Therapeutic Contact Lenses for Comfort Molecules Except where reference is made to the work of others, the work described in this thesis is my own or was done in collaboration with my advisory committee. This thesis does include proprietary or classi ed information. Maryam Ali Certi cate of Approval: Christopher B. Roberts Uthlaut Professor Chemical Engineering Mark E. Byrne, Chair Assistant Professor Chemical Engineering Ram B. Gupta Alumni Professor Chemical Engineering Jong Wook Hong Assistant Professor Mechanical Engineering George T. Flowers Interim Dean Graduate School Therapeutic Contact Lenses for Comfort Molecules Maryam Ali A Thesis Submitted to the Graduate Faculty of Auburn University in Partial Ful llment of the Requirements for the Degree of Master of Science Auburn, Alabama December 17, 2007 Therapeutic Contact Lenses for Comfort Molecules Maryam Ali Permission is granted to Auburn University to make copies of this thesis after August 1, 2012 at its discretion, upon the request of individuals or institutions and at their expense. The author maintains copyright and retains all publication rights. Signature of Author Date of Graduation iii Vita The author Maryam Ali was born in Rawalpindi, Pakistan on 25 December 1981. She attended the California Institute of Technology at Pasadena, California from 2001 to 2005 and received a Bachelor of Science in Chemical Engineering in 2005. She began work toward a Master of Science in Chemical Engineering at Auburn University at Auburn, Alabama in the Fall of 2005. She is joining The University of Texas at Austin in the Fall of 2007 for doctoral work in Biomedical Engineering. iv Thesis Abstract Therapeutic Contact Lenses for Comfort Molecules Maryam Ali Master of Science, December 17, 2007 (B.S., California Institute of Technology, 2005) 143 Typed Pages Directed by Mark E. Byrne Dry eye syndrome a ects nearly 15% of the population and can cause extreme dis- comfort that interferes with quality of life. Current treatment options include delivering comfort agents such as hyaluronic acid (HA) to the eye via eye drops, but low bioavail- ability continues to be a barrier to e ective treatment. We have designed a therapeu- tic hydrogel contact lens that can deliver HA to the eye at a therapeutic rate. Control over the release characteristics is improved through biomimetic imprinting, as functional monomers such as acrylamide,N-vinyl pyrrolidone and (diethylamino)ethyl methacrylate are added to the hydrogel structure. The di usion coe cients of hyaluronic acid, a long chain molecule, through the hydrogel can be controlled by varying the number of added functional monomers that interact with the drug molecule through memory sites. There is an inverse correlation between the total %-by-mass of functional monomers added to the hydrogel and the di usion coe cient. Increasing the variety of functional monomers lowered the di usion coe cient 1.5 times more than including a single type of functional monomer, and 1.6 times more than Nel lcon without added monomers. By optimizing the v functional monomer content of the hydrogel, we can deliver hyaluronic acid to the eye at a constant therapeutic rate of approximately 6 g per hour for 24 hours. vi Acknowledgments I would like to thank my advisor Dr. Mark E. Byrne for his guidance and mentoring during my graduate career. I also wish to thank my Advisory Committee members Dr. Ram B. Gupta, Dr. Jong Wook Hong, and Dr. Christopher B. Roberts for their encouragement. I am grateful to a number of people for technical assistance and access to equipment; in particular Dr. William Ravis, Dr. Y.Y. Lee, and Dr. Maria Auad. Special thanks to Shin Horikawa for his collaboration on the micro uidic device project, and to Dr. Mirna Mosiewicki De Ruiz for her incomparable help with the dynamic mechanical analyzer. My fellow lab mates have been a tremendous source of technical and moral support and to them I extend my gratitude. I also wish to thank my friends, both near and far, for being available with advice and fresh perspectives. Financial support for this work was provided by CIBA Vision Inc., Duluth, GA. I?d like to thank Dr. Lynn Winterton and Dr. John Pruitt for their support of this work. My greatest thanks are to my family for their encouragement throughout these past years. This work was made possible by the love and support of my parents. I also wish to thank my brother for his friendship and advice. vii Style manual or journal used Chicago Manual of Style together with the style known as \aums". Computer software used The document preparation package TEX, speci cally LATEX, and the integrated development environment TEXnic Center, together with the Auburn University thesis style- le aums.sty. The bibliography was prepared with the citation management software JabRef together with the Chicago Manual of Style BibTex style le ChicagoReedWeb written by Sarah Sugarman and distributed by Reed College. viii Table of Contents List of Figures xi List of Tables xiii 1 Introduction 1 2 Objective 4 3 Ocular Drug Delivery 7 3.1 Ocular Diseases and Impact . . . . . . . . . . . . . . . . . . . . . . . . . . 9 3.2 Barriers to Ocular Drug Delivery . . . . . . . . . . . . . . . . . . . . . . . 10 3.3 Strategies to Overcome Drug Removal at the Ocular Surface . . . . . . . . 17 3.4 Strategies for Permeation Enhancement through Ocular Membranes . . . . 26 3.5 Strategies to Delivery Drugs to the Posterior of the Eye . . . . . . . . . . . 29 4 Hydrogels and Imprinting 32 4.1 Di usion through a hydrogel . . . . . . . . . . . . . . . . . . . . . . . . . . 33 4.2 Theoretical model for di usion . . . . . . . . . . . . . . . . . . . . . . . . . 37 4.3 Equilibrium swelling theory . . . . . . . . . . . . . . . . . . . . . . . . . . 40 4.4 Rubber elasticity theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 4.5 Biomimetic Imprinting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 5 The Tear Film and Dry Eyes 44 5.1 The Tear Film . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 5.2 Etiology of Dry Eyes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 5.3 Treatment with Hyaluronic Acid . . . . . . . . . . . . . . . . . . . . . . . . 49 6 Lenses for Delivery of Comfort Molecules 52 6.1 Nel lcon A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 6.2 Hyaluronic Acid Binding Moieties . . . . . . . . . . . . . . . . . . . . . . . 54 6.3 Methods and Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.3.1 Synthesis of hydrogels . . . . . . . . . . . . . . . . . . . . . . . . . 58 6.3.2 Dynamic release studies . . . . . . . . . . . . . . . . . . . . . . . . 59 6.3.3 Heat stability studies . . . . . . . . . . . . . . . . . . . . . . . . . . 60 6.3.4 Tensile strength studies . . . . . . . . . . . . . . . . . . . . . . . . . 61 6.4 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61 6.4.1 Molecular transport and di usion coe cients . . . . . . . . . . . . . 61 6.4.2 E ects of heat sterilization . . . . . . . . . . . . . . . . . . . . . . . 77 6.4.3 Structural analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 ix 7 Drug Release Kinetics Under Physiological Flow 87 7.1 Micro uidic Platforms for Evaluating Drug Delivery Devices . . . . . . . . 88 7.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 7.2.1 Synthesis of molecularly imprinted hydrogel networks . . . . . . . . 89 7.2.2 Dynamic Therapeutic Release Studies . . . . . . . . . . . . . . . . . 90 7.2.3 Micro uidic Chip Design and Fabrication . . . . . . . . . . . . . . . 91 7.2.4 Dynamic Weight/Volume Swelling Studies and Partition Coe cients 92 7.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 8 Conclusion 105 Appendices 119 A Dynamic release of HA in various concentrations 120 B Dynamic release of HA of various sizes 122 C Tensile testing of Hydrogels 124 x List of Figures 3.1 Anatomy of eye surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 3.2 Anatomy of anterior chamber . . . . . . . . . . . . . . . . . . . . . . . . . 14 3.3 Anatomy of the retina . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 3.4 Illustration of drug concentration pro le in tear lm with application of eye drops . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 4.1 Illustration of mesh size within a hydrogel . . . . . . . . . . . . . . . . . . 35 4.2 Illustration of reptation model for transport through hydrogel . . . . . . . 36 5.1 Tear lm and related structures . . . . . . . . . . . . . . . . . . . . . . . . 46 5.2 Illustration of hyaluronic acid structure . . . . . . . . . . . . . . . . . . . . 50 6.1 Synthesis of Nel lcon A macromer from PVA . . . . . . . . . . . . . . . . 55 6.2 Comparison of functional groups on amino acids and acrylate monomers . 57 6.3 Cumulative release of HA from Nel lcon hydrogels . . . . . . . . . . . . . . 63 6.4 Cumulative release of HA from Nel lcon hydrogels with di erent %-by-mass of functional monomers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 6.5 Di usion coe cients versus %-by-mass functional monomer content . . . . 68 6.6 Cumulative release of HA from Nel lcon hydrogels with di erent proportions of functional monomers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70 6.7 Fractional release of HA from Nel lcon hydrogels with di erent proportions of functional monomers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 6.8 Cumulative release of HA from Nel lcon hydrogels with the same %-by-mass of DEAEMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73 6.9 24 hour release of HA from Nel lcon gels versus proportion of DEAEMA . 74 xi 6.10 Di usion coe cients versus proportion of DEAEMA . . . . . . . . . . . . . 75 6.11 Comparison of stability of HA solutions under heat sterilization conditions 78 6.12 Cumulative release of HA from Nel lcon hydrogels before and after heat- sterilization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79 6.13 Comparison of polymer volume fractions for various hydrogels . . . . . . . 81 6.14 Di usion coe cients versus polymer volume fraction for Nel lcon hydrogels with di erent %-by-mass of functional monomers . . . . . . . . . . . . . . 83 7.1 Illustration of micro uidic device for drug delivery evaluation . . . . . . . . 93 7.2 Cumulative release of ketotifen fumarate from HEMA hydrogels . . . . . . 96 7.3 Fractional release of ketotifen from various hydrogel lenses under in nite sink conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 7.4 Di usion coe cients versus polymer volume fraction in HEMA lenses for ketotifen release . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99 7.5 Fractional release of ketotifen under physiological ow conditions . . . . . . 101 A.1 Cumulative release of di erent concentrations of HA from Nel lcon hydrogels 121 B.1 Fractional release of various sizes of HA from Nel lcon hydrogels . . . . . . 123 C.1 Tensile test Nel lcon - sample 1 . . . . . . . . . . . . . . . . . . . . . . . . 125 C.2 Tensile test Nel lcon - sample 2 . . . . . . . . . . . . . . . . . . . . . . . . 125 C.3 Tensile test Nel lcon with HA - sample 1 . . . . . . . . . . . . . . . . . . . 126 C.4 Tensile test Nel lcon with HA - sample 2 . . . . . . . . . . . . . . . . . . . 126 C.5 Tensile test Nel lcon with HA - sample 3 . . . . . . . . . . . . . . . . . . . 127 C.6 Tensile test Nel lcon with functional monomers - sample 1 . . . . . . . . . 127 C.7 Tensile test Nel lcon with functional monomers - sample 2 . . . . . . . . . 128 C.8 Tensile test Nel lcon with functional monomers - sample 3 . . . . . . . . . 128 C.9 Tensile test Nel lcon with HA and functional monomers - sample 1 . . . . 129 C.10 Tensile test Nel lcon with HA and functional monomers - sample 2 . . . . 129 C.11 Tensile test Nel lcon with HA and functional monomers - sample 3 . . . . 130 xii List of Tables 3.1 Major diseases of the eye . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 6.1 Di usion and release order of HA from Nel lcon hydrogels with varying functional monomer amounts, all in [1:1:2] ratio . . . . . . . . . . . . . . . 66 6.2 Functional monomer content of hydrogels . . . . . . . . . . . . . . . . . . . 69 6.3 Di usion and release order of HA from Nel lcon hydrogels with varying functional monomer proportions . . . . . . . . . . . . . . . . . . . . . . . . 76 6.4 Equilibrium swelling parameters . . . . . . . . . . . . . . . . . . . . . . . . 82 6.5 Tensile parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 6.6 Mesh sizes of hydrogels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 7.1 Varying ketotifen release rates from AA-AM-NVP lenses under in nite sink conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 7.2 Summary of ketotifen di usion coe cients, orders of release and swelling data100 A.1 Di usion and release order of Nel lcon hydrogel with 6.5 and 40 mg HA/g Nel lcon . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120 xiii Chapter 1 Introduction Dry eye syndrome a ects nearly 50 million people in the United States to varying degrees. The disorder occurs when a patients? eyes are not adequately hydrated by the tears they produce, exposing the epithelia of their cornea and conjunctiva to desiccation. While ocular dryness is not immediately threatening to vision, the desiccation of the epithelia triggers a number of uncomfortable symptoms such as itchiness of the eye, a sensation of grittiness, light sensitivity, excessive watering, blurred vision and in ammation, all of which can signi cantly a ect the quality of a patient?s life. The customary treatment for dry eyes is the application of arti cial tears and comfort agents that hydrate the epithelia. Many arti cial tears and comfort agents increase the viscosity of the tears in the eye, preventing their drainage and evaporation, and increasing the moisture of the ocular surface. Arti cial tears are commercially available in the form of eye drops that need to be applied every 2 to 4 hours for relief. Unfortunately, eye drops have very low bioavailability in the eye because of the natural turnover of tears. Tear turnover is a barrier to drug delivery, not just for dry eye but for all therapeutics delivered to the surface of the eye, and 95% of the volume of eye drops can get ushed from the eye without reaching their target tissues. We need to develop a drug delivery technique that can deliver comfort molecules to the eye at a slow constant rate equivalent to the rate at which the eye needs them. Therapeutic contact lenses are devices that behave optically like conventional contact lenses but deliver needed medication at an appropriate rate. They contain drug that is released to the eye 1 slowly so that the fraction of drug reaching its target is increased. Therapeutic contact lenses also eliminate the need to apply eye drops multiple times a day, as a single lens delivers medication over an extended time period, from several hours to days. We have developed therapeutic contact lenses that deliver a comfort agent, hyaluronic acid (HA), to the eye at a rate comparable to the dosage regimen of HA eye drops. To control the release rate of the HA from the lens, we turn to a process called biomimetic imprinting that enhances the a nity of the HA for the hydrogel by introducing \memory sites" within the hydrogel that behave similarly to binding sites in HA-binding protein, a protein found naturally in the body. The increased a nity of the HA for the hydrogel slow down its di usion from the lens, leading to a more linear release rate during the rst 24 hours of contact lens wear. While biomimetic imprinting has previously been used to design contact lenses for delivering ocular allergy medication, the delivered drugs were all relatively small molecules, under 1000 Daltons in size. In contrast HA is a long chain molecule, a polysaccharide with a size of around 1 million Daltons. The application of the biomimetic imprinting process to HA thus introduces new challenges. A small molecule di using through the hydrogel material of a contact lens passes through the material with relatively little hindrance. A longer molecule is restrained at multiple points, as the tail end of the chain navigates through a path that the head end of the chain has already passed through. The increased hindrance lowers the HA?s rate of transport through the lens hydrogel material. In this study we explore the extent to which biomimetic imprinting can in uence the di usion of a molecule that is so constrained. Evaluation of the release characteristics of drug delivery devices, such as therapeutic contact lenses, is usually conducted in vitro before progressing to in vivo studies. Often 2 the device is seen to behave di erently in each study. Conventional in vitro evaluation of ocular drug delivery devices immerses the device in an arti cial lacrimal uid environment, with the uid regularly replaced so that the drug concentration in the uid outside the device is negligible. In such conditions, the driving force of the concentration ux is the concentration of drug inside the device and the nearly zero drug concentration in the bulk uid. Within the eye, the volumes of tear uid are very small and despite the tear turnover rate, there is always a non-negligible concentration of drug in the uid surrounding the device. In order to model the ow conditions in the eye and develop an improved technique for evaluation of drug delivery devices in the eye, we developed a micro uidic device that can contain a drug delivery device such as a therapeutic contact lens in a chamber, and ow arti cial lacrimal uid through the device at a rate equal to the ow rate of tears in the eye. Evaluation of a device in an environment that mimics the ocular ow conditions can provide a better understanding of how it will release drug in the eye, and enable researchers to optimize their design prior to conducting in vivo studies. 3 Chapter 2 Objective The objectives of this research were as follows: (1) to develop a therapeutic hydrogel contact lens for the controlled release of a long-chain drug molecule, speci cally hyaluronic acid, to the eye for the amelioration of dry eye related ocular discomfort induced by con- tact lens wear, and (2) to create a micro uidic device for the in vitro characterization of therapeutic contact lenses under physiological ow conditions. The speci c aims included: (1) the analysis of biological binding proteins in litera- ture to identify amino acids, and functionally similar acrylate monomers, with a poten- tial for a nity with the polysaccharide drug hyaluronic acid (aka HA or hyaluronan or hyaluronate); (2) the design and synthesis of a polymer hydrogel material with incorpo- rated moieties capable of chemical interactions with HA at multiple points; (3) the control of the in vitro release characteristics of HA, such as the di usion coe cient and cumu- lative released mass, by varying the number and diversity of functional monomers in the hydrogel; (4) the analysis of the structure of the hydrogel material; (5) the development of a micro uidic device used to release drug from a hydrogel lens in vitro under physiological conditions, and (6) the analysis of release of the drug ketotifen fumarate from a therapeutic lens placed in the micro uidic device. We hypothesized that increasing the diversity of functional monomers and increasing the chemical similarity of the memory sites with the biological binding site on HA-binding protein (HABP) would enhance the binding between the HA and the hydrogel and de- crease the di usion coe cient. Additionally, increasing the concentration of functional 4 monomers would increase the number of points of interaction between the hydrogel and the HA molecules, and consequently slow the di usion of the HA from the hydrogel. The studies undertaken in this project exemplify how biologically-inspired chemical modi cation of a hydrogel can tailor the release rate of long-chain molecules in a lacrimal environment. We have developed a polymeric hydrogel material synthesized with HA that can release the drug in the presence of arti cial lacrimal solution at physiological tempera- ture. The addition of biomimetically selected functional monomers has allowed us to tailor the delivery of HA to the eye at a constant rate in therapeutic amounts over the course of 24 hours. These modi cations do not change the mesh size or mechanical properties of the hydrogel signi cantly. The HA in the contact lens remains stable and delivers the drug at desired levels after heat sterilization treatment. In a parallel project, we hypothesized that placing a therapeutic hydrogel lens in an in vitro ocular ow environment would decrease the demonstrated rate of drug release. The conventional protocol for drug release measurement immerses the delivery device in an environment with well mixed conditions and signi cant volumes, where the drug concen- tration is negligible. Ocular conditions in contrast have small volumes of slow- owing uid. These may accumulate the released drug. The increased drug concentration in the uid surrounding the lens would lower the driving force of the drug transport. The release data collected from such a device can provide us with a better understanding of how therapeutic lenses will behave in the eye. The studies for this project demonstrate that conducting release studies on ocular drug delivery devices in ocular ow conditions extends their duration of release. Additionally, the release rate is shown to be zero-order, or concentration independent. In contrast, conventional release studies had indicated a Fickian release pro le. 5 Future developments related to the rst project will prepare the drug delivering contact lens for commercial manufacture and distribution, and ultimately provide relief to the 5 million people who seek treatment for dry eye syndrome,92 and the millions more who su er from ocular discomfort. Developments in the second project will involve the synthesis of the micro uidic device using materials such as silica glass to ensure a more robust design for future studies on the e ects of ow. Further re nements in the design can yield a system for superior assessment of drug release from ocular drug delivery devices. 6 Chapter 3 Ocular Drug Delivery Recent advances in genetics, neuroscience, and molecular biology are leading to un- precedented discovery of mechanisms underlying ocular disease and new therapeutics for treatments that increase quality of life.73 Equally important is the optimal delivery of ther- apeutics, which has been the subject of intense research and development that is continually pushing the boundaries currently delineated by traditional topical formulations. Topical formulations such as solutions and suspensions in the form of eye drops have been in use for centuries4 and are still the most common treatment approach used today.80 E ective drug administration rests on getting a su cient amount or concentration of drug to the site of action within a given time period. For the eye, the majority of drugs are administered topically and the rest are administered in a systemic manner. For anterior or front of the eye therapy, the majority of treatments require non-invasive, topically applied drugs. For posterior or back of the eye therapy, drugs are typically administered via systemic routes and also by intravitreal injection.72 All of these treatments have their own limitations that primarily involve the body?s natural mechanisms and barriers that impede the transport of molecules. However, it should be distinctly understood that quality of vision, being crucial to our evolutionary survival, translates to the eye doing an excellent job preventing foreign materials from crossing its barriers. Therefore any drug delivery mechanism we use needs to e ectively deliver medication without permanently weakening these protective barriers. 