed by FESEM within the first 2 weeks, it is likely that most of this swelling
occurs within the first several days of aqueous incubation and is the mechanism responsible
for the modest burst release observed early in the release profiles.
The magnitude of this initial burst release observed for PEG-PHS-PEG micelles was
significantly smaller than is typical of other micelle or nanoparticle systems[175,176].
Nevertheless, there was a difference in burst release behavior between sonicated and
67


extruded micelles. It is hypothesized that this difference results from a mild annealing
process concomitant with the sonication procedure, resulting from sample heating during
the extended sonication time.
Compared to other polymeric micelle systems, PEG-PHS-PEG micelles have several
distinct advantages. The theoretical ability to control the hydrophobicity of the micelle
core, while still being fully characterized, should allow the system to be tailored to a
specific drug molecule. This partial Boc deprotection strategy could also be used to
controllably introduce a positive charge to the micelle core, allowing the use of electrostatic
interactions to further modulate drug release kinetics or other favorable interactions. Also
under investigation are the biodegradation kinetics of the micelle system; however, it is
expected that the polymer would be fully biodegradable through the carbamate bonds
present throughout the PHS block.
Conclusions
The micelles developed in this work have the potential to significantly improve patient care
by allowing repeat injections of TA to be replaced by a single administration of TA-loaded
micelles. This approach would allow for continuous therapeutic levels of TA for well over
one year. The pre-cursor polymer was found to be fully biocompatible and contains
considerable levels of flexibility that have not yet been fully explored. In addition, the
extrusion process developed in this work has many benefits over other commonly-used
fabrication procedures and greatly simplifies the scale-up process required for
commercialization. This system will continue its translational path and is currently being
tested in animal models for in vivo biocompatibility.
68


CHAPTER III
A REVERSE THERMAE GEL SYSTEM FOR LOCALIZED OCULAR
DELIVERY OF TRIAMCINOLONE ACETONIDE
Posterior delivery of ocular therapeutics continues to be an unsolved clinical challenge. In
this work, a reverse thermal gelling polymer was developed to improve delivery of
triamcinolone acetonide (TA) to posterior targets. The polymer was synthesized by grafting
the thermo-sensitive polymer pol y(N-\ sopropyl acryl am i de) (PNIPAAm) to a novel
poly(urethane urea) backbone. The resulting copolymer was thoroughly characterized by
structural and thermo-mechanical assessments as well as in vitro cytotoxicity studies
against human retinal pigmented epithelial (RPE) cultures and in vivo biocompatibility
studies through intravitreal injections in rats. In vitro release experiments with TA were
performed with varying gel concentrations and TA loading fractions. The copolymer
transitions from a low-viscosity solution to a stable physical gel near 30C. No cytotoxic
effects were found against RPE cells in culture and intravitreal injections of the system in
rats showed no adverse reactions. Release kinetics were independent of initial polymer
concentration but did depend on TA loading fraction. All release profiles followed first-
order kinetics and samples loaded with 20 wt% TA sustained meaningful levels for 80
days. This novel reverse thermal gelling polymer system combines high biocompatibility,
favorable TA release kinetics and an advantageous delivery method to potentially improve
the posterior administration options for corticosteroids.
69


Introduction
Ocular drug delivery is a uniquely challenging field primarily due to the complex
anatomical and physiological barriers of the eye. While these barriers are especially
problematic for topical administration of drugs, it has persisted as the dominant delivery
mechanism of ocular drugs primarily due to its familiarity to patients and ophthalmologists
and ease of formulation[177]. However, as the pharmaceutical industry continues to shift
away from drugs that exhibit moderate to high aqueous solubilityand thus can be readily
formulated for topical administrationnew delivery systems will be required.
One class of materials that has recently received significant attention for ocular
applications is the in situ-gelling polymer system[78,106,178], Because they can be readily
deployed by injection to posterior periocular or intraocular target sites, these systems can
be used to overcome the aforementioned barriers and deliver drugs at or near the target
tissue by minimally-invasive methods. Generally, in situ-gelling polymer systems begin as
a low viscosity liquid that is injected through a small gauge needle or cannula and form a
physical or cross-linked gel via introduction of a specific stimulus.
Reverse thermal gels (RTG) are a class of in situ-gelling systems that undergo a reversible
solution-to-gel (sol-gel) phase transition by temperature change alone[108,179]. These are
advantageous in that they remove the need to introduce reactive species or an external
energy source to realize gelation of the system after deployment. Further, since most RTG
systems are composed of amphiphilic polymers and undergo sol-gel transitions based on
the temperature-dependent development of hydrophobic interactions, they are especially
well-suited for encapsulation and delivery of poorly soluble molecules[108].
70


RTGs have been explored for various tissue engineering and drug delivery
applications[91,180-182], RTGs developed for these applications are most commonly
comprised of amphiphilic block copolymer systems, such as poly(ethylene oxide)-
poly(propylene oxide)-poly(ethylene oxide) (poloxamers)[88,183] and polyethylene
glycol)-poly(lactic-co-glycolic acid)-poly(ethylene glycol)[106,184,185]. While these
systems are easy to synthesize and their thermal gelling characteristics are readily modified
by adjusting the block ratio of the resulting copolymer, the dependence of the thermal
gelling properties on the hydrophilic-hydrophobic balance limits their ability for
functionalization, e.g. with highly hydrophilic antibodies. To overcome this limitation, the
present work instead made use of poly(7V-isopropyl acrylamide) (PNIPAAm) to introduce
thermal gelling properties.
A novel RTG polymer has been developed that can specifically meet the needs of ocular
drug delivery applications. The novel chemistry of this RTG polymer is based on a
biomimetic poly(urethane urea) with functionalizable primary amines on each repeating
unit of the polymer. A fraction of these primary amines are used to graft the thermal-
sensitive PNIPAAm, which lends the copolymer reverse thermal gelling properties. This
graft co-polymer structure makes for a flexible system since it permits higher levels of
PNIPAAm content in the final system as compared to a linear block copolymer.
In this work, this novel RTG system was fully characterized and assessed for its in vitro
and in vivo biocompatibility and capacity to sustain drug release. Since this system has
most potential in posterior ocular applications where it can be used to overcome access-
based challenges, the poorly-soluble corticosteroid triamcinolone acetonide (TA) was
chosen as a model drug.
71


Materials and methods
Materials
Serinol, urea, hexamethylene diisocyante (HDI), trifluoroacetic acid (TFA), 4,4'-azobis(4-
cyanovaleric acid) (ACA), methanol, phenolphthalein, L-glutamine and dimethyl sulfoxide
(DMSO) were obtained from Sigma Aldrich (St. Louis, MO). Di-tert-butyl dicarbonate,
ethyl acetate, A-hydroxysuccinimide (NHS), N-(3-dimethylaminopropyl)-N'-
ethylcarbodiimide hydrochloride (EDC) and dimethyl sulfoxide-d6 (DMSO-d6) were
obtained from Alfa Aesar (Ward Hill, MA). Ethanol, hexane, diethyl ether and sodium
hydroxide (NaOH) were obtained from Fisher Scientific (Pittsburgh, PA).
Dimethylformamide (DMF) was obtained from BDH Chemicals (Poole, UK).
Dichloromethane (DCM) was obtained from J.T. Baker (Phillipsburg, NJ). N-
isopropylacrylamide (NIPAAm) was obtained from Acros Organics (Geel, Belgium).
Chloroform-d (CDCI3) was obtained from Millipore (Billerica, MA). Vybrant MTT cell
proliferation assay kit was obtained from Life Technologies (Carlsbad, CA). TA was
obtained from Fluka (St. Louis, MO). All other reagents were obtained from Thermo
Scientific (Waltham, MA).
Polymer synthesis
A-Boc serinol was synthesized through Boc group protection of serinol. Serinol (21.5
mmol) was dissolved in 20 mL of absolute ethanol and stirred at 4C. Di-tert-butyl
dicarbonate (26.0 mmol) was dissolved in 20 mL of absolute ethanol and added dropwise
to the serinol solution over a period of one hour, while maintaining 4C and constant
stirring. The solution was heated to 37C with vigorous stirring and reacted for one hour.
72


The ethanol was removed by rotary evaporation at 45C and 10 mbar vacuum and the solid
was re-dissolved in a 1:1 mixture of hexane and ethyl acetate by gentle heating. Additional
hexane was added until precipitation was observed and the resulting suspension was stored
at 4C overnight to allow recrystallization. Subsequent vacuum filtration yielded a white
flaky product.
The backbone poly(urethane urea) was synthesized by copolymerization of A-Boc serinol,
urea and HDI. A-Boc serinol (6.0 mmol) and (6.0 mmol) urea were lyophilized for 24 hours
and then dissolved in 6 mL of dry DMF. Under a nitrogen atmosphere, the reaction was
brought to 90C under moderate stirring and HDI (12 mmol) was added drop-wise through
the septum. The reaction was carried out for 7 days, after which it was cooled and
precipitated twice in diethyl ether and once in water. After lyophilization, a dry yellowish
product was obtained, poly(serinol hexamethylene urea) (PSHU), in 86.5% yield.
In order to expose primary amines for conjugation, Boc protective groups were removed
through a complete deprotection routine. PSHU (0.10 g) was dissolved in a 50:50 (v/v)
mixture of DCM and TFA (18 mL total volume) and reacted uncapped with rapid stirring
for 15 minutes. The solvents were immediately removed by rotary evaporation at 45C and
10 mbar vacuum and the product re-dissolved in DMF. Two precipitations in diethyl ether
and terminal rotary evaporation yielded a dry yellowish product, de-protected PSHU
(dPSHU).
Carboxylic acid-terminated PNIPAAm (PNIPAAm-COOH) was synthesized based on a
procedure described previously[180], NIPAAm (5.0 g) and ACA (0.060 g) were dissolved
in 25 mL dry methanol and bubbled with nitrogen for at least 30 minutes at room
temperature. The reaction was raised to 68C under moderate stirring and carried out for 3
73


hours. The product was precipitated twice in 60C water and dialyzed against 1 L dThO
for 48 hours in 3,500 Da molecular weight cut-off (MWCO) dialysis tubing (Spectrum
Labs, Inc., Rancho Dominguez, CA).
The RTG was synthesized by grafting PNIPAAm-COOH to dPSHU through carbodiimide
linking chemistry. Of the calculated 18 repeating units (and, thereby, primary amines) in
dPSHU, 25% were used for PNIPAAm conjugation. PNIPAAm-COOH (14.0 pmol), NHS
(42.0 pmol) and EDC (42.0 pmol) were dissolved in 3 mL DMF and reacted for 24 hours.
dPSHU (3.10 pmol) dissolved in 2 mL DMF was then added drop-wise and the conjugation
reaction proceeded for 24 hours. The product was added directly to a 50 kDa MWCO
dialysis tube and dialyzed for 48 hours to remove unreacted PNIPAAm. The product was
lyophilized to yield a white flaky product, PSHU-NIPAAm. The full synthetic route of
PSHU-NIPAAm is shown in Appendix C.
Polymer characterization
The molar mass of PNIPAAm was determined experimentally by titrating for the
carboxylic acid end groups. Approximately 0.050 g of PNIPAAm was dissolved in 10 mL
dH20 with 10 pL of phenolphthalein solution (2 wt% in absolute ethanol). The solution
was titrated to the end point by adding 0.01 N NaOH.
The molecular weight distribution of PSHU was determined by gel permeation
chromatography (GPC) on an ECOSEC system (Tosoh Biosciences, King of Prussia, PA)
with a TSKgel colum (Tosoh Biosceinces) using DMF with LiBr as the mobile phase.
Separations were made at 25C and detected by refractive index against poly(methyl
methacrylate) standards. Full results of the GPC analysis are shown in Appendix D.
74


Proton nuclear magnetic resonance (*H NMR) spectra were collected on an INOVA 500
MHz instrument (Varian) with a 5 mm triple resonance proton detector. Spectra were
collected in a mixed solvent of 5% (v/v) DMSO-d6 in CDCb at 25.0C and post-processed
in ACD ID NMR Processor software (Advanced Chemistry Development, Inc., Toronto,
ON). Elemental analysis was performed by Micro Analysis, Inc. (Wilmington, DE).
Fourier Transform Infrared (FT-IR) spectra were collected on a Nicolet 6700 (Thermo
Fisher Scientific, Waltham, MA) using polyethylene windowed cards.
Lower critical solution temperature (LCST) measurements were made on a Cary 100 UV-
visible spectrophotomer (Agilent Technologies, Inc., Santa Clara, CA) equipped with a
temperature-controlled 6-cell stage. Polymers were dissolved in PBS (pH 7.4) at 1 wt%
and transmittance readings were taken at 500 nm while the temperature was increased at
0.5 C/min.
Rheological studies were performed on a TA ARES rheometer (TA Instruments, New
Castle, DE) equipped with a peltier plate. A 5wt% solution of PSHU-NIPAAm dissolved
in PBS (pH 7.4) was placed between the plates and a small bead of silicon oil was applied
around the outer edge to prevent evaporation. Low-temperature strain sweeps with and
without silicon oil were used to ensure measurements were not affected by its presence.
All experiments were performed in the linear deformation region, as defined by preliminary
strain sweep experiments. Measurements were made at 1% strain and a frequency of 1 rad/s
while the temperature was swept from 20C to 40C at 0.5C/min.
75


In vitro and in vivo biocompatibility studies
In vitro biocompatibility studies were performed by directly exposing cells in culture to
the PSHU-NIPAAm RTG system in complete growth medium. ARPE-19 primary retinal
pigmented epithelial cells were cultured in complete growth medium containing 1:1
DMEM:F12, 10% FBS and 2 mM L-glutamine at 37C and 5% CO2. For cytotoxicity
studies, PSHU-NIPAAm was dissolved in the complete medium at 5 and 10 mg/mL. Cells
were seeded in 96-well plates at 5000 cells/well, grown for one day and the media were
replaced with the polymer samples. Positive controls of 5% DMSO in complete growth
medium and negative controls of complete growth medium alone were also included. At 1,
3 and 5 days, MTT assays were performed according to the manufacturers instructions.
All animal studies were performed according to guidelines set forth by the Institutional
Animal Care and Use Committee. Male Brown Norway rats (240-275 g; Charles River
Laboratories, Wilmington, MA) were housed in group cages and maintained on a 12 hour
light/12 hour dark cycle. Animals had free, continuous access to food and water throughout
the experiment. Rats received intravitreal injections of 2 pL of PSHU-NIPAAm in PBS
(solutions were filter sterilized before injection) in the left eyes and 2 pL of PBS (pH 7.4)
in the right eyes under isoflurane anesthesia. At 3 and 14 days (n=3 per time point), rats
were euthanized, both eyes were enucleated and immediately placed in Davidsons fixative
for 24 hours followed by neutral-buffered formalin for 4 hours and 70% ethanol overnight.
Eyes were then embedded in paraffin, sectioned and stained with hematoxylin and eosin
for histological analysis.
76


Release testing
In vitro release testing was performed on activated PSHU-NIPAAm gels with TA as a
model hydrophobic drug with relevance for posterior ocular administration. PSHU-
NIPAAm was dissolved in PBS (pH 7.4) to the desired concentration and TA was
suspended in this solution by brief sonication. A sample of 50 pL of this solution was then
pipetted into separate wells of a 24-well plate and the gel was activated by incubation at
37C. After several minutes, 3 mL of PBS was added and the plate was covered and
incubated with gentle agitation. At specified time intervals, the release medium was
removed for TA concentration determination and 3 mL of fresh PBS was added to the
wells. TA concentrations in the release media were measured by UV absorbance at 237 nm
against a pre-constructed standard curve (shown in Appendix B).
De-swelling studies
The de-swelling behavior of PSHU-NIPAAm gels was determined gravimetrically during
gel incubation in aqueous media. PSHU-NIPAAm gels of various concentrations were pre-
formed in 2 mL syringes (500 pL gel volume). After complete gelation, they were gently
transferred to 5 mL of pre-warmed PBS (pH 7.4) at 37C and maintained at this
temperature. At pre-determined time points, they were removed, weighed and returned to
PBS. The entire process was performed in a 37C chamber to maintain gel integrity.
Results and discussion
In this work, a novel RTG was designed, synthesized and characterized for application to
posterior ocular delivery of a poorly soluble corticosteroid. A biomimetic polymer, PSHU,
was grafted with the temperature-sensitive homopolymer PNIPAAm to produce a
77


copolymer with a high level of expected biocompatibility and a high capacity to control
release of hydrophobic drugs.
Polymer structural characterization
The structure of the backbone polymer, PSHU, before and after complete deprotection of
Boc protective groups was confirmed by *H NMR, GPC and elemental analysis. GPC
results indicated, for PSHU, Mn: 9.73 kDa, Mw: 12.5 kDa, Mz: 15.7 kDa and PI (Mw/Mn):
1.29.
The spectra of these polymers, PSHU and de-protected PSHU (dPSHU), are shown in
Figure 3.1. For PSHU, peaks associated with HDI were observed at 1.24 (-CO-NH-CH2-
CU2-CH2-CH2-I 1.40 (-CO-NH-CH2-CF/2-), 3.03 (-CF/2-CH2-NH-CO-NH), 3.11 (-CO-
NH-CF/2-CH2-), 5.28 (-CO-NF/-CH2-) and 5.53 (-CH2-NF/-CO-NH) ppm, peaks associated
with A-Boc serinol were observed at 1.34 ((C/Fifr-C-), 3.96 ((CH2)2-C7/-NH-), 4.05 (-0-
CF/2-CH-)and 4.81 (-CH-NF7-CO-) ppm and the peak associated with urea was observed
at 10.86 (-CO-N//-CO-NT/-CO-) ppm. Peak assignments for all protons were confirmed
through the use of an NMR simulation package (Advanced Chemistry Development, Inc.).
Based on calculated integral values of peaks positively associated with each monomer
(urea, A-Boc serinol and HDI), their molar ratios in the final copolymer were calculated to
understand the overall copolymer structure (see Appendix E). The molar ratios relative to
A-Boc serinol were found to be 2.13 for HDI and 0.13 for urea. Based on the monomer
ratio in the polymerization, the expected monomer ratio in the copolymer was 1:2:1 (N-
Boc serinol :HDI:urea). While the A-Boc serinol :HDI ratio was near the theoretical value,
the ratio of urea in the final copolymer was lower than expected, by a factor of 7.7. This
78


difference is likely attributed to differences in the reactivities of A-Boc serinol and urea to
HDI. Indeed, ureas have been reported to be 6.67 times less reactive than primary
hydroxyls to isocyanates[186] so the observed difference of molar ratios in the final
copolymer may be expected.
!H NMR spectra before and after the complete deprotection routine confirmed removal of
all Boc protective groups through disappearance of the peak at 1.34 ppm, thereby exposing
primary amine functionalities (Figure 3.1, Inset).
Figure 3.1: ^NMR spectroscopy confirmed the expected structure of PSHU. Inset. The
proton peak associated with Boc protective groups at 1.34 ppm in the base polymer (*) was
absent after the complete deprotection routine (**).
After complete deprotection, dPSHU contains one primary amine functionality per
repeating unit. This high density of functionalizable amine groups makes this polymer
particularly well-suited for further applications in tissue engineering (e.g. where these
groups could be used to covalently tether bioactive molecules) or specific drug delivery
applications (e.g. where covalent linking of the drug molecule to the delivery polymer
79


would be advantageous). While these have not been described herein, other work in our
lab has explored these possibilities[187].
Table 3.1: Elemental analysis results of the base polymers and resulting copolymer.
Polymer % C % H % N % O
PSHU 50.35 7.96 15.77 23.59
dPSHU 42.77 6.60 15.07 28.58
PNIPAAm 62.90 9.64 12.82 14.64
PSHU- NIPAAm 59.56 10.15 12.04 18.01
Elemental analysis of the PSHU and dPSHU starting polymers further confirmed the
expected structures, with results listed in Table 3.1 matching closely with theoretical values
(within 2%). Elemental analysis was also used to confirm the copolymer structure of
PSHU-NIPAAm. Using the results of dPSHU and PNIPAAm as bounds, the conjugation
ratio of PSHU-NIPAAm was calculated as 3.2 mol PNIPAAm/mol dPSHU. Given a Mn of
PNIPAAm of 28,128 Da (as determined from the end-group titration method), the Mn of
the resulting copolymer was calculated as 99,734 Da.
In addition, FT-IR was used to confirm successful conjugation of PNIPAAm to dPSHU
with these results shown in Figure 3.2. This was accomplished by monitoring three regions
of interest. The first, at 798 cm"1, was associated with theN-H wag of primary amines. This
peak was absent in PSHU but showed up after the deprotection routine in dPSHU and was
slightly shifted and reduced in intensity in PSHU-NIPAAm. The second, at 950 cm'1, was
associated with the O-H bend of carboxylic acids and was present in PNIPAAm-COOH
with a minute peak in PSHU-NIPAAm, indicating some small amount of unconjugated
PNIPAAm was present in the final copolymer. Finally, the broad peak near 1650 cm'1 was
associated with the C=0 stretch of amides. Of particular note is that this peak was slightly
80


shifted in dPSHU (1700 cm"1) and PNIPAAm (1650 cm'1) and indications of both amide
were present in the spectrum of PSHU-NIPAAm indicating that both species were present
in the final copolymer.
A
Wavenumbers, cm1
Figure 3.2: FT-IR analyses of the base polymers and the final copolymer suggested near-
perfect conjugation in PSHU-NIPAAm through the peaks at 1650-1700 cm'1 (A, amide
C=0 stretch), 950 cm"1 (B, carboxyl O-Hbend) and 798 cm'1 (C, primary amine N-H wag).
The presence of significant amide bonding in the PSHU backbone is expected to lend
PSHU-NIPAAm biodegradability through enzymatic and oxidative routes. Proteolytic
enzymes have been shown to be capable of breaking down amides, even within polymer
backbones[188]. Indeed, a polyurethane structurally similar to PSHU was previously
shown to degrade over several months in the presence of enzymatic activity [179],
Thermo-mechanical characterization
Investigation of thermal gelling characteristics of RTG systems have been reported by a
multitude of methods. The two most common methods are: a) measurement of the LCST
81


(or cloud point) by temperature-controlled spectroscopy; and b) measurement of the
temperature dependence of viscoelastic properties by rheometric analysis. While LCST
measurements provide a rapid, easy measurement of gelling characteristics, these are
typically performed at concentrations lower than are practically used and may be more
sensitive to polymer-solvent interactions and end-group effects. On the other hand, a
rheological study can more accurately assess the temperature-dependent development of
mechanical properties at a relevant polymer concentration.
-PNIPAAm-COOH - PSHU-NIPAAm Elastic Modulus, G' Viscous Modulus, G"
100
80
01
0
1 60
'E
w
I 40
I-
sS
20
0
20 22 24 26 28 30 32 34 36 38 40
Temperature (C)
muwmimuu
>1
-iiiiii .iwwwwt
Temperature (C)
Figure 3.3: Temperature-dependent gelling kinetics were assessed by LCST measurements
of PNIPAAm and PSHU-NIPAAm (left panel) and by rheometric analysis (right panel).
LCST values for both the PNIPAAm pre-cursor polymer and the PSHU-NIPAAm
copolymer were near 32.1 C and the sol-gel transition temperature was established to be
29.4 C.
The cloud points of PNIPAAm-COOH and PSHU-NIPAAm, as assessed by monitoring
the transmittance of a 1 wt% solution at 500 nm, are shown in Figure 3.3 (left panel). No
significant difference was noted between the two samples with a LCST of 32.1 C. While
previous reports have suggested hydrophilic end groups, such as a carboxylic acid, can
shift the LCST of PNIPAAm to higher temperatures, in the present polymer this end group
effect would likely be too small given its large molar mass (28.2 kDa). As such, the LCST
is observed near where that of true PNIPAAm has been reported. Similarly, the
82


difference that could be expected after conjugation to a hydrophobic backbone such as
PSHU would be mitigated by the molar mass imbalance between PNIPAAm and PSHU in
the final copolymer (3.2:1 PNIPAAm:PSHU).
Whereas LCST measurements were performed on dilute solutions (1 wt%), rheometric
analysis of the PSHU-NIPAAm copolymer was performed at a more application-relevant
concentration of 5 wt%. Under these conditions, a sol-gel transition temperature of 29.4 C
was observed, as shown in Figure 3.3 (right panel). The sol-gel transition temperature was
defined as the temperature at which the elastic modulus, G, surpassed the viscous
modulus, G . That this temperature, indicative of physical gel formation, is lower than the
LCST would be expected given the 5-fold difference in concentration used for each
respective analysis. At lower concentrations, the mean distance between polymer
molecules would be expected to be greater, thereby requiring more energy and time to
bring these molecules into interaction. At higher concentrations, these interactions can
begin to occur sooner and with less thermal energy, as observed in the lower sol-gel
transition temperature.
Finally, the range of PSHU-NIPAAm concentrations amenable to gel formation and
approximate gelation times are shown in Appendix F. Critically, that gel formation was
possible to concentrations as low 1.66 wt% provides a comfortable window below the
concentrations projected for typical use (5 to 10 wt%), ensuring gelation even if significant
dilution occurs after injection.
83


Cytotoxicity and biocompatibility assessments
5% DMSO Medium bRTG, 5 mg/mL RTG, 10mg/mL
***
kk
Day 1 Day 3 Day 5
Figure 3.4: ARPE-19 cells cultured in direct contact with PSHU-NIPAAm gels showed
no statistically significant difference in metabolic activity at all time points as compared to
cells cultured with complete growth medium alone at (* indicates p > 0.9, ** indicates p >
0.2). The only statistically significant difference was noted between the positive control
(cells cultured with 5% DMSO in complete growth medium) and experimental samples at
all time points (*** indicates p < 0.01).
To assess the cytotoxicity of the RTG, ARPE-19 cells were cultured in 96-well plates and
the RTG, dissolved in complete growth medium, was used to form a gel on top of the
cultured cells. Two PSHU-NIPAAm concentrations (5 and 10 mg/mL) were assessed;
these concentrations were chosen as sufficient to form a gel but low enough to be easily
recovered from the wells after the experimental time course. At 1, 3 and 5 days, the gel and
remaining medium were removed and cellular metabolic activity was assessed by MTT
assay. As shown in Figure 3.4, at all time points, no statistically significant difference was
84


noted between the negative control samples of cells cultured in complete growth medium
alone and either experimental concentration (p > 0.2). The only statistically significant
difference was observed between experimental samples and positive controls of cells
cultured with complete growth medium and 5% DMSO (p < 0.01). These data indicate
that the PSHU-NIPAAm gels are non-cytotoxic against cultured ARPE-19 cells.
Given these results, biocompatibility testing was extended to in vivo models. Rats received
intravitreal injections of 2 pL of PSHU-NIPAAm in PBS (pH 7.4) at a concentration of 2.5
wt%. The contralateral eye of each rat served as an internal control (intravitreal injection
of 2 pL of PBS, pH 7.4). At 3 and 14 days eyes were enucleated, fixed and sectioned for
histological analysis (n=3 rats per condition). Representative images of corneas (top row)
and retinae (bottom row) for control eyes and rats sacrificed at 3 and 14 days are shown in
Figure 3.5.
Figure 3.5: In vivo biocompatibility testing of PSHU-NIPAAm was carried out by
intravitreal injection in rats. Contralateral eyes served as internal controls and received
intravitreal injections of PBS, pH 7.4. Representative histological sections of corneas and
retinae from control eyes and rats sacrificed at 3 and 14 days are shown.
85


Eyes that received intravitreal injections of RTG showed no adverse reactions during the
14-day experimental time course. Visual assessments of rats indicated no signs of excessive
blinking, inflammation, hyperemia or lens or corneal opacity, as would be indicative of a
uveitic response. Histological analysis confirmed the lack of any signs of inflammatory
responses. Corneal and retinal sections were all clear of inflammatory markers including
foreign-body giant cells and mast cells. Significant macrophage infiltration was also absent
from all sections. Vitreous and aqueous humors were clear of infiltrating cells or other
markers of adverse reactions. Some tissue separation observed in the Day 3 section was
attributed to an artifact of the fixation/sectioning process as it was also present in the
control eyes for that time point.
In vitro release testing
In order to assess the ability of PSHU-NIPAAm gels to sustain release of a hydrophobic
small molecule drug, in vitro release kinetic testing was performed using the corticosteroid
TA. In the first experiment, the PSHU-NIPAAm concentration was fixed at 5 wt% and the
TA loading fraction was varied between 5 and 20 wt% (TA/RTG). In the second set of
experiments, the TA loading fraction was fixed at 10 wt% and the PSHU-NIPAAm
concentration was varied between 2.5 and 10 wt%.
In the first set of experiments, TA loading fractions up to 20 wt% showed no impact on the
ability of the RTG system to form and maintain a stable gel. Gelling times were
qualitatively observed to be similar between all samples. Release kinetics of TA from 5
wt% gels (shown in Figure 3.6, left panel) followed first-order kinetics (full modeling
details are provided in Appendix G) and were strongly dependent on the TA loading
fraction, with the rate of drug release increasing with TA loading fraction. The samples
86


Full Text

PAGE 1

A COMBINED REVERSE THERMAL GEL POLYME RIC MICELLE SYSTEM FOR SUSTAINED DELIVERY OF OPHTHALMIC DRUGS by AMIN FAMILI B.S., Lehigh University, 2009 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillm ent of the requirements for the degree of Doctor of Philosophy Bioengineering Program 2014

PAGE 2

ii This thesis for the Doctor of Philosophy degree by Amin Famili has been approved for the Bioengineering Program by Richard K. Benninger, Chair Malik Y. Kahook, Advisor Daewon Park, Advisor Tom Anchordoquy Jeff Stansbury February 11, 2014

PAGE 3

iii Famili, Amin (Ph.D., Bioengineering) A Combined Reverse Thermal Gel Polymeric Micelle System for Sustained Delivery of Ophthalmic Drugs Thesis dir ected by Professor Malik Y. Kahook and Assistant Professor Daewon Park ABSTRACT Delivery of drugs to the eye is a challenging endeavor due to the myriad anatomical and physiological barriers preventing intraocular absorption of molecules that either contac t the anterior surfaces of the eye or are present in the bloodstream. Topical administration the workhorse of ophthalmic drug delivery suffers from numerous drawbacks including poor intraocular bioavailability, excessive systemic absorption, need for frequ ent re administration and a reliance on patient adherence for therapeutic efficacy, among others. To overcome these limitations, researchers have explored various controlled release drug delivery systems with the ultimate goal of sustaining consistent, loc alized drug levels at the target tissue thereby maximizing therapeutic efficacy and minimizing unintended adverse effects. Among the options explored, in situ gelling polymeric systems may hold the most promise due to their ability to be administered direc tly at the target site by a minimally invasive injection and form a stable physical gel while conforming to the specific anatomy of that space. To date, c linical application of such systems has been hindered by their limited ability to sustain long term de livery of drugs. To overcome this limitation, we sought to develop a system comprising a reverse thermal gel (RTG) encapsulating drug loaded polymeric micelles as a combined system for the sustained local delivery of poorly soluble drugs. The polymers comp rising the RTG and the micelles both novel were first independently characterized as free standing systems. The

PAGE 4

iv combined system was then evaluated and was found to retain the injectability and in situ gelling characteristics of the thermal gel, but additio nally benefited from the superior ability of the entrapped micelles to encapsulate and sustain release of the poorly soluble corticosteroid triamcinolone acetonide (TA). Release of TA from the combined system was completely free of a burst release and was projected to continue at a steady rate for approximately twelve months. To our knowledge, this system is the first in the literature to achieve delivery time frames from an in situ gelling polymer beyond a few months and has the potential to significantly improve delivery of TA and, more broadly, clinical treatment of posterior ophthalmic diseases. The form and content of this abstract are approved. I recommend its publication. Approved: Malik Y. Kahook and Daewon Park

PAGE 5

v ACKNOWLEDGEMENTS I am grateful to all of the individuals that have helped me directly and indirectly in reaching this milestone It is a product of the countless teachers, professors, mentors, advisors, friends and family members that have contributed to my development and success and I am nothing without their contributions. My advisors, Drs. Malik Kahook and Daewon Park, have been invaluable in their support, advisement and guidance of my efforts. None of this would be possible without the m, and I am eternally grateful for their patience and unwavering backing. I also thank the members of the Department of Ophthalmology for their expertise and constant willingness to help, especially Dr. David Ammar who always took the time to help me accomplish and understand any task at hand. My lab mate s in the Translational Biomaterials Research Lab were always there when I needed them and helped me in more ways than they may realize. Two of my earliest career mentors and scientific role models Saurabh Palkar and Bill Baldy, put me on and pushed me alo ng the path that got me here today. None of what I have achieved would be possible without them and I am forever indebted to them for that. The friends I have made along the way have been a constant source of happiness, help and advice without which I woul my life and kept me going despite trials and tribulations, I express my deepest gratitude. Finally, I thank my parents, about whom enough cannot be said. It was their immeasurable efforts that instilled in me their curiosity, discipline, commitment and perseverance that have in turn, allowed me to progress in life. They have been fully supportive every step of the way and have sacrificed enormously for me. Thank you, from the bottom of my heart

PAGE 6

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ......................... 1 II. POLYMERIC MICELLES FOR SUSTAINED R ELEASE OF A POORLY SOLUBLE DRUG ................................ ................................ ........ 49 III. A REVERSE THERMAL GEL SYSTEM FOR LOCALIZED DELI VERY OF TRIAMCINOLONE ACETONIDE ................................ ... 6 9 IV. POLYMERIC MICELLES ENCAPSULATED IN A REVERSE THERMAL GEL A S AN INJECTABLE IN SITU GELL ING OCULAR DRUG DELIVERY SYSTEM ................................ ..................... 91 V. SYNTHESIS AND CHARACTERIZATION OF A BIODEGRADABLE POLY( N ISOPR OPYLACRYLAMIDE) BASED COPOLYMER ................................ ................................ ................. 102 VI. DISCUSSION, LIMITATIONS AND FUTURE DIR ECTIONS ................. 119 REFERENCES ................................ ................................ ................................ ................ 136 APPENDIX A. SYNTHETIC ROUTE O F PEG PHS PEG ................................ ................... 156 B. UV SPECTROMETRIC STANDARD CURVE FOR TRIAMCINOLONE ACETONIDE CONCENTRATION DETERMINATION ................................ ................................ ...................... 157 C. SYNTHETIC ROUTE OF PSHU NIPAAM ................................ ................. 158 D. PSHU GPC ANALYSIS ................................ ................................ ................ 159 E. PSHU MONOMER RATIO DETERMINATION ................................ ........ 160 F. PSHU NIPAAM MINIMUM GELLING CONCENTRATION DET ERMINATION ................................ ................................ ...................... 161 G. MODELING OF TA RELEASE FROM PSHU NIPAAM GELS ................ 162 H. RELEASE OF RANIBIZUMAB FROM PSHU NIPAAM GELS ............... 163

PAGE 7

vii LIST OF TABLES TABLE 1.1 INTRAOCULAR CONCENTRATIONS OF TOPICALLY ADMINISTERED TIMOLOL MALEATE ................................ .................. 8 1.2 BENEFITS OF CONTROLLED RELEASE DRUG DELIVERY S YSTEMS OVER TOPICAL OR SYSTEMIC ADMINISTRATION TECHNIQUES ................................ ................................ .............................. 17 2.1 PROPERTIES OF HIGH AND LOW MOLECULAR WEI GHT PEG PHS PEG BY GPC ANALYSIS ................................ ................................ ... 59 2.2 POTENTIAL AVENU ES FOR PEG PHS PEG MODIFICATION ............. 59 3.1 ELEMENTAL ANALYSIS RESULTS OF PSHU, DPSHU, PNIPAAM AND PSHU NIPAA M ................................ ............................... 80 E.1 1 H NMR PEAKS AND INTEGRATED AREAS USED IN PSHU MONOMER RATIO CALCULATIONS ................................ ...................... 160 F .1 CONCENTRATION DEPENDENCE OF PSHU NIPAAM GELLING ...... 161 G.1 FIT TO FIRST ORDER PARAMETERS FOR VARIOUS TA CONCENTRATION RELEASE PROFILES ................................ ................ 162