7 The rates at which drugs pass through or interact with the di erent ocular barriers is of signi cant interest to the eld of ocular drug administration. Systemic delivery of drugs to the eye is impeded by the blood-ocular barriers, which prevent transport from the blood to the eye interior.88 These barriers along with liver metabolism signi cantly limit the bioavailability of orally or intravenously administered drug. Drugs delivered topically to the ocular surface also face reduced drug transport, which is in uenced by lacrimation and tear turnover, nasolacrimal drainage, spillage from the eye, metabolic degradation, and non-productive adsorption/absorption. These protective mechanisms lead to poor drug absorption on the surface of the eye despite it being a very accessible organ to treat topically. As a result, ocular bioavailability of drugs applied topically to the eye is typically very poor with less than 1-7% of the applied drug being absorbed, and the rest entering the systemic circulation.47;120 From a clinical perspective, the challenge is to provide medication conveniently, non- invasively, and in therapeutically signi cant concentrations for long times with minimal transfer of drug to the systemic circulation - providing topical, targeted therapy to the eye. This can be best achieved by (i) extending the residence time or duration of drugs on the surface of eye and/or by (ii) increasing drug transport through ocular barriers such as the cornea, sclera, and conjunctiva. The concentration of drug reaching the desired site of action can be signi cantly improved by altering the kinetics of drug administration, removal, and/or absorption. 8 3.1 Ocular Diseases and Impact The US prescription ophthalmic drug market is valued at approximately $4.5 billion and growing at a 7% average annual growth rate.35 This is due to a number of factors such as an increase in the overall aging population and subsequent eye issues encountered, an increase in the incidence of disease and needed disease prevention due to an increase in surgical procedures and contact lens use, and an increase in the number of medicines prescribed since optometrists in most states can now directly prescribe most medications. 94 Highly debilitating diseases such as cataracts, retinal degenerative maladies such as macular degeneration and retinitis pigmentosa, diabetic retinopathy, glaucoma, and uveitis a ect a large number of the population and have signi cant economic impact.44 While the aforementioned diseases can lead to partial and complete blindness, other diseases such as dry eye, bacterial conjunctivitis, ocular allergy, and ocular in ammation typically do not lead to complete loss of vision, but they signi cantly a ect quality of life for a larger number of people and also have a considerable economic impact. Posterior drug candidates with a smaller target market have primarily been the work of specialty pharmaceutical companies with subsequent licensing, co-development, and manufacturing from large pharmaceutical companies. In certain respects, many reports highlight that this has led to a lack of ocular drug therapies especially for posterior eye disease.101 Table 3.1 outlines the major diseases of the eye, the number of US population a ected, as well as the treatment location within the eye. 9 Disease Ocular Location A ected US Population A ected or 2007 Ocular Market Refractive Error Anterior 75 Million (25% of general population) Cataracts Anterior 20.5 Million (54% of people over age 65) AMD Posterior 1.7 Million people over age 50 Retinal Degen- eration Posterior 5.3 Million (2.5% of people age 18 and older) Diabetic Macu- lar Edema Posterior 500,000 Diabetic Retinopathy Posterior 4.1 Million Glaucoma Anterior 2.2 Million (2% of people age 40 and older) Uveitis Anterior/ Posterior 346,000 Dry Eye Anterior 50 Million (15% of general population) Infection or Risk of Infection Anterior $740 Million Allergy Anterior 75 Million (25% of general population); $630 Million In ammation Anterior $500 Million Table 3.1: Major diseases of the eye 3.2 Barriers to Ocular Drug Delivery The eye is pharmacokinetically isolated from the rest of the anatomy and the site of drug action ultimately determines the strategy for successful topical delivery. Tear drainage and to some extent the absorption through the eyelids lead to less drug on the surface of the eye available to transport through ocular barriers such as the cornea, conjunctiva, and sclera. The cornea is a transparent, dome-shaped structure covering the front of the eye. It is contiguous with the conjunctiva, a delicate mucous membrane with a highly vascularized stroma that covers the sclera (the tough, opaque, white of the eye) and lines the inner surface of the eyelids. Figure 3.1 presents the anatomy of the ocular epithelia. The human eye surface holds a tear volume that ranges from 7.0-30.0 L with a tear turnover rate of 0.5-2.2 L /min.50;120 This translates to a therapeutically relevant drug residence time of under 5 minutes with complete exchange of tear volume in approximately 10 11 Figure 3.1: Anatomy of eye surface [A] is a lipophilic drug which cannot easily penetrate the tear film and is washed away. [B] is a lipophilic drug in the central cavity of a cyclodextrin molecule. The cyclodextrin solubilizes in the tear film and reaches the ocular epithelium. The lipophilic drug partitions out of the cyclodextrin and into lipid membrane of the epithelium. [C] is a hydrophilic drug that solubilizes in the tear film and reaches the epithelium. It cannot cross the epithelium transcellularly (because of lipid membrane) or paracellularly (because of tight junctions), and eventually washes away from the eye surface. [D] is a hydrophilic pro-drug which penetrates the epithelium transcellularly with the aid of a membrane transporter. Once in the ocular tissue, it is converted into the drug by enzymes. The corneal and conjunctival epithelia are contiguous and contain several layers of cells (not shown), the outermost layer features microvilli that interact with tear film mucins. Drugs that penetrate the epithelia can easily move between ocular tissues such as the corneal and conjunctival stroma, the sclera beneath the conjunctiva, the vascularized choroid, and the leaky endothelium. From there they can diffuse into the anterior chamber or laterally through the sclera to the eye posterior. ? ? 14 minutes assuming normal lacrimation and blinking rates since blinking aids in con- taminant removal and promotes a well-mixed tear uid. If the topical medication or the mechanical forces of the instilled drop irritate the eye, lacrimal secretion will increase and further dilute the dosage. The ocular tear system and the tear lm play a crucial role in maintaining an optically clear surface in the front of the eye. The bulk of the tear uid is a 6-7 m thick aqueous layer64 with dissolved oxygen, nutrients and proteins. The in- terface between this layer and the air comprises a 0.1 m thick layer of lipids that limits evaporative loss of the aqueous lm.121 Between the aqueous layer and the ocular epithelia (which are hydrophobic) exists a layer of hydrophilic mucins that maintain the integrity of the surface by trapping and removing foreign matter, and lubricate against the shearing force applied by blinking.65 The movement of uid in the eye depends on the ow of the aqueous phase, which is secreted by the lacrimal glands above the eye, spread over the eye surface through surface tension and blinking, and drains out of the eye through the lacrimal puncta with the aid of a pumping mechanism.32 Up to 95% of topically applied drug can get washed away from the eye surface within minutes.19 The rate determining barriers for transport through the cornea to the aqueous humor are the corneal epithelium, the stroma, and the endothelium. When a drug reaches the corneal or conjunctival epithelium, it needs to nd a path through the layers of cells. For a drug to take a transcellular path (i.e., through the cells), it needs to enter the cell ei- ther by facilitated transport or by di usion through the lipid bilayer. The former requires particular chemical interactions with transporters native to the cells while the latter re- quires lipophilicity and depends on the drug solubility, degree of ionization, and size, and on the cell membrane thickness. Both depend on the drug concentration gradient, and the e ective area. Lipophilic drugs can transport quickly through the transcellular pathway 12 but hydrophilic drugs, especially larger than 20,000 Daltons, have di culty.71 The para- cellular path (i.e., around the cells) is impeded by the presence of tight junctions. Stromal transport is approximately equivalent for all ocular drugs and relatively independent of drug partitioning, and the endothelium is only one cell layer in thickness with transport depending on partitioning behavior as in the epithelium. We refer the reader to an excellent review that compiles ocular tissue permeability measurements.111 Hydrophilic drugs have been demonstrated to transport through the outer layers of the conjunctiva more quickly than through the corneal epithelium. After conjunctival absorption, transport may include lateral di usion into the corneal stroma and, to a limited extent, arterial vessel uptake.81 Drug may also be secreted back to the surface via e ux proteins in the epithelia.71;88 After passing through the ocular barriers, the drug reaches the anterior chamber be- tween the cornea and the lens. Typically 3% of the instilled drug reaches this point.120 The aqueous humor is a clear ltrate of blood that is produced by the ciliary body, circu- lates through the anterior chamber at 1% per minute,112 and drains out via the trebecular meshwork. It delivers nutrients and anti-oxidants to the cornea and lens without interfering with visual clarity. The aqueous humor poses an additional impediment to topical drugs targeting the posterior of the eye. Any drug that di uses through the cornea will be at risk of dilution and ushing away via the aqueous humor. By this point, drugs delivered via the corneal route can get diluted to the point of ine cacy even before moving into the posterior segment. Figure 3.2 demonstrates the movement of aqueous humor through the anterior segment. Drug that reaches the sclera has another pathway at its disposal. It may di use laterally through the highly permeable sclera and reach the posterior segment of the eye.74 13 14 Figure 3.2: Anatomy of anterior chamber Hydrophilic and lipophilic drugs both pass from the permeable stroma and sclera into the anterior segment the choroid, and the posterior segment. They also penetrate to the ciliary body, transfer to the secreted aqueous humor and circulate around the anterior and posterior chamber before draining away through the trebecular meshwork. The tissues here support the retina and encase the vitreous humor, a highly viscous uid. Inside the sclera is a layer of vascularized tissue called the choroid and inside that is the retina, the tissue on which light falls to produce images. The retina consists of several layers of tissue which, relating to their importance to drug delivery, can be classi ed as neural tissue and the retinal pigmented epithelium (RPE). The choroid nourishes the outermost layers of the retina, including the outer one third of the neural tissue and the RPE. Bruch?s membrane is the innermost layer of the choroid. It also provides the basement membrane of the RPE. The RPE is a signi cant barrier to the transport of drug from the sclera (or systemically delivered drugs from the choroid) into the neural tissue and the vitreous humor. Another barrier is the endothelial cells of the retinal capillaries that are located among the retinal neural tissue. They prevent drugs from the circulatory system reaching the neural retina. The RPE and endothelial cells also bear e ux proteins that actively remove drugs from the retina. Together the RPE and the retinal endothelial cells form the blood-retinal barrier.88;134 Figure 3.3 shows the anatomy. The common alternatives to reach the posterior segment involve injecting the drug or inserting a drug delivery device into the vitreal cavity of the eye, or using a periocular route of delivery - applying the drug, carrier, or device within the eye surface barriers and relying on transscleral transport. Non-invasive methods are generally preferred because of the relative lack of patient discomfort and surgical complications such as endophthalmitis, hemorrhage, retinal detachment, and cataracts.13 In addition to the ocular barriers, ocular tissues contain metabolic enzymes to break down xenobiotics that manage to penetrate into the tissue. Thus any drugs that reach the interior of the eye are further depleted by the action of enzymes such as esterases, aldehyde and ketone reductases, and many others.33 15 16 Figure 3.3: Anatomy of the retina Topically delivered drugs diffuse through the sclera and systemically delivered drugs diffuse from the choroid vasculature in the posterior segment of the eye. The outermost layer of the retina is called the retinal pigment epithelium, a layer of tight-junctioned cells that prevent drugs from penetrating into the retina. Small lipophilic drugs penetrate the lipid membrane easily but large and hydrophilic drugs require assistance either from permeation enhancers or from transporters. When the drug reaches the neural retina, it acts upon the target cells. The retinal vasculature is lined with endothelial cells bound by tight-junctions to prevent blood borne drugs and pathogens from reaching the neural retina. Together, the retinal vascular endothelium and the RPE form the blood-retinal barrier. 3.3 Strategies to Overcome Drug Removal at the Ocular Surface The most common method for delivering drugs to the eye is through eye drop solutions administered to the eye surface. They are relatively simple to apply, are non-invasive, and most solutions are easy to prepare with low manufacturing costs. There are over 100 topical eye drop formulations on the market today. Patient compliance remains one of the biggest drawbacks of topical drop administra- tion with evidence suggesting a large percentage of patients with signi cant periods of ine ective drug concentration levels. The volume of instilled dose is also highly variable from application to application which depends on the squeeze or pressure force, the angle of administration, and the ability to resist blinking.122 These issues compound quick drug loss along with tear ow rate which washes the instilled dose from the eye within ten minutes. Also, the tear drainage rate has been shown to linearly increase with instilled volume.28; 120 Figure 3.4 highlights these e ects on the concentration pro le of topically instilled drug in the eye. Eye drop formulations typically contain preservatives to prevent pathogenic contami- nation, guarantee sterility, and, in some cases, stabilize the drug. Most multiple-use drops last for approximately one month and the longer the duration of use, the higher the prob- ability for contamination. Preservatives can be toxic to ocular tissue and providers try to optimize the contamination protection/toxicity ratio. In certain cases, preservatives have been shown to have ancillary bene ts with antibiotic medications18 as well as in other formulations acting as permeation enhancers.77 In preservative-free, single-use containers, the risk for contamination is great and good manufacturing practices must be assured. Typically, preservative-free formulations are single dose containers suited for patients with 17 18 Figure 3.4: Illustration of drug concentration profile in tear film with application of eye drops The concentration of drug delivered by eye drop to the tear film varies with time. When a drop is placed in the eye, the concentration spikes to its maximum level. The concentration eventually decreases as lacrimation, drainage and (to a lesser degree) absorption deplete the drug. The concentration drops below the level considered therapeutic before the next dose is applied. [A] If a dose is missed, the eye tissues spend excessive time without therapeutic levels of drug. [B] The maximum drug concentration in the eye from one application to another varies because of factors such as squeeze force on dropper and angle of dropper position. Patients may accidentally over-administer the drug as well. These variations in application can push the tear concentration to toxic levels. [C] An excessive volume of drug solution can increase the spillage and drainage from the eye (it is well documented that the drainage rate increases linearly with instilled volume). allergies or those with signi cant surgical concerns where preservative toxicity may interfere with healing. The physiochemical properties of drug such as hydrophilicity/ lipophilicity, degree of ionization, shape, and size a ect its ability to transport through ocular barriers. Typically, lipophilic drug properties increase the speed of the molecule through cell membranes, an in- creased degree of ionization of the drug decreases lipid solubility and subsequent membrane transport, and decreased drug radius or particle size increases transport. While hydrophilic drugs are formulated in solutions, lipophilic drugs are formulated in suspensions, which typically require resuspension prior to use. Suspensions have a much lower market share compared to solutions and face additional hurdles such as drug precip- itation and resuspension, as well as particle size and polydispersity issues, which can limit the amount of drug applied to the eye or the transport through ocular barriers. In recent years, smaller-sized particles within topical formulations have been studied for their ability to increase transport. These systems will be presented in this section since they have also been hypothesized to increase residence time. Micro or nanoemulsions97; 149 are highly stable systems containing hydrophobic organic phases, often in droplet form, dispersed within an aqueous continuous phase with amphiphilic interfacial lms. The dispersed phase contains lipophilic drug and the aqueous phase enables the microemulsion to e ectively mix with tear uid. A lipophilic formulation by contrast would wash out of the eye rapidly without reaching the epithelial tissue. Additionally, it is theorized that the lipophilic droplets adhere to the epithelium and increase their residence time.39;135 Particle sizes should be under 10 micrometers in diameter for maximum comfort.148 Also, it has been reported that submicron emulsions decease the susceptibility of drug to degradation. 109 19 Liposomes are microscopic vesicles made of concentric phospholipids bilayers with alternating lipophilic and aqueous compartments. Based on their structure, they can be categorized as small unilamellar vesicles, large unilamellar vesicles and large multilamellar vesicles. The cavities within the liposomes, lined by the polar \heads" of the phospholipids, can carry hydrophilic drugs. Lipophilic drugs can be solubilized within the bilayer among the hydrophobic \tails". The hydrophilic outer surface allows e ective dispersion in the tear lm. Liposomes also protect the drug from enzymatic degradation, and may have an increased residence time by binding to the epithelium.34;87 Nanosuspensions or colloidal suspensions are sub-micron colloidal dispersions of pure drug particles stabilized by surfactants, and have been used in order to increase the solu- bility of poorly soluble drugs and increase dissolution rates via increased surface area.107 Recent work highlights nanosuspensions of glucocorticoid drugs in comparison to solutions and micro-crystalline suspensions. In a rabbit model, nanosuspensions exhibited higher intensity of glucocorticoid action and higher extent of absorption with the viscosity of the nanosuspension playing an important role in increasing duration of action.75 Nano- and microspheres are sub-micron and micron-sized solid particles containing drug dispersed within a polymer. The spheres are suspended in aqueous solution to form eye drops. In one study biodegradable poly(lactide-co-glycolide) (PLGA) microsphere carriers for vancomycin were dispersed in the topical formulation carriers. In vivo results in rabbits measuring the aqueous humor concentration indicated a 2 fold increase in bioavailability over eye drops. Interesting enough, increasing the viscosity of the formulation by adding hydroxypropyl methylcellulose did not increase bioavailability.46 PLGA microspheres have also been used as carriers for gene delivery, for in vitro studies with human RPE cells and for in vivo studies with rats. In the latter, gene expression was observes in the RPE within 20 4 to 7 days.14 Ganciclovir was loaded into albumin protein nanoparticles for intravitreal injection, and no auto-immune response was noted.90 To overcome low drug bioavailability, topical formulations have remained marginally e ective to a large extent by the administration of small volumes of very high concentrations of drug multiple times on a daily basis. Thus many formulations attempt to deliver more drug and increase the driving force of the ux by delivering highly concentrated drug. This produces only a minor improvement, and can lead to toxic side e ects if improperly managed. Various improved methods have focused on increasing the residence time the drug spends on the surface of the eye before it is washed away by normal protective mechanisms. Viscosity enhancers such as methylcellulose, carboxymethylcellulose, hydroxypropyl- methylcellulose, hydroxymethylcellulose, polyvinylalcohol, and polyvinylpyrrolidone have been added to topical formulations to retain the drug on the eye surface for longer periods of time by increasing the viscosity of the tear uid and decreasing the tear drainage rate. These types of formulations typically are rated more comfortable compared to less viscous or saline based solutions and act as wetting agents lowering surface tension and increasing tear break-up time. Polysaccharides such as chitosan have also been that are mucoadhesive with the negatively charged mucin layer, increasing corneal residence time three fold.38 A considerable increase in viscosity produces ointments which invoke the smallest rate of drug loss, but signi cantly interfere with vision, are di cult to apply, and can be quite non-cosmetic. Thus, ointments are used to a much smaller extent than solutions and are typically used at night. Mucoadhesive polymers interact with the mucin layer of the tear lm and adhere to the ocular surface. Hyaluronan and other polymers have been used in this context but their weak interactions prevent true mucoadhesive behavior. A novel set of polymers known as 21 thiomers are synthesized by modifying polymers with thiol moieties.16 Through