PAGE 8

viii LIST OF FIGURES FIGURE 1.1 DIAGRAMMATIC REPRESENTATION OF THE EYE ............................ 2 1.2 SIMULATED KINETICS OF PULSATILE AND CONTROL LED RELEASE DRUG DELIVERY PARADIGMS ................................ ............ 9 1.3 ROUTES OF SUB CONJUNCTI VAL AND INTRAVITREAL INJECTIONS ................................ ......................... 1 1 1.4 DISTRIBUTION OF MARKETED AND INVESTIGATIONAL DRUGS WITHIN THE BCS CLASSIFICATION SYSTEM ....................... 14 1.5 MECHANISM OF LIQUID TO PHYSICAL GEL PHASE TRANSITION FOR BLOCK COPOLYMER BASED REVERSE THERMAL GELS ................................ ................................ ......................... 29 1.6 INFLUENCE OF DRUG HY DROPHOBICITY ON ENCAPSULATION EFFICIENCY IN PNIPAAM BASED REVERSE THERMAL GELS ................................ ................................ ...... 34 1.7 NANO CARRIER DRUG DELIVERY SYSTEMS AND THEIR MECHANISMS OF DRUG ENCAPSULATION ................................ ........ 43 2.1 1 H NMR ANALYSIS OF THE PEG PHS PEG COPOLYMER .................. 58 2.2 PEG PHS PEG IS NON CYTOTOXIC AGAINST RETI NAL PIGMENTED EPITHELIAL CELLS IN CULTURE ................................ ... 60 2.3 MICELLES PRODUCED BY A FILTER EXTRUSION METHOD ARE SIGNIFICANTLY MORE MONO DISPERSE THAN THOSE PRODUCED BY A SONICATION TECHNIQUE ................................ ...... 61 2.4 MICELLES PRODUCED BY BOTH PROCESSES WERE SPHERICAL AND POSS ESSED A CORE SHEL L MORPHOLOGY ....... 63 2.5 EFFICACY OF THE MICELLE PURIFICATION PROCESS IN REMOVING DMSO FROM THE FINAL PRODUCT ................................ 64 2.6 RELEASE OF TRIAMCINOLONE ACETONIDE FROM PEG PHS PEG MICELLES WAS NEAR ZERO ORDER AND HIGHLY SUSTAINED ................................ ................................ ................................ 66

PAGE 9

ix 3.1 1 H NMR SPECTRUM OF PSHU COPOLYMER AND CONFI RMATION OF BOC DEPROTECTION ................................ .......... 79 3.2 FT IR ANALYSIS CONFIRMED SUCCESSFUL CONJUGATION OF PNIPAAM COOH TO PSHU TO PRODUCE PSHU NIPAAM ........... 81 3.3 TEMPERATURE DEPENDENT GELLING KINETICS OF PSHU NIPAAM BY OPTICAL AND RHEOMETRIC ANALYSES ..................... 82 3.4 PSHU NIPAAM IS NON CYTOTOXIC AGA INST RETINAL PIGMENTED EPITHELIAL CELLS IN CULTURE ................................ ... 82 3.5 INTRAVITREAL INJECTIONS OF PSHU NIPAAM IN RATS DEMONSTRATED ITS IN VIVO BIOCOMPATIBILITY .......................... 84 3.6 RELEASE OF TRIAMCINOLONE ACETONIDE FROM PSHU NIPAAM GELS WAS DEPENDENT ON LOADING FRACTION BUT INDEPENDENT O F PSHU NIPAAM CONCENTRATION ............. 85 3.7 DE SWELLING KINETICS OF PSHU NIPAAM GELS ............................ 87 4.1 PEG PHS PEG MICELLES CAUSED NO ADVERSE REACTIONS AFTER INTRAVITREAL INJECTIONS IN RATS ................................ .... 96 4.2 RHEOLOGICAL ANALYSIS OF RTG ALONE AND THE COMBINED RTG MICELLE SYSTEM ................................ ...................... 97 4.3 RELEASE OF TR IAMCINOLONE ACETONIDE FROM THE COMBINED RTG MICELLE SYSTEM WAS SUPERIOR TO THE SAME FROM RTG OR MICELLES ALONE ................................ ............. 99 5.1 PROPOSED DEGRADATION ROUTES FOR PSHU NIPAAM COPOLYMERS ................................ ................................ ............................. 109 5.2 MASS LOSS OF PSHU NIPAAM GELS IN PBS AND CHOLESTEROL ESTERASE SOLUTIONS ................................ ............... 110 5.3 FT IR SPECTRAL ANALYSIS OF PSHU NIPAAM GELS IN PBS AND CHOLESTEROL ESTERASE SOLUTIONS ................................ ..... 111 5.4 QUANTIFICATION OF FT IR SPECTRA CONFIRM THAT CARBOXYLIC ACIDS WERE NOT GENERATED AND SIGNIFICANT AMIDE CLEAVAGE DID NOT OCCUR DURING INCUBATION IN CHOLESTEROL ESTERASE SOLUTIONS ................ 112

PAGE 10

x 5.5 MPA CONCENTRATION DICTATES HO PNIPAAM COOH MOLECUALR WEIGHT AND INCREASING MOLECULAR WEIGHTS DECREASE THE LCST ................................ ............................ 114 5.6 CONJUGATION OF HO PNIPAAM COOH TO PSHU YIELDED A COPOLYMER WITH AN LCST BELOW BODY TEMPERATURE ........ 1 1 5 5.7 ACCELERATED DEGRADATION T ESTING OF PSHU NIPAAM AND PSHU NIPAAM OH GELS IN HCL AND PAPAIN SOLUTIONS ................................ ................................ ................................ 116 6.1 HYDROLYTIC AND ENZYMATIC DEGRADATION PATHWAYS OF CARBOXYL DERIVED AMIDES ................................ ........................ 126 6.2 HISTOLOGICAL SECTIONS OF A CRUSHED OPTIC NERVE 7 DAYS AFTER RGD FUNCTIONALIZED PSHU NIPAAM INJECTI ON SHOWED OPTIC NERVE DAMAGE BUT NO FUNCTIONAL RECOVERY ................................ ................................ ....... 134 A.1 PEG PHS PEG FULL SYNTHETIC ROUTE ................................ .............. 156 B.1 STANDARD CURVE FOR DETERMINATION OF TA CONCENTRATION VIA UV SPECTROSCOPY ................................ ....... 157 C.1 PSHU NIPAAM FULL SYNTHETIC ROUTE ................................ ............ 158 D.1 RESULTS OF GPC ANALYSIS O F PSHU ................................ ................. 159 E.1 1 H NMR SPECTRUM OF PSHU WITH INTEGRAL VALUES FOR PEAKS USED TO CALCULAT E MONOMER RATIOS ........................... 160 G .1 FIRST ORDER KINETIC MODELING OF TA RELEASE FROM PSHU NIPAAM GELS ................................ ................................ ................. 162 H.1 SELECTIVE BOC REMOVAL FROM PSHU WAS ACHIEVED BY MILD ACID CATALYZED HYDROLYSIS ................................ ............... 163 H .2 PSHU NIPAAM GELS RE LEASED THEIR ENTIRE RANIBIZUMAB LOAD WITHIN 24 HOURS ................................ ............ 164

PAGE 11

xi ABBREVIATIONS ACA AZOBIS(4 CYANOVALERIC ACID) AMHP 2,2' AZOBIS[2 METHYL N (2 HYDROXYETHYL)PROPIONAMIDE] ANOVA ANALYSIS OF VARIANCE AUC AREA UNDER THE CURVE BCS BIOPHARMACEUTICS CLASSIFICATION SYSTEM BSA BOVINE SERUM ALBUMIN CMC CRITICAL MICELLE CONCENTRATION CMT CRITICAL MICELLIZATION TEMPERATURE DCM DICHLOROMETHANE DLS DYNAMIC LIGHT SCATTERING DMEM DULBECCO'S MODIFIED EAGLE'S MEDIUM DMF DIMETHYLFORMAMIDE DMSO DIMETHYL SULFOXIDE EDC N (3 DIMETHYLAMINOPROPYL) N ETHYLCARBODIIMIDE HYDROCHLORIDE

PAGE 12

xii FBS F ETAL BOV INE SERUM FESEM FIELD EMISSION SCANNING ELECTRON MICROSCOPY FT IT FOURIER TRANSFORM INFRARED GFAP G LIAL FIBRILLARY ACIDIC PROTEIN GPC GEL PERMEATION CHROMATOGRAPHY HDI HEXAMETHYLENE DIISOCYANATE HPMC HYDROXYPROPYL METHYLCELLULOSE IOP INTRAOCULAR PRE SSURE ISO INTERNATIONAL ORGANIZATION FOR STANDARDIZATION LCST LOWER CRITICAL SOLUTION TEMPERATURE MC METHYLCELLULOSE MPA 3 MERCAPTOPROPIONIC ACID MPEG METHOXYPOLYETHYLENE GLYCOL MTT 3 (4,5 DIMETHYLTHIAZOL 2 YL) 2,5 DIPHENYLTETRAZOLIUM BROMIDE MW MOLE CULAR WEIGHT

PAGE 13

xiii MWCO MOLECULAR WEIGHT CUT OFF NHS N HYDROXYSUCCINIMIDE NMP N METHYL 2 PYRROLIDONE NMR NUCLEAR MAGNETIC RESONANCE P/S PENICILLIN/STREPTOMYCIN PAA POLY(ACRYLIC ACID) PBS PHOSPHATE BUFFERED SALINE PCL POLY( CAPROLACTONE) PEG POLYETHYLENE GLYCOL PEO POLY(ETHYLENE OXIDE) PHS POLY(HEXAMETHYLENE ALT SERINOL) PI POLYDISPERSITY INDEX PLA POLY(LACTIC ACID) PLGA POLY(LACTIC CO GLYCOLIC ACID) PNIPAAM POLY( N ISOPROPYLACRYLAMIDE) PPO POLY(PROPYLENE OXIDE) PSHU POLY(SERINOL HEXAMETHYLENE UREA )

PAGE 14

xiv RI REFRACTIVE INDEX RPE RETINAL PIGMENTED EPITHELIAL RSD RELATIVE STANDARD DEVIATION RTG REVERSE THERMAL GEL TA TRIAMCINOLONE ACETONIDE TEM TRANSMISSION ELECTRON MICROSCOPY TFA TRIFLUOROACETIC ACID UV ULTRAVIOLET VPT VOLUME PHASE TRANSITION

PAGE 15

CHA PTER I INTRODUCTION and challenging endeavours facing the pharmaceutical scientist [1] The challenge is defined by a multitude of factors including: a) the unique physiological barriers present within the eye; b) the sensitive nature of intraocular tissues which are easily affected by exposure to foreign molecules; and c) t he unique anatomical structures of the eye which make accessing intraocular targets relatively difficult. These challenges exist because of and in spite of the apparent ease of external access to the eye. Because of this unique feature, the use of eye drop s (i.e. topical administration of ocular therapeutics) remains by far the single most commonly employed delivery modality. As will be explored herein, this modality inadequately addresses the needs and wants of the patient To this challenge, researchers h ave responded with great innovation and creativity. Most commonly, polymeric systems both established and novel have been employed for their relative ease of design, synthesis and characterization, their cost effectiveness and their superior flexibility. I n this work, two such polymeric drug delivery systems have been developed. The first, a reverse thermal gelling polymer, is an injectable, in situ gelling aqueous polymer that addresses the issues of access and target retention required of ocular drug deli very systems. The second, a polymeric micelle system, is an advantageous nano carrier system that addresses the challenges of delivery of poorly soluble drug molecules and their long term sustained release. The synergistic combination of these two systems a reverse thermal gel containing drug loaded polymeric micelles uniquely addresses the

PAGE 16

2 design criteria required of a translatable ocular drug delivery system and has the potential to significantly improve the treatment of several ocular diseases. In this i ntroduction, the challenges of ocular drug delivery will be presented followed by the role of ocular drug delivery systems in addressing these needs. Finally, the principles and mechanisms of reverse thermal gelling polymers and polymeric micelle systems w ill be described in addition to the limitations of current systems. Within these contexts, the solution proposed by the work undertaken herein will be explored. Ocular Drug Delivery Anatomical and physiological c hallenges The eye can broadly be separated i nto the anterior and posterior segments. The anterior segment is composed of the structures that lie anterior to the vitreous humor : the cornea, iris, ciliary body and lens, while the posterior segment includes the vitreous humor retina, choroid and optic nerve. These structures are illustrated in Figure 1. 1. [2] Figure 1. 1 : Diagrammatic representation of the eye with major structures labeled. [2]

PAGE 17

3 S imilarly, ocular diseases are generally classified as occurring primarily either in anterior or posterior segment tissues. In practice, there is a major difference in how anterior and posterior diseases are therapeutically addressed. These differences are based primarily on the significant difference in bioavailability at anterior and posterior tissues, respectively, of drugs adm inistered topically. In order to appreciate this difference, an analysis of the defining anatomical and physiological characteristics of anterior and posterior structures is necessary. The anterior segment, consisting of tissues that are directly exposed t o environmental conditions when the eye is open, has developed a multitude of barriers to ensure pollutants and potential toxins are not able to penetrate the eye and reach intraocular tissues where significant damage could occur At the foremost anterior surface of the eye is the tear film, which is a mucin containing liquid layer that forms a protective hydrophilic layer covering the cornea. The tear film has evolved to rapidly trap and clear pollutants from the anterior surface of the eye. As a result, its minimal volume of 7 10 L is turned over at a rate of approximately 10.5 % per minute [3] thereby reducing the r etention time of molecules on the ocular surfaces to only several minutes. Further, irritation of the ocular surface (e.g. by foreign objects or molecules that disrupt the tear film ) serves as an impetus to further increase lacrimation and thereby the tear film turnover rate. Beyond the tear film lies the cornea, which serves as the first mechanochemical barrier to ocular penetration of exogenous substances. It is primarily divided into three layers: the epithelium, stroma and endothelium. The anterior most layer the epithelium is a highly cellularized hydrophobic membrane which contains 90% of the total cells within the cornea. At the superficial most layer, these cells are connected by desmosomes and

PAGE 18

4 surrounded by tight junction complexes. As a result part icularly of these tight junctions, pericellular penetration of exogenous molecules through the epithelium is significantly retarded. Posterior to the corneal epithelium lies the stroma, which is a highly hydrated matrix of collagen fibrils. Comprising appr oximately 90% of the corneal thickness, the hydrophilic stroma presents a major challenge for hydrophobic drugs to partition across. The posterior most layer of the cornea is the endothelium, which provides minimal retardation of the passaging of substance s. Because of its location at the interface of the cornea and the aqueous humor the corneal endothelium is deliberately leaky to allow diffusion of macromolecules from the aqueous humor into the corneal stroma and vice versa. Based on the alternating hydr ophobic hydrophilic hydrophobic structure of the cornea, it can be understood that a drug molecule must be amphipathic in order to be able to partition across the entirety of the cornea. During their transit across the cornea, drug molecule s are also susce ptible to degradation and ad sorption via enzymes and proteins present in that tissue. The cornea is especially rich in esterases, which can hydrolyze susceptible drug molecules as they passage across the tissue [4] Specific esterases ( identi fied via study of the rabbit cornea ) include cholinesterase, acetylcholinesterase, carboxylesterase, acetylesterase, arylesterase and a non specific esterase [5] Many topical administration strategies have used the presence of these esterases to their advantage, designing pro drugs that are metabolized int o their active form upon hydrolysis. One example is latanoprost, which is administered as an isopropyl ester that is converted to its biologically active free acid form by endogenous enzymes. Since the isopropyl ester form also has better corneal permeabil ity than the free acid, this pro drug strategy improves intraocular bioavailability.

PAGE 19

5 Molecules that can successfully partition across the multitude of corneal barriers will reach the aqueous humor from which they can access and be distributed to surround ing tissues such as the iris, ciliary body, lens, vitreous humor and choroid/retina. However, several mechanisms also exist to prevent uninhibited access to these tissues Aqueous humor turnover and blood circulation in the anterior uvea are both major fac tors that physically remove drug molecules from the anterior chamber. Of a total aqueous humor volume of 300 L in humans half of this volume is turned over approximately every 0.77 hours [6] In additi on, it has been suggested that metabolic pathways contribute to drug clearance from the aqueous humor thereby accelerating clearance [6] However, counter to these mechanisms, drugs in the aqueous humor have the ability to reversibly bind to tissues, which can help prolong their residence times. In particular, the lens, vitreous humor and especially the pigmented uvea have been found to participate in this binding process. Non corneal drug absorption rou tes after topical administration can also be considered, e.g. through the conjunctiva or sclera. The conjunctiva in humans has 17 times greater surface area than the cornea [7] theoretically presenting a major potenti al route for drug absorption. However, drug penetration across the conjunctiva is severely limited. The primary mechanism for molecule exclusion is the presence of tight junctions in the superficial conjunctival epithelium, which act similarly to those in the corneal epithelium, but may be considered slightly more leaky [8] In addition, the presence of conjunctival cap illaries and lymphatic vessels efficiently clear molecules that do achieve some level o f absorption. The sclera on the other hand, represents a major absorption pathway. Due to its importance in maintaining the structural integrity of the globe, the sclera is a dense, disordered network of collagen fibrils. Significant amounts of proteoglyc ans and polysaccharides are also

PAGE 20

6 present, making the sclera a highly hydrated tissue similar to the corneal stroma. As a result, scleral permeability of small molecule drugs is generally significantly higher than corneal permeability, by a factor of 1.2 11 for several common molecules [9,10] Drugs administered by the other majo r administration route s systemic or oral administration also face a daunting challenge in reaching intraocular targets. Drugs administered by these techniques must reach ocular tissues from the bloodstream in order to affect treatment. In the anterior segm ent, the blood aqueous barrier limits passage of drugs, while the blood retinal barrier accomplishes this in the posterior segment. The blood aqueous barrier consists of two cell layers the endothelium of the blood vessels in the iris and ciliary body and the non pigmented ciliary epithelium expressing tight junctions that efficiently prevent the entry of molecules into the aqueous humor [11] The blood retinal barrier is maintained by retinal capillary endothelial cells and retinal pigmen ted epithelial (RPE) cells. The RPE cell layer is situated between the neural retina and the choroid and is chiefly responsible for regulating selective transport of molecules between photoreceptors in the retina and capillaries in the choroid. However, ti ght junctions between RPE cells heavily restrict intercellular transport. The role of the RPE cell layer cannot be understated in protecting the highly sensitive underlying neural retina. The highly vascularized choroid contains fenestrated capillaries, wh ich allow equilibration of plasma drug concentrations with t he choroidal extravascular space. The RPE cell layer is then tasked with excluding molecules in this space from reaching the neural retina. The previous l y mentioned tight junctions efficiently ser ve this purpose and make this transport mechanism almost entirely non productive [12]

PAGE 21

7 Conventional drug delivery modalities and their limitations The net effect of thes e anatomical and physiological barriers to reaching the eye is that traditional delivery techniques suffer from less than desirable distribution kinetics. For example, in the case of topical administration which accounts for 88% of the ocular therapeutics market less than 5% of the administered dose is able to reach the aqueous humor [13] The result is that the dose administered in each eye drop must be at least 20 fold higher than the concentration needed at the aqueous humor M uch of that excess drug is cleared through systemic absorption, especially by conjunctival vessels a nd the highly vascularized nasal mucosa, where drugs can end up after clearance through the nasolacrimal ducts [14] These mechanisms so rapidly and efficiently cl ear topically administered drugs that peak plasma concentrations are typically reached as soon as 5 minutes after topical instillation [15] Systemic absorption has been implicated as a factor in many observed side effects of topically administered drugs, especially when the patient has underlying cardiovascular or resp iratory diseases or is on other concomitan t medications [16 18] The estimated 5% of the instilled dose that is available for intraocular distribution is true only for anterior segment targets t he picture becomes even bleaker when posterior tissues are concerned. A s tudy in 1990 by Kyyrnen and Urtti [19] investigated the distribution of timolol maleate a highly soluble, small molecule beta blocker in various intraocular tissues after topical administration. Their findings are summarized in Table 1.1 They found that as the drug partitioned further to posterior targets, the concentrations at those tissues decreased by approximately one order of magni tude per barrier. Of the 125 g/g instilled dose, the concentration was 3 orders of magnitude lower in the vitreous humor (and

PAGE 22

8 therefore avai lable for posterior efficacy). For this reason, treatment of posterior ocular diseases has especially suffered from the under performance of topical administration. Table 1. 1 : Intraocular concentrations of topically administered timolol maleate [19] Location Concentration (g/g) Instilled dose 125 Cornea 17.38 Aqueous humor 2.16 Lens 0.121 Sclera 2.66 Vitreous humor 0.083 Another major drawback of topical or systemic administration regimens is the pulsatile nature of the drug concentration at the target tissue. When a dose of drug is administered by either method, the drug follows the classic pharmacokinetic paradigm of absorption distribution metabolism excretion. Because doses are administered at discrete time points at prescribed interval s tissue drug concentrations typically fluctuate with time This is shown schematically in Figure 1. 2 The reason this kinetic is less than ideal is that for each drug and target tissue, a therapeutic window exists based on a lower bound defined by the minimum effective concentration and a maximum bound defined by the maximum tolerated dose. For many drug dosage regimes, periods of time will be spent below and above these bounds, respectively. Tissue drug concentrations above the max imum tolerated dose are toxic to the tissue and result in many of the side effects attributed to specific drugs. Those below the minimum effective concentration are not therapeutic and result in loss of efficacy. Both scenarios result in less than ideal pa tient outcomes

PAGE 23

9 Perhaps the greatest challenge of topical or systemic administration regimens is the reliance on patient adherence in determining therapeutic outcomes. Patients that are not compliant with their dosing regimen have the potential to: a) s uffer continuing vision loss due to lost therapeutic efficacy [20,21] ; and b) cause an unnecessary change by their physician in their medical treatment plan as a result of perceived therapeutic inefficacy [22,23] As a res ult of the magnitude of these findings, scores of studies have evaluated all aspects of patient compliance to identify specific problematic behaviors One such study [23] found a multitude of common pro blems including patients compressing doses during the day and spacing them out at night, thus entirely missing doses and causing short or long term interruptions in the medication. These findings suggest a combination of motivations including convenience (taking more doses during the day and fewer at night) and Figure 1. 2 : Simulated kinet ics of pulsatile and controlled release drug administration paradigms demonstrate the inconsistent tissue drug concentrations in pulsatile (e.g. topical or systemic administration) techniques due to discrete time point dosing. In contrast, an ideal controlled release system should be able to maintain consistent tissue drug concentrations over a long period of time.

PAGE 24

10 forgetfulness (missing doses), but improvements were noted when education addressing the impact of compliance was provided to patients [24] Ho wever, a major challenge in many ocular diseases is that they are largely asymptomatic until the later stages, when significant irreversible damage has already set in. T his disconnect between medication and perceived symptoms may significantly contribute t o non compliance [25] Studies quantitating patient dosing behavior have further confirmed the magnitude of the challenge of non compliance. One study found that during a four to six week period, nearly 25% of patient s had at least one day per month without any administered doses and 15% of patients took less than one half of prescribed doses [26] Another found that over 3 months of prostaglandin analog treatment, the overall mean adherence rate was 71%, with nearly 45% of patients administering drops less than 75% of the time [27] The primary obstacles reported in patient surveys were situational/environmental factors (e.g. being away from home o r changes to daily routines) and complexities of the medication regimen [25] As a result, significant efforts have been invested in developing systems that can remove this burden from the patient and, thereby, improve medication efficacy and therapeutic outcomes, as will be detailed in later sections. Alternative delivery modalities One strategy that continues to receive significant attention is administration of the drug (by itself or encapsulated within a drug deliver y system) directly at an intra or periocular target. This method represents the simplest and most direct means of bypassing the myriad barriers preventing intraocular penetration and bioavailability of drug molecules. For example, if the drug can be admin istered directly to the aqueous humor (e.g. by means of a microneedle inserted through the cornea [28,29] ), the drug will avoid all of the

PAGE 25

11 permeability and cl earance concerns of the cornea and gain immediate access to intraocular targets. These alternative administration techniques are especially promising for improving posterior drug bioavailability for the reasons previously described. The techniques that are curren tly receiving the most attention are subconjunctival delivery, sub s capsule delivery and intravitreal delivery. Pictoral representations of these routes are shown in Figure 1.3. Subconjunctival delivery of drugs or drug delivery systems bypasses b oth the cornea and the conjunctiva, allowing improved drug levels both in the anterior space and the vitreous humor [30,3 1] Improved p osterior drug levels are likely related to transport by the uveoscleral outflow pathway [32] While subconjunctival delivery has proved promising for drug depot formation, only drug delivery implants with relatively high release rates Figure 1.3: A diagrammatic repr esentation of the location of sub conjunctival and intravitreal injection routes.

PAGE 26

12 were able to achie ve and maintain therapeutic choroidal and retinal drug levels [33,34] Amrite, et al. [35] found that neither nano nor micro particle formulations were able to sustain detectable levels of a fluorescent marker in the posterior tissues follow ing subconjunctival administration. The sub The major advantage of this space is that drugs or implan ts administered here are placed in direct contact with the sclera, allowing for transscleral access to the choroid and retina. Additionally, since the vitreous humor is not penetrated by this route, adverse effects associated with intravitreal injections ( e.g. retinal detachment and endophthalmitis) are far less likely. One drawback of sub retinal drug levels have not been demonstrated and in fact the distribution of drug across the retina likely depends on the location of injection and variations in scleral thickness [36] For this reason, sub injections may be better suited for localized lesion treatment. Fernandez, et al [37] implanted several biomaterials within this space and found that while hydrophilic materials were well tolerated, hydrophobic ones triggered inflammation and fibrosis, resulting in complete fibrotic encapsulation within 12 weeks Intravitreal administration is the most common approach for achieving therapeuti c posterior drug levels. The injection is performed by a minimally invasive intrusion through t he pars plana to reduce trauma. Ideally, use of a needle gauge of 25 or smaller is desired to provide a self healing wound (i.e. one that does not require suturi ng) [36] Nevertheless, t he rate of complications can be fairly high for intravitreal injections especially those requiring repeated administration including endophthalmitis, hemorrhage and retinal

PAGE 27

13 detachment [38] Because of the large volume of the vitreous humor, it is often used as a space to form depots of drug f or long term treatment courses. This strategy is particularly useful for poorly soluble drugs and those that can be crystallized as these will slowly dissolve upon depot formation in the v itreous [39,40] However, drugs that are not fully solubilized can cause ob structions in the visual field, making t hem less favorable to patients. Intravitreal injections have also been commonly employed for delivery of drug delivery systems, including nanoparticles [41] microparticles [42,43] bulk polymeric implants [44 46] and refillable devices anchored in the sclera [47,48] A major consideration for intravitreal devices, however, is th at drug diffusion rates are significantly lower in the viscous vitreous humor resulting in less even drug distribution throughout the posterior space [49,50] Drug molecule considerations The Biopharmaceutics Classification System (BCS) has been successfully employed for decades in the evaluation of new therapeutic molecule s It differentiates drug molecules b ased on two critical parameters: solubility and permeability (in particular, human intestinal permeability). These parameters fundamentally control the rate and extent of drug absorption and are also important determinants in the complexity of formulation and delivery of the molecule [51] Drugs are placed in one of four categories: high solubility and high permeability (Class 1), low solubility and high permeability (C lass 2), high solubility and low permeability (Class 3) or low solubility and low permea bility (Class 4). As drug formulation progresses, Class 1 drugs are cherry picked due to their ease of development, the result being that few Class 1 molecules remain f or development. Many of the molecules undergoing development today are less soluble or less permeable than

PAGE 28

14 ever before, necessitating the development of more sophisticated drug delivery methods. In 2007, a meta analysis of drug molecules and their BCS clas sifications was undertaken [52] In particular, this study looked at drugs that were currently marketed at that time versus those that were under investigation for marketing (new chemical entities, NCEs). These data are summarized in Figure 1. 4 This analysis highlighted a major shift in the industry from a lands cape dominated by Class 1 molecules to one dominated by Class 2 and Class 4 molecules. In short, this represented a move from generally soluble drug molecules to poorly soluble ones. One major consequence of this shift, especially as it relates to ocular therapeutics, is that it complicate s the picture of drug formulation and delivery. Because the vast majority of ocular drugs are delivered topically, they must be able to be readily formulated as such. Since topical formulations must be aqueous systems s olubility plays a critical role in this process. This limits the delivery options for poorly soluble drugs, which, as discussed earlier, are rapidly becoming a large part of the industry. Figure 1.4 : A major shift was noted in the distribution of drugs within the BCS classification system in 2007, with many more poorly soluble NCEs than m arketed drugs.

PAGE 29

15 Efforts to improve ocular bioavailability of poorly soluble drugs ha ve included many strategies, but only ophthalmic suspensions have had a major impact on marketed drugs However, even suspensions suffer from several drawbacks that limit more extensive clinical use. As described previously, patient compliance is a major c hallenge in therapeutic success. Due to the additional requirement that suspension eye drop bottles be shaken before dosing, they have been found to further reduce patient compliance [53] In addition, the actual gain in bioavailability from suspensions does not correlate with the additional amount of drug present in those systems. For instance, the drug hydrocortisone administered as a suspension with a drug conten t 33 fold higher than its solubility limit only increased intraocular bioavailability of the drug 5 fold, as compared to a saturated solution [54] An alternative strategy to improve the intraocular bioavailability of poorly soluble drugs is to form a water soluble derivative of the drug of interest. This is most often accomplished by for ming a salt of the parent molecule [55] The salt havi ng increased aqueous solubility, now allows greater concentrations to be instilled by a topical drop. However, it must be noted that this effect is at least partially counter ac ted by the now lower permeability of the drug. As was described earlier, the efficient corneal epithelium dictates that moderately lipophilic molecules have the best corneal permeability. Salt formation reduces the lipophilic nature of the molecule making it less permeable across the cornea. However, the increased driving concentration is generally sufficient to overcome this loss in permeability [56] Of course, not all drug molecules are amenable to salt formation so this strategy is limited in its applicability.

PAGE 30

16 Drug Delivery Systems A fundamentally different approach to this problem involves the application of drug delivery systems. These systems are designed to address the various shortcomings of traditional drug administration techniques and accomplish this task in significantly different ways, as will be described in this section. Motivation s and c onsiderations As discussed in previous sections, the major limitations of topical and systemic ophthalmic drug delivery are poor in traocular bioavailability, need for frequent re administration, inconsistent tissue drug concentrations and a strong dependence on patient compliance for therapeutic efficacy. The motivation for advanced drug delivery systems stems from the need to overcom e these barriers and thereby reduce patient and provider burdens and improve patient outcomes. The principle of controlled release drug delivery is that a system can be designed such that a reservoir of drug can be deployed near the tissue of interest and in a controlled fashion release drug from the reservoir over a period of time. Theoretically, this system would overcome the previously mentioned limitations in the ways described in Table 1.2 In order to achieve these goals, however, the system must be d esigned with several key criteria, many of which are not met by commercially available drug delivery solutions. For example, in order to improve intraocular bioavailability and bypass the complex anatomical and physiological barriers described earlier, the system must be packaged in a delivery system amenable to ophthalmic deployment. Given the compact and sensitive

PAGE 31

17 structures of the eye and the limited number and size of potential spaces, this task is more challenging in the eye than potentially any other tissue. Table 1. 2 : Controlled release drug delivery systems are theoretically able to overcome many of the limitations of topical or sys temic administration techniques Given these challenges, the biodegradability of the system also becomes a key criteria. Since potential spaces are few in number and small in size, if a controlled release system is deployed into one of these spaces, the system must be cleared from that space by the time another system needs to be delivered. Take for example a syste m that is deployed into the sub conjunctival space of a glaucomatous patient to deliver a drug that reduces intraocular pressure (IOP). Since glaucoma is a ch ronic disease, therapeutic levels of this system can provide such therapeutic levels for a maximum of 6 months, a replacement system would need to be administered at or near the 6 month time point. Since the sub conjunctival space is limited, the first system would need to be largely cleared from that space in order to allow deployment of the replacement system. This requirement then Limitation of topical/systemic administration Benefit of control led release system Poor intraocular bioavailability System can be deployed at or near the target tissue, overcoming physiological barriers Need for frequent re administration System can be loaded with sufficient drug to maintain therapeutic levels for mo nths or years Inconsistent tissue drug concentrations System can be tailored to achieve drug kinetics such that tissue drug levels are more consistent Dependence on patient compliance System can sustain drug release for long periods of time without patie nt/provider action

PAGE 32

18 necessitates that biodegradation o f the system be on a time scale similar to its maximum time frame for drug release. A successful controlled release system must also be able to precisely and reproducibly control the kinetics of drug release from the system. In order to allow a long usable life span for the system, a large reservoir of drug is typically loaded into the system. The larger this reservoir can be designed, the longer the system can theoretically provide therapeutic levels of drug release. In the ophthalmic field, this time peri od should ideally be 3 6 months o localized drug levels above the maximum tolerated dose also increases. The kinetics of drug release from (and biodegradation of) the syst em must be carefully controlled to prevent these potentially harmful possibilities. In order to understand the importance of release kinetics, a brief overview of the models and terminology associated with the to pic will be helpful. As drug molecules migra te from their position at the interior of the system to the surface and then into the surrounding medium, various factors can be involved in this process. The purpose of drug release kinetic modeling is to gain insight into which factors are dominant on th e time scale relevant to drug release. The first system that can be modeled is one in which the process of drug release from the system is independent of the drug concentration a zero order kinetic In this case, the concentration, C at time, t can be e xpressed as:

PAGE 33

19 where K 0 is the zero order rate constant and C 0 is the initial drug concentration. This kinetic ems where a bulk of drug is surrounded by a rate limiting membrane. process is directly proportional to the drug concentration, a first order kinetic. In this case, the co ncentration with time is expressed as: where K 1 is the first order rate constant. Note that in this case the process occurs as a proportion of the drug concentration existing at that time (and therefore the initial drug concentration) and so the release rate will decreas e with time (as the remaining drug concentration decreases) and will theoretically never reach completion. The Higuchi model was derived specifically to describe drug release from a planar matrix system and is based on the hypotheses that a) the initial dr ug concentration in the matrix is much greater than drug solubility in the release medium; b) drug diffusion takes place only swelling and dissolution are negligibl e; e) drug diffusivity is constant; and f) sink conditions are always maintained in the release medium. The simplified equation describing the Higuchi model is: where Q is the amount of drug released in time t and K H is the Higuchi rat e constant.