di-sul de linkages with the native mucins of the epithelium, they become covalently anchored to the ocular surface. Mucoadhesive polymers can be applied directly to the eye as a vehicle for drug, or they can be used to attach inserts to the eye. A thiolated polyacrylate insert has been shown to deliver uorescein for 8 hours.70 In situ gels and mucoadhesive polymers have both been designed to incorporate microspheres and liposomes for extended release. Timolol maleate encapsulated in chitosan was compared to timolol gel in rabbit eyes and demonstrated similar ability to lower the IOP at half the concentration of drug.2 We direct the reader to the following mucoadhesive reviews.55;86 In-situ gel forming systems are liquid for ease of application, but undergo phase transi- tions and acquire a gel-like consistency when they encounter the physiological environment of the eye. They are mixed with the desired drug and instilled into the cul-de-sac (i.e., the pocket underneath the lower eye lid) where they gel into a substance that withstands removal by tear circulation without interfering with vision. Currently they can deliver a fairly uniform dosage over the course of about 6 hours.84 Gellation can be triggered by a change in pH (e.g., Carbopol R with methylcellulose, 126) by the presence of mono or divalent ions (Gelrite R - a gellan-gum polysaccharide125) and by a change in temperature (e.g., Pluronic R F127 with Pluronic R F68,138 poly(N- isopropylacrylamide) and chitosan25). Research has focused on combinations of the afore- mentioned triggering mechanisms to decrease liquid viscosities, to optimize the phase tran- sition and gain better control over gellation times, to extend drug release, as well as decrease the proportion of polymer needed in a dosage83;144. We direct the readers to the following reviews.51;113;127 22 Recently, soft hydrogel contact lenses have been demonstrated as extended drug deliv- ery carriers for the eye. New methodologies, greater understanding of polymeric structural properties, as well as network formation have produced a number of developments that are considerably di erent than past e orts which involved taking a conventional lens and soak- ing it within a concentrated drug solution. Delivering medications via contact lenses has been a prevailing notion since the inception of using hydrophilic, crosslinked polymer gels on the surface of the eye. In fact, the rst patent in the eld from Otto Wichterle in 1965 141 states that \medicinally active substances such as antibiotics may be dissolved in the aqueous constituent of the hydrogels to provide medication over an extended period ...via di usion." The biggest obstacle to this rationale is maintaining a signi cant concentration of drug within the uid to have a therapeutically relevant e ect, which is ultimately lim- ited by the solubility of the drug. This has been the primary reason why drug release from contact lenses has not become a clinical or commercial success. One promising technique is to create contact lenses with therapeutically relevant drug loading and extended release is to produce a macromolecular framework with memory for the drug during polymer synthe- sis. This technology has roots to a eld termed molecular imprinting, which has primarily concentrated on highly crosslinked polymer matrices for separation and sensing. For molecularly imprinted hydrogel contact lenses, it has been shown that the exten- sion of release for weakly crosslinked systems has a strong dependence on the monomer to template (M/T) ratio63 as well as the diversity and number of interactions of the recognition site.137 Drug such as timolol,6 ketotifen fumarate,137 and antibiotics7 have demonstrated in-vitro extension of release using these methods. For example, biomimetic hydrogel con- tact lenses have been developed for the enhanced loading and extended release of the 23 anti-histamine, ketotifen fumarate,137 which exhibited an extended release pro le for du- ration of 5 days with three distinct rates of release. Multiplicity of monomer-template interactions was achieved with four functional monomers chosen from an analysis of his- tamine ligand-binding pockets which led to signi cantly enhanced loading and duration of release compared to less functionalized systems at a constant M/T ratio. Considering these systems maintain the mechanical and optical properties of contact lenses, there is high potential for therapeutic contact lenses based on these types of technology to deliver a number of ocular therapeutics without the need for multiple eye drops. In-vivo validation of the most recent systems is currently under study, and the imprinting process is not as e ective with lipophilic drugs due to solubility constraints during hydrophilic gel formation. We direct the reader to the following reviews on hydrogel imprinting.5;23;62 Increasing the drug reservoir within contact lenses has also included nanoparticulate and liposomal laden lenses,57 and ion exchange hydrogels,132 with duration of drug release during in-vitro and in-vivo experiments shown to be less than 1 day. Nanoparticulate laden lenses have shown promise within in-vitro studies and demonstrate 55% of drug released in 3 days. These techniques have concerns such as inadequate drug loading at therapeu- tically relevant concentrations for long release times; and for lens dispersed nanoparticles, decreased mechanical stability induced by grain boundaries, reduced optical clarity, and longer and more costly production schemes. Recently lenses have been demonstrated in the literature to deliver PVA chains as a moisturizing agent to counteract ocular discomfort. 142 Ocular inserts can also deliver drug to the eye while avoiding the need for repeated eye drops. Soluble inserts such as collagen shields have been used as corneal bandages and drug delivery carriers and are produced from porcine scleral tissue. Typically, they 24 are soaked in solutions of drug and dissolve in the eye at characteristic rates, but they have had poor control over release and poor comfort since they are not individually t for patients. They also interfere with vision and cannot be inserted or removed by the patient, and have self-expelled from the eye in many cases. Collagen shields release drug for hours and modi cation of the collagen lm has been shown to produce longer release rates up to several days,136 Ocular inserts are placed in the eye, deliver drug until depleted, and (unless they are biodegradable) are removed at the end of the release period. Ocusert R (Alza Corp., FDA approved in 1974) consists of a small wafer of drug reservoir enclosed by two di usion controlling membranes, which is placed in the corner of the eye and provides extended release of an anti-glaucoma agent for approximately 7 days with an increased release rate in the rst 7 hours.10;133 It must be removed at the end of the release period. Lacrisert R (Merck), which is also placed in the lower eyelid, is a cellulose-based polymer insert used to treat dry eyes110 administered once-a-day and degradable. However, inserts have not found widespread use due to increased price over conventional treatments, occasional noticed or unnoticed expulsion from the eye,115 and potential for fragmentation and membrane rupture with a burst of drug being released.72 Gel forming inserts have also been produced from high molecular weight poly(ethylene oxide) (PEO) with drug release controlled by surface erosion. For the delivery of o oxacin in rabbits, inserts were placed in the lower eyelid and demonstrated a 2 fold increase in drug residence time in the aqueous humor, a 3.8 fold increase in aqueous humor drug concentration, and approximately a 10 fold increase in bioavailability over Exocin R eye drops. The increases were attributed to PEO-enhanced permeability and/or increased tear viscosity.31 25 Bioadhesive ophthalmic drug inserts (BODI R ) are homogeneous extruded mixtures of polymer and drug, shaped into rods 5 mm long a 2 mm in diameter and placed in the cul-de-sac. Animal tests have been conducted in canines,11 delivering the antibiotic gentamicin over 7 days. The bacteriological cure rate was similar to that from eye drops, with the added advantage of ease of use - one deposition of the insert as opposed to 21 instillations of eye drops. Another ocular insert under development is the OphthaCoil, a thin coiled stainless steel wire coated with a drug-containing hydrogel. The coiled structure is intended to provide shape and exibility, the ends are capped to protect the eye from the wire edges, and the coil interior can be used as a drug reservoir. Release of cipro oxacin has been measured in vitro for over ve hours. The release time can potentially be increased by modifying the hydrogel coating and the polymer in the drug reservoir.108 The anatomy of the eye has also been altered to increase residence time. For example, a mechanical technique for increasing drug residence time in the eye is to block the lacrimal puncta with punctal plugs. The tears produced in the eye cannot drain, and accumulate in the eye, so any instilled drug is not washed away. This technique has demonstrated results when used with the drug timolol in glaucoma patients.13 3.4 Strategies for Permeation Enhancement through Ocular Membranes Drugs that reach the ocular surface need to penetrate the ocular epithelium, but the epithelium presents barriers that few drugs can easily overcome. For hydrophilic drugs tran- scellular transport is di cult unless facilitated by a limited range of transporters present on the corneal and conjunctival epithelial cells. The intercellular spaces have tight junctions that resist paracellular transport. 26 Lipophilic drugs can di use through the cell membranes with relative ease. However, as mentioned earlier, they cannot pass through the tear lm and reach the epithelium as easily as hydrophilic drugs can. We here have a dilemma - drugs that reach the epithelium with ease have trouble penetrating it and vice versa. Very few drugs have a high solubility in water as well as good partitioning in lipids. It would be useful if we could use a hy- drophilic vehicle to bring the drug to the cell membrane and then have it di use through the membrane through a lipophilic vehicle. Pairing an ionic drug with its counter-ion has been shown to improve ocular pene- tration. The cationic timolol, when paired with anionic sorbic acid, has a two-fold higher penetration into the aqueous humor than when delivered alone [80]. Cyclodextrins are ring-shaped oligosaccharides that can sequester lipophilic drugs within their central hydrophobic cavities.76 The hydrophilic shells solubilize in the tear lm and carry the drug to the epithelium, where the latter partitions into the cell membrane and penetrates the epithelium. This strategy has been demonstrated successfully with a num- ber of drugs including pilocarpine, which demonstrated a four-fold increased permeation in rabbit corneas after the addition of hydroxypropyl beta-cyclodextrin.3 Novel methylated cyclodextrins have lipophilic properties that let them di use through the cell membranes in addition to their action as solubilizers of lipophilic drugs in aqueous environments. Dexam- ethasone has been delivered to the eye posterior as a topical eye drop by complexation with randomly methylated cyclodextrins.85 Cyclodextrins also have anti-irritant properties. Another option is to chemically modify the drug into a less therapeutic but more pen- etrable form, so that after it penetrates the cornea it can be converted into the therapeutic form by enzymes in the eye. The modi ed form is known as a prodrug. A water-soluble 27 prodrug of cyclosporine A is produced by esteri cation of the drug with a moiety contain- ing a phosphate group. The prodrug has improved bioavailability and penetration, and conversion in the eye back to the drug is about 6% in 3 minutes.79 Aside from modifying the drug, researchers have increased permeation by modifying the epithelial cells. The cell membranes can be made more porous by disrupting the lipid bilayers with surfactants such as polyoxyethylene 20 stearyl ether.114 Chelating agents like ethylenediamine tetraacetic acid (EDTA) sequester calcium ions and consequently loosen the tight junctions, opening up the paracellular pathway.53 Recently studies indicate a cytotoxic e ect from many permeation enhancers and absorption promoters, but the use of fetal bovine serum can ameliorate this.27 An interesting fact is that absorption promoters have been shown to promote penetration of peptide drugs through the corneal epithelium more than through the conjunctival epithelium. This may allow control over the pathway and extent of drug penetration through the epithelia.118 A third option is to transiently modify the structure of the epithelium so that its permeability increases just long enough to deliver the drug. Iontophoresis, which can be transcorneal or transscleral, delivers drug to the eye close to an electrode with potential equal to the charge of the drug.36;95 The circuit is completed by touching the grounded electrode to another part of the body. The resulting electric eld forces the drug through the epithelium. Gentamicin has recently been delivered to rabbit eyes through a drug-loaded hydrogel probe.37;45 Phonophoresis is a similar technique that uses ultrasound to transiently increase the porosity of the epithelial membranes. It has been used to enhance the permeability of the drug betaxolol 4.4 times through rabbit corneas in vitro.146 28 3.5 Strategies to Delivery Drugs to the Posterior of the Eye Drugs delivered to the posterior of the eye can follow a number of routes.47 While topically applied drug may penetrate the conjunctiva and sclera, it generally gets diluted and eliminated to a sub-therapeutic dosage. The more common topical alternatives to reach the posterior segments involve either injection of a drug, a drug delivery carrier or a drug delivery device into the vitreal cavity of the eye; or by delivering periocularly: following a transscleral route to the back of the eye and allowing it to penetrate the RPE. Additionally, some drugs are delivered through systemic circulation with oral or intravenous sources. Most systemically delivered drugs reach the ocular posterior in minute amounts, and there is a risk of systemic toxicity.128 Non-invasive methods are generally preferred because of the relative lack of patient discomfort and surgical complications such as endophthalmitis, hemorrhage, retinal detach- ment, and cataracts.78 The least invasive method would be delivery to the ocular surface with eye drops. While there is a tremendous challenge involved in overcoming all the oc- ular barriers from the tear lm to the aqueous humor or RPE, progress is being made in delivering increasing amounts of drug to the posterior from surface delivered sources. A likely drug candidate would have high partitioning in both water and lipids. Methylated cyclodextrins solubilize in both phases and could potentially improve the penetration of any drug sequestered in their central cavities. Their action has been demonstrated for dexamethasone.85 More often, intraocular delivery involves repeated injections of the drug directly into the vitreal cavity. The wet form of age-related macular degeneration (AMD) and diabetic macular edema (DME) are commonly treated through intravitreal injections of anti-VEGF 29 antibody fragments such as ranibizumab or pegaptanib.98;119 As the procedure is invasive, there may be side e ects such as infection at the injection site, intraocular pressure (IOP) increase,43 cataract formation,129 or retinal detachment.100 In addition, the injections are needed as often as once a month. This not only causes discomfort and inconvenience to the patient, but also increases the chances of developing side e ects. Newer developments include sustained release implants that are inserted into the vit- real cavity, such as Vitrasert (ganciclovir) for CMV retinitis. While this technique is also surgically invasive, the implant many only need to be inserted once every few years. This dramatically reduces the risk of side e ects and limits patient discomfort to one surgical procedure annually or less often. Other implants in the market or in late-stage clinical trials are Retisert (Bausch&Lomb) delivering uocinolone acetonide to treat chronic non- infectious uveitis93 and Medidur (Alimera Sciences) delivering the same drug for diabetic macular edema. While the challenges for delivery to the posterior of the eye are greater than for other parts of the eye, researchers are making progress. One signi cant area of research is the study of RPE membrane transporters to facilitate penetration through this barrier. The RPE, similarly to the conjunctiva, consists of cell layers bounded by tight junctions. The paracellular transport route is di cult to bypass, so for delivering non-lipophilic drugs transcellularly it can be advantageous to use native membrane transporters. RPE trans- porters exist for amino acids, peptides, monocarboxylic acids, nucleosides, folate and or- ganic cations.88 Studies are currently being done in animal models. Iontophoresis and sonophoresis are also used to penetrate the sclera near the back of the eye. Coulomb controlled iontophoresis (CCI) deliver speci c dosages of drug more accurately. Research on probes has been done to improve e cacy and safety. A probe 30 coated with a hydrogel containing a gentamicin is being developed and has been tested on rabbits.37 In additional to drugs and proteins, iontophoresis can be used to delivery nucleic acids for gene therapy.95 Subconjunctival injections can deliver drug into the sclera while bypassing the epithe- lial barriers. The drug can di use laterally through the sclera and reach the choroid and retina. Cisplatin has been delivered to rabbit retinas successfully by this mechanism. Bet- ter results were achieved when the drug was delivered within a collagen matrix rather than in a bu er solution.49 A new technique involves delivering drug systemically, and using light energy to localize the drug in the target tissue. Known as light-targeted delivery (LTD), the procedure begins with liposome encapsulated drug injected intravenously. The encapsulation reduces systemic toxicity. As the liposomes circulate through the body, a light beam is directed into the pupil and directed at ocular tissue such as the choroid neovasculature (CNV) in cases of AMD. The light beam gently warms the RPE, the CNV and the choroid capillaries to 40 C, prompting the liposomes to melt and release the drug to the local area. We suggest a good review discussing the procedure and its applications.147 Scleral plugs are devices that are surgically implanted into the sclera and deliver drug to the sclera for extended time periods. They have the advantage over injections of having a higher capacity, but the implantation procedure is more invasive. There have been successful animal studies involving this device, including the treatment of uveitis in rabbits with a plug that delivered tacrolimus.116 31 Chapter 4 Hydrogels and Imprinting Hydrogels are three- dimensional water- swollen networks of polymer chains with crosslinks that are generally covalent or ionic. The chemical nature of the polymer chains determines the unique properties of each type of hydrogel. A hydrogel can be cationic, an- ionic or neutral depending on the functional groups present on the polymer backbones. The soft material is used in many applications in which soft, exible and hydrated structures are needed.103 Hydrogels are commonly used in drug delivery devices because their high water content makes them biocompatible with tissues in the body. Additionally, they can be designed to be physiologically responsive and alter their properties in response to environmental changes. For instance, oral delivery devices are made of hydrogels that con ne insulin while in the highly acidic stomach but release it in the alkaline intestine. The pH change triggers a transition in the mesh size of the hydrogel. Some physical crosslinks weaken in the deprotonated environment of the intestine, the mesh size increases, and the insulin di uses out. Still other hydrogels are designed to respond to temperature, osmolarity, electromagnetism and other conditions. The structural characteristics of hydrogels have been modeled extensively in the lit- erature. We are able to conduct tests that reveal macroscopic structural properties of hydrogels, and deduce a number of microscopic properties. A few of the theories describing hydrogel structure are described below. 32 4.1 Di usion through a hydrogel Di usion of particles is the net transport of particles from an area of high concentration to low concentration, driven by the concentration gradient in order to bring the system to thermodynamic equilibrium. It can be described by Fick?s laws of di usion. Di usion is seen when a soluble drug is added to a solvent such as water. At the microscopic level, each individual drug molecule is undergoing Brownian motion. However because there are more molecules in high concentration areas, they are more likely to move toward low concentration areas than vice versa. The movement of a molecule in water is relatively unconstrained, and we can assume that the separation between drug molecules is great enough that they do not interact with one another. The drug can di use through water fairly rapidly, and di usion coe cients of small molecules in water are in the 105 cm2/sec range. When a molecule is placed within a hydrogel, its random motion is constrained by the presence of the polymer chains. If the length of polymer chain between crosslinks is long, and the spaces within the network large enough relative to the molecule, the latter can move through the solvent without interacting with the hydrogel. In such a case the di usion coe cient would be similar to that in a free solvent. However as the chain length between crosslinks shrinks, the space between chains or the \mesh" becomes smaller. As the mesh size approaches the diameter of the molecule, the latter is more likely to be hindered sterically by the presence of the hydrogel. Because of this the di usion coe cient of the molecule through the system decreases. The mesh size of a hydrogel is schematically shown in Figure 4.1. 33 If a small molecule is constrained in its transport through a hydrogel, a long-chain molecule encounters even more constraint. The long chain can be entangled within the hydrogel and the movement of the front end through the network requires that the rest of the chain be free to move. As longer chains have more length to be entangled with the surrounding hydrogel network, they have lower di usion coe cients. The pathway that the long-chain molecule follows through the hydrogel can be modeled as a tube whose boundaries are de ned by the entangled polymers surrounding the long-chain molecule. It can also be viewed as the sequence of pores that each section of the chain must pass through. One model that is e ective in describing the movement of long-chain molecules through a hydrogel network is the reptation model. Reptation is the movement by which snakes, worms and other legless organism undergo locomotion. In the context of molecular trans- port, it describes how long-chain molecules can slowly pass through an entangled environ- ment one small segment at a time. A relaxed long-chain molecule has a shorter end-to-end distance, and tube length, than the contour of the chain. If the tube diameter is wide enough, small sections of the chain can undergo movement normal to the tube length. Such \loops" can travel laterally along the length of the chain and, upon reaching the chain \head", translate it forward by a small distance, as shown in Figure 4.2. If the mesh is larger, the size of the perpendicular loop can be larger and the head can translate by a greater amount leading to faster di usion. If there is a nity between the chain and the hydrogel because of chemical interactions, the free energy needed for dissociation leads to slower traveling loops and decreased di usion. Increasing the number of a nity points between the two can decrease di usion. 