PAGE 34

20 The final relevant model is the Korsmeyer Peppas model, which is described by the equation: where M t /M is the fraction of drug released at time t K KP is the Korsmeyer Peppas rate constant and n is the relea se exponent. This release exponent can be used to characterize the release mechanism, where n = 0.5 indicates Fickian diffusion, 0.45 < n non Fickian transport, n = 0.89 indicates Case II transport and n > 0.89 indicates Super Case II tran sport. These relationships hold true for the first 60% of drug release. Polymeric drug delivery systems Within the field of drug delivery, polymer based systems are by far the most commonly investigated. Polymers are long chain molecules that consist of a large number of small repeating units [57] By selecting appropriate repeating units and thereby controlling the final chemistry of the molecule, polymers can be designed with a wide variety of properties, as required by the specific application. This immense flexibility is the primary reason polymers are so commonly employed as biomaterials in general and drug delivery systems in particular For a polymer system to be amenable to application as a drug deliver y system, it must meet several critical design criteria. First and chief a mong these is biocompatibility, which defines the interactions between the host tissue and the implanted material This interaction is two way and involves both the effect of the mat erial on the host tissue and that of the host tissue on the material. Most importantly, the implanted material must not cause

PAGE 35

21 significant toxicity, inflammation, infection, tumorigenesis or any other adverse effect in the patient [58] In addition, the drug delivery system must be biodegradable. This entails both that the device should be cleared by biological mechanisms with in a pre defined timeframe, but also that its degradation products meet all of the biocompatibility criteria set forth earlier [59] In a drug delivery system, the ideal situation is that biodegradation occurs on a similar timescale to complete drug release. While a drug delivery system will need to meet all of the additional criteria common to biomaterials in general [60] one additional criterion is that any interactions between the drug delive ry vehicle and the encapsulated drug molecules do not alter the therapeutic efficacy of the drug. This is especially crucial for structure sensitive drugs such as antibodies and proteins, which can be degraded rapidly in the presence of certain stimuli [61 64] In situ gelling injectable systems One specific class of polymeric drug delivery systems that are especially useful as ophthalmic drug delivery systems is the in situ gelling i njectable system. This type of system is characterized by a transition from a low viscosity liquid to a physically or chemically cross linked gel upon introduction of a specific stimulus From a practical standpoint, this system is advantageous in that its initial liquid state allows it to be deployed by a minimally invasive injection through a syringe immediately at the desired anatomical location. Upon injection, the system is designed to deploy to its gel state allowing it to be retained at the injection site, resisting dilution or washing away by local fluid flows. Transition to a gel state also allows entrapment of incorporated drug molecules and control over their release from the system. In this way, release kinetics similar to bulk polymer systems ca n be attained but with the added benefit of injectability. In situ gelling systems

PAGE 36

22 commonly employed in drug delivery systems can be categorized into two major modes of sol gel behavior: in situ gelling systems and in situ polymer precipitation. In situ g elling systems are generally achieved by injecting either: a) a pre mixed monomer and initiator mixture or solution that polymerizes upon injection; or b) polymers that have reactive functiona lities amenable to crosslinking with or without a cross linking agent The impetus driving the polymerization or cross linking reaction upon injection can be provided by thermal energy (e.g. using body temperature or an external heat sour ce), the application of light (e.g. using a n ultraviolet ( UV ) probe) or the presen ce of ions An advantage of these type s of system s is that the cross linking density can be relatively precisely controlled by modifying the ratio of components at injection. Since highly crosslinked systems are theoretically possible, the kinetics of drug release can be favorably controlled. In addition, the pre crosslinked system can be composed of sufficiently low molecular weight (MW) polymers to make a low viscosity solution amenable to injection. An example of a temperature driven system was developed by Moore, et al. [65] in which an acrylate terminated cop olymer of lactide and caprolactone is mixed with a thermal initiator (benzoyl peroxide or N,N dimethyl p toluidine) immediately prior to administration. Upon heating to body temperature, a redox reaction drives polymerization of the system, which entraps pre mixed drug molecules. While release kinetics of flurbiprofen from the system were favorable, the system suffers from several distinct disadvantages. First is the toxicity concerns of direct injection of free radical producing species (i.e. the initiat ors), which have commonly been shown to be tumorigenic [66] As a result, initiator concentrations must be kept low, which consequently results in long gelation times. In the previously mentioned system, full gelation was reported to take

PAGE 37

23 between 5 and 30 minutes. During this time, encapsulated drug molecules can readily diffuse out of th e lightly crosslinked system resulting in a large burst of drug release within this time. Finally, many of the polymerization and/or cross linking reactions used in these systems are highly exothermic in nature resulting in significant heating of surroundi ng tissues. One such system reported local tissue temperatures as high as 94C [67] which can cause significant tissue necrosis around the injection site [68] Photo polymerizable systems may overcome some of the disadvantages of thermal initiated systems [69,70] The mechanism of injection and crosslinking is very similar between the two systems, but the additional control provided by photo polymerizable systems makes them more practical for controlled release systems. The prima ry advantage is that curing times can be much shorter than in thermal initiated systems, allowing for full gel formation within several seconds when lasers are employed for curing [71] While this eliminates the issue of high initial bur st release, the rapid curing and extent of polymerization has been reported to result in significant shrinkage and brittleness in the material In one application, a photo polymerized polyethylene glycol co poly(lactic acid) (PEG PLA) hydrogel was investig ated for controlled release of various therapeutic proteins [72] While drug release was favorable wit h minimal burst and diffusion controlled kinetics, complete drug release was realized within only 5 days for the molecules stu died. In addition, the concerns surrounding direct injection of reactive species remain in these systems. One final class of in si tu crosslinked systems are those driven by ion mediated gelation. These systems almost exclusively use the natural class of polymer s called alginates, which form a gel upon contact with divalent cations (e.g. calcium ions) [73] Alginates are natural

PAGE 38

24 polysaccharide polymers isolated from brown seaweed that can undergo selective ion binding to form inotropic hydrogels. The ratio of mannuronate to guluronic acid determines the physicochemical and swelling properties of the resulting gel, which also dictate drug release kinetics from the system [74] While most physiological tissues do not contain sufficient calcium content to drive gelation of alginate systems, the higher calcium content of lacrimal fluid permits gelation [73] A study by Cohen, et al. [75] found that an alginate system could efficiently entrap pilocarpine and form a gel within the conjunctival cul de sac. This system administered in rabbits was a ble to sustain lower IOP for up to 24 hours after instillation as opposed to 6 hours for free pilocarpine. As such, the system could reduce the frequency of administration from 4 times to twice a day. Despite these findings, two major limitations have prev ented more widespread application of alginate based systems. The first is the potential immunogenicity of the gel s [76] and the second is the long timeframe for in vivo biodegradation (no noticeable degradation was observed after 3 months when implanted subcutaneously in rats) [77] Drug delivery systems based on in situ precipitating polymers have also received significant attention in recent investigations [78] These systems are founded on the ability of certain polymers to be injected in solution form but then precipitate from that solution upon a secondary mechanism such as: a) solvent removal; b) change in pH ; or c) change in temperature In situ precipitating polymers based on solvent removal employ a water in soluble polymer that is dissolved in a water miscible organic solvent. Upon injection, the solvent diffuses into the surrounding medium and water diffuses into the system. B ecause the polymer is insoluble in water, it precipitates once enough water has ent ered the system and forms a

PAGE 39

25 p hysically entangled polymer matrix. Because the organic solvent phase diffuses into surrounding tissues, physiologically compatible solvents must be chosen to prevent cytotoxic responses. Examples of such solvents include N met hyl 2 pyrrolidone (NMP), propylene glycol and dimethyl sulfoxide (DMSO), among others [73] Because of the solvent exchange process responsible for gelation, drug release kinetics and burst release behavior are stron gly dependent on the partition coefficient of the drug [79] While long term drug release with minimal burst behavior can be achieved, careful selection of solvent, polymer, polymer molecular weight, polymer concentration and excipients is required for e ach drug molecule of interest [80] Since high polymer starting c oncentrations are more favorable [81] starting viscosities of these systems tend to be higher and require larger needles for injection (e.g. 22G) [80] Additionally, solvent toxicity remains a major concern in these systems. While solvents such as NMP and DMSO are gen erally considered non cytotoxic at lower concentrations [82,83] some studies have indicated mytotoxicity [84] and ocular specific low dose cytotoxicity [85] As a result, use of these solvents in in situ forming polymer implants will r emain controversial. Reverse Thermal Gels Principles and mechanism of operation As a way to circumvent solvent related concerns, polymers that precipitate in situ by temperature change alone have been heavily investigated [86] In such systems, an entropically driven phase separation process is triggered as the system reaches a certain temperature. The thermodynamic driving force for this phenomenon is the release of structured, bound water from the polymer resulting from inter actions between groups present in the backbone. This process results in phase separation between the water and

PAGE 40

26 polymer and formation of a physically crosslinked polymer matrix. The temperature at which this phase separation occurs is termed the lower criti cal solution temperature (LCST). Since the process is driven only by a thermodynamic competition between hydration of the polymer backbone (at T < LCST) and interactions between polymer molecules (at T > LCST), the process is fully reversible and the solut ion state can be recovered by cooling. For this reason, these systems are termed reverse thermal gels (RTGs). From a drug delivery perspective, the utility of an RTG as a controlled release system will depend on the physicochemical properties of the drug m olecule of interest. For hydrophilic small molecules, the expulsion of water associated with gelation of RTG systems will result in a concomitant expulsion of the water soluble drug. Applications in which immediate release of a bolus of drug are desired wo uld benefit from this process. However, most controlled release systems are intended for applications where an extended period of drug release is desirable. In these cases, application of an RTG would favor delivery of hydrophobic drugs. At the LCST, devel opment of hydrophobic interactions between polymer molecules drives collapse and dehydration of the system. Since two phases are formed the polymer phase, which is now overwhelmingly hydrophobic, and the aqueous phase the drug will be forced to partition b etween these phases based on its partition coefficient. The more hydrophobic the drug, the larger the amount of drug that will partition into the polymer matrix as it is formed. In addition, more hydrophobic drugs will partition out of the hydrophobic poly mer matrix at a slower rate, thereby increasing the delivery time frame.

PAGE 41

27 Gelation of RTG systems is a complex phenomenon that relies on the dynamics of several processes including heat transfer into and through the RTG from the surrounding medium, the abi lity of water to diffuse out of the system and into the surrounding medium and the associated build up of hydrostatic pressure within the gel and polymer diffusion and relaxation processes associated with collapse of the system [86] To date, no models exist that comprehensively take into account all of the factors. The major advantage of an RTG system over thermal or photo crosslinked systems is that no reactive species are present, which significantly reduces toxicity concer ns. In addition, for many RTG systems, the temperature induced phase separation process is sharp and rapid, allowing complete gelation within several seconds after injection. This rapid gelation prevents much of the burst release behavior observed in in si tu crosslinked system. RTG chemistries Various polymers exhibit thermally induced phase separation in aqueous solutions making them suitable for application as RTG s These can be categorized as: a) polymers with ether groups including poly(ethylene oxide) (PEO), random copolymers of PEO and poly(propylene oxide) (PPO), PEO PPO PEO block copolymers (known as poloxamers or Pluronics), poly(lactic co glycolic acid) (PLGA) PEO PLGA block copolymers, alkyl PEO block copolymers and poly(vinyl methyl ether); b) p olymers with hydroxyl functionalities including poly(hydropropyl acrylate), hydroxypropyl cellulose, methylcellulose, hydroxypropyl methylcellulose and poly(vinyl alcohol) derivatives; c) polymers with substituted amide groups including poly( N substituted acrylamides) [ such as poly( N isopropylacrylamide) (PNIPAAm) ] poly( N acryloyl pyrrolidine), poly( N

PAGE 42

28 acryloyl piperidine) and poly(acryl L amino acid amides); and d) other polymers including poly(methacrylic acid). Block copolymer based chemistries are perha ps the most commonly employed for drug delivery applications due to their ease of synthesis and flexibility of the block chemistry. The general principle behind these polymers is to achieve a balance between hydrophobic and hydrophilic segments within the polymer. At lower temperature s (e.g. room temperature), the polymers are sufficiently hydrophilic to allow solvation. With increasing temperature (e.g. near body temperature), hydrophobic domains aggregate to minimize the hydrophobic surface area contactin g the aqueous medium, which reduces the amount of structured water surrounding the hydrophobic domains and maximizes the solvent entropy. As a result, a stable physical gel can be formed. This process is illustrated diagrammatically in Figure 1. 5 O ne of t he most common classes of block copolymer based RTG chemistries is the poloxamer, which is constructed from a hydrophobic PPO block grafted at both ends with hydrophilic PEO blocks (PEO PPO PEO) [87 90] Because of th eir commercial availability and favorable thermal gelling properties, poloxamer based systems have been investigated for myriad tissue engineering [91,92] and drug delivery [93 95] applications. Most poloxamer systems actually exhibit cloud point te mperatures well above body temperature, but can be made to form physical gels by injecting high concentrations of the polymers (e.g. 20 wt%). The dynamic underlying this principles is that poloxamer based solutions at lower concentrations form micelle nano structures and there is no t sufficient polymer present for these nano structures to coalesce into a cohesive physical gel [96] Only at higher polymer concentrations is the polymer content per unit volume sufficient to form

PAGE 43

29 long scale polymer structures amenable to c omplete gelation. As a result, injections require larger gauge needles and may be hyperosmolar to the surrounding tissue at administration [73] both of which consequences are detrimental to their clinical utility. Nevertheless, several studies have shown the potential for poloxamer based drug delivery systems. Veyries, et al. [97] demonstrated that a 25 wt% poloxamer solution could prolong the residence time of vancomycin for administratio n in high infection risk areas (e.g. post surgical). In vivo high concentrations were maintained near the site of injection for 24 hours and therapeutic concentrations were maintained for 8 days. Further, the poloxamer system d id not alter vancomycin activity and was well tolerated after injection. Another study by Miyazaki, et al. [98] evaluat ed a poloxamer system for sustained anti tumoral activity using mitomycin C. Again using a 25 wt% system, they found that tumor bearing mice had longer life spans after treatment with the drug loaded poloxamer system as compared to free drug. Most importan tly, controlled release properties of the gel system allowed administration of drug concentrations that would be toxic on their own. This property is of great value in anti cancer therapies. Figure 1.5 : Block copolymers containing hydrophobic (red) and hydrophilic (blue) segments can undergo a reversible solution to gel phase transition at elevated temperature. The gel state is characterized by formation of hydrophobic domains (dashed circles) within the physical matrix, which can entrap hydrophobic drug molecules.

PAGE 44

30 Despite these positive findings, further application of poloxamer based systems has been hindered by several critical drawbacks, most of which are driven by the need to employ high concentration solutions to achieve body temperature gelation. One physiologically observed effect is that intraperitoneal injections of polo xamer systems resulted in sustained hypertriglyceridemia and hypercholesterolemia [99] Both effects were found to b e related to stimulation of 3 hyd roxy 3 methylglutaryl co enzyme A reductase activity in the liver, presumably during metabolic clearance of the polymer [100] Another troublesome clinical finding for poloxamer based systems is that certain formulations inhibit P glycoprotein resulting in increased cellular uptake of certain drugs and thereby increasing their cytoto xicity [101 103] While the exact mechanism of this effect has not been discovered, it has been proposed that it may be related to the ability of poloxamer molecules to incorporate into cell membranes and induce changes in plasma membrane order, morphology and mechanical stability [104,105] In order to circumvent these clinical challenges, other block copolymer chemistries have also been investigated. Another relatively common formulation is that comprised of the common building blocks PLGA and PEG, synthesized as PEG PLGA PEG or PLGA PEG PLGA Jeong, et al. [106] investigated a PEG PLGA PEG system for in vitro release kinetics of both a hydrophilic (ketoprofen) and hydrophobic (spironolactone) model drug. Because of the formation of hydrophobic domains within the physical gel and the ability of these domains to control release of hydrophobic molecules, release kinetics were more sustained for the hydrophobic drug at 55 days to complete drug release compared to 6 10 days for the hydrophilic drug. While the PLGA block permits complete biodegrad ation of these systems within 3 4 months [107] like the poloxamer based sys tems, high polymer

PAGE 45

31 concentrations (greater than 20 wt%) are required for adequate gelation and drug loading levels beyond 0.25 wt% are generally not reported PLGA PEG PLGA is a commercially available thermal gelling polymer sold under the name ReGel. As w ould be expected, its physical properties are very similar to those of PEG PLGA PEG systems, but its increased overall hydrophobicity makes it better suited for controlled release of hydrophobic drugs. Zentner, et al. [108] demonstrated that the system could sustain in vitro rel ease of paclitaxel for 50 days with complete in vitro degradation of the system within approximately 10 weeks. The amphiphilic nature of the copolymer also helped solubilize the hydrophobic drugs cyclosporine A and paclitaxel by 400 to 2000 fold. However, the ReGel system was not an advantageous carrier of hydrophilic molecules. Complete in vitro release of several model proteins was observed within 10 15 days for all of the molecules studied. Additionally, as with previously described systems, high polymer concentrations (greater than 20 wt%) were still needed for gelation, limiting the delivery options of the system. PNIPAAm based RTGs In order to circumvent the challenges associated with block copolymer systems as described in the previous paragraphs many groups have turned to specifi c thermo sensitive homopolymers. While many examples were given at the beginning of this section, none have received more attention than PNIPAAm [109] In its simplest homopolymer fo rm, PNIPAAm is a water soluble polymer that exhibits a sharp, reversible LCST near 32C [110] positioning its sol gel transition point comfortably above room temperature a nd sufficiently below body temperature. While many polymers possess LCSTs even in this temperature range PNIPAAm is unique in its sharp, almost discontinuous,

PAGE 46

32 transition [111] This property makes it particularly well suited for in situ gelling applications where rapid gel formation is desirable. The LCST behavior of PNIPAAm is generally attributed to an entropically dr iven coil to globule transition [112 115] At temperatures below the LCST, PNIPAAm molecules exist in a solva ted state in a flexible and expanded coil conformation d ue to an abundance of hydrogen bonding between amide groups and surrounding water molecules. With increasing temperature, these coils collapse to a globule state due to increasing intramolecular hydro ge n bonding between amide groups and hydrophobic interactions between isopropyl groups. Fr o m an enthalpic point of view, this transition process is endothermic, indicating the polar amide groups prefer hydrogen bonding with water to existing at the non pol ar interior of the collapsed globule. From an entropic point of view, this transition decreases the conformational degrees of freedom possessed by the polymer molecules but increases the degrees of freedom for the water molecule s that were previously hydr ogen bonded to the amide groups. Additionally, the space occupied by the polymer molecules is decreased significantly in the globule state, which decreases the excluded volume and further increases the overall configurational energy of water molecules in t he system. Therefore, the coil to globule transition is driven by: a) dominance of the enthalpic term at low temperatures, which stabilizes the expanded coil state of PNIPAAm molecules through hydrogen bonding between amide groups and water molecules; and b) a transition to dominance by entropy terms at high er temperature driven by the increased configurational entropy of water molecules, which stabilize s the compact globule state.

PAGE 47

33 Practically, this process results in several critical changes in an aqueou s PNIPAAm solution that permit facile characterization and application. The first is that in dilute aqueous solutions, phase separation at the LCST causes significant turbidity, allowing LCST measurement by monitoring the change in optical density of a PNI PAAm solution with increasing temperature. The second is that once in the collapsed globule state, globules can aggregate to further increase the entropy of free water, thereby forming a physical gel. Thermodynamically, this aggregation process will occur either slowly at low temperatures or in dilute solutions or quickly at high temperatures or in concentrated solutions. The temperature dependence is driven by the increased driving force to maximize the configurational entropy of water molecules at higher temperatures and the concentration dependence is driven by the mean distance between globules, which must be overcome in order to permit aggregation. Because the coil to globule transition is driven by a loss of structured water previously bound to PNIPAA m molecules, which allows formation of a more compact globule state, significant water is liberated from the PNIPAAm phase during this transition. This volume based transition from expanded coil to compacted globule is termed the volume phase transition an d has been thoroughly characterized in previous studies [111] Of particular note is the estimation by Shibayama, et al. [116] that 13 water molecules are liberated from the PNIPAAm phase per monomer unit. The net effect of this phenomenon is that during gelation of PNIPAAm bas ed systems, a significant amount of water is expelled from the system resulting in a largely dehydrated physical gel that is predominantly hydrophobic in nature [112]

PAGE 48

34 From a drug delivery perspective, the combined effect of the expulsion of water content and the formation of a hydrophobic gel core presents a favorable environment for e ntrapment and controlled release of hydrophobic drugs and an unfavorable one for the same of hydrophilic drugs, as illustrated in Figure 1. 6 For a hydrophobic drug mo lecule, the expulsion of water and hydrophobic interactions driving physical gel formatio n will encourage drug molecules to be entrapped within the gel. In addition, after gel formation, the drug will be released slowly into the surrounding aqueous medium as it partitions out Figure 1.6 : During the sol gel transition of PNIPAAm based RTGs, a hydrophobic physical gel is formed concomitant with the expulsion of significant amounts of water. For a hydrophobic drug, most of the drug in the system will partition into the physical gel during this transition. However, for a hydrophilic molecule, most of the drug will be expelled along with the water resulting in low loading and a large burst of d rug release.

PAGE 49

35 of t he hydrophobic gel core. However, for a hydrophilic molecule, th e drug will follow water molecules out of the polymer phase of the system during the sol gel transition, resulting in poor entrapment efficiencies and a large burst of drug release concomitant with the transition. This theoretical difference in drug releas e capabilities has also been borne out in the literature. For example, Zhang, et al. [117] explored a PNIPAAm based system for controlled release of the water soluble protein bovine serum albumin (BSA). Even though BSA is a large molecule (MW: 69 kDa) model compound, the group found that 20 50% of the drug load was released within the first several hours, presumably due to the volume phase transition associated with PNIPAAm systems. Additionally, Han, et al. [118] investigated release of 5 fluorouracil from PNIPAAm chitosan hydrogels and also noted a large 40 60% burst within the first hour and complete release of the drug paylo ad within 24 hours. In general, release kinetics of hydrophilic drugs from PNIPAAm based systems exhibit a large burst release and complete drug release within 24 48 hours. Systems intended for controlled release of hydrophobic drugs have generally fared b etter. Wilson, et al. [119] studied release of the hydrophobic small molecu le drug colchicine from a PNIPAAm based thermal gel and found the system could sustain therapeutic drug levels for 14 days. Hoare, et al. [120] studied uptake and release behaviors of several different drug molecules and found that more hydrophobic drugs more rapidly and more completely partitioned into the hydrophobic cores of PNIPAAm based micro gels and that release rates for these hydrophobic drugs were slower in vitro These results indicate that PNIPAAm based drug delivery systems are well suited for encapsulation and controlled delivery of hydrophobic drugs.

PAGE 50

36 PNIPAAm based RTGs also have the d istinct advantage over block copolymer based systems of de coupling polymer hydrophobicity and thermal gelling characteristics. Because block copolymer systems rely on a specific hydrophobic hydrophilic balance to achieve thermal gelling, introduction of p articularly hydrophilic or hydrophobic molecules especially covalently tethered to the polymer itself stand to compromise thermal gelling characteristics. For example, Xun, et al. [121] synthesized a thermo caprolactone) co lactide] PEG caprolactone) co lactide] (PCLA PEG PCLA) triblock copolymer. Functionalization with the peptide KRGDKK was found to shift the rheological sol ge l transition temperature down by 5C and also significantly altered drug release kinetics from the system. While adjustments could be made to the polymer chemistry to take these changes into account, the magnitude of this effect will be different depending on a large number of parameters including: functionalized molecule, functionalization density, loaded drug molecule physicochemical properties, drug loading fraction, etc. This would result in a resource consuming optimization process and an inflexible sy stem. Since PNIPAAm based RTG chemistries are largely unaffected by these parameters, they are In Situ Gelling Systems f or Ophthalmic Drug Delivery Motivations In situ gelli ng systems are uniquely capable of overcoming challenging delivery scenarios to provide a minimally invasive deployment strategy for some of the most complex anatomical challenges. Perhaps nowhere is this of greater importance than in the field of ophthalm ology, where minimally invasive deployment, harmonization with the host tissue

PAGE 51

37 anatomical space and non reactivity with highly sensitive cells and tissues are of critical concern. For these reasons, RTGs present an especially advantageous optio n for ophtha lmic drug delivery. First, the ability of RTGs to be injected through small gauge needles make them more acceptable to patients and providers as it minimizes patient discomfort and reduces the complexity of the procedure, leading to lower risk of complicat ions and better patient outcomes. This will be especially true for RTG chemistries that are low viscosity solutions at room temperature and can be injected through 30 or 32G needles, as are commonly used in ophthalmic procedures (e.g. intravitreal injectio ns ). This requirement precludes most alginate and block copolymer systems as they require larger needles for injection. While larger needles (up to 22G) can be used for intravitreal injections without the need for suturing, such systems have not been as we ll accepted clinically. For example, the Ozurdex system was an extruded PLGA matrix containing dexamethasone that was injected into the vitreous humor for long term treatment of macular edema ( secondary to branch retinal vein occlusion or central vein occl usion ) and uveitis. Because the device was fashioned as a bulk polymer rod, it required an applicator with a 22G diameter. Many papers associate the large needle size to reported side effects including endophthalmitis, intraocular inflammation, increased I OP, retinal detachment and conjunctival hemorrhage [122] In addition, reports of complicati ons were found to be higher in patients that needed repeated implantations [123] In order to avoid the potential for such complications, an RTG that can be delivered us ing a much smaller gauge needle would be of clinical interest. Another advantageous characteristic of RTGs and of many in situ gelling polymers is their ability to conform to the anatomical confines of the delivery space. As described in

PAGE 52

38 previous sections, potential spaces within the eye are relatively small (e.g. 100 L for the subconjunctival space) and abnormally shaped. In addition, devices that are too large or obtrusive to the ocular anatomy can at the least cause an increase in IOP. As a result, poly meric bulk implants can be problematic in that they are of a fixed geometry and cannot adapt after implantation. In contrast, because RTGs are injected initially as a solution and take several seconds or more to fully gel, they can conform to the anatomica l space into which they are injected before gelation, allowing a more anatomically appropriate fit. Examples A number of previous works have investigated in situ gelling polymer systems for ophthalmic drug delivery. These systems have employed naturally de rived and synthetic polymers and have included ion pH and thermally activated gelation mechanisms. A brief review of some of the more notable systems is included to highlight their successes and failures and where room exists for improvement in the fie ld. Naturally derived cellulose derivatives such as methylcellulose (MC) and hydroxypropyl methylcellulose (HPMC) exhibit phase transition temperatures above body temperature (> 40C); however, in the presence of various salts this can be reduced to below body temperature [124] Bhowmik, et al. [124] employed this strategy to develop thermo reversible gels for ophthalmic delivery of ketorolac. They found th at increasing salt concentrations lowered the sol gel phase transition temperature and also increased the longevity of drug release from their system, from 180 to 240 min utes This increased drug release time frame was attributed to an increase in the visc osity of the system resulting from the addition of salts. However, practical application of this systems would undoubtedly be very challenging due to the requirement that 6 8 wt% salts be added to the

PAGE 53

39 system to achieve the desired properties. This makes th e system severely hyperosmolar, which could pose a serious detriment to the tissues at the site of administration [78] In order to overcome this problem, several groups investigated MC or HPMC in combina tion with poloxamers to both improve the gelation kinetics of MC/HPMC and the sustained release properties of poloxamer based systems. El Kamel [125] found that a solution containing 15 wt% poloxamer (F127) and 3% MC gave the slowest release kinetics of timolol maleate and, applied topically, increased the intraocular bioavailability of the drug 2.4 fold in rabbits, presumably du e to the increased retention time accorded by the higher viscosity as compared to free drug solutions. Paavola, et al. [126] had similar results with poloxamer HPMC and polo xamer MC composites for sustained release of lidocaine HCl and ibuprofen. However, release time frames beyond 2 3 days could not be achieved with any of the systems, making them applicable only to short term delivery applications. Desai, et al. [127] found that an F127 poloxamer mixed with MC, HPMC or PEG could significantly reduce the release rates of pi locarpine as compared to poloxam er alone. Gelation kinetics were also improved, which were attributed to an increase in the However, release time frames for this system were also limited, achieving only several hours of sustained release. Thrimawithana, et al. [128] investigated thermal gelling dispersions of MC and carrageenan for transscleral delivery of macr omolecules. While the system could transition from a dispersion to a gel below body temperature, rheological properties of the system were complex, exhibiting several transitio ns at various temperature. Sub T enon injections of the system loaded with a Cx 43 antisense oligonucleotide led to a significant reduction in Cx43 levels in the choroid

PAGE 54

40 of rats 24 hours after administration. The longevity of this effect was not evaluated, however. In the realm of synthetic polymers, poloxame rs have received a great d eal of attention for ocular drug delivery, as in other areas, due primarily to their commercial availability and well characterized history. Gupta, et al. [129] found t hat a poloxamer based temperature and pH sensitive (through addition of chitosan) gel could enhance delivery of timolol maleate and improve intraocular bioavailability. Srividya, et al. [130] found that a pH sensitive thermal g el based on poly ( acrylic acid ) (PAA) and HMPC could sustain delivery of ofloxacin for up to 8 hours. Ma, et al. [131] employed a poloxamer PAA copolymer to sustain release of gatifloxacin for up to 12 hours wi th favorable gelation kinetics. For posterior applications, Lee, et al. [132] investigated the F127 poloxamer for transscleral sustained delivery of dexamethasone. In vitro studies determined that the system could provide up to 48 hours of delivery with good scleral permeation. Other block copolymers, especially those of PLGA and PEG have also received attention in this field. Duvvuri, et al. [133] explored PLGA PEG PLGA thermal gels for delivery of ganciclovir and found the system could maintain therapeutic release rates for approximately 10 days. Similarly, after intravitreal injections in rabbits, maintenance of vitreal drug concentrations was achieved fo r 14 days after injection [134] A different block copolymer chemistry, composed of PEO PCL PEO was developed by Wang, et al. [135] for intravitreal injection and delivery of the antibody bevacizumab. The polymer exhibited excellent biocompatibility both in vitro and after intravitreal injection in rabbits and sustained in vitro delivery of the large molecule for 11 days in vitro Similarly, Park, et al. [136] developed a PEGylated polyurethane block copolymer for intravitreal deli very of

PAGE 55

41 bevacizumab and found excellent in vitro and in vivo biocompatibility of the system, which could sustain release of the antibody for 17 weeks from a 20 wt% gel. Shortfalls of existing systems A major limitation of the above described systems is th at most were evaluated as a direct replacement for eye drops. As such, delivery timeframes were severely limited (to several hours) and the systems would be expected to suffer the same limitation of limited posterior bioavailability. None of the afore ment ioned studies investigated the impact of their systems on posterior drug levels. While some excellent systems were described that were designed specifically for intra or periocular administration and could provide therapeutic posterior drug levels releas e time frames of these systems were still fairly limited, with the longest system ac hieving 17 weeks of release. This falls short of the currently marketed Ozurdex system, which can provide 4 6 months of dexamethasone delivery. In order to achieve clinical relevance, the delivery time frame of RTGs intended for posterior delivery of ophthalmic drugs must be extended. As has been elucidated in previous sections, this problem has plagued nearly all RTGs that have been investigated, even outside the ophthalmic realm. What is clear is that hydrogel based systems (such as poloxamers) have an especially poor ability to sustain drug release due to their highly hydrated state even after gelation, which permits relatively free diffusion of drug molecules out of the R TG matrix. PNIPAAm systems have generally fared better in terms of release kinetics due to their formation of a dense, hydrophobic matrix, but their application to intra or peri ocular drug delivery scenarios could not be found in the literature. Neverthe less, even PNIPAAm based RTGs are limited to 1 2 months of delivery in most reported studies. As such, a novel approach is necessitated.

PAGE 56

42 Micelles as a secondary delivery vehicle One such approach is the incorporation of a nano carrier system within the rev erse thermal gel. Nano carrier systems as a class of drug delivery vehicles typically include nanoparticle s microparticles, micelles and liposomes. These are illustrated in Figure 1 .7 Nanoparticles and microparticles are bulk particles composed of non wa ter soluble polymers compacted into a particle form Drugs can be encapsulated within the particle during fabrication, which is typically realized through an oil in water emulsification process [137] Sustained drug release is achieved due to the time take n either: a) for the drug to partition out of the hydrophobic particle into the surrounding medium; or b) for the xtent to allow the drug to diffuse out of it. Their application in ophthalmic drug delivery is vast [138 140] but will not be extensively covered herein. Similarly liposomes have been applied for several ophthalmology specific applications [12,141,142] Liposomes are artificially fabricated vesicles composed of a lipid bilayer encapsulating an aqueous solution. They can be loaded with therapeutic molecules either in the interior of the particle (typically for hydrophilic molecules) or within the hydrophob ic segment of the lipid bilayer (for hydrophobic drug molecules). Liposomes are especially well suited for targeted drug delivery as they can be designed to release their contents only after fusion with a cell membrane (which can be mediated by cell specif ic antibodies/receptors) or to permit diffusion of molecules out of the liposome only after cellular internalization. Micelles possess some character similar to both nanoparticles and liposomes. Like liposomes, they are typically self assembled from amphi philic molecules. However, their final morphology is composed of a hydrophobic core (much like a nanoparticle)

PAGE 57

43 surrounded by hydrophilic segments, which make up the exteri or of the particle (commonly called the corona) Most micelle systems will self aggre gate into this structure in an aqueous environment when the correct conditions (typically of temperature and polymer/surfactant concentration) are met. The minimum concentration required for self aggregation is termed the critical micelle concentration (CM C) and the minimum temperature required is called the critical micellization temperature (CMT) [143] For a molecule to be amenable to this micellization process, it must poss ess both hydrophilic and hydrophobic character. For this reason, the two main categories of molecules that are used to form micelles are surfactants and amphiphilic copolymers, with copolymers being the most common choice. Amphiphilic block copolymers are typically synthesized either as A B or A B A block copolymers, where A is a hydrophilic polymer and B is a hydrophobic polymer. Upon micellization, the hydrophobic B segments aggregate to form the core of the micelle and the hydrophilic A segments form the corona which stabilizes the core shell structure. The thermodynamic driving force behind micellization of block copolymers is the decrease of free energy in the overall system due to removal of hydrophobic polymer segments from the surrounding aqueous me dium, allowing primarily hydrophilic segments to interface with water molecules [144] Because Figure 1.7 : Nano carrier systems typically employed for drug delivery applications include nanoparticles composed of bulk polymer (left), liposomes composed of a lipid bilayer surrounding an aqueous solution (center) and micelles with a hydro philic shell and hydrophobic interior (right). Typical modes of drug encapsulation are illustrated.

PAGE 58

44 micellization is driven by aggregation of hydrophobic polyme r segments and the final micelle structure is composed of a highly hydrophobic core, micelles are particularly advantageous carriers for poorly soluble drug molecules This property gives micelles a distinct advantage in: a) solubilization of poorly solubl e drug molecules thereby improving their bioavailability; b) reducing toxicity and other adverse effects due to high drug loading only within the micelle core; and c) substantial ly altering drug permeabilit y across physiological barriers [145] Sustained drug release kinetics are also generally more favorable from a micelle system The rate of drug release from a micelle will depend on several factors including the rate of diffusion of the drug out of the micelle core (this will also be influenced by the partition coefficient of the encapsulated drug molecule), micelle stability and the rate of degradation of the polymer comprising the micelle. If the micelles are stable and the rate of degradation is slow, the release rate will core, which can be driven by chemical, physical or electrostatic interactions. This suits micelles well for sustained release of hydrophobic drug molecules as their hydrophobic core will result i n relatively slow drug partitioning into the surrounding aqueous medium. However, these kinetics also imply that beyond micelle chemistry alone drug release behavior will also be dependent on the partition coefficient of the drug itself. Micelle fabricatio n and concomitant or subsequent drug loading are realized by one of several methods including dialysis, oil in water emulsification and film based methods. The dialysis method consists of dissolving the polymer and drug in a suitable organic solvent (one t hat is water miscible and solubilizes both the polymer and drug) and dialyzing this solution against water. As the solvent diffuses out of the dialysis chamber

PAGE 59

45 and is replaced by water, the polymer will aggregate to form micelles and a fraction of the drug will be encapsulated This method is perhaps the most simple of those available, but is commercially impractical due to the length of time required to ensure complete removal of the organic solvent (generally several days) [146] The oil in water emulsification technique is similar to that used to produce nanoparticles and comprises adding an organic phase consisting of the drug dissolved in a volatile solvent to an aqueous phase con sisting of the polymer dissolved in water followed by sonication and evaporation of the volatile organic solvent. This method is rapid and simple, but generally yields low drug encapsulation efficiency, broad micelle polydispersity and significant levels o f residual solvent, which are often toxic [146] Film based methods while subject to various modification s generally begin with dissolution of the polymer and drug in a volatile orga nic solvent followed by deposition of this solution onto a surface, evaporation of the solvent to produce a polymer drug film and rehydration of the film in water. Because solvents that are generally considered to be less cytotoxic (e.g. DMSO or NMP) are n on volatile, they cannot be used with this procedure and so residual solvent levels remain a major concern. Micelles have been fairly extensively investigated for purposes of ophthalmic drug delivery D ue to the reasons discussed previou sly, they have most commonly been employed for delivery of solubility limited molecules and those with poor ocular permeability. For example, cyclosporin A (CsA) is challenging to formulate for topical ophthalmic administration due to poor solubility and permeability Di Tom maso, et al. [147] developed PEG hexylsubstituted poly(lactide) di block copolymer based micelles for topical administration of CsA These micelles showed excellent CsA encapsulation and

PAGE 60

46 ocular compatibility and signif icantly increased intraocular bioavailability of the drug as compared to topical eye drops. Li, et al. [148] investigated a micelle system for improved solubilization and intraocular bioavailability of the poorly soluble drug diclofe nac. The micelle chemistry employed was a di block copolymer consisting of PEG and PCL. The micelles were ocular compatible and increased penetration of the drug across rabbit cornea by 17 fold compared to free drug and increased the area under the curve ( AUC) of the drug by 2 fold, also compared to free drug. Gupta, et al. [149] developed a crosslinked micelle based on a copolymer of NIPAAm, acrylic acid and N vinylpyrr olidone crosslinked by methylene bis acrylamide. The produced micelles were both pH and temperature se nsitive and could sustain release of the poorly soluble drug ketorolac for 10 hours with a 2 fold increase in trans corneal permeation ex vivo Civiale, et al. [150] developed polyhydroxyethylaspartamide based micelles and demonstrated a 2 fold increase in intraocular bioavailability of dexamethasone in vitro Finally, Lin, et al. [151] evaluated poloxamer coated chitosan micelles for improved topical penetration of metipranolol. They demonstrated a prolonged reduction of IOP in rabbits after topical administration of the micelles as compared to standard eye drops. Be cause micelles are well suited to solubilization and sustained release of poorly soluble drugs, they present an excellent system to supplement the release capabilities of an RTG. By encapsulating pre formed micelles (pre loaded with drug) within an RTG, th e micelle s can now provide the drug release limiting dynamics in addition to the RTG, which should result in longer term sustained release of the drug. One additional setback of using micelles alone is that they are prone to dissociation upon dilution (e.g after injection) and can therefore face major stability challenges. By encapsulating the micelles within the RTG

PAGE 61

47 system, they are protected from this dilution effect because they are physically entrapped within the RTG matrix, which as discussed previous ly is largely dehydrated upon gelation. Finally, the RTG presents an advantageous secondary vehicle for micelle delivery because of its in situ gelation, which will prevent micelles from being carried away by local fluid flows after injection. This effect has been shown to severely limit micelle retention at the injection site, especially after intravitreal injection where micelles were found to escape into the aqueous humor and be cleared systemically within 14 days [152] To acetonide (TA) was chosen as a model poorly soluble drug. TA has gained attention for ophthalmic indications in recent years by use as an intravitreal injection for treatment of exudative age related macular degeneration, diabetic macular edema, proliferative diabetic retinopathy, and branch and central retinal v ein occlusion, among others. These ocular diseases are characterized by a) a pathological proliferation of intraocular cells such as RPE cells in proliferative vitreoretinopathy and vascular cells in proliferative diabetic retinopathy, which are generally accompanied and stimulated by intraocular inflammatory processes ; or b) defects in the blood retinal barrier due to capillary leakage, which results in the accumulation of fluid in the intra or sub retinal spaces [153] Corticosteroids have been shown to be effective in reducing intraocular inflammation, tightening capillary vessel wall s and at sufficient concentrations suppressing both neovascularization and the proliferation of endothelial and fibroblast cells [154] In order to achieve sufficiently high concentrations of corticosteroids in posterior tissues direct intravitreal injections were explored and found to be highly effective [155] However, the use of soluble corticosteroids allowed for the ir rapid clearance from the vitreous humor resulting in the need for frequent

PAGE 62

48 re injections. To overcome this problem, large dose injections of the poorly soluble, crystalline corticosteroid TA (solubility at 37C: 24 g/mL [156] log P: 2.53 [157] ) were employed as a method to form an intravitreal depot of the corticosteroid, which could sustain efficacious drug levels for a longer period of time [158] This method continues to suffer from the use of a large single dose of the drug, which has been attributed with many side effects, but especially a spike in IOP which can persist for weeks t o months if untreated [159] In addition, the current treatment paradigm requires intravitreal TA injections to be repeated every 2 5 months, which can be burdensome for both the provider and the patient. The use of a controlled release system, such as the combined RTG micelle system, may improve the intraocular pharmacokinetics of TA administration by providing for a slow, continuous release of the drug ove r a longer time frame than is currently achievable. By minimizing the initial concentration spike of the drug (as is seen in intravitreal injections) and improving the longevity of a single administration, this system may reduce side effects such as IOP sp ikes and reduce the burden of treatment, potentially resulting in improved clinical outcomes. The ability for such a combined polymeric micelle/RTG system to encap sulate and sustain release of TA will be explored within this thesis.

PAGE 63

49 CHAPTER II POLYMERIC MICELLES FOR SUSTAINED RELEASE OF A POORLY SOLUBLE CORTICOSTEROID Poor aqueous solubility is a major limiting factor in the administration of existing and newly developed drugs. To overcome this challenge, a polymeric micelle system was developed based on a novel PEG capped poly(ester urethane) tri block copolymer for efficient encapsulation and long term delivery of a model poorly soluble corticosteroid. Low and high molecular weight variants of the base polymer designed to modulate drug release kinetics w ere characterized by GPC and 1 H NMR spectroscopy to confirm their structural properties. The polymer was found to be highly biocompatible based on in vitro cytotoxicity testing against human retinal pigmented epithelial cultures. Micelles were fabricated f rom the low and high molecular weight variants using both an established sonication technique and a newly developed filter extrusion process. Micelles produced by both techniques possessed the expected core shell structure, as confirmed by electron microsc opy. However, those produced by the extrusion technique were significantly more monodisperse in diameter with a relative standard deviation of 12.3% compared to 42.3% for those produced by the sonication technique. Release kinetics of the corticosteroid tr iamcinolone acetonide were dependent on the molecular weight of the base polymer, but all conditions were capable of maintaining zero order release kinetics for over one year with minimal burst release. This novel micelle system is well suited for extended delivery of poorly soluble small molecule compounds, and the filter extrusion process developed herein provides a rapid fabrication technique that produces a highly

PAGE 64

50 monodisperse micelle population and eliminates the need for additional sterilization steps Introduction Poor aqueous solubility is a major limiting factor in the administration of existing and newly developed drugs. Nearly 40% of new drug candidates exhibit poor solubility leading to low bioavailability, large intra and inter subject variabil ity a nd poor dose proportionality [160] While strategies can be employed to facilitate solubilization of these compounds (such as mi cronization and salt formation) [55] these methods do not typicall y improve the duration of therapeutic activity and, as such, require frequent re administration to maintain therapeutically effective levels of the drug in tissue. To simultaneously improve delivery and duration of therapeutic activity of poorly soluble dr ugs, advanced drug delivery systems have been developed that encapsulate the compound in a nanocarrier system. Such systems that have garnered the most attention include liposomes, nanoparticles and micelles; however, application of these systems continues to be hampered by various limitations. Liposomal systems tend to have limited drug loading capacity since higher drug loads can destabilize the membrane and affect overall stability of the system [161] Nanoparticulate systems most commonly formulated from poly(lactic co glycolic acid) (PLGA) or similar biodegra dable polymers typically produce acidic byproducts during biodegradation [162] leading to concerns of local inflamm ation [163] and drug stability [61] Micellar systems can overcome many of these limitations due to their unique core shell structure. Typically fabricated from di or tri block amphiphilic copolymers, mi celles self

PAGE 65

51 aggregate into nanostructures possessing a hydrophobic core and hydrophilic shell. This structure is highly advantageous for use as an advanced drug delivery system because poorly soluble drugs can be efficiently entrapped in the hydrophobic co re at high concentrations while the hydrophilic shell ensures high biocompatibility an d prevents particle aggregation [164] In this work, a novel micellar system has been developed using a functionalizable amphiphilic copolymer, which lends significant flexibility both in the optimization of the physi c o chemical prope rties of the micelles and in its conjugation with targeting ligands or other biomolecules. In addition, a novel filter extrusion process was developed and characterized for fabrication of the micelles, which possesses several advantages over current fabric ation techniques such as sonication. While this process is commonly employed in fabrication of liposomes, to our knowledge this is the first time such a process has been reported for the facile fabrication of polymeric micelles. In order to assess drug rel ease kinetics, the corticosteroid triamcinolone acetonide (TA, molar mass 434.5 g/mol) was employed as a model drug. TA has limited aqueous solubility but has gained extensive popularity recently, especially for ocular indications such as uveitis, macular edema, vitreoretinopathy and choroidal neovascularization secondary to age related macular degeneration [165] Current clinical administr ation techniques (such as intravitreal or periocular injection) rely on injecting large doses of a suspension of the drug to avoid the need for frequent re administration, but this method carries a high incidence of highly damaging side effects including e levated intraocular pressure, endophthalmitis, cataract and retinal detachment [166]

PAGE 66

52 The micelle system developed herein could significantly improve the clinical use of TA by acting as an injectable long term sustained release system. The synthesis, characterization and in vitro biocompatibility testing of the precursor polymer are presented in addition to characterization and in vitro drug release testing with TA of micelles fabricated by both a common sonication technique and an adapted filter extrusion technique. Materia l and m ethods Materials Dimethylformamide (DMF), hexamethylene diisocyanate (HDI), methoxypolyethylene glycol (mPEG 550), DMSO and serinol were obtained from Sigma Aldrich (St. Louis, MO). Di tert butyl di carbonate and ethyl acetate were obtained from Alf a Aesar (Ward Hill, MA). TA was obtained from Fluka (Sigma Aldrich, St. Louis, MO). Hexane was obtained from EMD Chemicals (Philadelphia, PA) and diethyl ether from Fisher Scientific International, Inc. (Waltham, MA). All organic solvents were anhydrous an d chemicals were used as received. Phosphate buffered saline (PBS), fetal bovine serum (FBS), penicillin/streptomycin (PS) and Dulbecco's Modified Eagle's Medium nutrient mixture F 12 (1:1 DMEM:F 12) were obtained from Thermo Scientific (Logan, UT). Vybran t MTT Cell Proliferation Assay Kit was obtained from Molecular Probes (Eugene, OR). Synthesis of N Boc serinol Serinol (21.5 mmol) was dissolved in 20 mL of absolute ethanol and stirred at 4C. Di tert butyl dicarbonate (26.0 mmol) was dissolved in 20 mL of absolute ethanol and added dropwise to the serinol solution over a period of one hour, while maintaining 4C and constant stirring. The solution was heated to 37C with vigorous stirring and reacted for

PAGE 67

53 one hour. The ethanol was removed by rotary evapor ation at 45C and 10 mbar vacuum and the solid was re dissolved in a 1:1 mixture of hexane and ethyl acetate by gentle heating. Additional hexane was added until precipitation was observed and the resulting suspension was stored at 4C overnight to allow r ecrystallization. Subsequent vacuum filtration yielded a white flaky product. Synthesis of poly(hexamethylene alt serinol) (PHS) block N Boc serinol (1.31 mmol) was weighed out and lyophilized for 12 hours at 45C and 0.045 mbar. In a reaction vessel, the N Boc serinol was dissolved in 3 mL dry DMF and heated to 80C under gentle stirring and a nitrogen atmosphere. Then HDI (1.31 mmol) was added drop wise and the reaction was carried out for either 3 or 5 days to produce low or high molecular weight PHS, r espectively. After the specified time, additional HDI (2.62 mmol) was added to cap each end of the polymer with an isocyanate group and terminate the reaction. After 24 hours, the product was rapidly precipitated twice in diethyl ether. Conjugation of mPEG The isocyanate terminated PHS from the previous reaction was immediately re dissolved in 3 mL dry DMF and heated to 80C under gentle stirring and a nitrogen atmosphere. An excess of mPEG (5 mmol) was lyophilized and added to the reaction. The PEGylation reaction was carried out for 12 hours and the resulting product, polyethylene glycol block poly(hexamethylene alt serinol) block polyethylene glycol (PEG PHS PEG), was purified by three precipitations in diethyl ether and then dried completely by extended rotary evaporation at 50C and 10 mbar vacuum. The full synthetic route of PEG PHS PEG can be found in Appendix A.

PAGE 68

54 Micelle f abrication by sonication PEG PHS PEG and TA were dissolved in 2 mL DMSO at 2.5 and 0.25 wt% respectively (polymer/DMSO and drug/DMSO ). This solution was then added dropwise to a beaker containing 40 mL of purified water (milliQ or equivalent) submerged in an ultrasonic bath (VWR International, West Chester, PA; 48W RF power). The resulting emulsion was sonicated for 10 min. Removal of DMSO was carried out by centrifugation at 4500 rcf for 5 min, pouring off the supernatant and then re suspending the micelles in purified water. This DMSO extraction procedure was carried out 3 times. The resulting micelles were either used immediately or lyophilized at 45C and 0.045 mbar to produce a dry product. Micelle Fabrication by extrusion PEG PHS PEG and TA were dissolved in DMSO at 2.5 and 0.25 wt% respectively (polymer/DMSO and drug/DMSO). This solution was then added to purified water (milliQ o r equivalent) at a 1:20 organic:aqueous phase ratio and loaded into one end of the extrusion apparatus (Avestin, Inc., Ottowa, ON). A silver membrane filter (100 nm average pore size, Sterilitech Corp., Kent, WA) was used as the extrusion filter. The emuls ion was passed through the filter 11 times and then transferred to a centrifuge tube. Removal of DMSO was carried out by centrifugation at 4500 rcf for 5 min, pouring off the supernatant and then re suspending the micelles in purified water. This DMSO extr action procedure was carried out 3 times. The resulting micelles were either used immediately or lyophilized at 45C and 0.045 mbar to produce a dry product.

PAGE 69

55 Residual DMSO quantification After each iteration of the above described DMSO extraction proced ure, the supernatant was analyzed by UV spectroscopy to determine the DMSO concentration by measuring the absorbance at 208 nm against a pre constructed standard curve. This yielded the amount of residual DMSO left in the system as a function of purificati on step. In addition, a secondary method was employed to directly measure the amount of residual DMSO contained in fabricated micelles. After three and five purification iterations, micelles were suspended in dH 2 O and sonicated for 1 hour at elevated tempe rature in order to lyse the micelles and permit release of entrapped DMSO. After centrifugation, the supernatants in these samples were also measured by UV spectroscopy to determine the residual DMSO concentration. The measured DMSO amount was divided by t he total micelle mass in each sample to express these data as the percent DMSO content remaining in the system. Polymer and micelle characterization 1 H NMR spectra were collected on an Inova 500 MHz NMR spectrometer (Varian, Inc., Palo Alto, CA) at 25C wi th a Nalorac NTR probe using CDCl 3 as the solvent. Spectra are displayed in ppm using the solvent peak as an internal reference. Molecular weight data were collected on an EcoSEC GPC (Tosoh Biosciences, King of Prussia, PA) with a refractive index (RI) det ector using DMF/LiBr as a mobile phase and poly(methyl methacrylate) standards. Particle size data was determined by dynamic light scattering (DLS) on a Nicomp 380 ZLS (PSS, Port Richey, FL) with micelles suspended in PBS at a concentration of 1.0 mg/mL. A ll DLS data are plotted using an intensity weighted distribution. For particle morphology

PAGE 70

56 assessments, dry micelles were sputter coated with 5nm Au and examined by field emission scanning electron microscopy (FESEM) on a JSM 7401F (JEOL USA, Peabody, MA). Cytotoxicity testing Cytotoxicity tests were carried out per ISO 10993 5 using an extract exposure procedure against human retinal pigmented epithelial (ARPE 19) cells and assessed by MTT assay. Cells were maintained at 37C and 5% CO 2 in 1:1 DMEM:F 12 sup plemented with 5% FBS and 1% PS. To prepare extractions, the polymer was added to complete media at 5 wt% and incubated at 37C for 48 hours. Dilutions of the extraction were prepared by diluting the extraction with complete medium. For cytotoxicity tests, cells were seeded at 10,000 cells/well and incubated for 24 hours at which point the media was replaced with the extraction medium. After 24 hours of further incubation, cellular metabolic activity was ions with absorbance read at 540 nm on a Synergy 4 Multi Mode Microplate Reader (BioTek, Winooski, VT). Data were normalized to the absorbance of control wells incubated in pure media. Statistical significance was assessed by ANOVA (MATLAB, MathWorks, Nati ck, MA) at a significance level of 0.05. In vitro TA release testing Immediately after fabrication, micelles were suspended in 1 mL sterile PBS (pH 7.4) and inserted into a 3500 MWCO dialysis chamber (Float A Lyzer G2, Spectrum Labs, Rancho Dominguez, CA) floating in a conical tube containing 14 mL of sterile PBS. The tubes were incubated at 37C on a rocking platform and samples were pulled periodically by removing 5 mL of the release medium outside of the dialysis chamber and replacing it with

PAGE 71

57 fresh PBS. The concentration of TA was determined by UV/visible spectroscopy (GENESYS 10S UV Vis, Thermo Scientific, Logan, UT) at 23 7 nm against a pre constructed standard curve (see Appendix B). Results and d iscussion Polymeric micelles have been effectively employ ed in myriad drug delivery applications and are particularly advantageous in their capacity for stabilized encapsulation of hydrophobic compounds [167] Micelles are character ized by a hydrophobic core and hydrophilic shell structure, wherein hydrophobic drugs can be encapsulated at the core while the hydrophilic shell provides solubility and protection against in vivo clearance mechanisms. The block co polymer structure most c ommonly employed in micelle design to achieve the contrasting core and shell properties lends great flexibility in tuning the micellar properties to the molecule of interest. Polymer synthes is and characterization The structure of N Boc serinol was confir med by 1 H NMR spectroscopy (500 MHz, CDCl 3 block copolymer PEG PHS PEG was synthesized in low and high molecular weights by the methods described earlier. Its structure was confirmed by 1 H NMR as shown i n Figure 2.1 : (500 MHz, CDCl 3 (2H, s), 4.17 (1H, t), 4.09 (4H, d), 3.64 (4H, t), 3.40 (3H, s), 3.18 (4H, t), 1.51 (4H, quin), 1.45 (9H, m), 1.35 (4H, quin). Peak assignments were corroborated by NMR modeling (Advanced Chemistry Development). Result s of GPC analyses of the low and high molecular weight copolymers are shown in Table 2.1

PAGE 72

58 The PEG PHS PEG copolymer was used as is in this work, but we predict that it can also be extensively modified to tailor it to the specific molecule being investiga ted. Specific methods of modification and their predicted impact on micelle performance are summarized in Table 2.2 The PHS block itself a copolymer benefits primarily from the inclusion of Boc protected serinol within the repeating unit. With all serino l primary amines protected, the frequent Boc groups contribute significant hydrophobicity to the PHS block and undoubtedly also influence micelle core density due to steric effects. As these are selectively removed, a) the hydrophobicity and steric effects of the PHS block are attenuated, allowing for a denser but less hydrophobic micelle core; and b) primary amines are exposed. Figure 2.1 : 1 H NMR (500 MHz, CDCl 3 ) analysis of the PEG PHS PEG copolymer confirmed the structure of the hydrophobic PHS block (N Boc serinol: B, G, H and J and HDI: A, C, D and I) and successful capping by mPEG (E and F). Peak assignments were corroborated by NMR modeling (Advanced Chemistry Development).

PAGE 73

59 Table 2.1 : Properties of high and low molecular weight PEG PHS PEG by GPC analysis. This latter outcome opens up an array of possibilities in that the primary amines can be used to conjugate antibodies or other targeting moieties for local ized micelle effect [1 64,168] This procedure would involve a partial or complete deprotection of the Boc groups of PEG PHS PEG after polymer synthesis is complete and then conjugation of the biomolecule of interest after micelles are fabricated. Additionally, as the fraction of Boc groups that are removed during the deprotection routine can be precisely controlled (data not shown), a balance can be found between retaining sufficient hydrophobicity to stabilize the micellar structure and exposing enough amines to allow conjugat ion. Investigation of this phenomenon is currently underway. Table 2.2 : Potential a venues for PEG PHS PEG modification Modification Predicted impact on micelle performance PHS block molecular weight Modification of micelle core hydrophobicity and density Selective Boc group deprotection in PHS block Reduced hydrophobicity of micelle core, increased potential for hydrogen bonding with drug, introduction of moderate positive charge to core Selective Boc group replacement Modification of micelle core hydrop hobicity, direct tethering of therapeutics or other biomolecules Poly(methylene) chain length in diisocyanate monomer Modification of micelle core density PEG block molecular weight Resistance to rapid in vivo clearance and protein adsorption Polymer M n [ k Da] M w [ k Da] PI Hydrophobic: hydrophilic ratio a) Low MW PEG PHS PEG 25 .3 48 .2 1.91 42.9 High MW PEG PHS PEG 32.5 92 .9 2.86 83.4

PAGE 74

60 Cytotoxic ity testing Cytotoxicty of the copolymer was assessed in vitro per ISO 10993 5 against ARPE 19 cells using the MTT assay for mitochondrial activity. As shown in Figure 2.2 mitochondrial activity between negative control samples incubated in pure media a nd experimental samples incubated in media with PEG PHS PEG extracts at all concentrations were statistically indifferent (p=0.265). Positive control samples incubated in complete media with 10% DMSO were the only samples that were statistically different (p=1.65 x 10 5 ). Micelle fabrication Drug loaded micelles were fabricated by two methods: sonication and extrusion. The sonication procedure is similar to a nanoparticle fabrication procedure in which an oil in Figure 2.2 : The PEG PHS PEG copolymer was non cytotoxic against ARPE 19 cells as assessed by MTT mitoc hondrial activity assay. Positive control samples (10% DMSO in culture media) were the only samples that were significantly different (p<0.0001); all experimental groups were statistically insignificant from the negative control sample of pure culture medi a (p=0.265). Data were normalized to the media only control sample; means + standard deviations shown (n=5).

PAGE 75

61 water emulsification is formed, followed by sonication to produce nano scale particles. The extrusion procedure also starts with an oil in water emulsification but uses serial passage through a filter instead of sonication to produce the final particles. A comparison of the size distribution of part icles fabricated from high MW PEG PHS PEG by the two techniques is shown in Figure 2.3 Micelles fabricated by the sonication technique had a mean diameter of 357.4 nm with a 42.3% relative standard deviation (RSD). In stark contrast, those obtained by extrusion through a 100 nm pore size filter had a mean diameter of 217.5 nm with a 12.3% RSD, indicating a significantly more monodisperse population. That the mean diameter of extruded micelles is larger than the average pore size of the filter material u sed can be explained by: a) the oil phase of the emulsification forming cylindrical micelles as they Figure 2.3: The size distribution of micelles obtained by extrusion was significantly more monodisperse than that obtained by sonication (12.3% re lative standard deviation (%RSD, micelles at 217.5 nm compared to 3 57.4 nm for sonicated micelles. Both populations were produced using high MW PEG PHS PEG. Plot shown for intensity weighted distributions.

PAGE 76

62 pass through the filter pores, which then collapse to spheres larger in diameter than the pore size after passing through the filter; and b) potential coal escence of micelles after extrusion but before DMSO is removed. Beyond producing a more monodisperse population of micelles, which will reduce batch to batch variability during the fabrication process, the filter extrusion process also carries other advant ages over current techniques. First, fabrication time is greatly reduced; the process takes under one hour, whereas a typical dialysis procedure takes over a day to complete. Second, the size of micelles produced by this process can be readily controlled b y simply by changing the average pore size of the extrusion filter membrane. Third, unlike the sonication procedure, the extrusion technique contains no steps that would be potentially damaging to structure sensitive molecules such as proteins or antibodie s. Finally, the greatest advantage of the extrusion process is that so long as the filter membrane pore size is less than 200 nm the micelles are pre sterilized and need not undergo secondary sterilization procedures, as would be the case with other fabric ation techniques. These secondary sterilization techniques can be especially damaging to polymeric systems as they can increase the toxicity of or alter the physi c o chemical characteristics of the final system [168 170] In addition to characterization by DLS, micelle morphology was examined by FESEM and TEM. TEM micrographs confirmed the core/shell structure that was expected, showing the higher contrast hydrophobic core surrounded by the lower contrast PEG shell (Figure 2.4 Panel D). Additionally, FESEM was used to investigate sonicated samples immediately after fabrication (Figure 2.4 Panel A) and at 2 weeks after incubation in PBS (Figure 2.4 Panel B) and extruded samples immediately after fabrication (Figure 2.4 Panel C). At all

PAGE 77

63 time points and by both methods, micelles had high sphericity. The differences noted in polydispersity by DLS were also evident in the micrographs; extruded samples were significantly more monodisperse. There was also a large difference noted in micelle diameters between the 0 and 2 week samples. As determined by post hoc image analysis, micelles immediately after fabrication had a mean d iameter of 369 nm, correlating fairly closely with DLS results. However, the mean diameter increased drastically within 2 weeks to 962 nm, indicating significant swelling of the micelles when incubated in PBS. Figure 2.4 : Micelle size and morphology assessed by FESEM after fabrication by sonication (A) and extrusion (C) corroborated DLS data and also demonstrated sphericity of produced particles. Micelles fabricated by sonicati on and incubated in PBS for 2 weeks (B) had significantly larger diameters than post fabrication samples, indicating significant swelling of the particles upon incubation. TEM imaging (D) confirmed the core/shell structure expected of micelles with a poly( ester urethane) core and PEG shell.

PAGE 78

64 Residual DMSO levels Solvents often used in pharmaceutical formulations such as DMSO and NMP are generally considered non cytotoxic at lower concentrations [82,83] H owever, some studies ha ve indicated mytotoxicity [84] and ocular specific low dose cyto toxicity [85] of DMSO, which raise concerns about its use in biomedical applications. In order to address concerns related to the use of DMSO in our micelle fabrication technique, we evaluated the efficacy of our DMSO removal process. The amount of DMSO removed at each iteration of the purification process is shown in Figure 2.5 Of the 55 mg of DMSO used in the initial process, 50.7 1.57 mg was removed in the first step, 1.71 0.184 mg was removed in the second step and 0.035 0.002 mg was removed in the third step. No DMSO was detectable in the supernatants at the fourth and fifth steps. These data confirm that the purification process used in the micelle fabrication process efficiently removes residual DMSO within three iterations. Figure 2.5: The amount of DMSO removed per purification iteration decreased with each additional iteration. The vast majority was removed in the first iteration with smaller amounts removed in the second and third itera tions. The amount of DMSO in the supernatant at the fourth and fifth steps was below detectable levels. Means and standard deviations of n=3 samples are shown.

PAGE 79

65 Samples of micelle s were also evaluated at three and five iterations directly for the amount of DMSO remaining within the product. After the third iteration DMSO accounted for 0.303 0.260% of the system and this value dropped to 0.137 0.030% after the fifth iteration. These findings confirm that the current purification process efficiently lowers DMSO to low levels within 3 5 iterations. Additionally, both of these residual DMSO values are below the 1.0% minimum concentration found to cause ocular cytotoxicity by Galvao, et al [85] Release of triamcinolone aceto nide TA was loaded into micelles during fabrication by co dissolution with PEG PHS PEG in the organic phase of the emulsion. The TA loading ratio was fixed at 10 wt% and the effect of varying the PEG PHS PEG molecular weight was examined by in vitro releas e testing. As shown in Figure 2.6 all samples showed varying levels of burst release within the first several days of testing, but for all samples this effect involved less than 6.0% of total drug load and was as low as 2.0% for samples produced by sonica tion. The release rate of TA from extruded micelles was inversely related to molecular weight: the average release rate for low and high molecular weight extruded micelles was 0.321% and 0.117%/day, respectively, compared to 0.175%/day for those produced b y sonication. These release rates represented significantly slower release than free drug, which was released at 6.49%/day. Excluding the initial burst period, R 2 fit to zero order was 0.939 and 0.943 f or low and high molecular weight extruded micelles, re spectively, and 0.986 for sonicated samples. All of the micelle systems were capable of sustaining near zero order release of TA for extended periods of time. Assuming consistent release rates were to be maintained, extrapolated zero order kinetics indica tes that low and high molecular weight extruded

PAGE 80

66 micelles would continue releasing TA for 350 52.5 and 1220 267 days, respectively, compared to 15 2.72 days for free drug. These projected release time frames from PEG PHS PEG micelles represent a signi ficant improvement over free drug and previous systems intended for TA delivery [165,171,172] and could potentially significantly improve the clinical administration of TA. To put this release behavior in perspective, Ozurdex is a currently marketed PLGA intravitrea l implant for the delivery of dexamethasone. Dexamethasone is a corticosteroid very similar in structure and indications to TA with a molar mass of 392.5 (as compared to 434.5 for TA). Even though Ozurdex is a large scale polymer implant, its 6 month sust ained dexamethasone delivery timeframe [173] is far shorter than what PEG PHS PEG micelles were able to achieve with TA. Further, since Ozurdex is a bulk polymer implant, Figure 2.6 : Release of TA from extruded and sonicated PEG PHS PEG micelles was near zero order for all conditions. TA release rate was inversely related to the molecular weight of PEG PHS PEG. Means standard deviations shown for n=3 samples.

PAGE 81

67 the PEG PHS PEG micelles will enable a much more facile injectable delivery, thereby reducing the trauma and potentially t he side effects associated with the procedure. Specifically compared to current administration techniques for TA, micelle systems have a further advantage over suspensions in that they are less obstructive to vision. Suspensions of TA which are opaque by n ature have been reported to cause transient obstructions in the visual fields of patients for several days or longer after administration [174] until the drug settles at the inferior inner limiting membrane of the retina. Nanoparticulate systems such as the micelles developed i n this work avoid this clinical drawback due to their small size and hydrated state (i.e. PEG shell), which would be expected to render them invisible to the patient upon injection. Mechanistic insights from release kinetics To better understand the capabi lities of the PEG PHS PEG micelle system, an analysis of the time dependent release kinetics and FESEM micrographs can elucidate the mechanisms driving TA release from the micelle core. Based on these data, the mechanism of drug release appears to be a com bination of swelling of the micelle core and simple diffusion. Based on the release kinetics in the first several days and the significant diameter increase observed by FESEM within the first 2 weeks, it is likely that most of this swelling occurs within t he first several days of aqueous incubation and is the mechanism responsible for the modest burst release observed early in the release profiles. The magnitude of this initial burst release observed for PEG PHS PEG micelles was significantly smaller than i s typical of other micelle or nanoparticle systems [175,176] Nevertheless, there was a difference in burst release behavior b etween sonicated and

PAGE 82

68 extruded micelles. It is hypothesized that this difference results from a mild annealing process concomitant with the sonication procedure, resulting from sample heating during the extended sonication time. Compared to other polymeric micelle systems, PEG PHS PEG micelles have several distinct advantages. The theoretical ability to control the hydrophobicity of the micelle core, while still being fully characterized, should allow the system to be tailored to a specific drug molecule. T his partial Boc deprotection strategy could also be used to controllably introduce a positive charge to the micelle core, allowing the use of electrostatic interactions to further modulate drug release kinetics or other favorable interactions. Also under i nvestigation are the biodegradation kinetics of the micelle system; however, it is expected that the polymer would be fully biodegradable through the carbamate bonds present throughout the PHS block. Conclusions The micelles developed in this work have the potential to significantly improve patient care by allowing repeat injections of TA to be replaced by a single administration of TA loaded micelles. This approach would allow for continuous therapeutic levels of TA for well over one year. The pre cursor p olymer was found to be fully biocompatible and contains considerable levels of flexibility that have not yet been fully explored. In addition, the extrusion process developed in this work has many benefits over other commonly used fabrication procedures an d greatly simplifies the scale up process required for commercialization. This system will continue its translational path and is currently being tested in animal models for in vivo biocompatibility.

PAGE 83

69 CHAPTER III A REVERSE THERMAL GEL SYSTEM FOR LOCALIZED OCULAR DELIVERY OF TRIAMCINOLONE ACETONIDE Posterior delivery of ocular therapeutics continues to be an unsolved clinical challenge. In this work, a reverse thermal gelling polymer was developed to improve delivery of triamcinolone acetonide (TA) to poster ior targets. The polymer was synthesized by grafting the thermo sensitive polymer poly( N isopropylacrylamide) (PNIPAAm) to a novel poly(urethane urea) backbone. The resulting copolymer was thoroughly characterized by structural and thermo mechanical assess ments as well as in vitro cytotoxicity studies against human retinal pigmented epithelial (RPE) cultures and in vivo biocompatibility studies through intravitreal injections in rats. In vitro release experiments with TA were performed with varying gel conc entrations and TA loading fractions. The copolymer transitions from a low viscosity solution to a stable physical gel near 30C. No cytotoxic effects were found against RPE cells in culture and intravitreal injections of the system in rats showed no advers e reactions. Release kinetics were independent of initial polymer concentration but did depend on TA loading fraction. All release profiles followed first order kinetics and samples loaded with 20 wt% TA sustained meaningful levels for 80 days. This novel reverse thermal gelling polymer system combines high biocompatibility, favorable TA release kinetics and an advantageous delivery method to potentially improve the posterior administration options for corticosteroids.

PAGE 84

70 Introduction Ocular drug delivery is a uniquely challenging field primarily due to the complex anatomical and physiological barriers of the eye. While these barriers are especially problematic for topical administration of drugs, it has persisted as the dominant delivery mechanism of ocular drugs primarily due to its familiarity to patients and ophthalmologists and ease of formulation [177] However, as the pharmaceutical industry continues to shift away from drugs that exhibit moderate to high aqueous solubility and thus can be readily formulated for topical administration new delivery systems will be requ ired. One class of materials that has recently received significant attention for ocular applications is the in situ gelling polymer system [78,106,178] Because they can be readily deployed by injection to posterior periocular or intraocular target sites, these systems can be used to overcome the aforementioned barriers and deliver drugs at or near the target tissue by minimally invasive methods. Generally, in situ gelling polymer systems begin as a low viscosity liquid that is injected through a small gauge need le or cannula and form a physical or cross linked gel via introduction of a specific stimulus. Reverse thermal gels (RTG) are a class of in situ gelling systems that undergo a reversible solution to gel (sol gel) phase transition by temperature change alon e [108,179] These are advantageous in that they remove the need to introduce reactive species or an external energy source to realize gelation of the system after deployment. Further, since most RTG systems are composed of amphiphilic p olymers and undergo sol gel transitions based on the temperature dependent development of hydrophobic interactions, they are especially well suited for encapsulation and delivery of poorly soluble molecules [108]

PAGE 85

71 RTGs have been explored for various tissue engineering and drug delivery applications [91,180 182] RTGs developed for these applications are most commonly compri sed of amphiphilic block copolymer systems, such as poly(ethylene oxide) poly(propylene oxide) poly(ethylene oxide) (poloxamers) [88,183] and poly(ethylene glycol) poly(lactic co glycolic acid) poly(ethy lene glycol) [106,184,185] While these systems are easy to synthesize and their thermal gelling chara cteristics are readily modified by adjusting the block ratio of the resulting copolymer, the dependence of the thermal gelling properties on the hydrophilic hydrophobic balance limits their ability for functionalization, e.g. with highly hydrophilic antibo dies. To overcome this limitation, the present work instead made use of poly( N isopropylacrylamide) (PNIPAAm) to introduce thermal gelling properties. A novel RTG polymer has been developed that can specifically meet the needs of ocular drug delivery appli cations. The novel chemistry of this RTG polymer is based on a biomimetic poly(urethane urea) with functionalizable primary amines on each repeating unit of the polymer. A fraction of these primary amines are used to graft the thermal sensitive PNIPAAm, wh ich lends the copolymer reverse thermal gelling properties. This graft co polymer structure makes for a flexible system since it permits higher levels of PNIPAAm content in the final system as compared to a linear block copolymer. In this work, this novel RTG system was fully characterized and assessed for its in vitro and in vivo biocompatibility and capacity to sustain drug release. Since this system has most potential in posterior ocular applications where it can be used to overcome access based challeng es, the poorly soluble corticosteroid triamcinolone acetonide (TA) was chosen as a model drug.