34 35 Figure 4.1: Illustration of mesh size within a hydrogel The space within a hydrogel network between crosslinked polymer chains is known as the mesh size (?). Molecules significantly smaller than ? diffuse easily through the hydrogel. Molecules with size on the same order as ? may be slowed by the steric hindrance they encounter from the polymer chains. Molecules significantly larger than ? cannot penetrate through the hydrogel. 36 Figure 4.2: Illustration of reptation model for transport through hydrogel A long-chain molecule within a hydrogel or other entangled polymer can be said to be confined within a ?tube? whose boundaries are defined by the surrounding polymer which restricts the movement of the long-chain molecule. When the long-chain molecule is in a relaxed state, its contour does not exactly match the tube. Rather, it has loops that can slide or ?reptate? along the length of the tube. As the loop reaches the ?head? of the long-chain molecule, the latter moves forward by a small length. This process enables the transport of the long-chain molecule through the hydrogel. 4.2 Theoretical model for di usion We can mathematically model the transport of a particle through a hydrogel.30 We consider the case in which the hydrogel is shaped like a slab. The aspect ratio of the exposed surface diameter to the thickness is greater than 10 so we can assume di usion is occurring in one dimension. The hydrogel is immersed in an aqueous environment when the concentration of the di using particle is negligible in the bulk uid outside the hydrogel. By application of Fick?s Second Law, and assuming the given initial and boundary conditions, @C @t = D @2C @t2 (4.1) C(x;t) = C0 when t = 0 (4.2) @C @t = 0 while t> 0 and x = 0 (4.3) C = Cs while t 0 and x = L=2 (4.4) 37 We describe a system in which a planer hydrogel undergoes one-dimensional di usion over time in an environment in which the drug concentration is always e ectively zero, where Co represents the initial drug concentration (assumed to be uniform) in the homogeneous gel, x represents the distance from the central length-wise axis of the hydrogel to the surface, C is the concentration of the drug within the gel at any given position and time, Cs is the concentration at the surface of the gel, D represents the constant di usion coe cient which is independent of position and concentration, t is time, and L is the thickness of the gel. At x = 0, the ux of the particle is e ectively zero. The solution of the PDE is given by,30 C C0 Cb C0 = 1 4 1X n=0 ( 1)n 2n+ 1e (2n+1)2 2D 4L2 t cos (2n+ 1) x 2L (4.5) An e ective way of comparing the release kinetics from di erent gels is to compare the fractional release of the drug at time t relative to the total drug released at in nite time, or Mt=M1. Mt is the total cumulative mass of therapeutic released at time t, and M1 is the total cumulative mass of therapeutic released at in nite time. Mt M1 = 1 1X n=0 8 (2n+ 1)2 2e (2n+1)2 2D 4L2 t (4.6) The above expression can be expressed in terms of error functions. 38 Mt M1 = 4 Dt L2 12 " 1 12 + 2 1X n=1 ( 1)nierfc nL 2pDt !# (4.7) At short times ( Mt=M1 < 0.65) the expression can be simpli ed to Mt M1 = 4 Dt L2 12 (4.8) By plotting the fractional release of HA versus (t0:5=L), we can calculate the di usion coe cient from the slope. We can also measure how well the data matched a Fickian release pro le by the empirical Power Law equation: Mt M1 = kt n (4.9) By plotting the log of fractional release versus the log of time and calculating the slope, we can determine the order of release of the particle from the hydrogel. The order of release and the slope of the plot, n, are related by order =jn 1j. For Fickan release, the order is 0.5. For time-independent release, the order is 0. 39 4.3 Equilibrium swelling theory Hydrogels in water are subject to a number of thermodynamic in uences. As the polymers solvate in water they tend to elongate, but the chemical bonds making up the chain resist the elongation. The free energy of mixing and the opposing free energy of elasticity both contribute to the total change in free energy. G = Gmix + Gel (4.10) Based upon this relationship, Peppas and Merrill104 have developed a model describing the relationship between the average molecular weight between crosslinks (Mc) and the polymer volume fraction in the swollen state (v2;s) in an swollen network of crosslinked polymers synthesized in the presence of a solvent. The model allows us to calculate the Mc, a microscopic parameter, from experimentally determined values such as the v2;s, polymer volume fraction in the relaxed state (v2;r) and speci c density of the polymer (v), and known quantities such as V1 (molar volume of water), molecular weight of uncrosslinked polymer chains (Mn) and the Flory polymer-solvent interaction parameter ( 1). 1 Mc = 2 Mn (v=V1)[ln(1 v2;s) +v2;s + 1v22;s] v2;s v2;s v2;r 1 3 1 2 v 2;s v2;r (4.11) 40 The mesh size ( ) can be calculated from the Mc by the following equation 4.12. Cn is the rigidity factor (8.3,56), Mr is the molecular weight of the polymer repeating unit (44 mol/g) and l is the length of the carbon-carbon bond (1.54 A). = 132;s 2C nMc Mr !1 2 l (4.12) Lower Mc, and higher v2;s, corresponds to a smaller mesh size. In general, a hydrogel with a smaller mesh size will manifest a lower di usion coe cient for solvent or other molecules di using through it. When all other factors are held constant, gels with lower Mc and higher v2;s tend to display lower di usion coe cients and vice versa. 4.4 Rubber elasticity theory Hydrogels demonstrate elastomeric behavior. When a constant stress is applied to a hydrogel, it undergoes deformation and reaches an equilibrium strain. Under such condi- tions, the sample is undergoing changes in the Helmholtz free energy A = U TS (4.13) Under isothermal conditions, assuming the volume of the sample does not change, the relationship in equation 4.14 exists. 41 f = @A @L ! T;V (4.14) From this relationship, Sillman and Peppas and Merrill104 deduced for a hydrogel prepared in the presence of solvent, a relationship between the normal stress applied( ) , the elonga- tion ratio ( ), the ratio of polymer volume fractions of swollen and relaxed gels (v2;s=v2;r), the ideal gas constant (R), the temperature of experimental conditions(T), speci c poly- mer volume (v), the molecular weight of uncrosslinked polymers (Mn), and the molecular weight between crosslinks (Mc). This model allows us to measure the relationship between and experimentally and calculate the Mc. = RT 1 vMc ! 1 2McM n ! 1 2 v 2;s v2;r !1 3 (4.15) The Mc enables us to calculate the mesh size according to equation 4.12. 4.5 Biomimetic Imprinting Biomimetic imprinting is a technique for synthesizing hydrogels and conferring certain unique physical properties to it in the process.62 It is based on the premise that synthesizing a hydrogel by crosslinking its constituent monomers in the presence of template molecules can alter the microstructure of the gel. Consequently, we can change the way the hydrogel interacts with the environment and with molecules di using through it. 42 To prepare a hydrogel by biomimetic imprinting, the constituent monomers are mixed with a selected molecule and allowed to reach equilibration. The monomers complex with the template molecules to lower the free energy of solvation in the mixture. If the hydrogel is crosslinked with the monomers in these optimum con gurations, then the hydrogel retains a molecular \memory" of the template molecule which persists even if the template is washed away. This molecular memory enhances the a nity of the hydrogel for the drug used as the template. The selection of monomers involved in the biomimetic imprinting process is critical for e ective imprinting. This involves identifying binding molecules and other biological agents that have evolved to bind with the template molecule. The active site of the binding molecule is examined for amino acids that are critical for binding to the template molecule. Analogous acrylate monomers are selected by comparing the chemistry of the acrylates to the critical amino acids. The acrylate monomers with functional groups most similar to the critical amino acids are likely to form the best memory sites within the hydrogel. This technique has been demonstrated for the molecule ketotifen fumarate,137 a small molecule. 43 Chapter 5 The Tear Film and Dry Eyes 5.1 The Tear Film The tear lm covers the exposed surface of the eye, and it comprises distinct layers. The epithelium of the eye secretes mucins, both membrane-bound and free oating. They form a dense layer over the epithelial microvilli called the glycocalyx (0.01-0.07 m),64;99 and these perform a number of functions - providing the hydrophobic epithelium a hydrophilic surface, lubrication against the strong shearing e ects of blinking, and regular cleanup of the eye surface. The concentration of mucins decreases toward the anterior section of the aqueous-mucin layer (4-9 m thick), which keeps the eye surface bathed in water with dissolved oxygen and antibacterial enzymes (Tear uid does not supply glucose. This is provided by interstitial uid for conjunctiva, and aqueous humor for cornea).139 At the interface between the tear lm and the atmosphere, there is a thin (0.1 m) layer or lipid molecules - non-polar lipids form a coating that retards the evaporation of tear water to the air, and polar lipids form an interface between the aqueous layer and the hydrophobic non-polar lipids. The lipids are secreted by the Meibomian glands in the eyelids. Tear uid turnover can be classi ed into three phases: secretory, distributional and excretory. In the secretory phase the aqueous component and some mucins of tear uid are produced by the lacrimal glands via the lacrimal ducts above the eye at a baseline rate of 1 L/min. The tear- lm turnover rate is 16% per minute,112 and the lm?s thickness is 4-9 m. The cul-de-sac behind the lower lid can hold 7-9 L normally, and up to 30 L if 44 blinking is avoided. It creates a reservoir of tear uid increasing the residence time of tears in the eye. The pH of tears ranges between 6.5 and 7.6. The distributional phase consists of the blink mechanism, which compresses the lipid layer within the palpebral ssure and redistributes the mucins across the epithelial surface. When the eye opens after the blink, the tear lm is spread out across the eye surface. The high surface tension between the aqueous surface and the air prompts the lipid to spread over the aqueous layer, forming a stable tear lm. The excretory phase involves the removal of tear uid through the puncta or openings of the lacrimal duct. The drainage process is rapid and highly e cient - 90% of all tear uid drains through the lacrimal duct at 25-50 L per 90 seconds. The uid ows through the puncta into lacrimal sacs behind the nose, encouraged by the negative pressure in the sac and the "pumping" e ect of blinking. It passes via the lacrimal duct into the vascularized nasal meatus in the nose, where the uid and any drugs it contains are absorbed. The remaining 10% of uid evaporates between blinks, at a rate of 7% per minute.121 If the eye is held open and immobilized, the time taken for the tear lm to break up (BUT) is 15-50 seconds for a normal eye, and less than 10 seconds for a dry eye. The tear lm and the related anatomical structures are shown in Figure 5.1. Air at the surface of the eye may also be considered an "ocular uid" when we consider the transport of oxygen. Di usion from the atmosphere is the only mechanism for the delivery of respiratory oxygen to corneal tissue, because optical clarity requires that the tissue be non-vascularized. The rate of di usion of oxygen to the eye is estimated to be 7.8 L/cm2 per hour.21 45 46 Figure 5.1: Tear film and related structures The tear film on the eye surface consists of roughly three layers: the mucus layer secreted by goblet cells, the watery layer secreted by the lacrimal gland and the oily layer secreted by the Meibomian glands. Tears secreted from glands flow over the eye. They are spread over the eye surface through blinking mechanisms, accumulate under the lower lid, and drain from the eye via the lacrimal puncta. They pass through the lacrimal duct into the lacrimal sac and are absorbed into the body through vasculature in the nose. 5.2 Etiology of Dry Eyes Keratoconjunctivits sicca, commonly referred to as \dry eyes", is a condition where the conjunctiva and cornea are not enclosed with a healthy amount or quality of tear uid. 15 Hyperosmolarity of the tear uid48 and inadequate hydration of the epithelial surface produce symptoms of itching, burning and excessive watering, and can lead to in amma- tion, corneal ulcers, bacterial conjunctivitis, corneal perforation, and loss of vision. A wide range of factors can be held responsible for dry eyes from congenital malfunction of tear glands, to autoimmune disorders such as Sjogren?s syndrome,42 to environmental risk fac- tors such as contact lens wear or air quality.58;102;143 Nonetheless, many of the causes share mechanistic pathways. Furthermore, multiple forms of dry eye syndrome can coexist in one patient, with or without causal links. In normal eyes a thin aqueous lm is con ned within the boundaries of the eyelids over the hydrophilic mucin-covered epithelial surface. The lm is stable because the surface tension of the uid lm plus the interfacial tension at the solid- uid boundary is lower than surface tension of solid surface. Between blinks the surface tension increases as the tear lm thins from evaporation. The lm becomes unstable and breaks up in spots, exposing the epithelium to air. The blinking process replenishes the tear lm, redistributes mucins over the eye surface and removes any foreign particles. During each blink while the eye is closed, the surface tension at the lipid-air interface drops as the exposed surface area decreases. When the eye opens, the tension increases rapidly and the lm thins and spreads over the eye surface. Over time, because of evaporation and drainage, the lm thickness continues to decrease until the lm destabilizes and breaks. The interval from the blink to the lm destabilization is called the break-up time (BUT), and the exposure to air triggers the next 47 blink. In normal eyes, the BUT is 15 to 50 seconds long. In dry eyes, the BUT decreases to less than 10 seconds121 as the eye attempts to compensate for the rapid evaporation. Dry eye syndromes can be classi ed as aqueous tear de cient (ATD) or evaporative dry eye (EDE). The former occurs when insu cient volumes of aqueous tears are produced by the lacrimal glands. ATD is further classi ed as Sjogren?s syndrome related (where abnormal autoantibodies attack the glands) or non-Sjogren?s syndrome related (generally milder, with etiology not fully determined but usually multifactorial).106 EDE occurs when the tear lm evaporates from the eye faster than it is secreted, leading to hyperosmolarity and desiccation. It is caused by thin lipid layers (from malfunctioning Meibomian glands), reduced blink rate, poor lid-globe congruity, unfavorable environments and disruptions in the tear lm continuity.41 Dry eyes are also developed by users of contact lenses and people who have undergone Laser-Assisted in Situ Keratomileusis (LASIK surgery).130 Many contact lens users complain of ocular discomfort that makes contact lens wear irritating or even painful.40 This discomfort has been linked to EDE. The tear lm is disrupted by the edges of the lens and the increased surface tension at the boundary. It has also been demonstrated that50 continued contact lens wear leads to decreased blinking rate. However, some studies show that the aqueous layer becomes more stable if the lens gets coated by mucins.69 The mucins make the lens surface more hydrophilic, lowering the aqueous-lens interfacial tension. Ocular dryness triggered by lens-wear is known as contact lens-induced dry eye (CLIDE). In one survey 76.8% of current contact lens users reported ocular dryness, with 26.8% reporting frequent to constant symptoms.26 There are economic impacts of this phenomenon - rstly the productivity lost by people who need to spend time and resources to manage the condition, and secondly the loss of sales by contact lens manufacturers when people stop using contact lenses.145 48 Current techniques for managing dry eye symptoms vary depending on the etiology. Most people with non-Sjogren-related dry eye and contact lens induced dry eye apply rewetting drops and arti cial tears to their eyes via eye drops. Some deliver lipids41 to make the lipid layer more substantial. Others apply viscosity enhancers to slow tear drainage and hold moisture close to the epithelia. But eye drop usage is inconvenient when needed multiple times a day, because it is followed by a time interval in which vision is blurred and activities such as driving and reading must be interrupted. Nonetheless eye drop formulations such as hyaluronic acid and carboxymethyl cellulose are commonly used as arti cial tears. Patients of Sjogren?s syndrome need anti-in ammatory drugs to control their dry eye. 5.3 Treatment with Hyaluronic Acid Hyaluronic acid (HA) is a polysaccharide that is fast becoming a preferred treat- ment for dry eye syndrome as an arti cial tear solution. It is an unbranched non-sulfated glycosaminoglycan composed of units of the saccharides D-glucuronic acid and D-N-acetyl- glucosamine, linked together via alternating -1,4 and -1,3 glycosidic bonds, as shown in Figure 5.2. Its conjugate base is hyaluronate, with a deprotonated carboxylic group on the glucuronic acid. The dimer unit has a molecular weight of about 415 Da, but the entire chain can be 25,000 dimer units in length. HA is polydispersive, and depending on the location of origin the average size can range from 5000 Da to 20 Million Da in vivo. HA is found throughout the body as a natural lubricant, predominantly in the synovial uid in skeletal joints and in the vitreous humor of the eye. It is also a major component of many extracellular matrices. Where it does not constitute the bulk of the material, it 49 50 Figure 5.2: Illustration of hyaluronic acid structure Hyaluronic acid is a long-chain molecule, specifically a polysaccharide consisting of repeating units of glucuronic acid and N-acetylglucosamine. The glucuronic acid has a carboxylic groups that deprotonates and interacts with cationic monomers such as (diethylamino)ethyl methacrylate and amino acids such as lysine and arginine. HA is found in various lengths depending on the source, usually in the 10 6 Da range. acts as a structural sca old for other materials such as chondroitin sulfate proteoglycans. The largest amount exists in the skin tissue, more so in the epidermis than the dermis. Concentrated HA has strong viscoelastic properties and displays shear-thinning behav- ior. When subjected to fast and short ow, the chains tangle and demonstrate elasticity with low viscosity. In contrast during slow and extended ow, the chains partially separate and align, and their interactions lead to more viscous behavior. The viscosity enhances the tear stability60 and slows tear removal131 by slowing the ow rate. The shear thinning behavior has a lubricating e ect. In the eye in particular, it prevents the shearing force of the eyelid from damaging the epithelium during blinking.1 In normal eyes, the healthy composition of the tears is enough to prevent this damage. HA is also mucoadhesive and interacts with ocular mucins when delivered to the surface of the eye. The mucins are proteoglycans, and the glycan components are similar to HA in structure and function. The glycans interact with the HA, and the HA behaves as part of the mucin layer, e ectively increasing the mucin layer?s thickness. When a contact lens is present in the eye, HA may cover the lens as an \arti cial mucin" and counteract the tear lm destabilization that occurs in the presence of the lens. Because HA is hygroscopic, it retains water close to the ocular surface and reduces dehydration.96 It also encourages corneal wound healing24 by promoting epithelial cell migration.52 For the above reasons and its high biocompatibility, HA has been used as a topically delivered arti cial tear solution for over twenty years.124 Because of its proven track record we have selected HA as a comfort molecule for delivery to the eye from a therapeutic contact lens for the treatment of ocular discomfort and some cases of dry eye syndrome.8; 9;22 51 Chapter 6 Lenses for Delivery of Comfort Molecules There is a strong unmet need for the sustained delivery of dry eye treatments without recourse to the use of eye drops. Dry eye syndrome causes great discomfort to people and compels many su erers to abandon the use of contact lenses until symptoms are alleviated. Currently patients manage their symptoms through the use of topical delivery of arti cial tears, moisturizers and comfort agents, with or without concomitant use of contact lenses. Instillation of eye drops is imprecise in untrained hands, and can be a nuisance for people who have to interrupt daily activities to apply them. The recommended amount of eye drops is not enough for some people, and patients have been known to apply arti cial tears 12 or more times a day.29 The development of a device that can deliver comfort agents in a sustained manner over the course of several hours to days without interfering with contact lens use would greatly ease the su ering of dry eye patients. Furthermore, a contact lens that delivers the comfort agent to the eye directly would simplify the procedure considerably. In this study we design such a therapeutic contact lens that delivers the comfort molecule hyaluronic acid (HA) to the eye in a sustained and tailorable fashion. 