PAGE 86

72 Materials and m ethods Materials azobis(4 cyanovaleric acid) (ACA), methanol, phe nolphthalein, L glutamine and dimethyl sulfoxide (DMSO) were obtained from Sigma Aldrich (St. Louis, MO). Di tert butyl dicarbonate, ethyl acetate, N hydroxysuccinimide (NHS), N (3 dimethylaminopropyl) ethylcarbodiimide hydrochloride (EDC) and dimethyl sulfoxide d6 (DMSO d6) were obtained from Alfa Aesar (Ward Hill, MA). Ethanol, hexane, diethyl ether and sodium hydroxide (NaOH) were obtained from Fisher Scientific (Pittsburgh, PA). Dimethylformamide (DMF) was obtained from BDH Chemicals (Poole, UK). Dic hloromethane (DCM) was obtained from J.T. Baker (Phillipsburg, NJ). N isopropylacrylamide (NIPAAm) was obtained from Acros Organics (Geel, Belgium). Chloroform d (CDCl 3 ) was obtained from Millipore (Billerica, MA). Vybrant MTT cell proliferation assay kit was obtained from Life Technologies (Carlsbad, CA). TA was obtained from Fluka (St. Louis, MO). All other reagents were obtained from Thermo Scientific (Waltham, MA). Polymer synthesis N Boc serinol was synthesized through Boc group protection of serinol. Serinol (21.5 mmol) was dissolved in 20 mL of absolute ethanol and stirred at 4C. Di tert butyl dicarbonate (26.0 mmol) was dissolved in 20 mL of absolute ethanol and added dropwise to the serinol solution over a period of one hour, while maintaining 4C and constant stirring. The solution was heated to 37C with vigorous stirring and reacted for one hour.

PAGE 87

73 The ethanol was removed by rotary evaporation at 45C and 10 mbar vacuum and the solid was re dissolved in a 1:1 mixture of hexane and ethyl acetate by gentle heating. Additional hexane was added until precipitation was observed and the resulting suspension was stored at 4C overnight to allow recrystallization. Subsequent vacuum filtration yielded a white flaky product. The backbone poly(urethane urea) w as synthesized by copolymerization of N Boc serinol, urea and HDI. N Boc serinol (6.0 mmol) and (6.0 mmol) urea were lyophilized for 24 hours and then dissolved in 6 mL of dry DMF. Under a nitrogen atmosphere, the reaction was brought to 90C under moderat e stirring and HDI (12 mmol) was added drop wise through the septum. The reaction was carried out for 7 days, after which it was cooled and precipitated twice in diethyl ether and once in water. After lyophilization, a dry yellowish product was obtained, p oly(serinol hexamethylene urea) (PSHU), in 86.5% yield. In order to expose primary amines for conjugation, Boc protective groups were removed through a complete deprotection routine. PSHU (0.10 g) was dissolved in a 50:50 (v/v) mixture of DCM and TFA (18 m L total volume) and reacted uncapped with rapid stirring for 15 minutes. The solvents were immediately removed by rotary evaporation at 45C and 10 mbar vacuum and the product re dissolved in DMF. Two precipitations in diethyl ether and terminal rotary eva poration yielded a dry yellowish product, de protected PSHU (dPSHU). Carboxylic acid terminated PNIPAAm (PNIPAAm COOH) was synthesized based on a procedure described previously [180] NIPAAm (5.0 g) and ACA (0.060 g) were dissolved in 25 mL dry methanol and bubbled wi th nitrogen for at least 30 minutes at room temperature. The reaction was raised to 68C under moderate stirring and carried out for 3

PAGE 88

74 hours. The product was precipitated twice in 60C water and dialyzed against 1 L dH 2 O for 48 hours in 3,500 Da molecular weight cut off (MWCO) dialysis tubing (Spectrum Labs, Inc., Rancho Dominguez, CA). The RTG was synthesized by grafting PNIPAAm COOH to dPSHU through carbodiimide linking chemistry. Of the calculated 18 repeating units (and, thereby, primary amines) in dPSH U, 25% were used for PNIPAAm conjugation. PNIPAAm COOH (14.0 mol), NHS (42.0 mol) and EDC (42.0 mol) were dissolved in 3 mL DMF and reacted for 24 hours. dPSHU (3.10 mol) dissolved in 2 mL DMF was then added drop wise and the conjugation reaction proce eded for 24 hours. The product was added directly to a 50 kDa MWCO dialysis tube and dialyzed for 48 hours to remove unreacted PNIPAAm. The product was lyophilized to yield a white flaky product, PSHU NIPAAm. The full synthetic route of PSHU NIPAAm is show n in Appendix C. Polymer characterization The molar mass of PNIPAAm was determined experimentally by titrating for the carboxylic acid end groups. Approximately 0.050 g of PNIPAAm was dissolved in 10 mL dH 2 O with 10 L of phenolphthalein solution (2 wt% in absolute ethanol). The solution was titrated to the end point by adding 0.01 N NaOH. The molecular weight distribution of PSHU was determined by gel permeation chromatography (GPC) on an ECOSEC system (Tosoh Biosciences, King of Prussia, PA) with a TSKge l colum (Tosoh Biosceinces) using DMF with LiBr as the mobile phase. Separations were made at 25C and detected by refractive index against poly(methyl methacrylate) standards. Full results of the GPC analysis are shown in Appendix D.

PAGE 89

75 Proton nuclear magnet ic resonance ( 1 H NMR) spectra were collected on an INOVA 500 MHz instrument (Varian) with a 5 mm triple resonance proton detector. Spectra were collected in a mixed solvent of 5% (v/v) DMSO d6 in CDCl 3 at 25.0C and post processed in ACD 1D NMR Processor s oftware (Advanced Chemistry Development, Inc., Toronto, ON). Elemental analysis was performed by Micro Analysis, Inc. (Wilmington, DE). Fourier Transform Infrared (FT IR) spectra were collected on a Nicolet 6700 (Thermo Fisher Scientific, Waltham, MA) usin g polyethylene windowed cards. Lower critical solution temperature (LCST) measurements were made on a Cary 100 UV visible spectrophotomer (Agilent Technologies, Inc., Santa Clara, CA) equipped with a temperature controlled 6 cell stage. Polymers were disso lved in PBS (pH 7.4) at 1 wt% and transmittance readings were taken at 500 nm while the temperature was increased at 0.5 C/min. Rheological studies were performed on a TA ARES rheometer (TA Instruments, New Castle, DE) equipped with a peltier plate. A 5wt % solution of PSHU NIPAAm dissolved in PBS (pH 7.4) was placed between the plates and a small bead of silicon oil was applied around the outer edge to prevent evaporation. Low temperature strain sweeps with and without silicon oil were used to ensure measu rements were not affected by its presence. All experiments were performed in the linear deformation region, as defined by preliminary strain sweep experimen ts. Measurements were made at 1% strain and a frequency of 1 rad/s while the temperature was swept f rom 20C to 40C at 0.5C/min.

PAGE 90

76 In vitro and in vivo biocompatibility studies In vitro biocompatibility studies were performed by directly exposing cells in culture to the PSHU NIPAAm RTG system in complete growth medium. ARPE 19 primary retinal pigmented epithelial cells were cultured in complete growth medium containing 1:1 DMEM:F12, 10% FBS and 2 mM L glutamine at 37C and 5% CO 2 For cytotoxicity studies, PSHU NIPAAm was dissolved in the complete medium at 5 and 10 mg/mL. Cells were seeded in 96 well p lates at 5000 cells/well, grown for one day and the media were replaced with the polymer samples. Positive controls of 5% DMSO in complete growth medium and negative controls of complete growth medium alone were also included. At 1, 3 and 5 days, MTT assay All animal studies were performed according to guidelines set forth by the Institutional Animal Care and Use Committee. Male Brown Norway rats (240 275 g; Charles River Laboratories, Wilmington MA) were housed in group cages and maintained on a 12 hour light/12 hour dark cycle. Animals had free, continuous access to food and water throughout the experiment. Rats received intravitreal injections of 2 L of PSHU NIPAAm in PBS (solutions were filt er sterilized before injection) in the left eyes and 2 L of PBS (pH 7.4) in the right eyes under isoflurane anesthesia. At 3 and 14 days (n=3 per time point), rats fo r 24 hours followed by neutral buffered formalin for 4 hours and 70% ethanol overnight. Eyes were then embedded in paraffin, sectioned and stained with hematoxylin and eosin for histological analysis.

PAGE 91

77 Release testing In vitro release testing was performed on activated PSHU NIPAAm gels with TA as a model hydrophobic drug with relevance for posterior ocular administration. PSHU NIPAAm was dissolved in PBS (pH 7.4) to the desired concentration and TA was suspended in this solution by brief sonication. A sampl e of 50 L of this solution was then pipetted into separate wells of a 24 well plate and the gel was activated by incubation at 37C. After several minutes, 3 mL of PBS was added and the plate was covered and incubated with gentle agitation. At specified t ime intervals, the release medium was removed for TA concentration determination and 3 mL of fresh PBS was added to the wells. TA concentrations in the release media were measured by UV absorbance at 237 nm against a pre constructed standard curve (shown i n Appendix B). De swelling studies The de swelling behavior of PSHU NIPAAm gels was determined gravimetrically during gel incubation in aqueous media. PSHU NIPAAm gels of various concentrations were pre formed in 2 mL syringes (500 L gel volume). After co mplete gelation, they were gently transferred to 5 mL of pre warmed PBS (pH 7.4) at 37C and maintained at this temperature. At pre determined time points, they were removed, weighed and returned to PBS. The entire process was performed in a 37C chamber t o maintain gel integrity. Results and d iscussion In this work, a novel RTG was designed, synthesized and characterized for application to posterior ocular delivery of a poorly soluble corticosteroid. A biomimetic polymer, PSHU, was grafted with the tempera ture sensitive homopolymer PNIPAAm to produce a

PAGE 92

78 copolymer with a high level of expected biocompatibility and a high capacity to control release of hydrophobic drugs. Polymer structural characterization The structure of the backbone polymer, PSHU, before an d after complete deprotection of Boc protective groups was confirmed by 1 H NMR, GPC and elemental analysis. GPC results indicated, for PSHU, M n : 9 .73 k Da, M w : 12 .5 k Da, M z : 15.7 k Da and PI (M w /M n ): 1.29 The spectra of these polymers, PSHU and de protecte d PSH U (dPSHU), are shown in Figure 3.1 For PSHU, peaks associated with HDI were observed at 1.24 ( CO NH CH 2 CH 2 C H 2 C H 2 ), 1.40 ( CO NH CH 2 C H 2 ), 3.03 ( C H 2 CH 2 NH CO NH), 3.11 ( CO NH C H 2 CH 2 ), 5.28 ( CO N H CH 2 ) and 5.53 ( CH 2 N H CO NH) ppm, peaks a ssociated with N Boc serinol were observed at 1.34 ((C H 3 ) 3 C ), 3.96 ((CH 2 ) 2 C H NH ), 4.05 ( O C H 2 CH )and 4.81 ( CH N H CO ) ppm and the peak associated with urea was observed at 10.86 ( CO N H CO N H CO ) ppm. Peak assignments for all protons were confirmed through the use of an NMR simulation package (Advanced Chemistry Development, Inc.). Based on calculated integral values of peaks positively associated with each monomer (urea, N Boc serinol and HDI), their molar ratios in the final copolymer were calcul ated to understand the overall copolymer structure (see Appendix E ) The molar ratios relative to N Boc serinol were found to be 2.13 for HDI and 0.13 for urea. Based on the monomer ratio in the polymerization, the expected monomer ratio in the copolymer w as 1:2:1 ( N Boc serinol:HDI:urea). While the N Boc serinol:HDI ratio was near the theoretical value, the ratio of urea in the final copolymer was lower than expected, by a factor of 7.7. This

PAGE 93

79 difference is likely attributed to differences in the reactiviti es of N Boc serinol and urea to HDI. Indeed, ureas have been reported to be 6.67 times less reactive than primary hydroxyls to isocyanates [186] so the observed difference of molar ratios in the final copolymer may be expec ted. 1 H NMR s pectra before and after the complete deprotection routine confirmed removal of all Boc protective groups through disappearance of the peak at 1.34 ppm, thereby exposin g primary amine functionalities (Figure 3.1, Inset). After complete deprot ection dPSHU contains one primary amine functionality per repeating unit. This high density of functionalizable amine groups makes this polymer particularly well suited for further applications in tissue engineering (e.g. where these groups could be used to covalently tether bioactive molecules) or specific drug delivery applications (e.g. where covalent linking of the drug molecule to the delivery polymer Figure 3.1 : 1 H NMR spectroscopy confirmed the expected structure of PSHU. Inset : The proton peak associate d with Boc protective groups at 1.34 ppm in the base polymer (*) was absent after the complete deprotection routine (**).

PAGE 94

8 0 would be advantageous). While these have not been described herein, other work in our lab has explore d these possibilities [187] Table 3.1 : Elemental analysis results of the base polymers and resulting copolymer. Polymer % C % H % N % O PSHU 50.35 7.96 15.77 23.59 dPSHU 42.77 6.60 15.07 28.58 PNIPAAm 62.90 9.64 12.82 14.64 PSHU NIPAAm 59.56 10.15 12.04 18.01 Elemental analysis of the PSHU and dPSHU starting polymers furt her confirmed the expected structures, with results listed in Table 3.1 matching closely with theoretical values (within 2%). Elemental analysis was also used to confirm the copolymer structure of PSHU NIPAAm. Using the results of dPSHU and PNIPAAm as boun ds, the conjugation ratio of PSHU NIPAAm was calculated as 3.2 mol PNIPAAm/mol dPSHU. Given a M n of PNIPAAm of 28,128 Da (as determined from the end group titration method), the M n of the resulting copolymer was calculated as 99,734 Da. In addition, FT IR was used to confirm successful conjugation of PNIPAAm to dPSHU with these results shown in Figure 3.2 This was accomplished by monitoring three regions of interest. The first, at 798 cm 1 was associated with the N H wag of primary amines. This peak was absent in PSHU but showed up after the deprotection routine in dPSHU and was slightly shifted and reduced in intensity in PSHU NIPAAm. The second, at 950 cm 1 was associated with the O H bend of carboxylic acids and was present in PNIPAAm COOH with a minu te peak in PSHU NIPAAm, indicating some small amount of unconjugated PNIPAAm was present in the final copolymer. Finally, the broad peak near 1650 cm 1 was associated with the C=O stretch of amides. Of particular note is that this peak was slightly

PAGE 95

81 shifted in dPSHU (1700 cm 1 ) and PNIPAAm (1650 cm 1 ) and indications of both amide were present in the spectrum of PSHU NIPAAm indicating that both species were present in the final copolymer. The presence of significant amide bonding in the PSHU backbone is ex pected to lend PSHU NIPAAm biodegradability through enzymatic and oxidative routes. Proteolytic enzymes have been shown to be capable of breaking down amides, even within polymer backbones [188] Indeed, a polyurethane structurally similar to PSHU was previously shown to degrade over several months in the presence of enzym atic activity [179] Thermo mechanical characterization I nvestigation of thermal gelling characteristics of RTG systems have been reported by a multitude of methods. The two most common methods are: a) measurement of the LCST Figure 3.2 : FT IR analyses of the base polymers and the final copolymer suggested near perfect conjugation in PSHU NIPAAm through the peaks at 1650 1700 cm 1 (A, amide C=O stretch), 950 cm 1 (B, carboxyl O H bend) and 798 cm 1 (C, primary amine N H wag).

PAGE 96

82 (or cloud point ) by temperature controlled spectroscopy; and b) measurement of the temperature dependence of viscoelastic properties by rheometric analysis. While LCST measurements provide a rapid, easy measurement of gelling characteristics, these are typically performe d at concentrations lower than are practically used and may be more sensitive to polymer solvent interactions and end group effects. On the other hand, a rheological study can more accurately assess the temperature dependent development of mechanical pro perties at a relevant polymer concentration. The cloud points of PNIPAAm COOH and PSHU NIPAAm, as assessed by monitoring the transmittance of a 1 wt% solution at 500 nm, are shown in Figure 3.3 (left panel). No significant difference was noted between the two samples with a LCST of 32.1 C. While previous reports have suggested hydrophilic end groups, such as a carboxylic acid, can shift the LCST of PNIPAAm to higher temperatures, in the present polymer this end group effect would likely be too small given its large molar mass (28.2 kDa). As such, the LCST Figure 3.3 : Temperature dependent gelling kinetics were assessed by LCST measurements of PNIPAAm and PSHU NIPAAm (left panel) and by rheometric analysis (right panel). LCST values for both the PNIPAAm pre cursor polymer and the PSHU NIPAAm copolymer were near 32.1 C and the sol gel transition temperature was established to be 29.4 C.

PAGE 97

83 difference that could be expected after conjugation to a hydrophobic backbone such as PSHU would be mitigated by the molar ma ss imbalance between PNIPAAm and PSHU in the final copolymer (3.2:1 PNIPAAm:PSHU). Whereas LCST measurements were performed on dilute solutions (1 wt%), rheometric analysis of the PSHU NIPAAm copolymer was performed at a more application relevant concentra tion of 5 wt%. Under these conditions, a sol gel transition temperature of 29.4 C was observed, as shown in Figure 3.3 (right panel). The sol gel transition temperature was LCST would be expected given the 5 fold difference in concentration used for each respective analysis. At lower concentrations, the mean distance between polymer molecules would be expected to be greater, thereby requiring more energy and time to bring these molecules into interaction. At higher concentrations, these interactions can begin to occur sooner and with less thermal energy, as observed in the lower sol g el transition temperature. Finally, the range of PSHU NIPAAm concentrations amenable to gel formation and approximate gelation times are shown in Appendix F Critically, that gel formation was possible to concentrations as low 1.66 wt% provides a comfortab le window below the concentrations projected for typical use (5 to 10 wt%), ensuring gelation even if significant dilution occurs after injection.

PAGE 98

84 Cytotoxicity and biocompatibility assessments To assess the cytotoxicity of the RTG, ARPE 19 cells were c ultured in 96 well plates and the RTG, dissolved in complete growth medium, was used to form a gel on top of the cultured cells. Two PSHU NIPAAm concentrations (5 and 10 mg/mL) were assessed; these concentrations were chosen as sufficient to form a gel but low enough to be easily recovered from the wells after the experimental time course. At 1, 3 and 5 days, the gel and remaining medium were removed and cellular metabolic activity was assessed by MTT assay. As shown in Figure 3.4 at all time points, no st atistically significant difference was Figure 3.4 : ARPE 19 cells cultured in direct contact with PSHU NIPAAm gels showed no statistically significant difference in metabolic activity at all time points as compared to cells cultured with complete growth medium alone at (* indicates p > 0.9, ** indicates p > 0.2). The only statistically significant difference was noted between the positive control (cells cultured with 5% DMSO in complete growth medium) and experimental samples at all time points (*** indicates p < 0.01).

PAGE 99

85 noted between the negative control samples of cells cultured in complete growth medium alone and either experimental concentration (p > 0.2). The only statistically significant difference was observed between experime ntal samples and positive controls of cells cultured with complete growth medium and 5% DMSO (p < 0.01). These data indicate that the PSHU NIPAAm gels are non cytotoxic against cultured ARPE 19 cells. Given these results, biocompatibility testing was ex tended to in vivo models. Rats received intravitreal injections of 2 L of PSHU NIPAAm in PBS (pH 7.4) at a concentration of 2.5 wt%. The contralateral eye of each rat served as an internal control (intravitreal injection of 2 L of PBS, pH 7.4). At 3 and 14 days eyes were enucleated, fixed and sectioned for histological analysis (n=3 rats per condition). Representative images of corneas (top row) and retinae (bottom row) for control eyes and rats sacrificed at 3 and 14 days a re show n in Figure 3.5 Figure 3.5 : In vivo biocompatibility testing of PSHU NIPAAm was carried out by intravitreal injection in rats. Contralateral eyes served as internal controls and received intravitreal injections of PBS, pH 7.4. Representative histological sections of corneas and retinae from control eyes and rats sacrificed at 3 and 14 days are shown.

PAGE 100

86 Eyes t hat received intravitreal injections of RTG showed no adverse reactions during the 14 day experimental time course. Visual assessments of rats indicated no signs of excessive blinking, inflammation, hyperemia or lens or corneal opacity, as would be indicat ive of a uveitic response. Histological analysis confirmed the lack of any signs of inflammatory responses. Corneal and retinal sections were all clear of inflammatory markers including foreign body giant cells and mast cells. Significant macrophage infilt ration was also absent from all sect ions. Vitreous and aqueous humor s were clear of infiltrating cells or other markers of adverse reactions. Some tissue separation observed in the Day 3 section was attributed to an artifact of the fixation/sectioning proc ess as it was also present in the contro l eyes for that time point. In vitro release testing In order to assess the ability of PSHU NIPAAm gels to sustain release of a hydrophobic small molecule drug, in vitro release kinetic testing was performed using t he corticosteroid TA. In the first experiment, the PSHU NIPAAm concentration was fixed at 5 wt% and the TA loading fraction was varied between 5 and 20 wt% (TA/RTG). In the second set of experiments, the TA loading fraction was fixed at 10 wt% and the PSHU NIPAAm concentration was varied between 2.5 and 10 wt%. In the first set of experiments, TA loading fractions up to 20 wt% showed no impact on the ability of the RTG system to form and maintain a stable gel. Gelling times were qualitatively observed to be similar between all samples. Release kinetics of TA from 5 wt% gels (shown in Figure 3.6 left panel) followed first order kinetics (full modeling details are provided in Appendix G ) and were strongly dependent on the TA loading fraction, with the rate of drug release increasing with TA loading fraction. The samples

PAGE 101

87 with the highest loading fraction, 20 wt%, were able to sustain meaningful levels of TA release for 80 days. In the second experiment, the influence of PSHU NIPAAm gel concentration on TA re lease kinetics was assessed. As shown in Figure 3.6 (right panel), the release kinetics again followed first order kinetics, as in the previous experiment. However, there was no statistically significant difference between the amount of TA released at any time point between the 2.5, 5 and 10 wt% samples (p > 0.2) This is a surprising result given the expected impact of gel concentration (which should affect RTG matrix density) on the ability of TA to partition out of the PSHU NIPAAm gel. Reports of other R TG systems indicated an as expected dependence of small molecule release kinetics on gel concentration. In order to understand these data, an experiment was designed to assess whether the observation that PSHU NIPAAm gels de swell upon incubation in an aqu eous environment could be a mechanism driving this effect. The de swelling behavior of Figure 3.6 : The release of TA from PSHU NIPAAm gels followed first order kinetics and was dependent on the TA loading fraction (left panel) but independent of the PSHU NIPAAm gel concentration (right panel). Me an and standard deviations of n=5 samples plotted.

PAGE 102

88 PNIPAAm based gels is based on the volume phase transition (VPT), as has been described in other works. This VPT is based on the coil to globule transition that PNIPAAm molecules undergo with increasing temperature [189] and is driven by an entropically driven liberation of water molecules (Shibayama, et al. estimated that 13 water molecules are liberated per monomer unit [116] ). In order to understand the magnitude of this effect on the PSHU NIPAAm system and its impact on drug release kinetics, the de swelling behavior of PSHU NIPAAm gels was monitored gravimetrically as a function of time for the RTG concentrations used in the release kinetics experiment, with these data shown in Figure 3.7 All samples rapidly lost significant water content wit hin the first several hours of incubation at 37 C and reached Figure 3. 7 : De swelling kinetics of PSHU NIPAAm gels at 2.5, 5 and 10 wt% were monitored gravimetrically at 37 C as a function of time. All samples rapidly lost water content wit hin the first 8 hours and reached an equilibrium state by 16 hours. After 40 hours there was no statistically significant difference in the gel concentration between all samples (p > 0.15). Mean and standard deviations of n=5 samples plotted.

PAGE 103

89 an equilibrium concentration by approximately 16 hours. This result indicates that the majority of the water content of PSHU NIPAAm gels is liberated within the first several hours of incubatio n. As a result, the gels rapidly become more concentrated in this initial period, forming a denser polymer matrix. Further, by the 40 hour experimental time point and thereafter, there was no statistically significant difference (p > 0.15) in the calculate d RTG concentration of PSHU NIPAAm gels despite their different starting concentrations (2.5, 5 and 10 wt%). By the 40 hour time point, the mean RTG concentration had increased to 82.1 +/ 4.9 wt%. This result indicates that PSHU NIPAAm gels rapidly achiev e an equilibrium de swollen state within approximately 20 hours of incubation and this de swollen state is independent of the starting RTG concentration. These data help to explain the results of the second release kinetics experiment, in which TA release kinetics were found to be independent of the starting RTG concentration. Within the first 20 hours of incubation, all of the gels would achieve the similar de swollen equilibrium concentrations, thereby negating the effect of starting concentration on rele ase kinetics. In practical terms, this result has the interesting consequence of de confounding the concentration of injected RTG and the release kinetics of the in situ formed gel. Instead, the starting RTG concentration could be selected based on the dos e of TA that is desired for administration. Clinically, this system represents an improvement over the current administration paradigm for TA, which involves high dose intravitreal injections of TA suspensions [165] These high dose suspension injections are intended to allow formation of an intravitreal depot of TA, which can provide long term therapeutic benefit. However, this

PAGE 104

90 administrat ion paradigm also has several drawbacks including: a) a high incidence of potentially serious side effects such as elevated intraocular pressure, endophthalmitis, cataract and retinal detachment [190] ; and b) transient obstructions in the visual fields of patients due to the opaque nature of the suspension, which can last as long as several days [174] Injection of the RTG, for example to a periocular target, would mitigate these drawbacks due to its limited burst release behavior and sustained TA release capability. For perspective, this system can be compared to the currently marketed implant Ozurdex (Allergan, Inc.), a PLGA intravitreal implant for the delivery of dexamethasone. Dexamethasone is a corticosteroid similar in structure and indications to TA with a molar mass of 392.5 (as compared to 434.5 for TA). While the release kinetics between the two systems are similar [173] the PSHU NIPAAm is differentiated from Ozurdex in two major areas. First, the magnitude of the burst release from the PSHU NIPAAm system is much less than in Ozurdex [191] which would be expected to result in a more favorable side effect profile from the thermal gel. Second, the PSHU NIPAAm thermal gel has a major advant age in its delivery characteristics. Whereas Ozurdex uses a 22 gauge applicator, which is on the border of what can be used for intravitreal injection without suturing, the PSHU NIPAAm system can be injected through needles down to 32 gauge, as is more com mon in intravitreal injections and would likely be better tolerated by treated patients.

PAGE 105

91 CHAPTER IV POLYMERIC MICELLES CONTAINING TRIAMCINOLONE ACETONIDE ENCAPSULATED IN A REVERSE THERMAL GEL AS AN INJECTABLE, IN SITU GELLING OCULAR DRUG DELIVERY SYSTEM Reverse thermal gels are a class of in situ gelling polymers that reversibly transition from a low viscosity solution at room temperature to a physical gel at body temperature. As ocular drug delivery systems, they are attractive for their injectability, i n situ gelling and high biocompatibility. To date, their application has been limited by their inability to sustain release of therapeutic drug levels for longer than 1 3 months To overcome this limitation, the present work explores the encapsulation of d rug loaded polymeric micelles within a reverse thermal gel to improve sustained drug release kinetics The combined system maintained the thermal gelling behavior of the reverse thermal gel, but release kinetics of the poorly soluble corticosteroid triamci nolone acetonide were superior in the combined system compared to the thermal gel or micelles alone. This novel combined system is especially well suited for long term posterior delivery of poorly soluble drugs, which remains an unmet clinical need. Introd uction Reverse thermal gels (RTGs) are a class of in situ gelling polymers that undergo a reversible phase transition from liquid to physical gel by temperature change alone. As compared to other in situ gelling polymer chemistries, RTGs are differentiated in that they do not require injection of reactive species (e.g. monomers or cross linkers) or introduction of outside energy sources (e.g. light probes) to realize gelation. Because of this property

PAGE 106

92 they are attractive systems for applications near sensit ive tissues or complex anatomical locations because they can be injected directly at the site of interest via a small gauge needle and rapidly form a stable physical gel upon heating to body temperature. For this reason, they have been extensively investi gated as injectable drug delivery systems [73,78,178] Despite their favorable injectability and in situ gel forming properties, RTG systems are limited in this application by a general inability to sustain long term drug release. In principle, this limitation can be understood as a consequence of the high water content of these systems, which permits relatively rapid efflux of drug from the gelled matrix. This water based drug efflux mechanism dictat es that RTGs are generally better at sustaining release of poorly soluble drugs [120] but su stained release time frames are still generally limited to less than 2 3 months [80,86,125] In order to improve the sustained delivery characteristics of RTGs, a small number of efforts have investigated encapsulating nano carriers (e.g. nanoparticles, liposomes or micelles) within the thermal gel as a secondary drug carrier [192 196] Because of their small size, these nano carriers have a minimal effect on thermal gelation kinetics and do not sacrifice the injectability of the system. However, because they efficiently solubilize, encapsulate and control release of poorly soluble drugs [143] they present an advantageous adjunct to RTGs. In this work, we describe the combination of two previously described systems a micelle and an RTG for sustained delivery and local administration of the poorly sol uble corticosteroid triamcinolone acetonide (TA). TA was chosen as a model drug due to its poor aqueous solubility, limited current delivery options [166] and need to achieve high posterior bioavailability for therapeutic efficacy. As such, the combined RTG micelle

PAGE 107

93 system de scribed herein may provide a platform for an injectable, in situ gelling, long term option for TA delivery Materials and m ethods Materials Triamcinolone acetonide was obtained from Fluka Chemicals (Sigma Aldrich, St. Louis, MO). P ho sphate buffered saline buffered formalin, paraffin, hematoxylin and eosin were obtained from Thermo Scientific (Logan, UT). All other reagents were obtained and used as previously described. Polymer synthesis TA loaded polyethylene glycol block poly(hexamethylene alt serinol) block polyethylene glycol (PEG PHS PEG) micelles were fabricated as previously described using a filter extrusion process (see Chapter II). The RTG, composed of a poly(urea urethane) backbone polymer grafted wi th poly( N isopropylacrylamide) (PNIPAAm) was also synthesized as previously described (see Chapter III) The copolymer was dissolved in PBS (1X, p H 7.4) to form a 5 wt% solution. For the combined system, micelles were physically mixed with the RTG solutio n by brief vortexing. The micelle loading concentration used throughout this study was 10 wt% (micelle mass/total RTG mass including water). Micelle in vivo biocompatibility testing All animal studies were performed according to guidelines set forth by the Institutional Animal Care and Use Committee. Male Brown Norway rats (240 275 g; Charles River

PAGE 108

94 Laboratories, Wilmington, MA) were housed in group cages and maintained on a 12 hour light/12 hour dark cycle. Animals had free, continuous access to food and wa ter throughout the experiment. Rats received intravitreal injections of 2 L of PEG PHS PEG micelles (containing 0.2 mg micelles) suspended in PBS in the left eyes and 2 L of PBS (pH 7.4) alone in the right eyes while under isoflurane anesthesia. At 3 and 14 days (n=3 rats per time point) rats were euthanized and both eyes were enucleated and immediately placed buffered formalin for 4 hours and 70% ethanol overnight. Eyes were then embedded in paraff in, sectioned and stained with hematoxylin and eosin for histological analysis. Thermo mechanical analysis Rheological studies were performed on a TA ARES rheometer (TA Instruments, New Castle, DE) equipped with a peltier plate. A 5 wt% solution of PSHU NI PAAm dissolved in PBS (pH 7.4) with or without micelles (10 wt%) was placed between the plates and a small bead of silicon oil was applied around the outer edge to prevent evaporation. Low temperature strain sweeps with and without silicon oil were used to ensure measurements were not affected by its presence. All experiments were performed in the linear deformation region, as defined by preliminary strain sweep experimen ts. Measurements were made at 50 % strain and a frequency of 50 rad/s while the temperat ure was swept from 20C to 40C at a rate of 0.5C/min. In vitro release testing Three different release conditions were compared within this study: micelles alone, RTG alone and the combined RTG micelle system. For micelles alone, micelles suspended in

PAGE 109

95 PB S were placed in a dialysis chamber and release samples were taken from outside the chamber at intervals. The release medium (PBS, 1X, pH 7.4) was replaced with fresh medium each time. For the RTG alone, 50 L of RTG solution with TA (10 wt%) was dropped i nto a 24 well plate and incubated at 37 C for one minute to form a gel. Pre warmed release medium (PBS, 1X, pH 7.4) was then added to fill the well and samples were taken at intervals as before For the combined system, this same procedure was followed but with 50 L of RTG with suspended micelles. The concentration of TA in the release medium was determined by UV absorbance at 237 nm against a pre constructed standard curve. Results and d iscussion Characterization of the combined RTG micelle system was car ried out in comparison to the RTG and micelle systems alone. Specifically, thermo mechanical behavior of the combined system was evaluated to ensure that the original properties of the RTG were not significantly affected In addition, drug release properti es of the combined system were evaluated to quantitate the impact on release properties of the micelles after encapsulation in the RTG. In addition, in vivo biocompatibility data of the micelles is presented to compliment the similar data previously report ed for the RTG PEG PHS PEG micelles in vivo biocompatibility Corneal and retinal histological sections were prepared to evaluate the biocompatibility of the micelle system after intravitreal injection. One eye of each animal received an injection of micel les while the contralateral eye served as a control and received an equal volume injection of PBS. Both eyes of all animals were free of adverse reactions during the 14 day

PAGE 110

96 experimental time course. No animals displayed excessive blinking, inflammation hy peremia or lens or corneal opacity, indicating absence of a uveitic response. Histological analysis confirmed these observation s with sections clear of inflammatory markers including foreign body giant cells and mast cells. Representative corneal (top row) and retinal (bottom row) sections are shown in Figure 4.1 Significant macrophage infiltration was also absent from all sections and vitreous and aqueous humo rs were clear of infiltrating cells or other markers of adverse reactions. These findings demonst rated the biocompatibility of this micelle chem istry and fabrication technique and, coupled with similarly positive in vivo findings with the RTG, suggest that the combined system is highly biocompatible with ocular tissues. Figure 4.1: In vivo biocompatibility testing of PEG PHS PEG micelles was carried out by intravitreal injection in rats. Contralateral eyes served as internal controls and received intravitreal injections of PBS Representative corneal and retinal histological sections from control eyes and rats sacrificed at 3 and 14 days are shown. Relevant tissues include: (A) corneal epithelium; (B) corneal stroma; (C) corneal endothelium; (D) aqueous humor; (E) choroid; (F) retinal pigmented epithelium; (G) neural retina; (H) vitre ous humor; and (I) lens.

PAGE 111

97 Thermo mechanical analysis Rheology studies were used to assess the impact of micelle encapsulation on gelation kinetics of the RTG. Specifically of interest were: a) whether the presence of micelles altered the rate of gelation (i.e. accelerate gelation by providing nucleation sit e s or hinder gelation by acting as physical barriers to PNIPAAm chain collapse); and b) whether the presence of micelles altered the mechanical properties of the formed gel. Rheological analysis of the RTG alone and with suspended micelles is shown in Figu re 4.2. These representative data indicate that micelle incorporation did not alter the sol gel transition temperature or rate of gel formation. However, there was a small increase in the elastic modulus at low and high temperatures as a result of micelle suspension in the system. Figure 4.2: Gel formation kinetics of the RTG alone and with suspended micelles were similar as assessed by rheological analysis. While the sol gel transition temperature was the same between samples, there was a small difference in the elastic moduli in the liquid and gel states.

PAGE 112

98 Overall, micelle encapsulation did not significantly affect the rheological properties of the system. This finding can be attributed to the a) relatively small size of micelles, which is likely too small to disrupt gel formation o f the surrounding RTG system; and b) the relatively low concentration of micelles in the overall system. In future studies, the maximum concentration of micelles that can be encapsulated without a significant impact on gelation kinetics will be investigate d. In vitro TA release kinetics As previously reported by our group, release kinetics of TA from the PEG PHS PEG micelle system are nearly zero order and sustained for several months. This is accomplished through use of a high molecular weight hydrophobic polyurethane middle block in the micelle forming copolymer which forms a dense, hydrophobic micelle core. The hydrophobic TA molecule then slowly partitions out of this core, proving long term release with minimal burst. Because these micelles are pre for med with TA encapsulated in the core before deployment within the RTG, we did not expect any significant alteration of the release kinetics after encapsulation. Results of in vitro release testing of TA from the RTG alone, micelles alone and the combined R TG micelle system are shown in Figure 4. 3 TA release from the combined RTG micelle was significantly slower than the RTG alone (left panel) indicating that the beneficial effect of micelle encapsulation occurred as expected. Within 60 days, the RTG alone released 84.7% of its total TA load, indicating it had neared the end of its therapeutic time frame. However, within the same period, the combined RTG micelle system had released only 12.8% of its drug load, indicating significant time remaining in its th erapeutic utility. As a result, this system would stand to significantly increase the therapeutic time

PAGE 113

99 frame of a single administration over the RTG alone and thereby reduce the frequency of administrations to sustain therapeutic efficacy. In order to u nderstand the impact of encapsulation within the RTG on micelle release properties, release kinetics of TA from micelles alone and those encapsulated within the RTG were al so compared directly (Figure 4.3 right panel). Release of TA from the micelles alon e was characterized by two distinct phases. The first phase a mild burst of drug release occurred over the first 4 days and accounted for 5.3% of drug release. After 4 days, the second phase of release behavior was established and was characterized by zero order release kinetics with a rate of 0.391%/day (or 6.45 g TA/day per 16.5 mg of micelles ). Similarly, the RTG micelle system exhibited a two phase release behavior. The first phase was characterized by an increasing release rate over the first 6 days. The second phase which was established by day 8, was zero order with a TA release rate of 0.237%/day, which was 40% slower than that of the micelles alone. This slower release rate is possibly related to the RTG materials surrounding the micelles acting i n the form of a second Figure 4.3: The combined RTG micelle system significantly improved release kinetics over both the RTG alone (left panel) and micelles alone (right panel) with a short initial lag in the release rate foll owed by achievement and maintenance of zero order kinetics.