6.1 Nel lcon A This study was sponsored by CIBA Vision, Inc., a major producer of contact lenses and lens care products, to formulate a version of their agship daily disposable contact lens 52 Focus R DAILIESTMthat alleviates ocular discomfort via HA release and can be worn by patients with some types of dry eye syndrome. Daily disposable lenses are one class of contact lenses currently in the market. They are inexpensive and designed to forego the need for daily cleaning and storage of contact lenses by the user. A new lens is worn daily while the old lens is discarded. Daily disposable lenses are recommended for patients with ocular allergies, to prevent the accumulation of dust and other allergens on the lens.61 Other classes of lenses include extended wear lenses that can be worn continuously in the eye for thirty days without removal, and are gaining popularity for their ease of use. In contrast annual and quarterly lenses, which need to be removed for cleaning every night, are growing less popular. The material that we have chosen for the fabrication of the contact lens is a commercial formulation known as Nel lcon A.17 It is primarily used for the synthesis of daily-wear dis- posable contact lenses, although such lenses have been known to be worn for multiple days. 91 Nel lcon A is a hydrogel material consisting of biocompatible polyvinyl alcohol (PVA) polymers in aqueous solution, photocrosslinked into a network structure. A novel feature of Nel lcon A is the lack of monomers in the prepolymerization sol, eliminating the need to purify the prepared lens. In fact, almost every component required for polymerization is tethered to the PVA chains so that small unreacted monomers do not remain in the nal hydrogel. This includes the crosslinkers, the initiator, and the tint where applicable. PVA contains hydroxyl (OH ) groups attached to a repeating polymeric backbone in the 1,3 position. It is synthesized by the acid hydrolysis of polyvinyl acetate. The 1,3 hydroxyl groups are perfectly positioned to undergo cyclic acetal formation upon reaction with aldehydes, and this is a highly useful mechanism for the attachment of components necessary for hydrogel formation. 53 The synthesis of Nel lcon A macromer was performed by CIBA Vision, Inc. (Du- luth, GA) according to a two step procedure.142 First a diacetal with acrylate functional- ity, N-acryloylaminoacetaldehyde-dimethylacetal (NAAADA), was synthesized by reacting aminoacetaldehyde with acryloyl chloride in a low-temperature alkaline aqueous solution. After neutralization and extraction, the crude product was puri ed through molecular distillation. In the second step, PVA was transacetylized with NAAADA resulting in PVA chains with a well controlled number of pendant crosslinking acrylate groups per macromer chain.17 A number of reactions were taking place at this acid catalyzed stage: the crosslinker?s acetal was hydrolyzed to aldehyde, the aldehyde reacted with the PVA, and some acetate groups remaining on the PVA from its synthesis were converted to hydroxyl groups, as shown in Figure 6.1. The added crosslinker was also modi ed by attaching the initiator Irgacure 2959. Atmospheric oxygen was used as the stabilizer for the acrylate. To make tinted lenses, a commercial dye such as Remazol Brilliant Blue RTMwas also activated and attached to PVA through acetylization. The reaction was quenched by neutralizing with alkali. The photopolymer was then puri ed by ultra ltration to the desired purity and concentration. 6.2 Hyaluronic Acid Binding Moieties Within the human body, HA binds to various receptors, the most signi cant is the cell-surface glycoprotein CD44. Using molecular modeling and site-speci c mutagenesis, researchers have identi ed the amino acid residues most responsible for the binding of CD44 to HA.12 Residues deemed critical for HA binding were tyrosine-42, arginine-78 and 54 55 Figure 6.1: Synthesis of Nelfilcon A macromer from PVA Nelfilcon A is synthesized using poly(vinyl alcohol) (PVA) as a starting material. N-acryloyl- aminoacetaldehyde-dimethylacetal (NAAADA) is reacted with the PVA through transacetylization under acidic aqueous conditions. The product is a PVA macromer with pendant acrylate groups at well-defined intervals. The trans-acetylization process can be used to attach initiator and tint to the macromer. The product is purified by diafiltration. tyrosine-79. Residues considered important for HA binding were lysine-38, arginine-41, lysine-68, asparagine-100, asparagine-101 and tyrosine-105. Based on this analysis we sought acrylate monomers with functional groups that bore similarities to tyrosine, arginine, lysine and asparagine. Tyrosine contains a 4-hydroxy- phenyl group which features aromatic behavior with some hydrogen bonding capability. Arginine and lysine have amine groups which bear positive charges when protonated. As- paragine possesses an amide group for additional hydrogen bonding. Acrylate monomers with similar chemical behavior are acrylamide (AM), N-vinyl pyrrolidone (NVP) and (diethylamino)ethyl methacrylate (DEAEMA). AM shares an amide group with asparagine. NVP, an aromatic lactam, can be seen as an analog to tyrosine for its aromaticity and hydrogen-binding capability. Finally, DEAEMA is a cationic acrylate, and is similar to arginine and lysine because of its positive charge. The structures are shown in Figure 6.2. We hypothesized that these monomers, if incorporated into the Nel lcon A network, would non-covalently interact with HA and increase the a nity of the molecule for the hy- drogel, thereby giving us an additional level of control over the release rate. The DEAEMA was expected to form an ionic bond with the carboxylic groups on the glucuronic acid units, and the AM and NVP to form general hydrogen-bonds with varied groups on both glu- curonic acid and acetylglucosamine. The increased ability to tailor the release kinetics would enable us to design the optimum formulation for the desired product. We also employed the principle of biomimetic imprinting, as described in Section 4.5. We hypothesized that when the functional monomers are added to the HA-Nel lcon mixture and allowed to equilibrate with HA, the monomers would prefer to be spatially arranged in a low energy con guration. Such a con guration would favor electrostatic and polar 56 57 Figure 6.2: Comparison of functional groups on amino acids and acrylate monomers For the biomimetic imprinting of hyaluronic acid, we select acrylate monomers that bear chemical similarity to the amino acids found on the binding site of hyaluronic acid binding protein CD44. Acrylamide and asparagines both have amide moieties, N-vinyl pyrrolidone and tyrosine have hydrogen bonding capability while (diethylamino)ethyl methacrylate is positively charged, like arginine and lysine. interactions between the monomers and the HA, much like the interactions between amino acids and HA in the CD44 binding site. When the gel is crosslinked the monomers would be immobilized in these favorable con gurations, creating sites within the network with a stronger a nity for HA than areas with the same chemical composition and random con guration. 6.3 Methods and Materials 6.3.1 Synthesis of hydrogels To prepare 6.5 mg HA/g Nel lcon hydrogels, 5 g of Nel lcon A macromer(CIBA Vi- sion, Inc.) was mixed with 32.5 mg of hyaluronic acid sodium salt(Streptococcus equi,Fluka) in a 15 mL centrifuge tube. Functional monomers acrylamide (Aldrich), N-vinyl pyrroli- done (Polysciences, Warrington, PA) and (diethylamino)ethyl methacrylate (Aldrich) were added to prepare imprinted hydrogels. The mixture was repeatedly stirred, centrifuged (30 minutes to 1 hour at a time, a minimum of 4 times), and rested overnight to dissolve the HA in the prepolymer until homogeneous. The mixture was nally centrifuged for 5 to 10 minutes to remove air bubbles. For the benchmark studies in the appendix, the concentration and size of HA were varied to study release characteristics. For the former, hydrogels were prepared at con- centrations of 2 mg HA/g Nel lcon, and 40 mg HA/g Nel lcon. For the latter, 6.5 mg HA/g Nel lcon hydrogels were prepared using HA of molecular weights 50 kDa and 100 kDa (Genzyme Pharmaceuticals). Moulds for the hydrogel lenses were prepared. PTFE Te on R (Scienti c Commodities Inc., Lake Havasu, AZ) spacers of 5 mil thickness were constructed by cutting sheets into 2" 58 by 1.5" frames with a 1" by 1" central space. Spacers were a xed to 2" by 1.5" microscope slides. Between 125 to 200 mg of the prepolymer mixture was pipetted into the central space carefully to avoid the introduction of air bubbles, and the mould was closed by placing a second microscope slide on top, sandwiching the prepolymer between the slides and within the spacer. The mould is clamped with binder clips reserved for this purpose. The mould was exposed to ultra-violet light from a UV light source (Novacure 2100, Exfo, Mississauga, Canada). The intensity of delivered light was 10.5 mW/cm2 (8.5 mW/cm2 for the benchmark studies included in the appendix) measured by radiometer (International Light IL400A). Duration of exposure was 15 seconds for hydrogels without functional monomers, and 45 seconds for hydrogels with functional monomers. The ex- posure times were determined with a Q-100 modulated di erential photo calorimeter (TA Instruments, New Castle, DE), measuring the reactio progression. The mould was opened and the hydrogel was covered with a small volume (2 to 5 mL) of water to soften and release the material. After 5 minutes, the hydrogel was peeled from the mould and cut into a disk with a cork borer (size no. 4, 14 mm diameter). To prepare strips for tensile studies, the Te on spacer was cut with an inner space of dimensions (6 cm by 3 cm). Crosslinking took place in a light source (Dymax UV ood light) at an intensity of 10 to 12 mW/cm2. The hydrogel was cut with a clean blade to strips 6 to 10 mm wide. 6.3.2 Dynamic release studies Dynamic release studies were conducted on the hydrogels to measure the release of HA. Prepared lenses were placed in 50 mL centrifuge tubes (in triplicate) with 20 mL of arti cial lacrimal solution 6.78 g/L NaCl, 2.18 g/L NaHCO3, 1.38 g/L KCl, 0.084 g/L 59 CaCl2.2H2O, pH 859), and incubated at 35 C on an orbital shaker (Stovall Life Sciences, Greensboro, NC) at a rotation speed of 20 to 30 rpm. After measured time intervals, the lenses were extracted and deposited into fresh lacrimal solution. The samples with released HA were stored at 4 C until assayed with a sandwich ELISA assay kit (Corgenix, Denver, CO). The assay kit had a detection range between 20 to 800 ng/mL, and some samples were diluted to prevent signal saturation. Benchmark release studies in the appendix were carried out using the protocol de- scribed previously.142 6.3.3 Heat stability studies The hydrogels were subjected to simulated sterilization conditions to determine if the release characteristics would be a ected. Nel lcon hydrogels were prepared with 6.5 mg HA/ g Nel lcon, no added monomers, and placed in 2 mL microcentrifuge vials with 6.5 mg/mL solution of HA in DI water to prevent partitioning out of the HA. The pH of the solution was adjusted to 11. The vials were heated to 120 C for 40 minutes and then cooled in a room temperature water bath. The lenses were removed from the vials, blotted to remove excess HA from the surface, and studied for their release kinetics. The e ect of heat conditions on HA solution was also assessed. In 2mL microcentrifuge vials, 1 mL samples of HA solutions of 500 ng/mL and 10 g/mL were heated to 105 C and 121 C respectively. The 500 ng/mL samples were heated for 0 min, 5 min, 30 min and 60 min while the 10 g/mL samples were heated for 15 min, 30 min, 45 min and 60 min. The sample vials were quenched in a room temperature water bath and assayed with an ELISA assay kit for HA. 60 Hydrogel swelling studies After release in lacrimal solution, hydrogel lenses were dried in air for a minimum of 24 hours and then in a vacuum oven (VWR) at 30 C and 28 in.Hg to remove moisture until the weight change in the lenses was less than 0.1% (at least 5 days). The gels were then weighed in air and in heptane, a non-solvent, using a microbalance (Sartorius). The lenses were equilibrated in DI water overnight and the fully swollen lenses were weighed in air and heptane. Swollen lenses without HA were synthesized and weighed after equilibrating in water overnight. Hydrogels in the relaxed state were synthesized and removed from mould without exposure to water, then weighed immediately. These were again weighed after dehydration. 6.3.4 Tensile strength studies Hydrogels prepared in strips (in triplicate) were mounted on a dynamic mechanical analyzer (RSA III, TA Instruments) at a gauge length of 30 to 35 mm, and extended at a rate of 4 mm/min. The gels were fully hydrated through the experiment, and hydration was maintained with an aerosol di user. 6.4 Results and Discussion 6.4.1 Molecular transport and di usion coe cients Hydrogels made with 6.5 mg HA/g demonstrated a concentration dependent release pro le. Figure 6.3 shows the cumulative mass of HA released from the hydrogel over a 5 day period. The release rates can be classi ed into three general zones. Initially HA is released over the rst 2 hours at a rate of around 24 g per hour. Over a 24 hour period 61 after that, we see a linear release pro le delivering 4 g per hour. After that the release rate gradually decreases until very low amounts are releasing for the last 3 days. We can subject the release pro le to the analysis described in Section 4.2 and determine the order of release and di usion coe cient of the HA from this hydrogel. From equation 4.8 we can plot Mt=M1 versus (t=L2)12 and calculate the di usion coe cient D by setting the slope (k) equal to 4 q D= . The order of release is n 1 when n is equal to the slope of Log[Mt=M1] versus Log[t]. The calculated di usion coe cient is 5:69 10 10, and the order of release is 0:61, close to the order of Fickian release, 0:5. The decrease in di usion coe cient is in uenced by two factors. First, the hydrogel network presents a steric barrier to the Brownian movement of particles through the solvent. For a particle to di use through a hydrogel, it needs to pass through the open space or mesh between crosslinked polymer chains. The smaller the mesh, the less space the particle has to di use. Second, our drug is not a simple particle but a long-chain molecule, behaving like a series of particles joined together. The chain-like nature of HA restricts the path that each constituent \particle" can pass through to di use through the chain - each must follow the one in front of it. The motion of a long-chain molecule through a hydrogel mesh can be described by the reptation model as described in Section 4.1. The HA does not slide through the mesh in one smooth motion. Rather the \tail" end of the polymer moves slightly forward to form a loop, and it is the loop that travels along the length of the chain, with none of the individual units traveling a large distance. When the chain reaches the \head" of the chain the loop disappears and as a result the entire chain undergoes a small displacement through the hydrogel. 62 63 y = 11.837x + 9.777 R 2 = 0.8866 y = 4.0424x + 46.888 R 2 = 0.9761 y = 0.1531x + 170.86 R 2 = 0.6 0 50 100 150 200 250 0 20 40 60 80 100 120 140 Time (hours) Cumu lative Ma ss Released (micr ograms) Figure 6.3: Cumulative release of HA from Nelfilcon hydrogel The prepolymer formulation used to make this hydrogel contained only the Nelfilcon macromer and hyaluronic acid. The release profile demonstrates Fickian kinetics over 5 days, with three distinct release rates. The initial rate lasts about 6 to 10 hours with a release rate of 12 ?g/ hr. The intermediate region demonstrates a nearly linear release profile, delivering 4 ?g/ hr. After about 2 days, the release rate tapers off until it is negligible. To compare the release rate of HA from these hydrogel lenses with the therapeutic regime of HA eye drops, we considered the therapeutic regime of topically delivered HA arti cial tear eye drops. In particular, we examined AQuify R Long-Lasting Comfort Drops, a 0.1% solution of HA. According to the package insert, the recommended dosage is 2 drops upto 3-4 times daily. The volume of a typical eye drop is 20 L122 so if drops are delivered four times a day, the delivered dosage would be 40 g of HA every 6 hours or 6.67 g per hour. As the bioavailability of eyedrops is always less than 100% because some drug is always lost through lacrimal drainage, it appears the hydrogel lens can deliver therapeutic or near therapeutic amounts of HA over the rst couple of days. However, by incorporation of functional monomers to the system we may be able to exert an additional level of control over the release rate. In Section 4.5 we discussed the process of biomimetic imprinting, which allows tailoring of the release properties of a hydrogel by adding functional monomers and crosslinking in the presence of the drug molecule. This phenomenon has been demonstrated previously 137 for ketotifen fumarate, a relatively small molecule, in a formulation containing added functional monomers as 5% of the total prepolymer mixture. Using this percentage of functional monomers as a starting point, we modi ed the basic Nel lcon A macromer sol by adding acrylamide (AM), N-vinyl pyrrolidone (NVP) and (diethylamino)ethyl methacrylate (DEAEMA) as functional monomers and crosslinking the gels in the presence of HA. The monomers AM, NVP and DEAEMA were added in a ratio of 1:1:2 by moles, and together comprised 5% by mass of the prepolymer before addition of HA. Release studies carried out with hydrogels of this formulation resulted in a negligible amount of HA released in a 24 hour period. Theorizing that the functional monomer 64 content was too high, we reduced the monomer content to 1% by mass of the prepolymer. Again, negligible HA was released. To understand the mechanism by which the functional monomers were immobilizing the HA, we produced hydrogels with 1% by mass functional monomers and placed them in release solutions of lacrimal solution at di erent pH conditions. We found that the 1% imprinted hydrogels released negligible HA in pH 8 solution but released signi cantly increased amounts in pH 12 solution. It appeared that the excessive OH ions in the alkaline solution would deprotonate the DEAEMA, reducing the electrostatic interactions with HA (in hyaluronate form) and allow the HA to release from the hydrogel. However we are designing our system for ocular drug delivery and it needs to release HA at physiological pH. With evidence indicating that electrostatic (and potentially other non-covalent) inter- actions were responsible for the immobilization of HA in the hydrogel, we further reduced the functional monomer content of the hydrogel. We hypothesized that if each HA chain were interacting with fewer functional monomers, it would undergo reptation at an in- creased rate. Our hypothesis can be illustrated by making an analogy with the fabric hook-and-loop fasteners such as VelcroTM. VelcroTMconsists of two fabric surfaces, one with minute loops and the other with hooks. If one isolated hook is connected to an iso- lated loop, the bond between them can be severed with the application of a very small force. But if large fabric surfaces are brought together, the combined bonds between thousands to hooks and loops require much more applied force to dissociate. Similarly, if the HA (with a carboxylate functional group on each dimer) is in contact with fewer functional groups, it encounters less resistance while di using through the hydrogel. 65 Hydrogel func- tional monomer content Di usion coe - cient cm2=s Std. dev. R2 Order of Release Std. dev. R2 Nel lcon only 5.689 10 10 0.005 10 10 0.99 0.61 0.02 0.99 0.05% 4.923 10 10 0.007 10 10 0.98 0.55 0.05 0.96 0.125% 3.553 10 10 0.004 10 10 0.99 0.47 0.02 0.99 0.25% 1.797 10 10 0.001 10 10 0.99 0.50 0.02 0.99 Table 6.1: Di usion and release order of HA from Nel lcon hydrogels with varying func- tional monomer amounts, all in [1:1:2] ratio As we tailored down the functional monomer content of the prepolymer, we were successful in attaining HA release from a hydrogel containing 0.25% functional monomers by mass. Release studies were also carried out on hydrogels with 0.125% and 0.025% monomers by mass. The cumulative release rates are presented in Figure 6.4. For all other hydrogel compositions, we see a clear trend that increasing the functional monomer content in the prepolymer reduces the cumulative mass of HA released. We calculate the di usion coe cients and orders of release for these pro les according to the method described in Section 4.2. The decrease in total mass released indicates that as more monomers are included, more HA is prevented from di using out of the hydrogel. That is, a fraction of the HA is immobilized inside the hydrogel. The HA that does di use from the hydrogel has di u- sion coe cients that depend on the %-by-mass of monomer content in the hydrogel. The monomer content and di usion coe cient are strongly correlated, as shown in Figure 6.5. The graph indicates that hydrogels with monomer content less than 0.36% are likely to allow HA chains to di use through the mesh, while di usive release is not expected from hydrogels with monomer content much higher than 0.36%. This agrees with our data from the 1% and 5% monomer content hydrogels, which released negligible amounts of HA. 66 67 -50 0 50 100 150 200 250 0 20 40 60 80 100 120 140 Time (hours) Cumul a ti ve Ma ss Rel eas ed (mi c rog r a m s ) Figure 6.4: Cumulative release of HA from Nelfilcon hydrogels with different %-by- mass of functional monomers Dynamic release studies were conducted on a series of hydrogels prepared from prepolymers containing different %-by-mass of functional monomers: 0.05% (?), 0.125% (c), 0.25% ( ), 1% (?) and 5% (|). All were made with the same ratio of functional monomers: [AM : NVP : DEAEMA] ~ [1:1:2]. A hydrogel made with prepolymer containing no functional monomer is also shown (?). It appears from the release profiles that increasing the %-by-mass of functional monomers in the hydrogel decreases the release rate and the cumulative released mass of HA. 68 0.00 1.00 2.00 3.00 4.00 5.00 6.00 0 0.2 0.4 0.6 0.8 1 1.2 %-by-mass functional monomer content D i ff usion C o e ffi ci ent (x 10 10 ) cm 2 /sec 0.36 Figure 6.5: Diffusion coefficients versus %-by-mass functional monomer content There is a strong inverse correlation between the diffusion coefficients of HA from hydrogels against the %-by-mass functional monomer content in the prepolymer. The diffusion coefficient is expected to decrease to minute levels as the functional monomer content approaches 0.361%. This agrees with our data that demonstrates negligible release of HA from hydrogels with 1% and 5% functional monomer content. The above analysis demonstrates that the presence of the functional monomers tailors the di usion coe cient of HA releasing from the hydrogel. It is unclear at this point in the analysis whether the HA interacts with the hydrogel solely because of the electrostatic interaction between HA carboxylate groups and the protonated DEAEMA, or if the AM and NVP contribute to the interaction. To explore this, we produced a series of hydrogels containing 0.125% functional monomers by mass in the prepolymer mixtures, but contain- ing varying proportions of AM, NVP and DEAEMA. The compositions are summarized in Table 6.2. Relative proportion of monomers Functional monomer as %-by-mass of prepolymer mixture 0.125% [AM:NVP:DEAEMA] AM NVP DEAEMA All monomers 0.125% [0:0:1] 0 0 0.125% 0.125% 0.125% [1:1:2] 0.0313% 0.0313% 0.0624% 0.125% 0.125% [1:1:0] 0.0625% 0.0625 % 0 0.125% Table 6.2: Functional monomer content of hydrogels Release studies were conducted on these hydrogels as described in Section 7.2, and the cumulative release pro les are presented in Figure 6.6. The cumulative mass released tends to decrease as the proportion of DEAEMA is increased. This suggests that DEAEMA has a strong a nity for HA and immobilizes a larger fraction within the hydrogel. However if we compare the fractional release of HA from these hydrogels, a di erent phenomenon with a distinct trend is revealed, as seen in Figure 6.7. During the rst 0.6 fraction of the cumulative release, the di usion of HA from both 0.125% [1:1:0] and 0.125% [0:0:1] hydrogels is faster than di usion from the 0.125% [1:1:2] hydrogel. In other words, even if the mass of functional monomers in the prepolymer is kept 69 70 0 50 100 150 200 250 0 20 40 60 80 100 120 140 Time (hours) Cu mulativ e Mass Releas ed (micrograms) Figure 6.6: Cumulative release of HA from Nelfilcon hydrogels with different proportions of functional monomers Dynamic release studies were conducted on a series of hydrogels prepared from prepolymers containing different proportions of functional monomers [AM : NVP : DEAEMA], all at 0.125%-by- mass of prepolymer: [1:1:0] (?), [1:1:2] (c) and[0:0:1] (?). For comparison we also plot the release profiles of hydrogels with 0.25% [1:1:2] ( ) and hydrogels with no functional monomers (?). Cumulative mass released seems to decrease as the proportion (and hence total amount) of DEAEMA increases. 71 0 0.2 0.4 0.6 0.8 1 1.2 2040608010120140 Time (hours) F r actional Mass Re lease d Figure 6.7: Fractional release of HA from Nelfilcon hydrogels with different proportions of functional monomers Dynamic release profiles were normalized with the total amount of HA released by hydrogels prepared from prepolymers containing different proportions of functional monomers [AM : NVP : DEAEMA], all at 0.125%-by-mass of prepolymer: [1:1:0] (?), [1:1:2] (c) and[0:0:1] (?). For comparison we also plot the fractional release profiles of hydrogels with 0.25% [1:1:2] ( ) and hydrogels with no functional monomers (?). The diffusion coefficient appears to decrease with greater diversity of functional monomers in the prepolymer. constant, increasing the variety of functional monomers reduces the di usion coe cient of the HA. For comparison we also juxtapose the cumulative release pro les of 0.25% [1:1:2] and 0.125% [0:0:1] in Figure 6.8. They both contain the same amount of DEAEMA, but the former contains an additional 0.0625%-by-mass each of AM and NVP. The cumulative released mass from both hydrogels is the same, but the di usion coe cients vary. We can conclude from this that the release rates of HA can be controlled in two distinct ways - we can vary the cumulative mass released by varying the total amount of functional monomers added, and we can vary the di usion coe cient by varying the diversity of incorporated monomers. The two trends are made more apparent in Figures 6.9 and 6.10. The cumulative mass released for all 0.125% monomer compositions is compared in Figure 6.9. The composition with no DEAEMA, 0.125% [1:1:0], releases a high cumulative mass of HA, almost as much as is released by the Nel lcon hydrogel without any added monomers. As the proportion of DEAEMA is increased, the release amount decreases to the level released by 0.25% [1:1:2], which cantains all functional monomers. The di usion coe cients of HA from all 0.125% monomer compositions are com- pared in Figure 6.10. The compositions with no DEAEMA (0.125% [1:1:0]) and with only DEAEMA (0.125% [0:0:1]) have di usion coe cients close to that of Nel lcon with- out added monomers. The composition containing all monomers (0.125% [1:1:2]) has a signi cantly lower di usion coe cient: 1.5 times lower than the 0.125% [0:0:1], and 1.6 times lower than the Nel lcon without monomers. Also compare the di usion coe cients of 0.125% [0:0:1] and 0.25% [1:1:2], which contain the same amount of DEAEMA. Although the release equivalent cumulative masses of HA, their di usion coe cients di er by a factor 72 73 0 2 4 6 8 10 12 14 16 18 20 0 20 40 60 80 100 120 140 Time (hours) Cu mulativ e Mass Releas ed (micrograms) Figure 6.8: Cumulative release of HA from Nelfilcon hydrogels with the same %-by- mass of DEAEMA The hydrogels of composition 0.125% [0:0:1] (?) and 0.25% [1:1:2] ( ) contain the same amount of DEAEMA. They release similar cumulative masses of HA, but their diffusion coefficients are different. The composition with greater diversity of functional monomers has the lower diffusion coefficient. 74 0 20 40 60 80 100 120 140 160 180 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DEAEMA as a fraction of added monomers Cumulative release at 24 hr (micr ogram) Figure 6.9: 24 hour release of HA from Nelfilcon gels versus proportion of DEAEMA The amounts of HA released over 24 hours by various hydrogels containing %-by-mass of functional monomers 0.125% (c) and 0.25% ( ) were compared, along with hydrogels with no functional monomers (?). The three data points for 0.125% correspond to different proportions of DEAEMA. The 0 on the x-axis indicates [AM : NVP : DEAEMA}~[1:1:0], the 0.5 refers to [1:1:2] and the 1 refers to [0:0:1]. It is clear that increasing the amount of DEAEMA in the hydrogel decreases the cumulative mass of HA released. Furthermore, the 0.125% [0:0:1] and 0.25% [1:1:2] hydrogels contain the same amount DEAEMA, and release the same cumulative mass of HA. 75 3.55 4.82 1.80 5.69 5.31 0.00 1.00 2.00 3.00 4.00 5.00 6.00 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 DEAEMA as fraction of added monomers Diffusion Coefficient (x 10 10 ) cm 2 /sec 0.125% 0.25% Figure 6.10: Diffusion coefficients versus proportion of DEAEMA The diffusion coefficients of HA through various hydrogels containing %-by-mass of functional monomers 0.125% (c) and 0.25% ( ) were compared, along with hydrogels with no functional monomers (?). The three data points for 0.125% correspond to different proportions of DEAEMA. The 0 on the x-axis indicates [AM : NVP : DEAEMA}~[1:1:0], the 0.5 refers to [1:1:2] and the 1 refers to [0:0:1]. We can see that increasing the diversity of the functional monomers, by incorporating all of AM, NVP and DEAEMA in the hydrogel, lowers the diffusion coefficient. of 3. The di usion coe cients and orders of release of compositions discussed in relation to the monomer diversity comparison are summarized in Table 6.3. Hydrogel [AM:NVP:DEAEMA] Di usion coe cient cm2=s Std. dev. R2 Release order Std. dev. R2 Nel lcon only 5.689 10 10 0.005 10 10 0.99 0.61 0.02 0.99 0.125% [1:1:0] 4.824 10 10 0.010 10 10 0.99 0.57 0.03 0.97 0.125% [1:1:2] 3.553 10 10 0.004 10 10 0.99 0.47 0.02 0.99 0.125% [0:0:1] 5.306 10 10 0.015 10 10 0.96 0.66 0.01 0.99 0.25% [1:1:2] 1.797 10 10 0.001 10 10 0.99 0.50 0.021 0.99 Table 6.3: Di usion and release order of HA from Nel lcon hydrogels with varying func- tional monomer proportions We can explain the changes in di usion coe cients in the above experiments by refer- ring to the biomimetic imprinting process. In the prepolymer mixture containing the Nel l- con macromer, functional monomers and HA, the functional monomers position around the HA so that the interactions between the functional monomers and HA moieties decreases the free energy of the system. When the mixture is crosslinked, the functional monomers are incorporated into the hydrogel in these favorable con gurations through the pendant acrylate groups on the PVA chains. In this manner, the hydrogel is synthesized with mem- ory sites that have an a nity for HA. While interactions do occur between the HA and each functional monomer individually, the presence of all three monomers allows multiple moieties on the HA to interact with the hydrogel at one site. The multiple functional inter- actions are believed to increase the similarity of the interaction sites with the binding sites on HA-binding protein CD44, leading to enhanced a nity and lower di usion coe cients. The possibility remains that the addition of monomers resulted in a hydrogel network with a tighter mesh structure, and the di usion coe cient decreased because reptation takes longer as steric hindrance increases. To examine this possibility fully, we performed hydrogel swelling studies to obtain information about the mesh size. 76 6.4.2 E ects of heat sterilization In the manufacture process of Nel lcon A contact lenses, the crosslinked lenses are immediately sealed in foil blister-packs with 0.85 mL of bu er solution at pH 9. The sealed packs are then autoclaved at 121 C for 40 minutes. As HA is known to undergo denaturation when heated to high temperatures, we needed to assess the e ect of the high temperature sterilization procedure on the HA incorporated in the lenses. We heated aqueous solutions of HA (500 ng/mL and 10 g/mL) to temperatures above the boiling point of water for various time intervals. When the samples were assayed, the concentration of 1 million Da HA chains was lower for samples that had been heated longer. Interestingly, the % change in concentration over time was lesser for the more con- centrated solution (10 g/mL) than the more dilute solution (500 ng/mL). The change in concentration in the 10 g/mL solution was 30% over 60 minutes while the change in signal strength in the 500 ng/mL solution was 50% over 60 minutes. The percentage change in concentration versus time of heating is shown in Figure 6.11. This suggests that higher concentrations of HA may have protective e ects on the stability of the long-chain HA molecule. We conducted a dynamic release study to assess the heat e ects on the Nel lcon- HA hydrogel lenses. Comparing the heat-treated lenses with the control untreated lenses we measured similar release pro les, suggesting that heat-treatment does not denature the HA within the hydrogel. The release pro les are shown in Figure 6.12. 6.4.3 Structural analysis There are many theoretical models that enable us to obtain structural and con gu- rational information about a hydrogel from experimental data obtained by swelling and 77 78 0 10 20 30 40 50 60 10203040506070 Time (min) Percen tag e cha n g e i n c o n cen tratio n of l arge HA Figure 6.11: Comparison of stability of HA solutions under heat sterilization conditions When solutions of HA in water are heated to above 100?C, they undergo some heat degradation. In the less concentrated solution (500 ng/mL, ?) nearly 50% of the large HA degrades to shorter HA over 60 minutes, whereas in the higher concentration solution (10 ?g/mL, c), the degradation is only 30% in the same time. This suggests that higher concentrations have a protective effect on the large HA molecule. 79 0 50 100 150 200 250 0 20 40 60 80 100 120 140 Time (hours) Cumulative Mass Released (micr o gr ams) Figure 6.12: Cumulative release of HA from Nelfilcon hydrogels before and after heat-sterilization When hydrogels containing HA with Nelfilcon were heated to 121?C, the dynamic release profile of HA for such lenses (c) was similar to the release profile from hydrogels that did not undergo heat treatment (?). tensile studies. In particular we can obtain information about the mesh size of the hydro- gel and determine whether a tighter mesh is responsible for the decrease in the di usion coe cient of HA through imprinted hydrogels. The mesh size is related to the molecular weight between crosslinks according to equa- tion 4.12. The Nel lcon macromer mixture contains water as a solvent. Based on the discus- sion in Section fsec:EquilibriumSwellingTheory, the Peppas-Merrill model describes the relationship between the average molecular weight between crosslinks (Mc) and the equi- librium polymer volume fraction (v2;s) for a swollen hydrogel crosslinked in the presence of a solvent. Equilibrium swelling studies were performed on the hydrogels to determine the weight swelling and volume swelling ratios (q and Q) and the polymer volume fraction (v2;s). Q is the ratio of swollen to dry volumes of the hydrogel. q is the ratio of swollen to dry weights. v2;s is the ratio of the dry polymer volume to swollen polymer volume, and (1- v2;s) gives us the fractional water content of the hydrogel. These parameters were calculated for all hydrogels and are summarized in Table 6.4. We also illustrate some of the values in Figure 6.13. The hydrogels synthesized with HA generally have slightly higher Q and q than the hydrogels synthesized without HA. The former also have lower v2;s, indicating higher water content. Two factors in uence this di erence. First of all, the presence of HA in the pre- polymer mixture can in uence the formation of polymer chains and associated crosslinking points, making the polymer chains more mobile and increasing the hydrogels? capacity to hold water. Secondly, the residual HA in the hydrogels increases the hydrogels? capacity to hold water. 80 81 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 012345 Added %-by-mass of functional monomers Po lymer Vo lume F r actio n Figure 6.13: Comparison of polymer volume fractions for various hydrogels All hydrogels have an equilibrium polymer volume fraction that lies in the narrow range between fractions 0.23 and 0.3, and the hydrogels with released HA generally have a lower polymer volume fraction than hydrogels synthesized without HA. Hydrogel Weight swelling ratio (q) Volume swelling ratio (Q) Polymer volume fraction (v2;s) Nel lcon without added monomers Nel lcon without HA 3:06 0:04 3:35 0:07 0:30 0:01 Nel lcon with HA 3:42 0:09 4:09 0:04 0:24 0:01 Nel lcon with HA, sterilized 3:46 0:05 4:29 0:16 0:23 0:01 Nel lcon in relaxed state 2:96 0:04 3:63 0:14 0:28 0:01 Nel lcon with %-by-mass of functional monomers added (HA released) 0.05% 3:20 0:02 3:83 0:04 0:26 0:01 0.125% 3:34 0:05 4:10 0:06 0:24 0:01 0.25% 3:56 0:06 4:33 0:08 0:23 0:01 1% 3:46 0:02 4:28 0:12 0:23 0:01 5% 3:56 0:06 4:33 0:07 0:23 0:01 Nel lcon with 0.125% monomers [AM:NVP:DEAEMA] (HA released) 0.125% [0:0:1] 3:41 0:04 4:12 0:04 0:24 0:01 0.125% [1:1:2] 3:34 0:05 4:10 0:06 0:24 0:01 0.125% [1:1:0] 3:30 0:03 4:07 0:17 0:25 0:01 Nel lcon with %-by-mass of functional monomers added (No HA) Nel lcon, no monomers 3:06 0:04 3:35 0:07 0:30 0:01 1% 3:35 0:03 3:95 0:11 0:25 0:01 5% 3:24 0:02 3:61 0:13 0:28 0:01 Nel lcon with 0.125% monomers swollen in various pH conditions (HA released) pH 7 3:46 0:02 4:28 0:12 0:23 0:01 pH 12 3:39 0:03 4:26 0:10 0:24 0:01 Table 6.4: Equilibrium swelling parameters In general, all the hydrogels havev2;s falling within the range of 0.23 and 0.29 suggesting that the mesh size is similar for all the gels synthesized. In particular, we note that the swelling parameters of Nel lcon hydrogel with 1%-by-mass functional monomers remain the same at pH 7 and pH 12, indicating that the pH dependent increase in HA release described in Section 6.4.1 did not result in change in mesh size. Furthermore, the swelling parameters do not change for Nel lcon gels synthesized with HA despite heat sterilization. In Figure 6.14, we illustrate the relationship between the di usion coe cient and polymer volume fraction in gels with various %-by-mass of functional monomers. 82 83 0 1 2 3 4 5 6 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Polymer volume fraction Diffusion Coeffic i ent (10 ^ 10 cm^2/sec) Figure 6.14: Diffusion coefficients versus polymer volume fraction for Nelfilcon hydrogels with different %-by-mass of functional monomers The diffusion coefficients of hydrogels made with different %-by-mass of functional monomers are plotted against their polymer volume fractions: 0.05% (?), 0.125% (c), 0.25% ( ), and no added functional monomers (?). The diffusion coefficients vary significantly while the polymer volume fractions are limited to a narrow range. This indicates that changes in mesh size are not responsible for the changes in diffusion coefficients. The gure clearly reveals the narrow range of polymer volume fraction of the hydrogels. In contrast, the di usion constants of HA through these networks varies dramatically. The highest di usion coe cient (from the gel without functional monomers) is nearly 4 times higher than the lowest di usion coe cient (from the gel with 0.25% functional monomers). This is strong evidence that the di usion coe cients do not vary because of structural parameters such as the mesh size. To determine mesh size, further structural analysis was done through tensile testing of the sample, in which the hydrogel samples were extended at a constant rate and the tension on the sample was recorded. The extension versus applied force were plotted for four types of hydrogel samples: Nel lcon gels without added functional monomers or HA, Nel lcon with HA, Nel lcon with 0.25% by mass functional monomers, and Nel lcon with HA and 0.25% by mass functional monomers. Structural parameters for the hydrogels were obtained and summarized in Table 6.5. Hydrogel Young?s Modulus (MPa) Std. dev. Shear Modulus (MPa) Std. dev. Nel lcon 0.557 0.023 0.201 0.011 Nel lcon with HA 0.423 0.060 0.153 0.022 Nel lcon with 0.25% f.m. 0.535 0.032 0.195 0.015 Nel lcon with 0.25% f.m. and HA 0.550 0.015 0.203 0.009 Table 6.5: Tensile parameters Of the four gels, three have very similar structural parameters. Nel lcon with HA (no functional monomers) has moduli that are lower than the other hydrogels. This suggests that the presence of HA in the prepolymer without added functional monomers results in a network with longer chains between crosslinks. We base this on equation 4.15. 84 From this equation, we can calculate the molecular weight between crosslinks (Mc) in a hydrogel. The normal stress applied is , the elongation ratio is , the ratio of poly- mer volume fractions of swollen and relaxed gels is v2;s=v2;r 1, the ideal gas constant is R = 8:314472 cm3.MPa.K 1.mol 1, the temperature of experimental conditions is T = 298 K, speci c polymer volume is v = 0:909, the molecular weight of uncrosslinked polymers is Mn 50;000 Da, and the molecular weight between crosslinks, Mc, remains to be calcu- lated. This model allows us to measure the relationship between and experimentally and calculate the Mc. The slope of versus 1= 2, obtained from the tensile tests, is the Shear Modulus and enables us to calculate Mc. From Mc, we can calculate the size of the mesh between crosslinked hydrogel chains. The Mc and the mesh sizes are summarized in Table 6.6. Hydrogel Mc (g/mol) std.dev ( A) std.dev. Nel lcon 8.79 103 0.325 103 133 3 Nel lcon with HA 10.5 103 0.876 103 155 7 0.25% without HA 8.98 103 0.438 103 140 4 0.25% with HA 8.74 103 0.250 103 144 2 Table 6.6: Mesh sizes of hydrogels This con rms that the mesh of Nel lcon crosslinked with only HA has a more open network than the other hydrogels. The presence of HA along with Nel lcon macromer in the prepolymer in the absence of functional monomers appears to produce a hydrogel with a greater moleculer weight between crosslinks than Nel lcon macromer does alone. The HA seems to in uence the formation of polymer chains and associated crosslink points in the hydrogel resulting in a larger mesh size. This agrees with the results of the swelling studies, in which Nel lcon alone has a greater polymer volume fraction than Nel lcon with HA. However, the addition of functional monomers leads to a decrease in the molecular weight between crosslinks, indicating that the presence of the functional monomers decreases the 85 mesh size to an extent comparable with Nel lcon hydrogel with no added HA or functional monomers. 86 Chapter 7 Drug Release Kinetics Under Physiological Flow In a Fickian model of release kinetics, the release rate of a drug from the delivery device is proportional to the concentration gradient between the drug source and the surroundings. In practical terms this means that as the nite drug source is depleted, the rate of drug release decreases. A zero-order (i.e., independent of time) release rate is preferable because it would deliver medication at a constant rate for an extended time. The challenge is to use a nite drug source to achieve an extended zero-order release, and a number of strategies have been attempted in hydrogel drug delivery systems. Therapeutic contact lenses are swollen when inserted into the eye, and cannot ex- ploit the solvation-transition rate to control drug delivery. The other strategies also have drawbacks and cannot be easily applied to produce contact lenses. Our lab has recently demonstrated with molecular imprinting methods that one can also restrict and delay the transport of drug from the matrix via interaction of numerous functional groups with the template drug.137 In biomimetic imprinting, monomers chosen to mimic residues in the drug?s biological binding molecule are complexed non-covalently to the drug and crosslinked into a hydrogel matrix. The drug?s heightened interaction with these residue pockets slows its release from the hydrogel, exploiting a programmable memory within the polymer chains and not the free volume available for drug transport. This type of network formation - with a proper optimization of drug a nity relating to number and strength of functional monomer interactions, crosslinking structure, and mobility of polymer chains - has a strong 87 potential to in uence a number of hydrogel systems and add to the variables one can alter to control the release pro le. In e orts to understand the mechanisms behind release kinetics, various mathematical models of solvent penetration and solute release have been developed.20;82 Typically, in both modeling and experimental work, in nite sink conditions are assumed and accumulation of drug in the solution surrounding the hydrogel is considered to be negligible. This is appropriate for the majority of studied systems but for ocular drug delivery, considering the small tear volume and ow rates encountered in vivo, it does not realistically describe drug release kinetics. In these types of physiological situations, it is imperative that micro uidic models be used to characterize the release pro les. 7.1 Micro uidic Platforms for Evaluating Drug Delivery Devices Micro uidic platforms, typically dealing with 10 9 to 10 19 liters of small uid amounts, interface engineering, chemistry, and biology for conducting experiments at very small scales.89;140 For instance, solid-state silicon microchips can provide controlled release of single or multiple chemical substances on demand117 or with multi-pulse drug release from resorbable matrices.54 While there have been an increasing number of cases in the last few years of controlling drug release by the application of micro- and nanotechnology for drug administration, there has been very little to no use of micro- or nano uidic platforms in the evaluation of drug release devices. This paper describes the in vitro drug release kinetics of imprinted hydrogel contact lenses within in vivo physiological ow by developing and implementing a micro uidic chip that matches the steady state tear volume of the eye with ow rates within the physiological 88 range. This has not been demonstrated to date with any other contact lens drug delivery systems. We hypothesized that the physiological ow model of drug release would show that the therapeutic lens, under ow conditions similar to those in the human eye, would increase the release time and may provide a more linear and sustained release pro le. 7.2 Materials and Methods Acrylic acid (AA), acrylamide (AM), 2-hydroxyethylmethacrylate (HEMA), N-vinyl 2-pyrrolidone (NVP), azobisisobutyronitrile (AIBN) and ketotifen fumarate were purchased from Sigma-Aldrich (Milwaukee, WI). Polyethylene glycol (200) dimethacrylate (PEG200- DMA) was purchased from Polysciences, Inc. (Warrington, PA). All chemicals were used as received. Polymer and copolymer networks were made using various mixtures of above monomers (e.g. poly (AA- co- AM- co- HEMA- co- PEG200DMA), poly (AA- co- HEMA- co- PEG200DMA), poly (AM- co- HEMA- co- PEG200DMA), poly (AA- co- AM- co- NVP- co- HEMA- co- PEG200DMA)). 