PAGE 114

100 compartment through which the TA must partition before being released into the surrounding medium. Because the overall RTG system will possess some potential to also to the surrounding medium, the release rate from the combined system might be expected to be slightly slower. In addition, this slower release might be caused by physical confinement of the micelles within the RTG matrix. In a previous work, our group demo nstrated that micelles swelled significantly during incubation. This effect would be expected to moderately increase the drug release rate as water enters the micelle core allowing TA to more readily partition out. However, micelles encapsulated within the RTG may have less potential to swell as they are physically confined within a dense polymer matrix, thereby reducing th eir observed drug release rate. Additionally, the initial phase of drug release from the RTG micelle system, which occurred over the fir st 6 days, was nearly the inverse of the micelles alone. T he micelles exhibited a typical burst release characterized by a high ini tial release rate that decreased with time to finally reach a steady state, the RTG micelle system exh ibited no initial drug release. The release rate steadily increased over the first several days before finally achieving a steady state. This effect is likely caused by the ability of the RTG matrix to ng of the incubation period, all of the TA is encapsulated within the micelle and the RTG is free of drug. During the first several days, as TA is released from the micelles, the RTG matrix first absorbs the majority of this drug load until it reaches a sa turation level, at which point further drug released from the micelles can displace drug within the RTG matrix, which

PAGE 115

101 can then partition into the surrounding medium. Because this process would be expected to happen gradually, the first several days of rele ase from the RTG micelle system are characterized by an increasing drug release rate. This effect results in a dampening of the burst release characteristic of nano carrier systems and may also be useful in other applications where this initial burst is de trimental to the therapeutic course. Conclusions This novel RTG micelle system combines the advantageous features of an RTG (e.g. injectability, retention at the injection site, harmonization with host anatomy) with those of drug loaded micelle s (e.g. solu bilization of poorly soluble drugs, sustained drug delivery, high drug load capacity) resulting in a system with great translational potential. This system lends itself especially well to long term posterior drug delivery applications, where minimizing rep eated injections and improving drug delivery kinetics is of great clinical concern. While both the RTG and the micelles have been shown to be highly biocompatible in in vivo studies, future work will establish in vivo drug release properties of this novel combined system and explore encapsulation and delivery of other relevant drug molecules.

PAGE 116

102 CHAPTER V SYNTHESIS AND CHARACTERIZATION OF A BIODEGRADABLE POLY( N ISOPROPYLACRYLAMIDE) BASED COPOLYMER A critical stumbling block in the application of poly( N isopr opylacrylamide) (PNIPAAm) based reverse thermal gels in biomedical applications is the difficulty of PNIPAAm to be cleared by physiological degradation and clearance mechanisms In this work, we describe the synthesis and characterization of a novel graft copolymer based system designed to permit complete biodegradation using a PNIPAAm homopolymer. By modulating the lower critical solution temperature (LCST) of the PNIPAAm molecule before and after grafting to a poly(urea urethane) backbone, the system prod uces only degradation products that a re soluble at body temperature permitting efficient clearance and complete biodegradation. Introduction Biodegradable polymers play a critical role in myriad biomedical applications from tissue engineering to drug deliv ery to devices. In many of these applications, polymer degradation on a particular timescale is critical to the therapeutic pathway of the treatment. For this reason, several specific biodegradable polymer chemistries (especially polyesters such as poly(la ctic co glycolic acid), PLGA and poly( caprolactone), PCL) have received more attention that any others in the biomaterials realm [197 199] Their ability to degrade under in vivo conditions in a biocompatible manner and to produce degradation products that can be cleared by physiological mechanisms is of great advantage.

PAGE 117

103 On the o ther hand, utilization of otherwise advantageous polymers with less favorable degradation kinetics is often hindered by their inability to meet this critical criterion. One such exampl e is poly( N isopropylacrylamide) (PNIPAAm), which has been investigated as a component of many biomedical polymer systems [117,180,193] The main advantage of PNIPAAm is its ability to i mpart thermo sensitive gelling properties to otherwise non stimuli sensitive polymers or molecules. This thermo sensitivity is evidenced by the lower critical solution temperature (LCST) of PNIPAAm near 32C, which coincides with a transition of PNIPAAm mo lecules from an expanded coil state to a collapsed globule state [112] In PNIPAAm, th is transition is particularly sharp indeed almost discontinuous and results in rapid and complete gelation upon heating. The rate of gelation, along with the LCST lying just below body temperature, make PNIPAAm based systems extremely favorable for applica tion as injectable biomaterials. While it has been shown in vitro and in vivo to be highly biocompatible, the pure PNIPAAm homopolymer is non biodegradable. Generally, the non biodegradability of PNIPAAm is attributed to its LCST behavior, which dictates that PNIPAAm molecules are insoluble at physiological temperature s. As a result, PNIPAAm molecules cannot be readily cleared from the injection site and the injected matrix will remain, despite perceived molecular weight loss. Efforts to produce biodegrad able PNIPAAm based polymer systems have employed various strategies. Many groups conjugated PNIPAAm homopolymers to biodegradable segments in an effort to degrade the formed thermal gel [181,200] While these systems may exhibit molecular weight loss degradation of the cleavable segments would leave PNIPAAm homopolym ers behind, which, as discussed before, remain insoluble and would

PAGE 118

104 likely not be cleared from the site of injection. In order to permit complete biodegradation and clearance of the system, many groups have designed specific degradation pathways that result in a significant change in LCST. Specifically, this is generally achieved by cleavage of hydrophobic side chains of monomers co polymerized with NIPAAm [201,202] The degraded polymer being less hydrophobic will then exhibit a higher LCST and if this LCST is above body temperature will be solubiliz ed and c leared from the system. However, that such system rely on side chain hydrophobicity to control the LCST results in a relatively inflexible system. For example, conjugation of peptides or targeting antibodies to the polymer backbone as is highly desirable f or tissue engineering application would not be possible with such chemistries. In addition, the hydrophobic side chains were reported to significantly increase viscosity of those systems, which may limit their ability to be administered by small gauge need le injection To overcome these limitations, we developed a novel biodegradation scheme for a PNIPAAm based copolymer. The polymer is comprised of PNIPAAm homopolymers grafted to a hydrophobic poly(urea urethane) backbone. The polymer backbone poly(serinol hexamethylene urea) (PSHU), as has been reported previously [187] contains a high density of primary amine functionalities, only a fraction of which are used for PNIPAAm conjugation. As a result, the copolymer retains the ability for further functionalization. To permit complete biodegradation and physiological clearance, a nove l PNIPAAm homopolymer was developed with two hydrophilic end groups: a hydroxyl and a carboxylic acid resulting in a homopolymer with LCST above 37C Upon conjugation to the hydrophobic backbone polymer through the carboxylic acid terminals, the LCST of the system is shifted below 37C to permit gelation at body temperature. However, upon

PAGE 119

105 hydrolytic or enzymatic cleavage of the PNIPAAm homopolymer, the LCST of the PNIPAAm returns to that of the homopolymer, permitting solubilizati on at physiological condi tions. This degradable copolymer chemistry is compared to a previously reported RTG system which used a high molecular weight, low LCST PNIPAAm to demonstrate maintenance of thermal gelling properties while significantly improving degradation kinetics. Mat erials and m ethods Materials azobis(4 cyanovaleric acid) (A CA), methanol phenolphthalein 3 mercaptopropionic acid (MPA), cholesterol esterase (from porcine pancreas) and hyd rochloric acid (HCl) were obtained from Sigma Aldrich (St. Louis, MO). N hydroxysuccinimide (NHS) and N (3 dimethylaminopropyl) ethylcarbodiimide hydrochloride (EDC) were obtained from Alfa Aesar (Ward Hill, MA). H exane, diethyl ether, sodium hydroxide (NaOH) and 2 propanol were obtained from Fisher Scientific (Pittsburgh, PA). Dimethylformamide (DMF) was obtained from BDH Chemicals (Poole, UK). Dichloromethane (DCM) was obtained from J.T. Baker (Phillipsburg, NJ). N isopropylacrylamide (NIPAAm) was obta ined from Acros Organics (Geel, Belgium). 2,2' Azobis[2 methyl N (2 hydroxyethyl)propionamide] (AMHP) was obtained from Wako Chemicals (Richmond, VA). Papain was obtained from Worthington Biochemical Corp. (Lakewood, NJ).

PAGE 120

106 PNIPAAm s ynthesis Two PNIPAAm ch emistries we re used in the study. The first was a high molecular weight PNIPAAm with one controlled end group: a carboxylic acid (PNIPAAm COOH). The second was a lower molecular weight PNIPAAm with both end groups controlled: one a hydroxyl and the other a car boxylic acid (HO PNIPAAm COOH). PNIPAAm COOH was synthesized based on a procedure described previously [180] NIPAAm (5.0 g) and ACA (0.060 g) were dissolved in 25 mL of dry methanol and bubbled with nitrogen for 30 minutes at room temperature. The reaction was ra ised to 68C under moderate stirring and carried out for 3 hours. The product was precipitated twice in 60C water and dialyzed against 1 L dH 2 O for 48 hours in 3,500 Da molecular weight cut off (MWCO) dialysis tubing (Spectrum Labs, Inc., Rancho Dominguez CA). HO PNIPAAm COOH was synthesized by free radical polymerization using the hydroxyl terminated azo initiator AMHP and the carboxyl terminated chain transfer agent MPA. NIPAAm (4.42 mmol), AMHP (34.7 mo l) and MPA were dissolved in 10 mL of 2 propanol and the solution was bubbled with dry nitrogen for 30 minutes. The reaction was then heated to 85C and allowed to proceed under a nitrogen atmosphere for 48 hours. The reaction was terminated by cooling to room temperature followed by rotary evaporation t o remove the solvent. The product was dissolved in acetone and precipitated twice in hexane followed by terminal drying by extended rotary evaporation. Typical yields were greater than 90%.

PAGE 121

107 Molecular weight determination by end group titration The molecu lar weight of PNIPAAm COOH and HO PNIPAAm COOH were determined by titrating for the carboxylic acid end groups. Approximately 0.050 g of PNIPAAm was dissolved in 10 mL dH 2 O with 10 L of phenolphthalein solution (2 wt% in absolute ethanol). The solution wa s titrated to the end point by adding 0.01 N NaOH. LCST determination LCST measurements were made on a Cary 100 UV visible spectrophotomer (Agilent Technologies, Inc., Santa Clara, CA) equipped with a temperature controlled 6 cell stage. Polymers were dis solved in dH 2 O at 1 wt% and transmittance readings were taken at 500 nm while the temperature was increased at 0.5 C/min. PSHU s ynthesis The backbone poly(urethane urea) was synthesized by copolymerization of N Boc serinol, urea and HDI. N Boc serinol was synthesized as described previously [187] N Boc serinol (6.0 mmol) and (6 .0 mmol) urea were lyophilized for 24 hours and then dissolved in 6 mL of dry DMF. Under a nitrogen atmosphere, the reaction was brought to 90C under moderate stirring and HDI (12 mmol) was added drop wise through the septum. The reaction was carried out for 7 days, after which it was cooled and precipitated twice in diethyl ether and once in water. After lyophilization, a dry yellowish product was obtained in 86.5% yield. In order to expose primary amines for conjugation, Boc protective groups were remove d through a complete deprotection routine. PSHU (0.10 g) was dissolved in a 50:50 (v/v) mixture of DCM and TFA (18 mL total volume) and reacted uncapped with rapid stirring

PAGE 122

108 for 15 minutes. The solvents were immediately removed by rotary evaporation at 45C and 10 mbar vacuum and the product re dissolved in DMF. Two precipitations in diethyl ether and terminal rotary evaporation yielded a dry yellowish product, de protected PSHU (dPSHU). Conjugation of PNIPAAm to d PSHU The RTG copolymer was synthesized by gr afting PNIPAAm COOH or HO PNIPAAm COOH to dPSHU through carbodiimide linking chemistry. Of the calculated 18 repeating units (and, thereby, primary amines) in dPSHU, 25% were used for conjugation. PNIPAAm, NHS (3M excess) and EDC (3M excess) were dissolved in 3 mL of DMF and reacted for 24 hours. dPSHU dissolved in 2 mL of DMF was then added drop wise and the conjugation reaction proceeded for 24 hours. The product was added directly to a 12 14 kDa MWCO dialysis tube and dialyzed for 48 hours for purificati on The product was lyophilized to yield a white flaky product, PSHU NIPAAm (if conjugating PNIPAAm COOH) or PSHU NIPAAm OH (if conjugating HO PNIPAAm COOH). Degradation s tudies PSHU NIPAAm degradation was assessed during exposure to hyd rolytic and enzymat ic stresses and both gravimetrically and by spectroscopy The polymer was dissolved in PBS to make a 10 wt% solution and the gel was formed in the bottom of a vial. PBS (1X, pH 7.4) or cholesterol esterase solution (0.2 units/mL) was added on top of the ge l and the samples were incubated at 37C with gentle shaking. Every week both the PBS and esterase solutions were poured off and fresh pre warmed solution was added. At 1, 3 and 6 months certain samples were lyophilized and weighed to calculate mass loss.

PAGE 123

109 Accelerated degradation testing was carried out in a similar manner. PSHU NIPAAm and PSHU NIPAAm OH gels were formed at 10 wt% and incubated at 37C in either 1N HCl or papain enzyme solution (23.3 units/mL, 1 mL per sample). Mass loss was calculated gravi metrically at 1 and 3 days as above. Samples were additionally analyzed by Fourier Transform Infrared (FT IR) spectroscopy. Lyophilized samples were dissolved in isopropyl alcohol and dropped onto polyethylene windowed FT IR cards. After drying, the spectr a were collected on a Nicolet 6700 (Thermo Fisher Scientific, Waltham, MA). Results and d iscussion Figure 5.1: Proposed degradation routes for the PSHU NIPAAm and PSHU NIPAAm OH copolymers described herein.

PAGE 124

110 The d PSHU NIPAAm copolymer was designed to permit in vivo biodegradation, primarily by enzymatic pathways. The origi nal copolymer, shown in Figure 5.1 (S cheme A), utilizes a high MW PNIPAAm COOH (M n : 28,182 Da), which is attached to PSHU via amide linkages. This copolymer, d PSHU NIPAAm, was expected to degrade enzymatically through these amide linkages between PNIPAAm and PSHU and also through the hydrolyt ically and enzymatically labile [179] urethane bonds within the PSHU backbone. The hydrolytic and enzymatic degradation rates of d PSHU NIPAAm were experimentally evaluated respectively by long term incubation in: a) a neutral PBS (pH 7.4) b uffer ; and b) a cholesterol esterase solution. These data are shown in Figure 5.2 By the 6 month time point, d PSHU NIPAAm had lost 20.8 2.65% of its starting mass by incubation in PBS and 17.7 0.092% by incubation in cholesterol esterase solution, wit h no statistically significant difference between the groups (p=0.21). Degradation rates were relatively linear for both samples, with the PBS group losing an average mass of 3.68% per month and the cholesterol esterase group losing an average of 2.93% per month. Assuming Figure 5.2 : dPSHU NIPAAm samples incubated in PBS (pH 7.4) and CE solutions showed min imal degradation at 1, 3 and 6 month time points. Means and standard deviations of n=3 samples are plotted.

PAGE 125

111 linear kinetics, these data forecast an approximately 30 month window for complete degradation of the system FT IR analysis confirmed the minimal amount of degradation observed in the mass loss study. Representative spectra from dPSHU NIPAAm samples incubated in cholesterol esterase solution are shown in Figure 5.3 as collected at 1, 3 and 6 months. No major differences were noted between the spectra. Samples of dPSHU NIPAAm incubated in PBS solution showed the same peaks and intensitie s and are omitted here for clarity. In addition, indications of amide hydrolysis (e.g. generation of carboxylic acids (926 cm 1 ) or decreases in amide I (1649 cm 1 ) or II (1539 cm 1 ) bonding) were not observed, indicating PNIPAAm was not cleaved from PSHU during the incubation period. Quantification of the carboxylic acid and amide II peaks referenced to the PNIPAAm isopropyl peak at 1368 cm 1 is shown in Figure 5.4 There was n o significant change from Figure 5.3: FT IR analysis confirmed the lack of degradation in dPSHU NIPAAm samples. Representative spectra from samples incubated in cholesterol esterase solution are shown. Samples incubated in PBS were nearly identical and are omitted for clarity.

PAGE 126

112 baseline values Degradation products would be noted b y an increase in the carboxyl/isopropyl ratio or a decrease in the amide II/isopropyl ratio, neither of which were present. In order to understand these results, an appreciation for the LCST behavior of PNIPAAM and its physicochemical implications is req uired. Fundamentally, the LCST defines the temperature below which PNIPAAm exists in an expanded coil configuration and above which it exists as a collapsed globule. In the expanded coil state, PNIPAAm molecules are heavily solvated and exist in solution ( i.e. they are water soluble below the LCST). However, formation of the collapsed globule configuration results from the development of hydrophobic interactions and results in de solvation of the molecules and their precipitation from solution (i.e. they ar e water insoluble above the LCST). Figure 5.4: Quantification of FT IR spectra after incubation of PSHU NIPAAm gels in cholesterol esterase solution confirmed that neither were carboxyl ic acids generated nor was amide bonding significantly decreased. These results indicate a lack of degradation in dPSHU NIPAAm samples.

PAGE 127

113 This mechanistic understanding of PNIPAAm LCST behavior can be used to explain the minimal degradation kinetics observed for d PSHU NIPAAm gels at 37C in vitro In the presence of enzyme activity, the amide linkages betwe en d PSHU and PNIPAAm COOH would be expected to degrade, resulting in the reproduction of a primary amine functionality on PSHU and a carboxylic acid terminal on PNIPAAm. This PNIPAAm COOH molecule, now cleaved from the PSHU backbone, would be expected to r egain its homopolymer LCST behavior, which is shown in Figure 5.1, Scheme A S ince the LCST (32.5 C) of PNIPAAm COOH is below the incubation temperature (37C), these cleaved molecules remain water insoluble and cannot readily diffuse out of the d PSHU NIPA Am matrix for clearance. This results in the slow degradation kinetics observed for this system. In order to circumvent this problem, a PNIPAAm chemistry was developed that would be water soluble at body temperature after cleavage from the PSHU backbone (t o permit clearance from the matrix) but insoluble at body tem perature when attached to PSHU (to permit thermal gelation). In order to achieve this dynamic, the LCST of the homopolymer must be greater than 37C, but tethering to PSHU must lower the LCST be low 37C. LCST modulation of PNIPAAm has been reported previously [203,204] and is dependent on : a) the hydrophobicity of the PNIPAAm end group ; and b) the molecular weight of the polymer More hydrophobic end groups and higher molecular weights decrease the LCST. The magnitude of the LCST increase is dependent on the m olecular weight of the polymer as is typic al of end group driven effe cts with lower molecular weights displaying a larger increase in the LCST.

PAGE 128

114 In order to achieve a large increase in LCST, a new PNIPAAm chemistry was designed with hydrophilic end groups on both ends of the polymer. The first, a hydroxyl end group, is achie ved by use of hydroxyl terminated thermal initiator (AMHP) and the second, a carboxylic acid, is achieved by use of a carboxyl terminated chain transfer agent (MPA). Use of a chain transfer agent also permitted greater control over the molecular weight of PNIPAAm. The structure of this PNIPAAm molecule, HO PN IPAAm COOH, is shown in Figure 5. 5 along with resulting molecular weights from various polymerizations compared to theoretical values. As expected, molecular weight changes were achieved by merely chang ing the molar fraction of MPA in the reaction. LCST values of HO PNIPAAm COOH were dependent on molecular weight: 2.0 kDa transitioned at 4 9.4 C, 3.5 kDa transitioned at 4 3.0 C 5.8 kDa transitioned 40. 4 C, 13.8 kDa transitioned at 39.1C and 2 2 .0 kDa t ransitioned at 38.4C; as co mpared to PNIPAAm COOH at 28.2 k Da, which transitioned at 32.5C. Further, as shown in Figure 5.6, conjugation of HO PNIPAAm COOH to d PSHU yielded a copolymer with an LCST Figure 5.5 : Synthesized HO PNIPAAm COOH (structure shown in inset) molecular weight followed closely to theoretically c alculated values and was dependent on the molar fraction of MPA (left panel). The LCST of HO PNIPAAm COOH decreased with increasing molecular weight, but even at high molecular weight (greater than 20 kDa) the LCST did not fall below physiological temperat ure (dashed line).

PAGE 129

115 of 26.7 C, allowing the copolymer to gel below body temp erature. This result was expected based on two effects. The first is that conjugation of HO PNIPAAm COOH to d PSHU occurs through the carboxylic acid terminal, resulting in a loss of one of two hydrophilic end groups. The second is that d PSHU provides a hyd rophobic backbone, thereby acting as a high molecular weight hydrophobic end group, which is also known to reduce the LCST. In order to quantify the degradation kinetics of the new copolymer chemistry as compared to the previous one, accelerated in vitro degradation experiments were carried out in HCl and papain solutions. These conditions were meant to isolate hydrolysis mediated and enzyme mediated degradation pathways. PSHU NIPAAm gels in either solution showed no statistically significant decrease in mass between days 0 and 1 (p=0.89 for HCl and Figure 5.6: Conjugation of HO PNIPAAm COOH (MW: 13.8 kDa) with an LCST of 39.1 C to PSHU yielded a copolymer with an LCST of 26.7 C. While the copolymer is insoluble at body temperature permitting gelation, the cleaved HO PNIPAAm COOH m olecule would be soluble at body temperature permitting clearance from the system.

PAGE 130

116 p=0.21 for papain) or days 1 and 3 (p=0.70 for HCl and p=0.38 for papain). This result confirms that PSHU NIPAAm gels are unable to degrade or clear mass during degradation, even in exaggerated conditions. In contrast, PSHU NIPAAm OH gels degraded significantly over the first day of incubation in both acid and enzyme solutions (p < 0.0001 for both ) and those incubated in enzyme solution continued to degrade between days 1 and 3 (p=0.02). This result confirms t he expected solubilization of cleaved PNIPAAm from PSHU NIPAAm OH gels and provides direct evidence for the ability of these gels to clear degraded mass from the system. Figure 5.7: Accelerated degradation testing of PSHU NIPAAm and PSHU NIPAAm OH gels in HCl and papain solutions demonstrated the improved degradation and clearance dynamics of the PSHU NIPAAm OH chemistry. PSHU NIPAAm gels showed no statistically significant difference between any time points for either of the exposure conditions (p > 0.2) while all PSHU NIPAAm OH gels showed mass loss between days 0 and 1 (p < 0.0001) and th e enzyme incubated sample showed continued mass loss between days 1 and 3 (p < 0.05).

PAGE 131

117 This result allows for a fully biodegradable PNIPAAm based copolymer, with the expect ed degradation behavior shown in Figure 5.1 Scheme B. Because the copolymer possesses a n LCST below body temperature, it can form a physical gel in vivo With enzymatic cleavage of the amide bond between PSHU and PNIPAAm, the original HO PNIPAAm COOH homo polymer is recovered, which is now a soluble degradation product because of its LCST occurring above body temperature Because this product is soluble, it can diffuse out of the RTG matrix and be cleared by the body [205] As before, the PSHU backbone is expected to degrade via hydrolytic and enzymatic mechanisms. A potential pitfall of this approach is that it relies on enzymatic degradation pathways to cleave PNIPAAm from the PSHU backb one. However, many tissues such as the vitreous humor of the eye [206] have relatively low enzymatic activity levels, which would hinder biodegradation. Further, amides are not hydrolytically cleavable at neutral pH. A potential solution is to convert the d PSHU to PNIPAAm link age to a hydrolytically labile bond such an ester, as shown in Figure 5.1 Scheme C. This would be accomplished through use of a glycolic acid linker, which would be conjugated to those primary amines to which PNIPAAm will then be attached. Esterification between carboxyl terminated PNIPAAm and hydroxyl terminated glycolic acid will result in an ester linkage, which will permit hydrolytic degradation. Two problems would have to be overcome to permit this chemistry. First is the hydroxyl terminal of HO PNIPA Am COOH, which would compete with esterification through glycolic acid. This could be overcome by employing a different hydrophilic azo initiator that results in a hydrophilic end group other than a hydroxyl. The second expected problem would be t hat becau se only 25% of amines are typically used for PNIPAAm attachment

PAGE 132

118 and the remaining amines should be left available for future conjugation esterification between the hydroxyl terminal of PSHU glycolic acid and the carboxylic acid of PNIPAAm COOH will suffer from competition from the N acylation reaction between the primary amines of PSHU and the carboxylic acid of PNIPAAm COOH. This reaction would be favored in a typical esterification because of the higher nucleophilic nature of amines over hydroxyls. To ove rcome this barrier, an O selective acylation procedure can be used, as described by Ohshima, et al. [207] using a tetranuclear zinc catalyst. This catalyzed esterification would be expected to heavily favor O acylation resulting in esterification through the hydroxyls of the glycolic aci d linkers instead of the remaining amines. Conclusions In this work, a PNIPAAm based reverse thermal gelling copolymer was modified to permit complete biodegradation under physiological conditions. The newly developed PNIPAAm chemistry retains an LCST abov e body temperature even at high molecular weights Conjugation to a previously characterized poly(urea urathane) reduces the LCST below body temperature to permit in situ gelling after injection. When cleaved from the polymer backbone, this novel PNIPAAm is solubilized and can be efficiently cleared from the thermally gelled matrix. This completely biodegradable PNIPAAm system is of great interest to many applications, where an injectable in situ gelling polymer can be used to deliver drugs or provide a sc affold for cell delivery or tissue regeneration. Because the biodegradation and clearance mechanisms do not rely on a specific hydrophobic balance within the polymer chemistry, the system also retains a high amenability to functionalization, further enhanc ing its potential in these fields.

PAGE 133

119 CHAPTER VI DISCUSSION, LIMITATIONS AND FUTURE DIRECTIONS Discussion Delivery of ophthalmic therapeutics is hindered by challenges associated with ocular anatomy and physiology, drug physicochemical properties and patient complexities. These challenges have limited the advancement of treatments for ophthalmic diseases, but also spurred development of novel strategies to overcome them. By and large, nano carrier systems are the most heavily investigated owing to their ubiqu ity, ease of fabrication and a bility to be administered by injection. However, such systems are best suited for short term therapeutic courses as their residence time in various spaces tends to be on the order of hours to days [208,209] or up to a few weeks in the specific case of intravitreal injected particles [210] A key factor in this short residen ce time is the nano scale size of such particles, which allows them to be internalized by many cells through endocytic mechanisms. Internalization greatly improves cell targeting (thereby reducing off target drug concentrations, which can reduce side effec ts) and can more directly target an intracellular site of action [211] However, appropriate concern must also be given to the fate of internalized nano carriers. While a multitude of factors (e.g. particle size, surface charge, composition) play a critical role in determining the endocytic mechanism by which particles are internalized thereby al tering the endocytic pathway the fate of the se particles is the same [92] Generally, cellular uptake of nanoparticles is realized by enclosing them into endosomes. Initially early endosomes, they eventually mature into late endosomes and finally to lysosomes. Critically, entrapment in lysosomes is accompanied by a significant drop in

PAGE 134

120 pH [212] This acidification can catalyze hydrolytic pathway s used by many nano carrier chemistries (e.g. PLGA, PLA, PCL) for biodegradation and ultimately lead to acceleration of particle degradation and drug release. For this reason, nano carrier systems are generally best suited for highly targeted, short term t herapies. In addition, nano carriers injected into intra or peri ocular spaces are hypothesized to suffer short residence times [210] due to local fluid flows used physiologically as deliberate c learance mechanisms. When drugs or particles are injected into the vitreous humor through the sclera, they can be cleared by two mechanisms. The first mechanism is derived from the injection itself. When the sclera and vitreous body are punctured by the in jecting needle, outflow of fluid results in backward perfusion of water through the anterior vitreous body and out the hole created by the injection [213] In most cases, this outflow would be sufficient to cause reflux of some fraction of particles initially injected into the vitreous although this effect has not been quantified in the literature The second mechanism involves passage of particles from the vitreous humor into the aqueous chamber and subsequent clearance out of the eye. Normally, convective flows in the eye occur in the posterior direction, from the anterior chamber posterior through the vitreous humor [214] However, this physiological process can be disrupted by several changes to the eye, many of which are concomitant with disease states associat ed with requiring posterior therapeutic intervention These changes include patients with age related liquefaction of the vitreous humor (synchisis senilis) which reduces the vitreous viscosity and permits high levels of convective flow anterior and poste rior [215] and those having received a vitrectomy, which causes a significant decrease in vitreous viscosity due to its replacement with a synthetic alternative [216] These changes, which are not

PAGE 135

121 uncommon in elderly patients [217] can great ly increase the rate of nano carrier efflux after intravitreal injection. Similarly, injections to periocular spaces are also susceptible to fluid flows that will limit nano carrier residence times. For example, the subconjunctival space (the most common t arget for periocular injections) is exposed to blood flows from the conjunctiva and choroid and lymphatic flows through the conjunctiva and episclera [218] These flows have been shown to rapidly clear both small molecule drugs and macromolecules injected into this space [219] and would likely apply to nano carriers as well. Nevertheless, nano carrier drug delivery systems (especially micelles) are still attractive for their ability to : a) s olubilize and entrap poorly soluble drugs; b) contain relatively large amounts of drug in a small volume; and c) sustain drug release over long periods of time. As these are many of the same properties that RTGs lack and indeed those that limit their clini cal utility the combination of the two systems presents an opportunity for significant synergy. As the RTG will benefit from th ese advantageous properties of encapsulated micelles, so the micelles will benefit from prolonged retention at the site of inject ion vis vis protection from the physiological flows descri bed earlier. Besides independent characterization of the two systems for their distinct properties, two major questions arise related to their combination. First, how does suspension of micelles w ithin the RTG solution affect its gelation kinetics and the properties of the resulting physical gel? Second, how does encapsulation of micelles within the activated RTG matrix influence the kinetics of drug be ing released from the micelles? These major q uestions were addressed in Chapter IV through thermo mechanical characterization and in vitro drug release kinetics studies.

PAGE 136

122 The concern surrounding the impact on thermal gelling kinetics is derived from the presence of micelles which can be thought of as micro scale disruptions within the RTG system during gelation. Because the gelation process is associated with significa nt polymer rearrangement and compaction [220] this process may be hindered by the presence of micelles within the system. Indeed, a greater concern may be t he physicochemical characteristics of the micelles. Micelles are characterized by a core shell structure in which the outer shell is generally comprised of highly hydrophilic PEG molecules These PEG molecules critically: a) stabilize the hydrophobic core structure of the micelle by providing an interfacial layer with the surrounding water molecules; and b) prevent micelle aggregation [221] However, they may also serve as a disruption in the gelation mechanism of the RTG since this process is driven by the development of hydrophobic interactions and subsequent collapse of the network. However, thermo mechanical cha racterization (see Chapter IV) clearly demonstrated that even at a relatively high micelle concentrati on of 10 wt%, there was no significant alteration of thermal gelling kinetics. Despite the reasons given above for the potential avenues for disruption of gelation, these data can be explained by the relatively small size and number of micelles within the system. The average diameter of micelles produced by the extrusion process was 217.5 nm, which is on the scale of the average pore size observed in the RT G. Therefore, even during RTG collapse, it is hypothesized that the micelles are not on a sufficient size scale to cause significant disruption in the formation of the RTG matrix. Further a micelle concentration of 10 wt% indicates that the vast majority o f the system is composed of PSHU NIPAAm and so matrix collapse can continue to occur around the relatively small number of micelles present. An edge of failure analysis would

PAGE 137

123 be needed to determine at what concentration micelles begin to have a significant impact on gelation kinetics In vitro drug release kinetic testing confirmed that micelles encapsulated within the RTG had, if anything, improved drug release behavior over the RTG or micelles on their own. A major drawback of most polymeric drug delivery systems even those delivering poorly soluble drugs such as TA is the presence of an initial burst of drug release [171,176] In many cases, this burst of drug release can be a major source of side effects or other adverse reactions due to the rapid exposure of surrounding tissues to a high concentration of drug. The RTG micelle combined system effectively dampe ned this burst release due to the ability of the RTG to absorb this initial burst of TA as released from the micelles. This allows for a much more favorable drug release profile in which the concentration of drug is slowly built up over the first several days before achieving a steady state release rate. Indeed, this advantageous combination of a nano carrier and a thermo sensitive gel is not c ompletely novel and a limited number of published works have described similar systems. The first, described by Bochot, et al. [196] in 1998 describes encapsulation of liposomes within a poloxamer thermo sensitive gel. In order to stabilize the micelles during poloxamer gelation, 1,2 d istearoyl sn glyce ro 3 phosphatidylethanolamine N (poly(ethyleneglycol) 2000) (2000PEG DSPE) was incorporated into their composition, which protected the liposomal membranes from poloxamer intrusion and their subsequent destabilization. Nevertheless, it was found that lipos omes were sensitive to the thermal gelation procedure and required additional stabilizers to protect them from the surfactant like poloxamer molecules. This study did not include in vitro release testing of any model drugs to verify clinical relevance of t he system.

PAGE 138

124 In 1999, Barichello, et al. [194] described encapsulation of PLGA nanoparticles within a poloxamer (F127) gel for parenteral delivery of peptide and proteins. Using insulin as a model drug, they found that release of the molecule was reduced by nearly half when loaded into the nanoparticles within the gel as opposed to the gel itself. Further, they used in vivo experiments to demonstrate th at the times (t max ) required to achieve the maximum plasma insulin concentration were longest for the nanoparticle gel combined system and the gel itself. While long term release studies were not conducted, it was reported that 35% of insulin loaded in the nanoparticle gel combined system was released within 12 hours. Zhang, et al. [193] reported in 2004 on a PNIPAAm based cross linked hydrogel encapsulating PCL based microspheres loaded with ovalbumin. Drug release from the combined system was impressively sustained (30% at 55 days) and was found to correlate well with the microparticle content of the combined systems. However, due to the cross linked hydrogel nature of this system, the system is not injectable and so its clinical utility is relatively limited. In 2008, G ou, et al. [195] described encapsulation of biodegradable nanoparticles with a thermo sensitive gel. The nanoparticles were fabricated from a PCL PEG PCL tri block copolymer encapsulating the hydrophobic model drug honokiol and the thermo sensitive gel was a poloxamer (F127). Whereas the nanoparticles on their own exhibited a 40% burst release, this effect was diminished to 12% after encapsulation in the poloxamer gel. The nanoparticles did mildly impact thermal gelation kinetics of the poloxamer system, but this was mostly evident at high nanoparticle concentration. Long term drug release studies were not presented, but the combined system released approximately 50% of drug load within 7 days.