7.2.1 Synthesis of molecularly imprinted hydrogel networks Hydrogels of di ering compositions were synthesized in a temperature controlled, non- oxidative environment using free-radical UV photopolymerization. Polymer compositions consisted of 5 mole% crosslinking monomer and 95 mole% functional monomer (92 mole% of the backbone functional monomer, HEMA, and the balance 3 mole% as combinations of other functional monomers). The monomer to template ratio was optimized to achieve desired amount of drug loading. Typically, the reaction solutions consisting of monomers and template molecule were sonicated to produce a homogeneous mixture before adding the initiator. Solutions were 89 then allowed to equilibrate in darkness in order to facilitate non-covalent complexation at the molecular level. The solutions were transferred to a MBRAUN Labmaster 130 1500/1000 glovebox, where they were purged with nitrogen until oxygen levels were 0.1 to 10 ppm. The solutions were pipetted into trichloromethylsilane-treated glass moulds (6" 6"), separated by a Te on frame of varying thicknesses. The polymerization reaction was allowed to run for ten minutes with light intensity, measured using a radiometer (In- ternational Light IL1400A), equal to 40 mW/cm2 (Dymax UV ood light) at a constant temperature of 36 C. Control gels were prepared without the template molecule, following similar steps. The polymerized gels were removed from the nitrogen atmosphere, sub- merged into a deionized water bath (Millipore, 18.2 m cm, pH 6), and carefully peeled from the glass surface. Circular discs of 13.5 mm and 400 m or 700 m were cut with a cork borer. Polymer discs were then washed for several days with deionized water until ketotifen fumarate could no longer be detected by spectroscopic monitoring (Biotek UV-Vis Spectrophotometer). The lenses were then loaded with drug by equilibrating in ketotifen fumarate solution. 7.2.2 Dynamic Therapeutic Release Studies Kinetic release studies were conducted in arti cial lacrimal uid (6.78 g/L NaCl, 2.18 g/L NaHCO3, 1.38 g/L KCl, 0.084 g/L CaCl2.2H2O, pH 859). In the in nite sink studies, gels which had been reloaded with drug were placed in 30 ml of uid which was continuously agitated with a Servodyne mixer (Cole Palmer Instrument Co.) at 120 rpm. Preliminary experiments were conducted to determine the amount of uid needed to approximate in - nite sink analysis by comparing release rates for a xed amount of uid versus release rates when refreshing the uid at speci c various time intervals. Release of drug was monitored 90 at 268 nm by drawing 200 L of uid into a 96-well Corning Costar UV-transparent mi- croplate, and measurements were taken in a Synergy UV-Vis/Fluorescence/Luminescence Spectrophotometer (Biotek). Absorbances were recorded for three samples and averaged. Solutions were replaced after each reading. In the physiological ow studies, the drug- loaded disk was placed within the chamber of the micro uidic device. A KDS101 Infusion Pump from KD Scienti c (Holliston, MA) injected lacrimal uid into the chamber at 3 L/min, while an outlet line removed uid from the chamber at the same rate for collec- tion at regular time intervals. Release of drug for two samples was monitored similarly to the in nite sink case. 7.2.3 Micro uidic Chip Design and Fabrication The micro uidic chips were fabricated by soft lithography.67;68;66 Masks de ning mi- cro uidic features were designed with AutoCAD 2006 (Auto Desk) and photoplotted on transparencies at the resolution of 4000/5000 dots per inch (DPI) through a commercial printing company (CAD/ART Services, Bandon, Oregon). The transparent masks were used to fabricate microstructures of photoresists on a four-inch silicon wafer with pho- tolithography, an established method for semiconductor production. Two di erent pho- toresists, SU-8 2025 (Microchem Co., Newton, MA) for 50 m thick uidic channels and SU-8 2100 (Microchem Co.) for a 550 m very-thick chamber, were used. For the third step, based on the microstructures on a silicon wafer, the micro uidic layer was made out of transparent silicone polymer, polydimethylsiloxane (PDMS). A drug loaded lens was placed in the central chamber and the chip was sealed against a glass plate. To ensure reproducible ow rates and to limit non-speci c adsorption to the device, each device had a lifetime of three- ve runs or less, as determined by monitoring uid leakage (e.g., the 91 surface free energy decreases with time and the forces holding the chip and plate together weaken) and adsorption in separate experiments. All the processes for manufacturing the devices, including the chip design, mold fabri- cation, and chip fabrication were carried out in the Alabama Micro-electronics Science and Technology Center (AMSTC) and Nano/Micro uidics Laboratory. The structure is shown in Figure (7.1). The device is designed to mimic the ow rate of tears but does not fully reproduce other ocular conditions. While the device was operated at ambient temperature, ocular physiological temperature is 35 C and will increase the di usional transport. In the human eye, the mixing and ow of tears is complicated by the presence of contact lenses. The tear lm is thinner and varies in thickness by evaporation between blinks and tear breakup. These factors may also a ect drug release, but relative to tear ow rate these e ects are small.115;123 In future versions of the device, we plan to reproduce more of these conditions in vitro. 7.2.4 Dynamic Weight/Volume Swelling Studies and Partition Coe cients Recognitive and control gels were dried at room temperature for 24 hours, followed by vacuum drying (T=30 C, 28 in. Hg vacuum) until no change in dry weight was observed (i.e., less than 0.1 weight percent di erence). Dry samples manufactured with and without drug (n=3) were placed in a constant volume of deionized water at 25 C. The gels were weighed by removing the gels from the swelling media at speci c time points and blotting with absorbent, lint-free tissue to remove excess surface solvent. When the samples reached equilibrium water uptake the weight swelling ratio at equilibrium (q) (the weight of the swollen polymer divided by the weight of the dry polymer at equilibrium) was calculated. 92 93 Figure 7.1: Structure of microfluidic device for drug delivery evaluation (a) Schematic of the experimental set-up for contact lens drug delivery evaluation. The hydrogel is placed in the microfluidic chamber between the four posts, lacrimal fluid is flowed through the chip and drug release is measured. (b) Example of microliter/nanoliter flow control and near-plug flow profile in the microfluidic chip (two different food dyes were used). (c) Length of scale of microfluidic chip. The equilibrium volume swelling ratio (Q) was calculated as the ratio of the swollen gel volume at equilibrium to the volume of the dry polymer. The volume of the gel in the swollen or dry state was obtained by determining its weight in air and in n-heptane, a non-solvent for the polymer, and calculated using Archimedes buoyancy principle. The partition coe cients of the gels (calculated as the ratio of the drug concentration in the gel to the equilibrium drug concentration in solution) were obtained by immersing gels in ketotifen fumarate solution and obtaining the concentrations via mass balances. Aqueous solubility was measured by saturating ketotifen fumarate in deionized water and stirring overnight. The solution was adjusted to pH 7 and ltered. The concentration was measured by absorbance at 268 nm against a series of ketotifen standards. Log P octanol/water was calculated as the logarithm of the ratio of the equilibrium concentrations of ketotifen in octanol to ketotifen in water. A volume of 4 mL of a known concentration of aqueous ketotifen solution was shaken with 4 mL of octanol for 24 hours and then let rest for 24 hours. Concentration of ketotifen in octanol was obtained by mass balance. 7.3 Results and Discussion Among the in nite sink dynamic release studies, the most structurally functional net- work, poly(AA-co-AM-co-NVP-co-HEMA-PEG200DMA), exhibited an extended release pro le for a duration of 5 days within arti cial lacrimal solution (80% of drug was re- leased in approximately 4 days) (Figure 7.2). This system also demonstrated the high- est loading.137 Other less functionalized systems demonstrated controlled release for ap- proximately one day. The partition coe cients are 5.65, 7.13, 18.06 and 45.05 for poly 94 (AA-co-HEMA-PEG200DMA), poly (AM-co-HEMA-PEG200DMA), poly (AA-co-AM-co- HEMA-PEG200DMA) networks and poly (AA-co-AM-co-NVP-HEMA-PEG200DMA) net- works, respectively when the gels are placed in solutions of ketotifen fumarate in deionized water of concentration 0.4 mg/mL . Additionally, the bound concentrations of ketotifen within the gels are 5.1 10 3mmol/g, 7.4 10 2, and 4.9 10 2 mmol/g for poly (AA-co- HEMA-PEG200DMA), poly (AM-co-HEMA-PEG200DMA), poly (AA-co-AM-co-HEMA- PEG200DMA) networks and poly (AA-co-AM-co-NVP-HEMA-PEG200DMA) networks respectively. Ketotifen fumarate is hydrophilic with log10 octanol-water partition coe cient of -0.3, and an aqueous solubility of 3.4 mg/mL at pH 7 0:2 and 20 C. This indicates that molecular imprinting and multiplicity of interactions have a greater in uence on binding than general hydrophobic interactions. Figure Phase Release Rate (mg/min) Duration of Phase A 3:04 10 3 120 min B 2:97 10 4 120 - 2280 min (2 hrs to 2 days) C 1:34 10 4 2880 - 10620 min (2 days to 7.3 days) Table 7.1: Varying ketotifen release rates from AA-AM-NVP lenses under in nite sink conditions Within a Fickian di usion process, the fractional mass released depends linearly on t0:5/L at short times or fractional release less than 0.67 with a slope directly proportional to the di usion coe cient. Poly (AA-co-AM-co-NVP-co-HEMA-PEG200DMA) networks exhibited a ketotifen fumarate di usion coe cient of 5.57 10 10 cm2/sec, which was a factor of 9, 7.2, and 13.8 less than poly (AA-co-HEMA-PEG200DMA), poly (AM-co- HEMA-PEG200DMA), and poly (AA-co-AM-co-HEMA-PEG200DMA) networks, respec- tively (Figure 7.3, Table 7.2). These results show that therapeutic release can be tailored via the memory within the polymer chains, through the arrangement, type, and amount 95 96 Figure 7.2: Cumulative release of ketotifen fumarate from HEMA lenses in infinite sink conditions Cumulative mass released from the poly (AA-co-AM-co-NVP-co-HEMA-co-PEG200DMA) lens. Release rates were calculated for the three phases of release under infinite sink conditions. A was calculated 3.04 ? 10 -3 mg/min over the first 120 minutes, B was 2.97 ? 10-4 mg/min from 120 minutes to 2 days, and C was 1.34 ? 10-4 mg/min from 2 days to 7.3 days. of functionality. This is signi cant considering all hydrogels exhibited equilibrium volume swelling ratios that were in close agreement with one another suggesting similar structures available for free volume transport. Also, the monomers PEG200DMA and HEMA make up 97% of the feed monomers in each gel, and reaction analysis indicated that most of the double bonds within the systems reacted.137 Because the hydrogels were produced without solvent, the Flory-Rehner equation105 can be used with the experimentally derived equilibrium swollen polymer volume fraction to determine network structural parameters such as the molecular weight between adjacent crosslinks, and also the correlation length or network mesh size. Therefore the polymer volume fraction can be used as an indicator that correlates with structural parameters. Figure 7.4 highlights the signi cant di erence in the di usion coe cient of the di erent gels as compared to the equilibrium polymer volume fraction in the swollen state. Once again, the equilibrium polymer volume fraction in the swollen state was not statistically di erent between all the gels. In order to gauge the appropriateness of the t to a Fickian mechanism, the log of the fractional drug release was plotted against the log of time. The exponents of all gels in the in nite sink release indicated that they were in agreement with a Fickian di usion mechanism, where the values of n are approximately equal to 0.5 (Table 7.2). The fractional release at physiological ow rates for the poly (AA-co-AM-co-NVP- co-HEMA-PEG200DMA) networks shows that under physiological ow conditions, drug is released in a linear manner and at much lower concentrations than conventional in nite sink release studies suggested, indicating that such hydrogel lenses have the capacity to deliver sustained amounts of drug in a constant manner over an extended time period as shown in Figure 7.5. Results demonstrate a slower release of drug with a constant, zero-order rate 97 98 Figure 7.3: Fractional release of ketotifen from various HEMA lenses under infinite sink conditions Fractional release profiles of therapeutic contact lenses for poly (n-co-HEMA-co-PEG200DMA) networks in artificial lacrimal fluid at 25?C, where n is AA (?), AM (?), AA-co-AM (?) or AA-co- AM-co-NVP (?) imprinted networks. The abscissa is time normalized to the square of thickness, (due to slab geometry) as the thicknesses of the gels differed; 700 ?m (?, ?, ?) and 400 ?m (?). (n=3). 99 Figure 7.4: Diffusion coefficients versus polymer volume fraction in hydrogels releasing ketotifen The change in transport characteristics is related to the imprinting process rather than a change in overall polymer mesh structure for poly (n-co-HEMA-co-PEG200DMA) networks where n is AA (?), AM (?), AA-co-AM (?) or AA-co-AM-co-NVP (?). Network Type (n) Di usion coe cient ( 1010cm2=s) R2 Release order R2 Q v2;s AA1 50:2 4:8 0:958 0:459 0:041 0:833 1:547 0:027 0:647 0:011 AM1 40:4 1:9 0:989 0:620 0:028 0:966 1:544 0:051 0:648 0:021 AA-co-AM1 77:3 3:5 0:991 0:521 0:025 0:965 1:513 0:094 0:661 0:041 AA-co-AM- co-NVP1 5:57 0:31 0:983 0:406 0:022 0:946 1:646 0:136 0:608 0:050 AA-co-AM- co-NVP2 N/A N/A 0:981 :006 0:997 1:646 0:136 0:608 0:050 Superscripts: [1] In nite sink model, [2] Physiological ow model Table 7.2: Summary of ketotifen di usion coe cients, orders of release and swelling data of release for approximately 3 1/2 days (i.e., independent of concentration or time). For the physiological release case, using the empirical power law equation indicates the order of release n 1 is equal to 0:019 0:006 for the physiological ow case. Therefore, zero- order release is achieved by reducing the concentration gradient through the accumulation of ketotifen in the slow moving uid at simulated physiological tear turnover rates. The importance of matching physiological ow is crucial to the characterization of this delivery system. Compared to the in nite sink release pro le, the cumulative mass released under phys- iological conditions is reduced by a large amount and is dependent on the volumetric ow rates. In 3.5 days, 45 g is released at a constant rate of 12.9 g/day compared to approx- imately 1,200 g in the in nite sink release study which shows decreasing rates of release. This is a decrease of a factor of 27. These lenses are about 3-4 times thicker than conven- tional contact lenses and normalizing for the di erence in thickness, with fractional mass released being proportional to the inverse of the square of the thickness, yields 0.8 to 1.4 g/day release. Conventional ketotifen topical drops deliver approximately 1-1.5 g/day based on the recommended dosage regimen (assuming the maximum 7% bioavailability) 100 101 Figure 7.5: Fractional release of ketotifen fumarate from HEMA lenses under physiological flow The fraction of ketotifen fumarate released from poly(AA-co-AM-co-NVP-co-HEMA-co- PEG200DMA) lenses in artificial lacrimal fluid at 25?C via steady in vitro physiological flow at 3 ?L/min using the microfluidic device (sample size n=2). within a typical topical peak and valley pro le (i.e., ZaditorTM ophthalmic solution con- centration is 0.345 mg ketotifen fumarate/mL; assuming a typical eyedrop volume of 20 L per drop at the recommended frequency of administration of 2-3 drops per day yields 1-1.5 g/day at 7% bioavailability). Therefore, there is strong potential to release therapeuti- cally relevant concentrations of drug from a contact lens platform. However, it is important to note that these calculations should be taken as estimates since the lens release studies were conducted at a temperature of 25 C and not 35 C, the temperature of the eye. At physiological temperature the release of drug will be faster from the lens. This long, constant duration of release of a therapeutic dosage from a contact lens has not been demonstrated previously and is inherently linked to nite tear ow rates. Considering we know the total amount that is delivered from these lenses with the in nite sink study (depending the initial loading concentration), we have delivered less than 5% of the loaded drug in 3.5 days in the physiological ow case. Thus the potential to deliver for extended time periods much greater than a week could be possible with a contact lens platform, depending on the tolerability of the lenses in the eye. Furthermore, the slowed depletion of the drug concentration in the gel and the non-zero concentration of the drug in the surrounding uid produce e ectively zero-order release kinetics. A near zero-order release pro le could be maintained for a much greater amount of time, for at least a week. These studies demonstrate that it is imperative to evaluate hydrogel release kinetics with a non-zero bulk concentration if nite turnover conditions hold true for the in vivo environment. The results clearly demonstrate that release from imprinted lenses is further delayed in an in vitro environment by matching the ocular volumetric ow rates. This e ect may be due to two reasons - nite tear turnover rates which lead to signi cant boundary layers compared to the nite sink case, and molecular imprinting strategies which lead 102 to delayed release kinetics despite equivalent e ective volume of transport through the polymer chains. Within the device, both lens surfaces are in contact with the owing uid and the lens rests on the bottom surface of the device. The average velocity in the device is 5.5 10 4m/s considering a cross-sectional area of 8.69 105 m2 (560 m height by 1,600 m width) and a volumetric ow rate of 3 L/min. This translates to a Reynolds number of approximately 2-8 along the length of the lens (using physical properties of water at 25 ), indicating laminar ow. At these velocities mass transfer e ects in the uid cannot be neglected and signi cant boundary layers will exist. The micro uidic release rate is lower and more constant compared to in nite sink conditions. Within in nite sink conditions there is su cient uid volumes producing a maximum concentration driving force and stirring which disrupts boundary layers. It is clear that boundary layer e ects are important to the di erences in release comparing the micro uidic and the in nite sink cases. It is premature to ascertain the e ect of imprinting on these results, but the in nite sink case and polymer volume fraction analysis highlights a potential mechanism for delayed transport via imprinting. The mechanism of delayed transport due to imprinting is hypothesized to consist of multiple on-o binding interactions between ketotifen and the memory \sites" consisting of multiple functionality within the network. Thus as ketotifen moves through the network its transport is slowed down due to interactions with the polymer chains. Also, molecular imprinting strategies pull the technology into a clinically signi cant reality by enhancing the therapeutic loading, which is crucial for achieving therapeutically relevant delivery levels. 137 Further control of the tear ow rate with a partially non-wetted surface, mixing similar to the ow pro les induced by the lid blinking process, appropriate tear lm, and addition of protein and lipid to the arti cial lacrimal uid are warranted for further study in the 103 near future, and will ultimately con rm the high potential for drug releasing contact lenses. However the results from the device, as presented here, are a much better approximation to the actual ocular conditions than the in nite sink model. 104 Chapter 8 Conclusion The release data of hyaluronic acid from imprinted Nel lcon hydrogels reveals inter- esting relationships between the amount and variety of functional monomers added to the prepolymer. The cumulative mass of HA released and the di usion coe cient decrease with increasing functional monomer composition. The cumulative mass released under no functional monomer addition is 200 g, with a di usion coe cient of 5.689 10 10 cm2/s. When the functional monomer content increases to 0.25% of the prepolymer by mass, the mass released is 17.5 g with a di usion coe cient of 1.797 10 10 cm2/s. The HA mass released decreases as the proportion of DEAEMA increases, but the di usion coe cient is lowest in the presence of all functional monomers. The hydrogel containing DEAEMA exclusively has a cumulative release mass of 16.2 g and a di usion coe cient of 5.306 10 10 cm2/s. The hydrogel containing AM and NVP with no DEAEMA has a cumula- tive released mass of 138 g and a di usion coe cient of 4.824 10 10 cm2/s. The hydrogel with all three functional monomers has a cumulative released mass of 54 g and di usion coe cient of 3.553 10 10 cm2/s. Additionally, the di usion coe cient is negligible when functional monomer %-by-mass greater than 0.36%. The optimum composition for functional monomers would release the maximum amount of HA, so that less of the HA initially added to the system would be permanently entrapped. It would also have a low di usion coe cient so that the release rate over the rst 24 hours would be more linear. 105 Future work on the system could potentially optimize the amounts and proportions of the various functional monomers to determine what hydrogel composition would release higher amounts of HA with lower di usion coe cients in order to extend the release over longer periods of time. The optimum monomer content would be between 0.25% and 0.361% by mass of the prepolymer, while the fraction of DEAEMA among all functional monomers would be between 0 and 0.5. Before practical implementation of the HA releasing imprinted hydrogels, assurances need to be made that the monomers, particularly acrylamide, will not have a toxic e ect on the ocular tissues. On potential strategy is to synthesize oligomers of the required functional monomers, and then attach these to the PVA chains of the Nel lcon macromer. In keeping with the technique used to functionalize PVA, the oligomers could be modi ed with an aldehyde or ketone moiety on one of the oligomer terminals. A transacetylation reaction could attach the oligomers to the PVA as pendant groups at de ned intervals. The new macromer structure could be optimized by varying the intervals between oligomer attachments. Furthermore, the oligomers could be varied by length of the chain, and the nature and size of the repeating unit. A useful design goal would be to synthesize an oligomer that folds to form a binding site similar to that found in hyaluronic acid binding protein. Future improvements in the micro uidic device for simulation of ocular environment for drug delivery devices would bring the device closer to ocular conditions in terms of temperature, blink forces, protein content in the tear uid and other factors. 