PAGE 139

125 Finally, Boddu, et al. [192] in 2010 described doxorubicin loaded PLGA PEG folate micelles encapsulated in a PLGA PEG PLGA RTG. Doxorubicin release from micelles within the RTG was sustained for 300 hours, compared to 100 hours for micelles alone and 150 hours for the RTG alone. Micelles were found by microscopy to be stable after encapsulation in the RTG, but the impact of micelle encapsulation on thermal gelation was not quantified. One avenue that previous works did not explore b ut may be a major advantage of this sort of dual carrier system is the ability to simultaneously deliver two or more drugs. Most simply, this could be accomplished by loading one drug within the micelles and another within the RTG but could also be expande d to include more drugs if micelles loaded with different drugs were mixed within the system. The use of combination therapies in ophthalmology has expanded rapidly in recent years and may represent a new frontier in the treatment of various complex pathol ogies [222 224] Because the drug molecules included in combination therapies can often have vastly different physicochemical properties, mixing them within one system (suc h as in the RTG itself) can have unfavorable results as the drugs often have different amenities to the encapsulating polymer and may release at considerably different rates By tailoring different systems to the drug molecules of interest, drug release ki netics can be more carefully controlled so that each drug is released in such a way to maximize therapeutic efficacy. The RTG micelle system provides this control by allowing one drug to be encapsulated in the RTG and another (or more) in the micelles. The combined systems described previously possessed many attractive properties, but none met all of the requirements of a long term ophthalmic drug delivery solution. That is, no

PAGE 140

126 system combined : a) drug release on the order of months; b) injectability; c) ef ficient encapsulation of poorly soluble drugs; d) biodegradation; and e ) high biocompatibility. The system we have developed and described within this work is differentiated from those previously described by simultaneously meeting all of these criteria. Biodegradation is a key differentiating factor of our system over other RTGs, especially PNIPAAm based ones. While not comprehensively explored herein, our biodegradation strategy is theoretically based on a two step process involving: a) cleavage of PNI PAAm from the PSHU backbone via hydrolytic or enzymatic mechanisms; and b) breakdown of the PSHU backbone via hydrolytic and enzymatic mechanisms. In the current system, the first step relies on cleavage of the amide bond formed through carbodiimide based linking of PNIPAAm COOH to PSHU NH 2 Figure 6.1: Hydrolytic (proton catalyzed) and enzymatic degradation pathways of carboxyl derived amides. Y: nucleophilic component of enzyme, Z: electrophilic component of enzyme, H B: proton donor component of enzyme.

PAGE 141

127 Amide cleavage in biological environments may proceed through hydrolytic or enzymatic pathways. The hydrolytic pathway can be proton or HO catalyzed. Proton mediated amide hydrolysis occurs through protonation of the carbonyl O leading to polarization of the carbonyl group, which facilitates addition of a nucleophile (e.g. H 2 O). This process is shown in Figure 6.1. Because protonation of the carbonyl O begins the process, the hydrolytic degradation pathway is extremel y slow at neutral pH [225] Enzymatic hydrolysis of carboxyl der ived amides has been reported to be orders of magnitude more effective than chemical hydrolysis [225] This acceleration is caused by the presence of three catalytic features at the active sites of hydrolases, as demonstrated in Figure 6.1. These are: a) an electrophilic component that increases the polarization of the carbonyl O atom; b) a nucleophilic component that attacks the carbonyl C atom leading to formation of the tetrahedral intermediate; and c) a proton donor that transforms the amine moiety into a better leaving group [225] Acceleration by certain hydrolases increases the rate of amide hydrolysis by 10 9 to 10 10 fold over purely hydrolytic reactions [226,227] In several of the copolymer chemistries described herein, degradation relies on amide hydrolysis, especia lly in cleaving PNIPAAm from the PSHU backbone. Based on the relative rates of hydrolytic and enzymatic degradation (and the understanding that most ocular environments are neutral pH [228,229] ), it would be expected that enzymatic pathways would be the predominant mechanism of degradation. Indeed, research on poly(ester amide) chemistries have confirmed t hat amide groups are highly resistant to hydrolysis [230] and these systems degrade almost exclusively by ester hydrolysis. Even in the presence of proteolytic enzymes, ester groups were found to be the primary source of

PAGE 142

128 degradation, presu mably due to the rate of ester degradation being orders of magnitude larger than the rate of amide degradation [231] As discus sed previously, enzyme levels in ocular tissues have been reported to be relatively low compared to other tissues [206] This raises a concern in permitting efficient degradation of PSHU NIPAAm gels a fter injection to these spaces, as they would rely on the presence of proteolyt ic en zymes to catalyze this process. Indeed, if previous f indings based on poly(ester amid es) can be translated to this system, the rate of amide degradation in PSHU NIPAAm may still be too low to permit degradation on the time scale of interest for the pr esent application. If this is the case, the strategy laid out in Chapter V to convert the linkage between PSHU NH 2 and PNIPAAm COOH from an amide to an est er may be entirely necessary. Limitations The work presented herein has explored many facets of these two systems an RTG and a polymeric micelle and their advantageous combination. However, several avenues of investigation could n ot be covered within its scope. As such, these limitations should be seen as areas that require further investigation. While st rong characterization work has been presented on the novel PSHU NIPAAm polymer and its variants, several questions remain unanswered about this copolymer. For instance, the conjugation efficiency of PNIPAAm COOH to dPSHU has not been accurately quantified and some findings have indicated this value may be lower than expected. To date, this calculation has only been accomplished through interpolation of

PAGE 143

129 elemental analysis results, which provides a relatively imprecise result. A method to directly assay the c onjugation efficiency on a batch by batch basis would be of great value. In addition, the methods of polymer purification used throughout this work could be improved to yield a higher purity product This is especially true in the case of the dPSHU to PNI PAAm COOH grafting procedure. At the end of this synthesis, it is important to remove unreacted PNIPAAm to yield pure PSHU NIPAAm. In the present work, this was realized through extended dialysis. However, some findings suggested an amount of unreacted PNI PAAm remained in the system even after extended dialysis. This could be a homopolymers [232] To overcome this, PSHU NIPAAm purification would be better realized through chromatography, where separation by molecular weight would provide a much more efficien t and thorough process. A chromatography method could also feasibly separate PSHU NIPAAm molecules with low conjugation ratios (which are often solubility challenged) to improve solution properties of the resulting system. T he combined RTG micelle system h as been shown to effectively sustain delivery of the poorly soluble corticosteroid TA for several months (a nd potentially beyond one year). However, it is suspected that the delivery time frame and release kinetics of this system pair ed with other drug mol ecules will be strongly dependent on the physicochemical properties of the drug itself As was described earlier, both the RTG and the micelles developed in this work release hydrophobic drugs from internal hydrophobic regions of the polymers through a par tition based mechanism. As a result, the release kinetics will depend heavily on the partition coefficient of the drug. A more complete understanding of the system and its applicability as a platform drug delivery technology (i.e. one that can be

PAGE 144

130 applied t o a broad range of drug molecules) would be garnered by evaluating drug encapsulation and release kinetics of several drug molecules with a range of physicochemical properties. The proposed system may also struggle to encapsulate and sustain delivery of la rge molecule drugs such as proteins and antibodies, which have become more common in ophthalmology in recent years [233 235] While not the major focus of this work, the PSHU NIPAAm RTG was evaluated for its ability to sustain release of the antibody ranibizumab (see Appendix H ). Ranibizumab, loaded a t 10 wt%, was completely released within 24 hours. As described earlier, this is hypothesized to be caused by the rapid de swelling behavior of PNIPAAm based gels and may hold true for many hydrophilic drug. To date, only preliminary studies have been per formed on the ability of this system to release hydrophilic drugs, especially large molecules. These results were less favorable than those seen for hydrophobic drugs and deserve more investigation. In order to ic drug delivery, the ability to control the de swelling behavior of the RTG during gelation may be necessary As described in previous sections, this de swelling behavior has been previously reported as fundamental to PNIPAAm based thermal gels [111] and may be difficult to overcome. Drug release studies were also only performed in in vitro conditions within this work. Drug delivery systems often suffer fr om a lack of in vitro in vivo correlation in drug release kinetics [236] As a result, while in vitro studies provide a quick, efficient way to screen technologies for their amenability to deliver certain drugs, only in vivo studies can provi de an accurate measure of how the system will perform in that context. While in vivo efficacy

PAGE 145

131 studies were outside the scope of the present work, their absence does limit the conclusions that can be made about the translational potential of this system. Fi nally, biodegradation analyses, especially of the newly proposed PSHU NIPAAm OH system, are incomplete. Biodegradation continues to be a major challenge for PNIPAAm based systems, but the novel degradable graft copolymer structure proposed herein may provi de an entirely new avenue for biodegradable RTGs. However, this new system requires more thorough characterization work to prove that its properties are not adversely affected as compared to the previous PSHU NIPAAm chemistry. The limited characterization and degradation data available for PSHU NIPAAm OH limits the ability to make conclusions about its applicability and translational potential. Future d irections Polymer development PSHU and PSHU NIPAAm have been shown to meet the requirements of the present drug delivery system and possess several desirable properties including biocompatibility and a strong ability to be functionalized. Efforts to c ontinue improving these polymers may yield significant results. For example, the current PSHU synthetic route y ields a polymer with low urea content (see Appendix E ). The ratio of urea to urethane bonding within poly(urea urethanes) has been shown to directly affect mechanical properties such as tensile strength and elasticity [237,238] In these copolymers, areas of high urea bonding typical form hard segments du e to extensive urea ure a hydrogen bonding. In contrast, long methylene segments and carbamate bonds typical of urethane segments form soft segments and contribute less to overall mechanical properties. A more deliberate PSHU synthetic route

PAGE 146

132 which permits control over the ratio o f urea and urethane bonding would likely improve mechanical properties of the system and may aid development of a similar polymer with elastomeric properties. Swelling kinetics The rate and extent of de swelling of PSHU NIPAAm gels poses on e of the most s erious challenges of this system. This property could significantly impact the application of this system both as a drug delivery vehicle (specifically in application to hydrophilic molecules) and as a tissue engineering construct. Therefore, future effort s should consider strategies to limit this kinetic. This would involve designing a mechanism by which the RTG can retain water molecules during (but not hindering) development of hydrophobic interactions between PNIPAAm monomer units. Of course, this prese nts a significant challenge since PNIPAAm chain collapse and water retention ar e competing processes. For example, grafting a number of short PEG molecules to PSHU along with PNIPAAm would help retain water molecules, but the highly hydrophilic nature of P EG and its surrounding water molecules may prevent PNIPAAm molecules from aggregating and f orming a coherent physical gel. Determining whether a balance can be found that permits thermal gelation but prevents de swelling would be a valuable research path. Large molecule translation Large molecule drug delivery systems remain of major interest in the pharmaceutical sciences. The challenges are numerous and include designing a system that can prevent rapid diffusion of the hydrophilic antibodies, proteins or peptides from the system while preserving their delicate three dimensional superstructure [239,240] In the PSHU

PAGE 147

133 NIPAAm system, rapid efflux of these molecules was observed, presumably due to the severe de swelling behavior of the system. In order to improve the platform nature of this system, its a menability to large molecule delivery should be improved. One avenue would involve first solving the aforementioned de swelling problem, which may help retain large molecule drugs within the matrix for subsequent release. An alternative strategy would invo lve covalently linking the drug directly to the PSHU NIPAAm copolymer, thereby ensuring it is retained within the gel even if significant water efflux occurs. This could be readily accomplished due to the abundance of unconjugated primary amines available on the PSHU backbone, which could be used to link protein, antibody or peptide drugs through their carboxylic acid terminals. Release kinetics of covalently linked drugs from this system would then be dependent on the rate of cleavage of the bond between P SHU NIPAAm and the drug and subsequent diffusion of the molecule out of the RTG matrix. Evaluation of different conjugation chemistries, their effect on release kinetics and whether the system can protect the drugs from denaturation over an extended period of time would all be necessary characterization steps. In vivo experimentation To demonstrate the long term therapeutic potential of this system, in vivo efficacy studies would be beneficial. This could be accomplished, for example, against choroidal neov ascularization in rats. Experimental choroidal neovascularization is readily induced in rats by laser photocoagulation and intravitreal TA has been shown to be effective in preventing fibrovascular proliferation in this model, which is readily assessed by f undus and fluorescein angiography examinations [241] Long term evaluation of rats after intravitreal injections of the TA loaded RTG micelle system against this model would

PAGE 148

134 provide valuable in vivo over a long period of time. A head to head comparison against intravitreal TA injections would demonstrate the improved longevity of a single administration of the system. Tissue engineering applications The PSHU NIPAAm system has been described herein for its applicability to drug delivery scenarios. However, it may also hold great promise as a n injectable tissue engineering scaffold for cell delivery and/or o rganized regeneration of tissue. Its ability to simultaneously delivery therapeutics such as drugs or growth factors may be a significant benefit. Some preliminary in vivo experiments were completed in which the system was investigated for its ability to r egenerate injured optic nerve. After induction of an optic nerve crush model in rats (to simulate severe optic nerve damage or degeneration), RGD conjugated PSHU NIPAAm was injected into the optic nerve sheath at the site of injury. After 6 days, rats were euthanized and the eyes were sectioned and stained for g lial Figure 6.2: Functionalized PSHU NIPAAm was injected into the optic nerve sheath of rats after induction of an optic nerve crush model to promote functional recovery. (A) A representative histological secti on demonstrates thinning of the optic nerve at the crush site but a lack of adverse reactions from PSHU NIPAAm injection. (B) GFAP staining of the same optic nerve shows deletion of active astrocytes within the crush region indicating positive functional d amage from the optic nerve crush model but no functional recovery within the 7 day experiment.

PAGE 149

135 fibrillary acidic protein (GFAP) to image functional optic nerve damage. Representative sections of optic nerve are shown in Figure 6.2 and demonstrate induction of optic nerve damage (identifia ble by a narrowing of the optic nerve at the crush site and a lack of GFAP staining in this section, indicating loss of active astrocytes within the crush area). However, after 7 days no regeneration of functionality was observed in any sections. Neverthel ess, injection of the functionalized RTG at the injection site may promote regeneration on a longer time scale or with simultaneous delivery of a neurotropic growth factor. This work may also lead into other valuable tissue engineering applications.

PAGE 150

136 RE FERENCES [1] P.M. Hughes, A.K. Mitra, Overview of ocular drug delivery and iatrogenic ocular cytopathologies, in: A.K. Mitra (Ed.), Ophthalmic Drug Deliv. Syst., Marcel Dekker, Inc., New York, 1993: pp. 1 27. [2] Wikipedia: Eye, (n.d.). [3] H. Mochizuki, M. Yamada, S. Hatou, K. Tsubota, Turnover rate of tear film lipid layer determined by fluorophotometry., Br. J. Ophthalmol. 93 (2009) 1535 8. [4] K. Laine, K. Jrvinen, R. Mechoulam, A. Breuer, T. Jrvinen, Com parison of the enzymatic stability and intraocular pressure effects of 2 arachidonylglycerol and noladin ether, a novel putative endocannabinoid., Invest. Ophthalmol. Vis. Sci. 43 (2002) 3216 22. [5] M. Nakamura, E. Shirasawa, M. Hikida, Characterization o f esterases involved in the hydrolysis of dipivefrin hydrochloride., Ophthalmic Res. 25 (1993) 46 51. [6] R.D. Schoenwald, Ocular drug delivery. Pharmacokinetic considerations., Clin. Pharmacokinet. 18 (1990) 255 69. [7] M.A. Watsky, M.M. Jablonski, H.F. E delhauser, Comparison of conjunctival and corneal surface areas in rabbit and human., Curr. Eye Res. 7 (1988) 483 6. [8] A.J.W. Huang, S.C. Tseng, K.R. Kenyon, Paracellular permeability of corneal and conjunctival epithelia., Invest. Ophthalmol. Vis. Sci. 30 (1989) 684 689. [9] K.D. Kao, D.W. Lu, C.H. Chiang, H.S. Huang, Corneal and scleral penetration studies of 6 hydroxyethoxy 2 benzothiazole sulfonamide: a topical carbonic anhydrase inhibitor., J. Ocul. Pharmacol. 6 (1990) 313 20. [10] I. Ahmed, R.D. Gok hale, M. V Shah, T.F. Patton, Physicochemical determinants of drug diffusion across the conjunctiva, sclera, and cornea., J. Pharm. Sci. 76 (1987) 583 6. [11] J. Barar, A.R. Javadzadeh, Y. Omidi, Ocular novel drug delivery: impacts of membranes and barrier s., Expert Opin. Drug Deliv. 5 (2008) 567 81. [12] R. Gaudana, H.K. Ananthula, A. Parenky, A.K. Mitra, Ocular drug delivery, AAPS J. 12 (2010) 348 60. [13] I. Ahmed, The noncorneal route in ocular drug delivery., in: A.K. Mitra (Ed.), Ophthalmic Drug Deliv Syst., Marcel Dekker, Inc., New York, 2003: pp. 335 363.

PAGE 151

137 [14] C. Gray, Systemic toxicity with topical ophthalmic medications in children, Paediatr. Perinat. Drug Ther. 7 (2006) 23 29. [15] L. Salminen, Review: systemic absorption of topically applied ocu lar drugs in humans., J. Ocul. Pharmacol. 6 (1990) 243 9. [16] W.L. Nelson, F.T. Fraunfelder, J.M. Sills, J.B. Arrowsmith, J.N. Kuritsky, Adverse respiratory and cardiovascular events attributed to timolol ophthalmic solution, 1978 1985., Am. J. Ophthalmol 102 (1986) 606 11. [17] J.P. Diamond, Systemic adverse effects of topical ophthalmic agents. Implications for older patients., Drugs Aging. 11 (1997) 352 60. [18] C. Le Jeunne, Y. Munera, F.C. Hugues, Systemic effects of three beta blocker eyedrops: comp arison in healthy volunteers of beta 1 and beta 2 adrenoreceptor inhibition., Clin. Pharmacol. Ther. 47 (1990) 578 83. [19] K. Kyyrnen, A. Urtti, Improved ocular: systemic absorption ratio of timolol by viscous vehicle and phenylephrine., Invest. Ophthal mol. Vis. Sci. 31 (1990) 1827 33. [20] W.C. Stewart, R.P. Chorak, H.H. Hunt, G. Sethuraman, Factors associated with visual loss in patients with advanced glaucomatous changes in the optic nerve head., Am. J. Ophthalmol. 116 (1993) 176 81. [21] O. Kosoko, H .A. Quigley, S. Vitale, C. Enger, L. Kerrigan, J.M. Tielsch, Risk factors for noncompliance with glaucoma follow clinic., Ophthalmology. 105 (1998) 2105 11. [22] men. Difficulties in identifying the noncooperator., JAMA. 203 (1968) 922 6. [23] P.A. Granstrm, Glaucoma patients not compliant with their drug therapy: clinical and behavioural aspects., Br. J. Ophthalmol. 66 (1982) 464 70. [24] S.A. Taylor, S.M. Galbra ith, R.P. Mills, Causes of non compliance with drug regimens in glaucoma patients: a qualitative study., J. Ocul. Pharmacol. Ther. 18 (2002) 401 9. [25] G.F. Schwartz, Compliance and persistency in glaucoma follow up treatment., Curr. Opin. Ophthalmol. 16 (2005) 114 21. [26] M.A. Kass, D.W. Meltzer, M. Gordon, D. Cooper, J. Goldberg, Compliance with topical pilocarpine treatment, Am. J. Ophthalmol. 101 (1986) 515 23.

PAGE 152

138 [27] C.O. Okeke, H.A. Quigley, H.D. Jampel, G. Ying, R.J. Plyler, Y. Jiang, et al., Adheren ce with topical glaucoma medication monitored electronically the Travatan Dosing Aid study., Ophthalmology. 116 (2009) 191 9. [28] J. H. Park, M.G. Allen, M.R. Prausnitz, Polymer microneedles for controlled release drug delivery., Pharm. Res. 23 (2006) 100 8 19. [29] J. Jiang, H.S. Gill, D. Ghate, B.E. McCarey, S.R. Patel, H.F. Edelhauser, et al., Coated microneedles for drug delivery to the eye., Invest. Ophthalmol. Vis. Sci. 48 (2007) 4038 43. [30] V. Lee, V. Li, Prodrugs for improved ocular drug delivery, Adv. Drug Deliv. Rev. (1989). [31] K. Hosoya, V.H.L. Lee, K. J. Kim, Roles of the conjunctiva in ocular drug delivery: a review of conjunctival transport mechanisms and their regulation., Eur. J. Pharm. Biopharm. 60 (2005) 227 40. [32] E.Y. Kim, Z.G. Gao, J.S. Park, H. Li, K. Han, rhEGF/HP beta CD complex in poloxamer gel for ophthalmic delivery., Int. J. Pharm. 233 (2002) 159 67. [33] T.W. Kim, J.D. Lindsey, M. Aihara, T.L. Anthony, R.N. Weinreb, Intraocular distribution of 70 kDa dextran after subconjunc tival injection in mice., Invest. Ophthalmol. Vis. Sci. 43 (2002) 1809 16. [34] H. Kim, M.R. Robinson, M.J. Lizak, G. Tansey, R.J. Lutz, P. Yuan, et al., Controlled drug release from an ocular implant: an evaluation using dynamic three dimensional magnetic resonance imaging., Invest. Ophthalmol. Vis. Sci. 45 (2004) 2722 31. [35] A.C. Amrite, S.P. Ayalasomayajula, U.B. Kompella, Nano and Micro Particles Reach Retina Following Systemic but Not Subconjunctival Administration, Invest. Ophthalmol. Vis. Sci. 44 (2003) 4449. [36] A.L. Weiner, Drug Delivery Systems in Ophthalmic Applications, in: T. Yorio, A.F. Clark, M.B. Wax (Eds.), Ocul. Ther. Eye New Discov., Elsevier Academic Press, New York, NY, 2008: pp. 7 43. [37] V. Fernandez, P.D. Lamar, F. Fantes, S. Dub ovy, M. Murahara, H. Lau, et al., Tissue Interaction Of Differents Materials For A New Glaucoma Implant In Rabbits, Invest. Ophthalmol. Vis. Sci. 43 (2002) 3371. [38] acetonid e., Can. J. Ophthalmol. 40 (2005) 63 8.

PAGE 153

139 [39] H.J. Koh, L. Cheng, K. Bessho, T.R. Jones, M.C. Davidson, W.R. Freeman, Intraocular properties of urokinase derived antiangiogenic A6 peptide in rabbits., J. Ocul. Pharmacol. Ther. 20 (2004) 439 49. [40] L. Chen g, K. Hostetler, N. Valiaeva, A. Tammewar, W.R. Freeman, J. Beadle, et al., Intravitreal crystalline drug delivery for intraocular proliferation diseases., Invest. Ophthalmol. Vis. Sci. 51 (2010) 474 81. [41] N.B. Shelke, R. Kadam, P. Tyagi, V.R. Rao, U.B. Kompella, Intravitreal Poly(L lactide) Microparticles Sustain Retinal and Choroidal Delivery of TG 0054, a Hydrophilic Drug Intended for Neovascular Diseases., Drug Deliv. Transl. Res. 1 (2011) 76 90. [42] J.A. Cardillo, L.A.S. Melo, R.A. Costa, M. Skaf, R. Belfort, A.A. Souza Filho, et al., Comparison of intravitreal versus posterior sub triamcinolone acetonide for diffuse diabetic macular edema., Ophthalmology. 112 (2005) 1557 63. [43] S. Duvvuri, K. Gaurav Janoria, A.K. Mitr a, Effect of polymer blending on the release of ganciclovir from PLGA microspheres., Pharm. Res. 23 (2006) 215 23. [44] J.S. Myung, G.D. Aaker, S. Kiss, Treatment of noninfectious posterior uveitis with dexamethasone intravitreal implant., Clin. Ophthalmol 4 (2010) 1423 6. [45] P. Mruthyunjaya, D. Khalatbari, P. Yang, S. Stinnett, R. Tano, P. Ashton, et al., Efficacy of low release rate fluocinolone acetonide intravitreal implants to treat experimental uveitis., Arch. Ophthalmol. 124 (2006) 1012 8. [46] P. A. Pearson, T.L. Comstock, M. Ip, D. Callanan, L.S. Morse, P. Ashton, et al., Fluocinolone acetonide intravitreal implant for diabetic macular edema: a 3 year multicenter, randomized, controlled clinical trial., Ophthalmology. 118 (2011) 1580 7. [47] S.E. Varner, D. Guven, L. Lawin, A. Anderson, E. de Juan, Minimally Invasive Helical Implant for Intravitreal Drug Delivery, Invest. Ophthalmol. Vis. Sci. 45 (2004) 4555. [48] E. Sakurai, Scleral Plug of Biodegradable Polymers Containing Tacrolimus (FK506) for Experimental Uveitis, Invest. Ophthalmol. Vis. Sci. 44 (2003) 4845 4852. [49] M.S. Stay, J. Xu, T.W. Randolph, V.H. Barocas, Computer simulation of convective and diffusive transport of controlled release drugs in the vitreous humor., Pharm. Res. 20 (2003) 96 102.

PAGE 154

140 [50] L. Pitknen, M. Ruponen, J. Nieminen, A. Urtti, Vitreous is a barrier in nonviral gene transfer by cationic lipids and polymers., Pharm. Res. 20 (2003) 576 83. [51] G.L. Amidon, H. Lennerns, V.P. Shah, J.R. Crison, A theoretical basis for a biopharmaceutic drug classification: the correlation of in vitro drug product dissolution and in vivo bioavailability., Pharm. Res. 12 (1995) 413 20. [52] J. Baldoni, Role of BCS in drug development, in: AAPS/FDA Work. BE, BCS Beyond, n.d. [53] L. Apt, A. Henrick, L.M. Silverman, Patient compliance with use of topical ophthalmic corticosteroid suspensions., Am. J. Ophthalmol. 87 (1979) 210 4. [54] N.M. Davies, Biopharmaceutical Considerations In Topical Ocular Drug Delivery, Clin. Exp. Pharmacol. Physiol. 2 7 (2000) 558 562. [55] B.J. Aungst, Novel formulation strategies for improving oral bioavailability of drugs with poor membrane permeation or presystemic metabolism., J. Pharm. Sci. 82 (1993) 979 87. [56] D.G. Musson, A.M. Bidgood, O. Olejnik, Comparative corneal penetration of prednisolone sodium phosphate and prednisolone acetate in NZW rabbits., J. Ocul. Pharmacol. 7 (1991) 175 82. [57] S.L. Cooper, S.A. Visser, R.W. Hergenrother, N.M.K. Lamba, Classes of materials used in medicine: Polymers, in: B.D. Ra tner, A.S. Hoffman, F.J. Schoen, J.E. Lemons (Eds.), Biomater. Sci., Second Edi, Elsevier Academic Press, London, UK, 2004: pp. 67 79. [58] J.M. Anderson, G. Cook, B. Costerton, S.R. Hanson, A. Hensten Pettersen, N. Jacobsen, et al., Host Reactions to Biom aterials and their Evaluation, in: B.D. Ratner, A.S. Hoffman, F.J. Schoen, J.E. Lemons (Eds.), Biomater. Sci., Second Edi, Elsevier Academic Press, London, UK, 2004. [59] J. Kohn, S. Abramson, R. Langer, Bioresorbable and Bioerodible Materials, in: B.D. Ra tner, A.S. Hoffman, F.J. Schoen, J.E. Lemons (Eds.), Biomater. Sci., 2nd Ed, Elsevier Academic Press, San Diego, CA, 2004: pp. 115 127. [60] B.D. Ratner, A.S. Hoffman, F.J. Schoen, J.E. Lemons, Biomaterials Science: An Introduction to Materials in Medicine 2nd ed., Elsevier Academic Press, San Diego, CA, n.d. [61] T. Estey, J. Kang, S.P. Schwendeman, J.F. Carpenter, BSA degradation under acidic conditions: a model for protein instability during release from PLGA delivery systems., J. Pharm. Sci. 95 (2006) 1626 39.

PAGE 155

141 [62] G. Zhu, S.P. Schwendeman, Stabilization of proteins encapsulated in cylindrical poly(lactide co glycolide) implants: mechanism of stabilization by basic additives., Pharm. Res. 17 (2000) 351 7. [63] M. Diwan, T.G. Park, Pegylation enhances pr otein stability during encapsulation in PLGA microspheres., J. Control. Release. 73 (2001) 233 44. [64] K. Fu, A.M. Klibanov, R. Langer, Protein stability in controlled release systems., Nat. Biotechnol. 18 (2000) 24 5. [65] L.A. Moore, R.L. Norton, S.L. W hitman, R.L. Dunn, An injectable biodegradable drug delivery system based on acrylic terminated poly caprolactone, in: Annu. Meet. Soc. Biomater., CA, 1995. [66] M. Saleem, A. Alam, S. Ahmed, M. Iqbal, S. Sultana, Tephrosia purpurea Ameliorates Benzoyl Per oxide induced Cutaneous Toxicity in Mice: Diminution of Oxidative Stress, Pharm. Pharmacol. Commun. 5 (1999) 455 461. [67] G. Biehl, J. Harms, U. Hanser, [Experimental studies on heat development in bone during polymerization of bone cement. Intraoperative measurement of temperature in normal blood circulation and in bloodlessness]., Arch. Orthop. Unfallchir. 78 (1974) 62 9. [68] J.S. Temenoff, A.G. Mikos, Injectable biodegradable materials for orthopedic tissue engineering., Biomaterials. 21 (2000) 2405 12 [69] S. Lu, K.S. Anseth, Photopolymerization of multilaminated poly(HEMA) hydrogels for controlled release., J. Control. Release. 57 (1999) 291 300. [70] J. Hubbell, Hydrogel systems for barriers and local drug delivery in the control of wound healing, J Control. Release. 39 (1996) 305 313. [71] N.P. Desai, J.L. Hill, J.A. Hubbell, P.P. Chandrashekhar, A.S. Sawhney, Photopolymerizable biodegradable hydrogels as tissue contacting materials and controlled release carriers, 1993. [72] J.L. West, J.A. Hubbel l, Localized intravascular protein delivery from photopolymerized hydrogels, in: Proc. Int. Symp. Control. Rel. Bioact. Mater., 1995: pp. 17 18. [73] A. Hatefi, B. Amsden, Biodegradable injectable in situ forming drug delivery systems., J. Control. Release 80 (2002) 9 28. [74] H.H. Tnnesen, J. Karlsen, Alginate in drug delivery systems., Drug Dev. Ind. Pharm. 28 (2002) 621 30.

PAGE 156

142 [75] S. Cohen, E. Lobel, A. Trevgoda, Y. Peled, A novel in situ forming ophthalmic drug delivery system from alginates undergoing gelation in the eye, J. Control. Release. 44 (1997) 201 208. [76] Y. Suzuki, Y. Nishimura, M. Tanihara, K. Suzuki, T. Nakamura, Y. Shimizu, et al., Evaluation of a novel alginate gel dressing: cytotoxicity to fibroblasts in vitro and foreign body reaction in pig skin in vivo., J. Biomed. Mater. Res. 39 (1998) 317 22. [77] A.B. Lansdown, M.J. Payne, An evaluation of the local reaction and biodegradation of calcium sodium alginate (Kaltostat) following subcutaneous implantation in the rat., J. R. Coll. Surg. Edinb. 39 (1994) 284 8. [78] ocular drug delivery., Expert Opin. Drug Deliv. 9 (2012) 383 402. [79] N.H. Shah, A.S. Railkar, F.C. Chen, R. Tarantino, S. Kumar, M.H. Infeld et al., A biodegradable injectable implant for delivering micro and, 27 (1993) 139 147. [80] W. Lambert, K. Peck, Development of an in situ forming biodegradable poly lactide coglycolide system for the controlled release of proteins, J. Control. Release. 33 (1995) 189 195. [81] M.L. Radomsky, G. Brouwer, B.J. Floy, D.J. Loury, F. Chu, A.J. Tipton, et al., The controlled release of Ganirelix from the Atrigel injectable implant system, in: Proc. Int. Symp. Control. Rel. Bioact. Mater., 1993. [82] M.A. Royal s, S.M. Fujita, G.L. Yewey, J. Rodriguez, P.C. Schultheiss, R.L. Dunn, Biocompatibility of a biodegradable in situ forming implant system in rhesus monkeys., J. Biomed. Mater. Res. 45 (1999) 231 9. [83] D.M. Gordon, K.E. Kleberger, The Effect of Dimethyl S ulfoxide (DMSO) on Animal and Human Eyes, Arch. Ophthalmol. 79 (1968) 423 427. [84] H. Kranz, G.A. Brazeau, J. Napaporn, R.L. Martin, W. Millard, R. Bodmeier, Myotoxicity studies of injectable biodegradable in situ forming drug delivery systems., Int. J. P harm. 212 (2001) 11 8. [85] J. Galvao, B. Davis, M. Tilley, E. Normando, M.R. Duchen, M.F. Cordeiro, Unexpected low dose toxicity of the universal solvent DMSO., FASEB J. (2013). [86] A. Hoffman, Applications of thermally reversible polymers and hydrogels in therapeutics and diagnostics, J. Control. Release. 6 (1987) 297 305.

PAGE 157

143 [87] P. C. Chen, D.S. Kohane, Y.J. Park, R.H. Bartlett, R. Langer, V.C. Yang, Injectable microparticle gel system for prolonged and localized lidocaine release. II. In vivo anesthetic effects., J. Biomed. Mater. Res. A. 70 (2004) 459 66. [88] A. Paavola, J. Yliruusi, Y. Kajimoto, E. Kalso, T. Wahlstrm, P. Rosenberg, Controlled release of lidocaine from injectable gels and efficacy in rat sciatic nerve block., Pharm. Res. 12 (1995) 1997 2002. [89] M.D. Determan, J.P. Cox, S.K. Mallapragada, Drug release from pH responsive thermogelling pentablock copolymers., J. Biomed. Mater. Res. A. 81 (2007) 326 33. [90] J. Cleary, L.E. Bromberg, E. Magner, Diffusion and Release of Solutes in Pluronic g poly(acrylic acid) Hydrogels, Langmuir. 19 (2003) 9162 9172. [91] A. Gutowska, B. Jeong, M. Jasionowski, Injectable gels for tissue engineering., Anat. Rec. 263 (2001) 342 9. [92] Y.K. Han, J.W. Kwon, J.S. Kim, C. S. Cho, W.R. Wee, J.H. Lee, In vitro and in vivo study of lens refilling with poloxamer hydrogel., Br. J. Ophthalmol. 87 (2003) 1399 402. [93] J.D. Kretlow, L. Klouda, A.G. Mikos, Injectable matrices and scaffolds for drug delivery in tissue engineering., Adv. Drug Deliv. Rev. 59 (2007) 263 7 3. [94] K. Kazunori, K. Glenn S., Y. Masayuki, O. Teruo, S. Yasuhisa, Block copolymer micelles as vehicles for drug delivery, J. Control. Release. 24 (1993) 119 132. [95] T.R. Hoare, D.S. Kohane, Hydrogels in drug delivery: Progress and challenges, Polymer (Guildf). 49 (2008) 1993 2007. [96] F. Artzner, S. Geiger, A. Olivier, C. Allais, Interactions between poloxamers in aqueous solutions: Micellization and gelation studied by differential scanning calorimetry, small angle X ray scattering, and, Langmuir. ( 2007) 5085 5092. [97] M.L. Veyries, G. Couarraze, S. Geiger, F. Agnely, L. Massias, B. Kunzli, et al., Controlled release of vancomycin from poloxamer 407 gels., Int. J. Pharm. 192 (1999) 183 93. [98] S. Miyazaki, Y. Ohkawa, M. Takada, D. Attwood, Antitumo r effect of pluronic F 127 gel containing mitomycin C on sarcoma 180 ascites tumor in mice., Chem. Pharm. Bull. (Tokyo). 40 (1992) 2224 6. [99] Z.G. Wout, E.A. Pec, J.A. Maggiore, R.H. Williams, P. Palicharla, T.P. Johnston, Poloxamer 407 mediated changes in plasma cholesterol and triglycerides

PAGE 158

144 following intraperitoneal injection to rats., J. Parenter. Sci. Technol. 46 (n.d.) 192 200. [100] T.P. Johnston, W.K. Palmer, Mechanism of poloxamer 407 induced hypertriglyceridemia in the rat., Biochem. Pharmacol. 4 6 (1993) 1037 42. [101] D.W. Miller, E. V Batrakova, T.O. Waltner, Alakhov VYu, A. V Kabanov, Interactions of pluronic block copolymers with brain microvessel endothelial cells: evidence of two potential pathways for drug absorption., Bioconjug. Chem. 8 (n .d.) 649 57. [102] E. V Batrakova, H.Y. Han, Alakhov VYu, D.W. Miller, A. V Kabanov, Effects of pluronic block copolymers on drug absorption in Caco 2 cell monolayers., Pharm. Res. 15 (1998) 850 5. [103] Alakhov VYu, Moskaleva EYu, E. V Batrakova, A. V Kab anov, Hypersensitization of multidrug resistant human ovarian carcinoma cells by pluronic P85 block copolymer., Bioconjug. Chem. 7 (n.d.) 209 16. [104] Z. Zhang, M. Al Rubeai, C.R. Thomas, Effect of Pluronic F 68 on the mechanical properties of mammalian c ells., Enzyme Microb. Technol. 14 (1992) 980 3. [105] G. Wu, H. a. Khant, W. Chiu, K.Y.C. Lee, Effects of bilayer phases on phospholipid poloxamer interactions, Soft Matter. 5 (2009) 1496. [106] B. Jeong, Y.H. Bae, S.W. Kim, Drug release from biodegradable injectable thermosensitive hydrogel of PEG PLGA PEG triblock copolymers., J. Control. Release. 63 (2000) 155 63. [107] B. Jeong, Y.H. Bae, S.W. Kim, In situ gelation of PEG PLGA PEG triblock copolymer aqueous solutions and degradation thereof., J. Biomed. Mater. Res. 50 (2000) 171 7. [108] G.M. Zentner, R. Rathi, C. Shih, J.C. McRea, M.H. Seo, H. Oh, et al., Biodegradable block copolymers for delivery of proteins and water insoluble drugs., J. Control. Release. 72 (2001) 203 15. [109] D. Schmaljohann, Ther mo and pH responsive polymers in drug delivery., Adv. Drug Deliv. Rev. 58 (2006) 1655 70. [110] H.G. Schild, Poly(N isopropylacrylamide): experiment, theory and application, Prog. Polym. Sci. 17 (1992) 163 249. [111] S. Hirotsu, Y. Hirokawa, T. Tanaka, Vo lume phase transitions of ionized N isopropylacrylamide gels, J. Chem. Phys. 87 (1987) 1392.

PAGE 159

145 [112] G. Graziano, On the temperature induced coil to globule transition of poly N isopropylacrylamide in dilute aqueous solutions., Int. J. Biol. Macromol. 27 (20 00) 89 97. [113] K. Kubota, S. Fujishige, I. Ando, Single chain transition of poly (N isopropylacrylamide) in water, J. Phys. Chem. (1990) 5154 5158. [114] X. Wang, X. Qiu, C. Wu, Comparison of the Coil to Globule and the Globule to Coil Transitions of a S ingle Poly( N isopropylacrylamide) Homopolymer Chain in Water, Macromolecules. 31 (1998) 2972 2976. [115] E.I. Tiktopulo, V.N. Uversky, V.B. Lushchik, S.I. Klenin, V.E. Bychkova, O.B. Globule Transition in Homopolymers, Macromolecul es. 28 (1995) 7519 7524. [116] M. Shibayama, M. Morimoto, S. Nomura, Phase Separation Induced Mechanical Transition of Poly(N isopropylacrylamide)/Water Isochore Gels, Macromolecules. 27 (1994) 5060 5066. [117] X. Z. Zhang, D. Q. Wu, C. C. Chu, Synthesis, characterization and controlled drug release of thermosensitive IPN PNIPAAm hydrogels., Biomaterials. 25 (2004) 3793 805. [118] J. Han, K. Wang, D. Yang, J. Nie, Photopolymerization of methacrylated chitosan/PNIPAAm hybrid dual sensitive hydrogels as carri er for drug delivery., Int. J. Biol. Macromol. 44 (2009) 229 35. [119] S.J. Wilson, A. V Gorelov, Y.A. Rochev, F. McGillicuddy, K.A. Dawson, W.M. Gallagher, et al., Extended delivery of the antimitotic agent colchicine from thermoresponsive N isopropylacry lamide based copolymer films to human vascular smooth muscle cells., J. Biomed. Mater. Res. A. 67 (2003) 667 73. [120] T. Hoare, R. Pelton, Impact of microgel morphology on functionalized microgel drug interactions., Langmuir. 24 (2008) 1005 12. [121] W. X un, D. Q. Wu, Z. Y. Li, H. Y. Wang, F. W. Huang, S. X. Cheng, et al., Peptide functionalized thermo sensitive hydrogels for sustained drug delivery., Macromol. Biosci. 9 (2009) 1219 26. [122] E.M. Agency, Ozurdex Summary of Product Characteristics, Annex 1 2010. [123] R. Gallego pinazo, P. Hernndez martinez, A. Hervs ontiveros, S. Martnez castillo, R. Dolz marco, J.F. Arvalo, et al., Local Safety Concerns of Repited Dexamethasone Intravitreal Implant (Ozurdex) For Macular Diseases, J. Ocul. Dis. Ther. 1 (2013) 10 14.

PAGE 160

146 [124] M. Bhowmik, M.K. Bain, L.K. Ghosh, D. Chattopadhyay, Effect of salts on gelation and drug release profiles of methylcellulose based ophthalmic thermo reversible in situ gels., Pharm. Dev. Technol. 16 (2011) 385 91. [125] A.H. El Kame l, In vitro and in vivo evaluation of Pluronic F127 based ocular delivery system for timolol maleate., Int. J. Pharm. 241 (2002) 47 55. [126] A. Paavola, J. Yliruusi, P. Rosenberg, Controlled release and dura mater permeability of lidocaine and ibuprofen f rom injectable poloxamer based gels., J. Control. Release. 52 (1998) 169 78. [127] S.D. Desai, J. Blanchard, In vitro evaluation of pluronic F127 based controlled release ocular delivery systems for pilocarpine., J. Pharm. Sci. 87 (1998) 226 30. [128] T.R. Thrimawithana, S. a Young, C.R. Bunt, C.R. Green, R.G. Alany, In vitro and in vivo evaluation of carrageenan/methylcellulose polymeric systems for transscleral delivery of macromolecules., Eur. J. Pharm. Sci. 44 (2011) 399 409. [129] H. Gupta, S. Jain, R. Mathur, P. Mishra, A.K. Mishra, T. Velpandian, Sustained ocular drug delivery from a temperature and pH triggered novel in situ gel system., Drug Deliv. 14 (2007) 507 15. [130] B. Srividya, R.M. Cardoza, P.D. Amin, Sustained ophthalmic delivery of ofloxac in from a pH triggered in situ gelling system., J. Control. Release. 73 (2001) 205 11. [131] W. D. Ma, H. Xu, C. Wang, S. F. Nie, W. S. Pan, Pluronic F127 g poly(acrylic acid) copolymers as in situ gelling vehicle for ophthalmic drug delivery system., Int. J. Pharm. 350 (2008) 247 56. [132] S. B. Lee, D.H. Geroski, M.R. Prausnitz, H.F. Edelhauser, Drug delivery through the sclera: effects of thickness, hydration, and sustained release systems., Exp. Eye Res. 78 (2004) 599 607. [133] S. Duvvuri, K.G. Janoria A.K. Mitra, Development of a novel formulation containing poly(d,l lactide co glycolide) microspheres dispersed in PLGA PEG PLGA gel for sustained delivery of ganciclovir., J. Control. Release. 108 (2005) 282 93. [134] S. Duvvuri, K.G. Janoria, D. Pal, A .K. Mitra, Controlled delivery of ganciclovir to the retina with drug loaded Poly(d,L lactide co glycolide) (PLGA) microspheres dispersed in PLGA PEG PLGA Gel: a novel intravitreal delivery system for the treatment of cytomegalovirus retinitis., J. Ocul. P harmacol. Ther. 23 (2007) 264 74.

PAGE 161

147 [135] C. H. Wang, Y. S. Hwang, P. R. Chiang, C. R. Shen, W. H. Hong, G. H. Hsiue, Extended release of bevacizumab by thermosensitive biodegradable and biocompatible hydrogel., Biomacromolecules. 13 (2012) 40 8. [136] D. Pa rk, V. Shah, B.M. Rauck, T.R. Friberg, Y. Wang, An anti angiogenic reverse thermal gel as a drug delivery system for age related wet macular degeneration., Macromol. Biosci. 13 (2013) 464 9. [137] J. Panyam, V. Labhasetwar, Biodegradable nanoparticles for drug and gene delivery to cells and tissue., Adv. Drug Deliv. Rev. 55 (2003) 329 47. [138] T.W. Prow, I. Bhutto, S.Y. Kim, R. Grebe, C. Merges, D.S. McLeod, et al., Ocular nanoparticle toxicity and transfection of the retina and retinal pigment epithelium. Nanomedicine. 4 (2008) 340 9. [139] Y. Diebold, M. Calonge, Applications of nanoparticles in ophthalmology., Prog. Retin. Eye Res. 29 (2010) 596 609. [140] J. L. Bourges, S.E. Gautier, F. Delie, R.A. Bejjani, J. C. Jeanny, R. Gurny, et al., Ocular drug d elivery targeting the retina and retinal pigment epithelium using polylactide nanoparticles., Invest. Ophthalmol. Vis. Sci. 44 (2003) 3562 9. [141] S.K. Sahoo, F. Dilnawaz, S. Krishnakumar, Nanotechnology in ocular drug delivery., Drug Discov. Today. 13 (2 008) 144 51. [142] K.G. Janoria, S. Gunda, S.H.S. Boddu, A.K. Mitra, Novel approaches to retinal drug delivery., Expert Opin. Drug Deliv. 4 (2007) 371 88. [143] V.P. Torchilin, Targeted polymeric micelles for delivery of poorly soluble drugs., Cell. Mol. L ife Sci. 61 (2004) 2549 59. [144] M. Jones, J. Leroux, Polymeric micelles a new generation of colloidal drug carriers., Eur. J. Pharm. Biopharm. 48 (1999) 101 11. [145] V.P. Torchilin, Structure and design of polymeric surfactant based drug delivery syst ems., 2001. [146] I. Pepic, J. Lovric, J. Filipovic Grcic, Polymeric Micelles in Ocular Drug Delivery: Rationale, Strategies and Challenges, Chem. Biochem. Eng. Q. 26 (2012) 365 377. [147] C. Di Tommaso, A novel cyclosporin A micelle formulation for topica l ophthalmic applications: development, biocompatibility, corneal penetration, and efficacy studies, University of Geneva, 2011.

PAGE 162

148 [148] X. Li, Z. Zhang, J. Li, S. Sun, Y. Weng, H. Chen, Diclofenac/biodegradable polymer micelles for ocular applications., Nan oscale. 4 (2012) 4667 73. [149] A.K. Gupta, S. Madan, D.K. Majumdar, A. Maitra, Ketorolac entrapped in polymeric micelles: preparation, characterisation and ocular anti inflammatory studies., Int. J. Pharm. 209 (2000) 1 14. [150] C. Civiale, M. Licciardi, G. Cavallaro, G. Giammona, M.G. Mazzone, Polyhydroxyethylaspartamide based micelles for ocular drug delivery., Int. J. Pharm. 378 (2009) 177 86. [151] H. R. Lin, P. C. Chang, Novel pluronic chitosan micelle as an ocular delivery system., J. Biomed. Mater. Res. B. Appl. Biomater. 101 (2013) 689 99. [152] L. Xu, X. Xu, H. Chen, X. Li, Ocular biocompatibility and tolerance study of biodegradable polymeric micelles in the rabbit eye., Colloids Surf. B. Biointerfaces. 112 (2013) 30 4. [153] J.B. Jonas, I. Kreiss ig, R. Degenring, Intravitreal triamcinolone acetonide for treatment of intraocular proliferative, exudative, and neovascular diseases., Prog. Retin. Eye Res. 24 (2005) 587 611. [154] A.N. Antoszyk, J.L. Gottlieb, R. Machemer, D.L. Hatchell, The effects of intravitreal triamcinolone acetonide on experimental pre retinal 40. [155] R.O. Graham, Intravitreal Injection of Dexamethasone, Arch. Ophthalmol. 92 (1974) 149. [156] L.H. Block, R.N Patel, Solubility and dissolution of triamcinolone acetonide, J. Pharm. Sci. 62 (1973) 617 621. [157] A. Goundalkar, M. Mezei, Chemical modification of triamcinolone acetonide to improve liposomal encapsulation, J. Pharm. Sci. 73 (1984) 834 835. [158] P. M. Beer, S.J. Bakri, R.J. Singh, W. Liu, G.B. Peters, M. Miller, Intraocular concentration and pharmacokinetics of triamcinolone acetonide after a single intravitreal injection., Ophthalmology. 110 (2003) 681 6. [159] J.B. Jonas, R.F. Degenring, I. Kreissi g, I. Akkoyun, B.A. Kamppeter, Intraocular pressure elevation after intravitreal triamcinolone acetonide injection., Ophthalmology. 112 (2005) 593 8. [160] R.N. Gursoy, S. Benita, Self emulsifying drug delivery systems (SEDDS) for improved oral delivery of lipophilic drugs., Biomed. Pharmacother. 58 (2004) 173 82.

PAGE 163

149 [161] P.P. Desai, A. a. Date, V.B. Patravale, Overcoming poor oral bioavailability using nanoparticle formulations opportunities and limitations, Drug Discov. Today Technol. 9 (2012) e87 e95. [1 62] K. Fu, D.W. Pack, a M. Klibanov, R. Langer, Visual evidence of acidic environment within degrading poly(lactic co glycolic acid) (PLGA) microspheres., Pharm. Res. 17 (2000) 100 6. [163] M.S. Kim, H.H. Ahn, Y.N. Shin, M.H. Cho, G. Khang, H.B. Lee, An in vivo study of the host tissue response to subcutaneous implantation of PLGA and/or porcine small intestinal submucosa based scaffolds., Biomaterials. 28 (2007) 5137 43. [164] V.P. Torchilin, A.N. Lukyanov, Z. Gao, B. Papahadjopoulos Sternberg, Immunomice lles: targeted pharmaceutical carriers for poorly soluble drugs., Proc. Natl. Acad. Sci. U. S. A. 100 (2003) 6039 44. [165] S. Mansoor, B.D. Kuppermann, M.C. Kenney, Intraocular sustained release delivery systems for triamcinolone acetonide., Pharm. Res. 2 6 (2009) 770 84. [166] C.M. Jermak, J.T. Dellacroce, J. Heffez, G. a Peyman, Triamcinolone acetonide in ocular therapeutics., Surv. Ophthalmol. 52 (2007) 503 22. [167] H.M. Aliabadi, A. Lavasanifar, Polymeric micelles for drug delivery., Expert Opin. Drug Deliv. 3 (2006) 139 62. [168] D. Rickert, A. Lendlein, A.M. Schmidt, S. Kelch, W. Roehlke, R. Fuhrmann, et al., In vitro cytotoxicity testing of AB polymer networks based on oligo(epsilon caprolactone) segments after different sterilization techniques., J. Biomed. Mater. Res. B. Appl. Biomater. 67 (2003) 722 31. [169] sterilized by ethylene oxide, J. Pharm. Sci. 57 (1968) 12 17. [170] A.G. Hausberger, R.A. Kenley, P.P. DeLuca, Gamma irradiation effects on molecular weight and in vitro degradation of poly(D,L lactide CO glycolide) microparticles., Pharm. Res. 12 (1995) 851 6. [171] A. Sabzevari, K. Adibkia, H. Hashemi, A. Hedayatfar, N. Mohsenzadeh, F. Atyabi, et al., Polymeric triamcinolone acetonide nanoparticles as a new alternative in the treatment of uveitis: In vitro and in vivo studies., Eur. J. Pharm. Biopharm. 84 (2013) 63 71. [172] H. Krause, A. Schwarz, P. Rohdewald, Polylactic acid nanoparticles, a colloidal drug deli very system for lipophilic drugs, Int. J. Pharm. 27 (1985) 145 155.

PAGE 164

150 [173] S.S. Lee, P. Hughes, A.D. Ross, M.R. Robinson, Biodegradable implants for sustained drug release in the eye., Pharm. Res. 27 (2010) 2043 53. [174] D.H. McGee, O. Dembinska, M.M. Grue bbel, Evaluation of triamcinolone acetonide following intravitreal injection in New Zealand white rabbits., Int. J. Toxicol. 24 (2005) 419 25. [175] P. Lim Soo, L. Luo, D. Maysinger, A. Eisenberg, Incorporation and Release of Hydrophobic Probes in Biocompa tible Polycaprolactone block poly(ethylene oxide) Micelles: Implications for Drug Delivery, Langmuir. 18 (2002) 9996 10004. [176] R.S. Kadam, P. Tyagi, H.F. Edelhauser, U.B. Kompella, Influence of choroidal neovascularization and biodegradable polymeric particle size on transscleral sustained delivery of triamcinolone acetonide., Int. J. Pharm. 434 (2012) 140 7. [177] A. Urtti, Challenges and obstacles of ocular pharmacokinetics and drug delivery., Adv. Drug Deliv. Rev. 58 (2006) 1131 5. [178] M. Madan, A Bajaj, S. Lewis, N. Udupa, J.A. Baig, In situ forming polymeric drug delivery systems., Indian J. Pharm. Sci. 71 (2009) 242 51. [179] D. Park, W. Wu, Y. Wang, A functionalizable reverse thermal gel based on a polyurethane/PEG block copolymer., Biomateria ls. 32 (2011) 777 86. [180] H. Tan, C.M. Ramirez, N. Miljkovic, H. Li, J.P. Rubin, K.G. Marra, Thermosensitive injectable hyaluronic acid hydrogel for adipose tissue engineering., Biomaterials. 30 (2009) 6844 53. [181] J. P. Chen, T. H. Cheng, Thermo respo nsive chitosan graft poly(N isopropylacrylamide) injectable hydrogel for cultivation of chondrocytes and meniscus cells., Macromol. Biosci. 6 (2006) 1026 39. [182] D. Schmaljohann, Thermo and pH responsive polymers in drug delivery., Adv. Drug Deliv. Rev. 58 (2006) 1655 70. [183] G. Dumortier, J.L. Grossiord, F. Agnely, J.C. Chaumeil, A review of poloxamer 407 pharmaceutical and pharmacological characteristics., Pharm. Res. 23 (2006) 2709 28. [184] Y. Gao, F. Ren, B. Ding, N. Sun, X. Liu, X. Ding, et al., A thermo sensitive PLGA PEG PLGA hydrogel for sustained release of docetaxel., J. Drug Target. 19 (2011) 516 27.

PAGE 165

151 [185] A.A. Ghahremankhani, F. Dorkoosh, R. Dinarvand, PLGA PEG PLGA tri block copolymers as in situ gel forming peptide delivery system: effect of formulation properties on peptide release., Pharm. Dev. Technol. 13 (2008) 49 55. [186] Isocyanates, (n.d.). [187] D. Yun, A. Famili, Y.M. Lee, P.M. Jenkins, C.R. Freed, D. Park, Biomimetic poly(serinol hexamethylene urea) for promotion of neurite outg rowth and guidance., J. Biomater. Sci. Polym. Ed. (2013) 1 16. [188] B.D. Ratner, K.W. Gladhill, T.A. Horbett, Analysis of in vitro enzymatic and oxidative degradation of polyurethanes., J. Biomed. Mater. Res. 22 (1988) 509 27. [189] V.J. Jijo, K.P. Sharma R. Mathew, S. Kamble, P.R. Rajamohanan, T.G. Ajithkumar, et al., Volume Transition of PNIPAM in a Nonionic Surfactant Hexagonal Mesophase, Macromolecules. 43 (2010) 4782 4790. [190] C.M. Jermak, J.T. Dellacroce, J. Heffez, G. a Peyman, Triamcinolone acet onide in ocular therapeutics., Surv. Ophthalmol. 52 (2007) 503 22. [191] J. E. Chang Lin, M. Attar, A. a Acheampong, M.R. Robinson, S.M. Whitcup, B.D. Kuppermann, et al., Pharmacokinetics and pharmacodynamics of a sustained release dexamethasone intravitre al implant., Invest. Ophthalmol. Vis. Sci. 52 (2011) 80 6. [192] S.H.S. Boddu, J. Jwala, M.R. Chowdhury, A.K. Mitra, In vitro evaluation of a targeted and sustained release system for retinoblastoma cells using Doxorubicin as a model drug., J. Ocul. Pharma col. Ther. 26 (2010) 459 68. [193] X. Z. Zhang, P. Jo Lewis, C. C. Chu, Fabrication and characterization of a smart drug delivery system: microsphere in hydrogel., Biomaterials. 26 (2005) 3299 309. [194] J.M. Barichello, M. Morishita, K. Takayama, T. Nagai Absorption of insulin from pluronic F 127 gels following subcutaneous administration in rats., Int. J. Pharm. 184 (1999) 189 98. [195] M. Gou, X. Li, M. Dai, C. Gong, X. Wang, Y. Xie, et al., A novel injectable local hydrophobic drug delivery system: Bio degradable nanoparticles in thermo sensitive hydrogel., Int. J. Pharm. 359 (2008) 228 33. [196] A. Bochot, E. Fattal, J.L. Grossiord, F. Puisieux, P. Couvreur, Characterization of a new ocular delivery system based on a dispersion of liposomes in a thermos ensitive gel, Int. J. Pharm. 162 (1998) 119 127.

PAGE 166

152 [197] C. Wischke, S.P. Schwendeman, Principles of encapsulating hydrophobic drugs in PLA/PLGA microparticles., Int. J. Pharm. 364 (2008) 298 327. [198] R. a Jain, The manufacturing techniques of various drug loaded biodegradable poly(lactide co glycolide) (PLGA) devices., Biomaterials. 21 (2000) 2475 90. [199] T.G. Park, Degradation of poly(lactic co glycolic acid) microspheres: effect of copolymer composition, Biomaterials. 16 (1995) 1123 30. [200] H. Tan, C .M. Ramirez, N. Miljkovic, H. Li, J.P. Rubin, K.G. Marra, Thermosensitive injectable hyaluronic acid hydrogel for adipose tissue engineering., Biomaterials. 30 (2009) 6844 53. [201] Z. Ma, D.M. Nelson, Y. Hong, W.R. Wagner, Thermally responsive injectable hydrogel incorporating methacrylate polylactide for hydrolytic lability., Biomacromolecules. 11 (2010) 1873 81. [202] E. Rosellini, C. Cristallini, G.D. Guerra, N. Barbani, P. Giusti, Synthesis and characterization of a novel PNIPAAm based copolymer with h ydrolysis dependent thermosensitivity., Biomed. Mater. 5 (2010) 035012. [203] Y. Xia, N.A.D. Burke, H.D.H. Stver, End Group Effect on the Thermal Response of Narrow Disperse Poly( N isopropylacrylamide) Prepared by Atom Transfer Radical Polymerization, M acromolecules. 39 (2006) 2275 2283. [204] X. Qiu, T. Koga, F. Tanaka, F.M. Winnik, New insights into the effects of molecular weight and end group on the temperature induced phase transition of poly(N isopropylacrylamide) in water, Sci. China Chem. 56 (201 2) 56 64. [205] S. L, B. Li, B. Ni, Z. Sun, M. Liu, Q. Wang, Thermoresponsive injectable hydrogel for three dimensional cell culture: chondroitin sulfate bioconjugated with poly(N isopropylacrylamide) synthesized by RAFT polymerization, Soft Matter. 7 (20 11) 10763. [206] A. Vaughan Thomas, S.J. Gilbert, V.C. Duance, Elevated Levels of Proteolytic Enzymes in the Aging Human Vitreous, Invest. Ophthalmol. Vis. Sci. 41 (2000) 3299 3304. [207] T. Ohshima, T. Iwasaki, Y. Maegawa, A. Yoshiyama, K. Mashima, Enzyme like chemoselective acylation of alcohols in the presence of amines catalyzed by a tetranuclear zinc cluster., J. Am. Chem. Soc. 130 (2008) 2944 5. [208] S. Yang, J. Zhu, Y. Lu, B. Liang, C. Yang, Body distribution of camptothecin solid lipid nanoparticle s after oral administration, Pharm. Res. 16 (1999) 751 757.

PAGE 167

153 [209] S.C. Yang, L.F. Lu, Y. Cai, J.B. Zhu, B.W. Liang, C.Z. Yang, Body distribution in mice of intravenously injected camptothecin solid lipid nanoparticles and targeting effect on brain., J. Con trol. Release. 59 (1999) 299 307. [210] A.M. de Campos, Y. Diebold, E.L.S. Carvalho, A. Snchez, M.J. Alonso, Chitosan nanoparticles as new ocular drug delivery systems: in vitro stability, in vivo fate, and cellular toxicity., Pharm. Res. 21 (2004) 803 10 [211] H. Gao, Z. Yang, S. Zhang, S. Cao, S. Shen, Z. Pang, et al., Ligand modified nanoparticles increases cell uptake, alters endocytosis and elevates glioma distribution and internalization., Sci. Rep. 3 (2013) 2534. [212] M. Grabe, Regulation of Organ elle Acidity, J. Gen. Physiol. 117 (2001) 329 344. [213] D.M. Maurice, Flow of water between aqueous and vitreous compartments in the rabbit eye., Am. J. Physiol. 252 (1987) F104 8. [214] A. Narasimhan, C. Sundarraj, Effect of choroidal blood perfusion and natural convection in vitreous humor during transpupillary thermotherapy (TTT)., Int. J. Numer. Method. Biomed. Eng. 29 (2013) 530 41. [215] J. Sebag, Age related changes in human vitreous structure, Graefes Arch. Clin. Exp. Ophthalmol. 225 (1987) 89 93. [216] E. Stefnsson, T. Loftsson, The Stokes Einstein equation and the physiological effects of vitreous surgery., Acta Ophthalmol. Scand. 84 (2006) 718 9. [217] J. Sebag, Ageing of the vitreous., Eye (Lond). 1 ( Pt 2) (1987) 254 62. [218] V. P. Ranta, E. Mannermaa, K. Lummepuro, A. Subrizi, A. Laukkanen, M. Antopolsky, et al., Barrier analysis of periocular drug delivery to the posterior segment., J. Control. Release. 148 (2010) 42 8. [219] S.H. Kim, K.G. Csaky, N.S. Wang, R.J. Lutz, Drug elimination kinet ics following subconjunctival injection using dynamic contrast enhanced magnetic resonance imaging., Pharm. Res. 25 (2008) 512 20. [220] P.W. Zhu, D.H. Napper, The longer time collapse kinetics of interfacial poly(N isopropylacrylamide) in water, J. Chem. Phys. 106 (1997) 6492. [221] V.P. Torchilin, PEG based micelles as carriers of contrast agents for different imaging modalities., Adv. Drug Deliv. Rev. 54 (2002) 235 52. [222] M.I. Dorrell, E. Aguilar, L. Scheppke, F.H. Barnett, M. Friedlander, Combination angiostatic therapy completely inhibits ocular and tumor angiogenesis., Proc. Natl. Acad. Sci. U. S. A. 104 (2007) 967 72.

PAGE 168

154 [223] J. Bradley, M. Ju, G.S. Robinson, Combination therapy for the treatment of ocular neovascularization., Angiogenesis. 10 (2007) 141 8. [224] M.B. Sherwood, E.R. Craven, C. Chou, H.B. DuBiner, A.L. Batoosingh, R.M. Schiffman, et al., Twice daily 0.2% brimonidine 0.5% timolol fixed combination therapy vs monotherapy with timolol or brimonidine in patients with glaucoma or ocular hyp ertension: a 12 month randomized trial., Arch. Ophthalmol. 124 (2006) 1230 8. [225] B. Testa, J.M. Mayer, Catalytic mechanisms of hydrolytic enzymes, in: Hydrolys. Drug Prodrug Metab., VHCA and Wiley VCH, Zurich, Switzerland, 2003: pp. 47 80. [226] M. Phil ipp, M.L. Bender, Kinetics of subtilisin and thiolsubtilisin., Mol. Cell. Biochem. 51 (1983) 5 32. [227] P. Carter, J.A. Wells, Dissecting the catalytic triad of a serine protease., Nature. 332 (1988) 564 8. [228] M. V Andersen, Changes in the vitreous bod y pH of pigs after retinal xenon photocoagulation., Acta Ophthalmol. 69 (1991) 193 9. [229] H.F. Edelhauser, J.L. Ubels, Cornea and sclera, in: P.L. Kaufman, A. Alm (Eds.), [230] S. Yeol Lee, J.W. Park, Y.T. Yoo, S.S. Im, Hydrolytic degradation behaviour and microstructural changes of poly(ester co amide)s, Polym. Degrad. Stab. 78 (2002) 63 71. [231] I. Dupret, C. David, a Daro, Biodegradation of polyester amides using a pure strain of micro organi sms or papain II. Polymer, Polym. Degrad. Stab. 67 (2000) 505 513. [232] R.J. McMahon, Affinity Precipitation, in: R.J. McMahon (Ed.), Avidin Biotin Interact. Methods Appl., Humana Press, Totowa, NJ, 2008: p. 49. [233] J.F. Arevalo, L. Wu, J.G. Sanchez, M. Maia, M.J. Saravia, C.F. Fernandez, et al., Intravitreal bevacizumab (Avastin) for proliferative diabetic retinopathy: 6 months follow up., Eye (Lond). 23 (2009) 117 23. [234] Z.F. Bashshur, A. Bazarbachi, A. Schakal, Z. a Haddad, C.P. El Haibi, B.N. Nour eddin, Intravitreal bevacizumab for the management of choroidal neovascularization in age related macular degeneration., Am. J. Ophthalmol. 142 (2006) 1 9.

PAGE 169

155 [235] P.J. Rosenfeld, R.M. Rich, G.A. Lalwani, Ranibizumab: Phase III clinical trial results., Ophth almol. Clin. North Am. 19 (2006) 361 72. [236] J. Emami, In vitro in vivo correlation: from theory to applications., J. Pharm. Pharm. Sci. 9 (2006) 169 89. [237] C.S. Paik Sung, T.W. Smith, N.H. Sung, Properties of Segmented Polyether Poly(urethaneureas) Based of 2,4 Toluene Diisocyanate. 2. Infrared and Mechanical Studies, Macromolecules. 13 (1980) 117 121. [238] J. L. Sormana, J.C. Meredith, High Throughput Discovery of omolecules. 37 (2004) 2186 2195. [239] S. Frokjaer, D.E. Otzen, Protein drug stability: a formulation challenge., Nat. Rev. Drug Discov. 4 (2005) 298 306. [240] V.H.L. Lee, Peptide and Protein Drug Delivery, Marcel Dekker, Inc., New York, NY, 1991. [241] T .A. Ciulla, M.H. Criswell, R.P. Danis, T.E. Hill, Intravitreal triamcinolone acetonide inhibits choroidal neovascularization in a laser treated rat model., Arch. Ophthalmol. 119 (2001) 399 404. [242] S.K. Li, M.R. Liddell, H. Wen, Effective electrophoretic mobilities and charges of anti VEGF proteins determined by capillary zone electrophoresis., J. Pharm. Biomed. Anal. 55 (2011) 603 7.

PAGE 170

156 APPENDIX A SYNTHETIC ROUTE OF PEG PHS PEG Figure A.1: The PEG PHS PEG block copolymer is synthesized by: 1) synthesi s of the PHS polyurethane middle block; 2) end capping of PHS with reactive isocyanates; and 3) conjugation of mPEG (550 Da) to both ends via the isocyanate end groups.

PAGE 171

157 APPENDIX B UV SPECTROMETRIC STANDA R D CURVE FOR TRIAMCINOLONE ACETONIDE CONCENTRATION DE TERMINATION Figure B.1: This standard curve, constructed by measuring the UV absorbance (237 nm) of several solutions of known TA concentration, was used throughout this wo rk for TA concentration calculations. The standard curve showed strong linearity (R 2 =0.9994) and was usable to a minimum concentration of 0.1 g/mL.

PAGE 172

158 APPENDIX C SYNTHETIC ROUTE OF PSHU NIPAAM Figure C .1: The PSHU NIPAAm graft copolymer is synthesized by: 1) synthesis of the poly(urea urethane) PSHU; 2) removal of Boc groups at each repeating unit to expose primary amine functionaliti es; and 3) conjugation of PNIPAAm through carbodiimide linking.

PAGE 173

159 APPENDIX D PSHU GPC ANALYSIS Figure D .1: GPC analysis of PSHU indicated a relatively narrow MW distribution with M n : 9 .73 k Da; M w : 12 .5 k Da; M z : 15.7 k Da; and M w /M m : 1. 29 indicating a we ll designed synthetic route.

PAGE 174

160 APPENDIX E PSHU MONOMER RATIO DETERMINATION Table E.1 : 1 H NMR peaks and integrated areas used in PSHU monomer ratio calculations Peak ppm Assignment Number of protons Integral Value Integral/Proton 11.1 Urea, K 2 1.00 0. 5000 4.00 N Boc serinol, F+G 5 19.54 3.907 3.10 HDI, D+E 4 31.40 7.849 1.25 HDI, A 4 34.06 8.511 Based on these data, the typical PSHU molecule can be taken as containing 22.5 units of N Boc serinol, 47.8 of HDI and 2.92 of urea (using PSHU M w = 12.5 k Da). Figure E.1: 1 H NMR spectra of PSHU with identified peaks and associated integral area values as used for m onomer ratio calculations.

PAGE 175

161 APPENDIX F PSHU NIPAAM MINIMUM GELLING CONCENTRATION DETERMINATION Table F .1 : Qualitative assessment of PSHU NIPAAm gelation indicated a minimum gelling concentration of 1.66 wt% and concentrations near 5 10 wt% possessing a good balance of viscosit y and gelling time. Concentration (wt %) Qualitative viscosity at 23 C Stable Gel Formation at 37 C Approximate Gelling Time (s) 20 Very high Yes 5 10 Slightly above water Yes 10 5.0 Yes 20 3.3 Yes 40 2.5 Yes 50 2.0 Ye s 50 1.66 Yes 60 1.43 No 1.25 No

PAGE 176

162 APPENDIX G MODELING OF TA RELEASE FROM PSHU NIPAAM GELS TA release kinetics from PSHU NIPAAm gels was fit against a first order model as shown in Figure G .1. The fit to first order coeffic ients are shown in Table G Table G .1 : Fit to first order parameters for the various TA concentration release profiles. TA Concentration (wt %) Fit to first order (R 2 ) 5.0 0.951 10 0.967 15 0.989 20 0.992 Figure G.1: Release of TA from PSHU NIPAAm gels followed closely to first order kinetics for all loading concentrations.

PAGE 177

163 APPENDIX H RELEASE OF RANIBIZUMAB FROM PSHU NIPAAM GELS In order to assess the ability of PSHU NIPAAm gels to sustain delivery of a large molecule drug, the monoclonal antibody fragment ranibizumab was employed. In order to maximize the potential of the gel to encapsulate ranibizumab and achieve a long term delivery target, PSHU NIPAAm copolymers were synthesized with varying levels of de protected primary amines. These positively charged amines were hypothesized to permit electrostatic interact ions with the negatively charged antibody [242] and improve drug release kinetics. In order to accomplish this, a procedure was developed that could selectively remove a fraction of Boc protective gr oups from PSHU (in contrast to the typical deprotection routine which removes all Boc groups), as shown in Figure H .1. Boc group deprotection was assessed by 1 H NMR analysis by monitoring the Boc methyl peak at 1.34 ppm. After partial deprotection PSHU with deprotection ratios of 25, 50, 75 and 100% were selected and conjugated with the same amount of PNIPAAm COOH to achieve a fixed 25% grafting ratio. Ranibizumab was then loaded at 10 wt% in 5 wt% gels and release was assessed using a similar exp erimental protocol as used for TA in vitro release experiments, but measuring ranibizumab concentrations by UV spectroscopy at 280 nm. Results of this in vitro testing are shown in Figure H .2 and indicate that all gels released their entire loading of rani bizumab within 24 hours, independent of the deprotection ratio. Figure H.1: Selective Boc group removal was achieved by varying the amount of TFA in the deprotection re action per PSHU content. This reaction permitted control over the amount of primary amines exposed on PSHU molecules.

PAGE 178

164 Figure H.2: PSHU NIPAAm gels with varying deprotection levels (hypothetically modulating the extent of electrostatic interactions between the gel and ranibizumab) all released their entire ranibizumab load within 24 hours.