106 Bibliography [1] Acosta, M C, J Gallar, and C Belmonte. \The in uence of eye solutions on blinking and ocular comfort at rest and during work at video display terminals." Experimental Eye Research 68, 6: (1999) 663{669. Clinical Trial. [2] Aggarwal, Deepika, and Indu P. Kaur. \Improved pharmacodynamics of timolol maleate from a mucoadhesive niosomal ophthalmic drug delivery system." International Journal of Pharmaceutics 290, 1-2: (2005) 155{159. [3] Aktas, Yesim, Nursen Unlu, Mehmet Orhan, Murat Irkec, and A Atilla Hincal. \In u- ence of hydroxypropyl beta-cyclodextrin on the corneal permeation of pilocarpine." Drug Development and Industrial Pharmacy 29, 2: (2003) 223{230. Comparative Study. [4] Albert, Daniel M., and Diane D. Edwards. The History of Ophthalmology. Cambridge MA: Blackwell Science, 1996. [5] Alvarez-Lorenzo, Carmen, and Angel Concheiro. \Molecularly imprinted polymers for drug delivery." Journal of Chromatography B 804, 1: (2004) 231{245. [6] Alvarez-Lorenzo, Carmen, Haruyiki Hiratani, Jose Luis Gomez-Amoza, Ramon Martinez- Pacheco, Consuelo Souto, and Angel Concheiro. \Soft contact lenses capable of sustained delivery of timolol." Journal of Pharmaceutical Sciences 91, 10: (2002) 2182{2192. [7] Alvarez-Lorenzo, Carmen, Fernando Yanez, Rafael Barreiro-Iglesias, and Angel Concheiro. \Imprinted soft contact lenses as nor oxacin delivery systems." Journal of Controlled Re- lease 113, 3: (2006) 236{244. [8] Aragona, P, F Ferreri G Di Stefano, R Spinella, and A Stilo. \Sodium hyaluronate eye drops of di erent osmolarity for the treatment of dry eye in Sjgren?s syndrome patients." The British journal of ophthalmology 86, 8: (2002) 879{884. [9] Aragona, Pasquale, Vincenzo Papa, Antonio Micali, Marcello Santocono, and Giovanni Milazzo. \Long term treatment with sodium hyaluronate-containing arti cial tears reduces ocular surface damage in patients with dry eye." British Journal of Ophthalmology 86, 2: (2002) 181{184. Clinical Trial. [10] Armaly, Monsour F., and K.R. Rao. \The E ect of Pilocarpine Ocusert with Di erent Release Rates on Ocular Pressure." Investigative Ophthalmology 12, 7: (1973) 491{496. [11] Baeyens, V, O Felt-Baeyens, S Rougier, S Pheulpin, B Boisrame, and R Gurny. \Clinical evaluation of bioadhesive ophthalmic drug inserts (BODI) for the treatment of external ocu- lar infections in dogs." Journal of Controlled Release 85, 1-3: (2002) 163{168. Comparative Study. [12] Bajorath, J, B Green eld, S B Munro, A J Day, and A Aru o. \Identi cation of CD44 residues important for hyaluronan binding and delineation of the binding site." Journal of Biological Chemistry 273, 1: (1998) 338{343. 107 [13] Bartlett, Jimmy D., Keith Boan, David Corliss, and Ian B. Gaddie. \E cacy of silicone punctal plugs as adjuncts to topical pharmacotherapy of glaucoma{a pilot study. Punctal Plugs in Glaucoma Study Group." Journal of the American Optometric Association 67, 11: (1996) 664{668. Clinical Trial. [14] Bejjani, Riad Antoine, David BenEzra, Hagit Cohen, Jutta Rieger, Charlotte Andrieu, Jean-Claude Jeanny, Gershon Gollomb, and Francine F Behar-Cohen. \Nanoparticles for gene delivery to retinal pigment epithelial cells." Molecular Vision 11: (2005) 124{132. [15] Berkow, Robert, Mark H. Beers, Robert M. Bogin, and Andrew J. Fletcher. The Merck Manual of Medical Information. Merck and Co., 1997. [16] Bernkop-Schnurch, Andreas. \Thiomers: a new generation of mucoadhesive polymers." Advanced Drug Delivery Reviews 57, 11: (2005) 1569{1582. [17] Bhler, Niklaus, Hans-Peter Haerr, Manfred Hofmann, Christine Irrgang, Andreas Mhlebach, Beat Mller, and Fredrich Stockinger. \Nel lcon A, a New Material for Contact Lenses." Chimia 53: (1999) 269{274. [18] Blondeau, J.M., P. Hedlin, and S.D. Borsos. \The antimicrobial activity of Gati oxcin (GAT) with or without Benzalkonium Chloride (BAK) against ocular bacterial pathogens." In Association for Research in Vision and Ophthalmology Annual Meeting. Fort Lauderdale, FL, USA, 2005. [19] Bourlais, Chrystele Le, Liliane Acar, Hosein Zia, Pierre A. Sado, Thomas Needham, and Roger Leverge. \Ophthalmic drug delivery systems{recent advances." Progress in Retinal Eye Research 17, 1: (1998) 33{58. [20] Brazel, Chris S, and Nicholas A. Peppas. \Modeling of drug release from swellable poly- mers." European journal of pharmaceutics and biopharmaceutics 49, 1: (2000) 47{58. [21] Brennan, Noel A. \Corneal oxygenation during contact lens wear: comparison of di usion and EOP-based ux models." Clinical and Experimental Optometry 88, 2: (2005) 103{108. Comparative Study. [22] Brignole, Francoise, Pierre-Jean Pisella, Benedicte Dupas, Vincent Baeyens, and Christophe Baudouin. \E cacy and safety of 0.18with moderate dry eye syndrome and super cial keratitis." Graefes Arch Clin Exp Ophthalmol 243, 6: (2005) 531{538. [23] Byrne, Mark E, Kinam Park, and Nicholas A Peppas. \Molecular imprinting within hydro- gels." Advanced Drug Delivery Reviews 54, 1: (2002) 149{161. [24] Camillieri, Giovanni, Claudio Bucolo, Settimio Rossi, and Filippo Drago. \Hyaluronan- induced stimulation of corneal wound healing is a pure pharmacological e ect." Journal of Ocular Pharmacology and Therapeutics 20, 6: (2004) 548{553. Comparative Study. [25] Cao, Y, C Zhang, W Shen, Z Cheng, LL Yu, and Q Ping. \Poly(N-isopropylacrylamide)- chitosan as thermosensitive in situ gel-forming system for ocular drug delivery." Journal of Controlled Release 215: (2001) 101111. JOURNAL ARTICLE. 108 [26] Chalmers, Robin L, and Carolyn G Begley. \Dryness symptoms among an unselected clinical population with and without contact lens wear." Cont Lens Anterior Eye 29, 1: (2006) 25{30. Comparative Study. [27] Chetoni, P, S Burgalassi, D Monti, and M F Saettone. \Ocular toxicity of some corneal pene- tration enhancers evaluated by electrophysiology measurements on isolated rabbit corneas." Toxicology In Vitro 17, 4: (2003) 497{504. In Vitro. [28] Clark, Abbot F., and Thomas Yorio. \Ophthalmic Drug Discovery." Nature Reviews Drug Discovery 2, 6: (2003) 446{459. [29] Clegg, John P., Julian F. Guest, Almut Lehman, and Andrew F. Smith. \The annual cost of dry eye syndrome in France, Germany, Italy, Spain, Sweden and the United Kingdom among patients managed by ophthalmologists." Ophthalmic Epidemiology 13, 4: (2006) 263{274. Comparative Study. [30] Crank, J. The Mathematics of Di usion. Oxford University Press, 1975. [31] DiColo, G, S Burgalassi, P Chetoni, M P Fiaschi, Y Zambito, and M F Saettone. \Gel- forming erodible inserts for ocular controlled delivery of o oxacin." International Journal of Pharmaceutics 215, 1-2: (2001) 101{111. [32] Doane, Marshall G. \Blinking and the mechanics of the lacrimal drainage system." Oph- thalmology 88, 8: (1981) 844{851. [33] Duvvuri, Sridhar, Soumyajit Majumdar, and Ashim K. Mitra. \Role of Metabolism in Ocular Drug Delivery." Current Drug Metabolism 5, 6: (2004) 507{515. [34] Ebrahim, Shehab, Gholam A Peyman, and Paul J Lee. \Applications of liposomes in ophthalmology." Survey of Ophthalmology 50, 2: (2005) 167{182. [35] Elder, Melissa. \US Market for Prescription Ophthalmic Drugs, Devices, Diagnostics, and Surgical Equipment." Technical report, Business Communications Research., 2006. [36] Eljarrat-Binstock, Esther, and Abraham J Domb. \Iontophoresis: A non-invasive ocular drug delivery." Journal of Controlled Release 110, 3: (2006) 479{489. [37] Eljarrat-Binstock, Esther, Frederik Raiskup, David Stepensky, Abraham J Domb, and Joseph Frucht-Pery. \Delivery of gentamicin to the rabbit eye by drug-loaded hydrogel iontophoresis." Investigative Ophthalmology and Visual Science 45, 8: (2004) 2543{2548. [38] Felt, O, P Furrer, J M Mayer, B Plazonnet, P Buri, and R Gurny. \Topical use of chitosan in ophthalmology: tolerance assessment and evaluation of precorneal retention." International Jouornal of Pharmaceutics 180, 2: (1999) 185{193. Comparative Study. [39] Fialho, S.L., and A. da Silva-Cunha. \New vehicle based on a microemulsion for topical ocular administration of dexamethasone." Clinical and Experimental Ophthalmology 32, 6: (2004) 626{32. 109 [40] Foulks, Gary N. \What is dry eye and what does it mean to the contact lens wearer?" Eye Contact Lens 29, 1 Suppl: (2003) 96{100. [41] Foulks, Gary N. \The correlation between the tear lm lipid layer and dry eye disease." Survey of Ophthalmology 52, 4: (2007) 369{374. [42] Fox, Robert I. \Sjogren?s syndrome." Lancet 366, 9482: (2005) 321{331. [43] Frenkel, Ronald E P, Lakshmi Mani, Allison R Toler, and Max P C Frenkel. \Intraocular pressure e ects of pegaptanib (Macugen) injections in patients with and without glaucoma." American Journal of Ophthalmology 143, 6: (2007) 1034{1035. [44] Friedman, D. S., N. Congdon, J. Kempen, and J. M. Tielsch. \Vision Problems in the U.S.: Prevalence of Adult Vision Impairment and Age-Related Eye Disease in America." Technical report, National Eye Institute and Prevent Blindness America, 2002. [45] Frucht-Pery, Joseph, Frederik Raiskup, Hadas Mechoulam, Mervyn Shapiro, Esther Eljarrat-Binstock, and Abraham Domb. \Iontophoretic treatment of experimental pseu- domonas keratitis in rabbit eyes using gentamicin-loaded hydrogels." Cornea 25, 10: (2006) 1182{1186. [46] Gavini, Elisabetta, Patrizia Chetoni, Massimo Cossu, Maria Gemma Alvarez, Marco Fab- rizio Saettone, and Paolo Giunchedi. \PLGA microspheres for the ocular delivery of a peptide drug, vancomycin using emulsi cation/spray-drying as the preparation method: in vitro/in vivo studies." European Journal of Pharmaceutics and Biopharmaceutics 57, 2: (2004) 207{212. Comparative Study. [47] Geroski, Dayle H., and Henry F. Edelhauser. \Drug delivery for posterior segment eye disease." Investigative ophthalmology & visual science 41, 5: (2000) 961{964. [48] Gilbard, J.P. \Human tear lm electrolyte concentrations in health and dry-eye disease." International Ophthalmology Clinics 34: (1994) 27{36. [49] Gilbert, Jake A, Amanda E Simpson, David E Rudnick, Dayle H Geroski, Thomas M Jr Aaberg, and Henry F Edelhauser. \Transscleral permeability and intraocular concentrations of cisplatin from a collagen matrix." Journal of Controlled Release 89, 3: (2003) 409{417. In Vitro. [50] Glasson, M.J., F. Stapleton, L. Keay, and M.D.P.Willcox. \The e ect of short term contact lens wear on the tear lm and ocular surface characteristics of tolerant and intolerant wearers." Contact lens & anterior eye : the journal of the British Contact Lens Association 29, 1: (2006) 41{47. Comparative Study. [51] Gokulgandhi, Mitan R., Dharmesh M. Modi, and Jolly R. Parikh. \In Situ Gel Systems for Ocular Drug Delivery: A Review." Drug Delivery Technology 7, 3: (2007) 30{37. [52] Gomes, J A P, R Amankwah, A Powell-Richards, and H S Dua. \Sodium hyaluronate (hyaluronic acid) promotes migration of human corneal epithelial cells in vitro." British Journal of Ophthalmology 88, 6: (2004) 821{825. 110 [53] Grass, G M, R W Wood, and J R Robinson. \E ects of calcium chelating agents on corneal permeability." Investigative Ophthalmology and Visual Science 26, 1: (1985) 110{113. [54] Grayson, Amy C., Insung S. Choi, Betty M. Tyler, Paul P. Wang, Henry Brem, Michael J. Cima, and Robert Langer. \Multi-pulse drug delivery from a resorbable polymeric microchip device." Nature Materials 2, 11: (2003) 767{72. [55] Greaves, J.L., and C.G. Wilson. \Treatment of diseases of the eye with mucoadhesive delivery systems." Advanced Drug Delivery Reviews 11, 3: (1993) 349{383. [56] Gudeman, Linda F., and Nikolaos A. Peppas. \pH-Sensitive membranes from poly(vinyl alcohol) / poly(acrylic acid) interpenetrating networks." Journal of Membrane Science 107: (1995) 239{248. [57] Gulsen, Derya, and Anuj Chauhan. \Dispersion of microemulsion drops in HEMA hydrogel: a potential ophthalmic drug delivery vehicle." International Journal of Pharmaceutics 292, 1-2: (2005) 95{117. Comparative Study. [58] Gupta, Suresh K, Viney Gupta, Sujata Joshi, and Radhika Tandon. \Subclinically dry eyes in urban Delhi: an impact of air pollution?" Ophthalmologica 216, 5: (2002) 368{371. [59] Haeringen, N.J. Van. \Clinical Biochemistry of Tears." Survey of Ophthalmology 26, 2: (1981) 84{96. [60] Hamano, T, K Horimoto, M Lee, and S Komemushi. \Sodium hyaluronate eyedrops enhance tear lm stability." Japanese Journal of Ophthalmology 40, 1: (1996) 62{65. [61] Hayes, Valerie Y, Cristina M Schnider, and Jane Veys. \An evaluation of 1-day disposable contact lens wear in a population of allergy su erers." Contact Lens and Anterior Eye 26, 2: (2003) 85{93. [62] Hilt, J Zachary, and Mark E Byrne. \Con gurational biomimesis in drug delivery: molecular imprinting of biologically signi cant molecules." Adv Drug Deliv Rev 56, 11: (2004) 1599{ 1620. [63] Hiratani, Haruyuki, Yuri Mizutani, and Carmen Alvarez-Lorenzo. \Controlling drug re- lease from imprinted hydrogels by modifying the characteristics of the imprinted cavities." Macromol Biosci 5, 8: (2005) 728{733. [64] Holly, F J, and M A Lemp. \Tear physiology and dry eyes." Survey of Ophthalmology 22, 2: (1977) 69{87. [65] Holly, Frank J. \Tear Film Physiology." American Journal of Optometry and Physiological Optics 57, 4: (1980) 252{257. [66] Hong, Jong Wook, Yan Chen, W. French Anderson, and Stephen R. Quake. \Molecular Biology on a Micro uidic Chip." Journal of Physics: Condensed Matter 18, 18: (2006) S691{S701. 111 [67] Hong, Jong Wook, and Stephen R. Quake. \Integrated Nanoliter Systems." Nature Biotech- nology 21, 10: (2003) 1179{1183. [68] Hong, Jong Wook, Vincent Studer, Giao Hang, W French Anderson, and Stephen R Quake. \A nanoliter-scale nucleic acid processor with parallel architecture." Nature Biotechnology 22, 4: (2004) 435{439. Letter. [69] Hori, Yuichi, Pablo Argueso, Sandra Spurr-Michaud, and Ilene K Gipson. \Mucins and contact lens wear." Cornea 25, 2: (2006) 176{181. Comparative Study. [70] Hornof, Margit, Wim Weyenberg, Annick Ludwig, and Andreas Bernkop-Schnurch. \Mu- coadhesive ocular insert based on thiolated poly(acrylic acid): development and in vivo evaluation in humans." Journal of Controlled Release 89, 3: (2003) 419{428. [71] ichi Hosoya, Ken, Vincent H.L. Lee, and Kwang-Jin Kim. \Roles of the conjunctiva in ocular drug delivery: a review of conjunctival transport mechanisms and their regulation." European Journal of Pharmaceutics and Biopharmaceutics 60, 2: (2005) 227{240. [72] Hughes, Patrick M, Orest Olejnik, Joan-En Chang-Lin, and Clive G Wilson. \Topical and systemic drug delivery to the posterior segments." Advanced Drug Deliv Reviews 57, 14: (2005) 2010{2032. [73] Institute, National Eye. \Progress in Eye and Vision Research 1999-2006." Technical report, National Institute of Health, US Department of Health and Human Services, 2006. [74] Jiang, Jason, Dayle H. Geroski, Henry F. Edelhauser, and Mark R. Prausnitz. \Measure- ment and Prediction of Lateral Di usion within Human Sclera." Investigative Ophthalmol- ogy and Visual Science 47, 7: (2006) 3011{3016. [75] Kassem, MA, AA Abdel Rahman, MM Ghorab, MB Ahmed, and RM Khalil. \Nanosus- pension as an ophthalmic delivery system for certain glucocorticoid drugs." International Journal of Pharmaceutics JOURNAL ARTICLE. [76] Kaur, Indu Pal, Sonia Chhabra, and Deepika Aggarwal. \Role of cyclodextrins in oph- thalmics." Current Drug Delivery 1, 4: (2004) 351{360. [77] Kaur, Indu Pal, and R Smitha. \Penetration enhancers and ocular bioadhesives: Two new avenues for ophthalmic drug delivery." Drug development and industrial pharmacy 28, 4: (2002) 353 { 369. [78] Kurz, Daryl, and Thomas A Ciulla. \Novel approaches for retinal drug delivery." Ophthal- mology Clinics of North America 15, 3: (2002) 405{410. [79] Lallemand, F, O Felt-Baeyens, S Rudaz, A R Hamel, F Hubler, R Wenger, M Mutter, K Besseghir, and R Gurny. \Conversion of cyclosporine A prodrugs in human tears vs rabbits tears." European Journal of Pharmaceutics and Biopharmaceutics 59, 1: (2005) 51{56. Comparative Study. 112 [80] Lang, John C. \Ocular Drug Delivery Conventional Ocular Formulations." Advanced Drug Delivery Reviews 16, 1: (1995) 39{43. [81] Lee, Vincent H. L. Ophthalmic Drug Delivery Systems, Marcel Dekker, 1993, chapter Pre- corneal, Corneal and Postcorneal Factors, 1{27. [82] Lin, Chien-Chi, and Andrew T. Metters. \Hydrogels in controlled release formulations: Network design and mathematical modeling." Advanced Drug Delivery Reviews 58, 12-13: (2006) 1379{408. [83] Lin, Hong-Ru, K C Sung, and Wen-Jong Vong. \In situ gelling of alginate/pluronic solutions for ophthalmic delivery of pilocarpine." Biomacromolecules 5, 6: (2004) 2358{2365. [84] Lin, Hong-Ru, and K.C. Sung. \Carbopol/pluronic phase change solutions for ophthalmic drug delivery." Journal of Control Release 69, 3: (2000) 379{388. [85] Loftsson, Thorsteinn, Hakon H. Sigurdsson, Dagny Hreinsdottir, Fifa Konradsdottir, and Einar Stefansson. \Dexamethasone delivery to posterior segment of the eye." Journal of Inclusion Phenomena and Macrocyclic Chemistry 57, 1-4: (2007) 585{589. [86] Ludwig, Annick. \The use of mucoadhesive polymers in ocular drug delivery." Advanced Drug Delivery Reviews 57, 11: (2005) 1595{1639. [87] Mainardes, Rubiana Mara, Maria Cristina Cocenza Urban, Priscila Oliveira Cinto, Najeh Maissar Khalil, Marco Vinicius Chaud, Raul Cesar Evangelista, and Maria Palmira Da on Gremiao. \Colloidal carriers for ophthalmic drug delivery." Current Drug Targets 6, 3: (2005) 363{371. [88] Mannermaa, Eliisa, Kati-Sisko Vellonen, and Arto Urtti. \Drug transport in corneal epithe- lium and blood-retina barrier: emerging role of transporters in ocular pharmacokinetics." Advanced Drug Delivery Reviews 58, 11: (2006) 1136{1163. [89] Manz, A., D.J. Harrison, E.M.J. Verpoorte, J.C. Fettinger, A. Paulus, H. Ludi, and H.M. Widmer. \Planar Chips Technology for Miniaturization and Integration of Separation Tech- niques into Monitoring Systems: Capillary Electrophoresis on a Chip." Journal of Chro- matography 593, 1-2: (1992) 253{258. [90] Merodio, Marta, Juan Manuel Irache, Fatemeh Valamanesh, and Massoud Mirshahi. \Oc- ular disposition and tolerance of ganciclovir-loaded albumin nanoparticles after intravitreal injection in rats." Biomaterials 23, 7: (2002) 1587{1594. [91] Michaud, Langis, and Claude J. Giasson. \Overwear of contact lenses: increased severity of clinical signs as a function of protein adsorption." Optometry and Vision Science 79, 3: (2002) 184{192. [92] Miljanovic, Biljana, Reza Dana, David A Sullivan, and Debra A Schaumberg. \Impact of dry eye syndrome on vision-related quality of life." American Journal of Ophthalmology 143, 3: (2007) 409{415. 113 [93] Mohammad, Dina A, Burgunda V Sweet, and Susan G Elner. \Retisert: is the new advance in treatment of uveitis a good one?" Annals of Pharmacotherapy 41, 3: (2007) 449{454. [94] Murphy, John. \You?re Only In Up to Your Knees." Technical report, Review of Optometry Annual Index, 2000. [95] Myles, Marvin E, Donna M Neumann, and James M Hill. \Recent progress in ocular drug delivery for posterior segment disease: emphasis on transscleral iontophoresis." Advanced Drug Delivery Reviews 57, 14: (2005) 2063{2079. [96] Nakamura, M, M Hikida, T Nakano, S Ito, T Hamano, and S Kinoshita. \Characterization of water retentive properties of hyaluronan." Cornea 12, 5: (1993) 433{436. In Vitro. [97] Naveh, N, S Muchtar, and S Benita. \Pilocarpine incorporated into a submicron emulsion vehicle causes an unexpectedly prolonged ocular hypotensive e ect in rabbits." Journal of Ocular Pharmacology 10, 3: (1994) 509{520. [98] Ng, Eugene W M, and Anthony P Adamis. \Anti-VEGFaptamer (pegaptanib) therapy for ocular vascular diseases." Annals of the New York Academy of Science 1082: (2006) 151{171. [99] Nichols, B A, M L Chiappino, and C R Dawson. \Demonstration of the mucous layer of the tear lm by electron microscopy." Investivative Ophthalmology and Visual Science 26, 4: (1985) 464{473. [100] Nicolo, M, D Ghiglione, and G Calabria. \Retinal pigment epithelial tear following intravit- real injection of bevacizumab (Avastin)." European Journal of Ophthalmology 16, 5: (2006) 770{773. Case Reports. [101] Pagnini, Francois. \Ocular Drugs: Light at the End of the Tunnel." Technical report, IMS Global Insights, 2004. [102] Paschides, CA, M Stefaniotou, J Papageorgiou, P Skourtis, and K Psilas. \Ocular surface and environmental changes." Acta Ophthalmologica Scandinavica 76, 1: (1998) 74{7. [103] Peppas, N A, P Bures, W Leobandung, and H Ichikawa. \Hydrogels in pharmaceutical formulations." European Journal of Pharmaceutics and Biopharmaceutics 50, 1: (2000) 27{46. [104] Peppas, Nicholas A. Hydrogels in Medicine and Pharmacy. CRC Press, 1987. [105] Peppas, Nicholas A., and E.W. Merrill. \Crosslinked poly(vinyl alcohol) hydrogels as swollen elastic networks." Journal of Applied Polymer Science 21, 7: (1977) 1763{1770. [106] P ugfelder, S.C. \Di erential diagnosis of dry eye conditions." Advances in Dental Research 10: (1996) 9{12. [107] Pignatello, Rosario, Claudio Bucolo, Piera Ferrara, Adriana Maltese, Antonina Puleo, and Giovanni Puglisi. \Eudragit RS100 nanosuspensions for the ophthalmic controlled delivery of ibuprofen." European Journal of Pharmaceutical Sciences 16, 1-2: (2002) 53{61. 114 [108] Pijls, Rachel T., Lars P.J. Cruysberg, Rudy M.M.A. Nuijts, Aylvin A. Dias, and Leo H. Koole. \Capacity and tolerance of a new device for ocular drug delivery." International Journal of Pharmaceutics [Epub ahead of print]. JOURNAL ARTICLE, OphthaCoil. [109] Prankerd, RJ, and VJ Stella. \The use of oil-in-water emulsions as a vehicle for parenteral drug administration." Journal of parenteral science and technology 44, 3: (1990) 139{149. [110] Prause, J.U. \Treatment of keratoconjunctivitis sicca with Lacrisert." Scandinavian Journal of Rheumatology (Supplement) 61: (1986) 261{263. Clinical Trial. [111] Prausnitz, Mark R., and Jeremy S. Noonan. \Permeability of cornea, sclera, and conjunc- tiva: a literature analysis for drug delivery to the eye." Journal of Pharmaceutical Sciences 87, 12: (1998) 1479{1488. [112] Robinson, James C. Ophthalmic Drug Delivery Systems, Marcel Dekker, 1993, chapter Overview of Ocular Drug Delivery and Iatrogenic Ocular Cytopathologies, 29{57. [113] Ruel-Gariepy, Eve, and Jean-Christophe Leroux. \In situ-forming hydrogels{review of temperature-sensitive systems." European Journal Pharmaceutics and Biopharmaceutics 58, 2: (2004) 409{426. [114] Saettone, M. Fabrizio, Patrizia Chetoni, Riccardo Cerbai, Gabriela Mazzanti, and Laura Braghiroli. \Evaluation of ocular permeation enhancers: in vitro e ects on corneal trans- port of four beta-blockers, and in vitro/in vivo toxic activity." International Journal of Pharmaceutics 142: (1996) 103{113. [115] Saettone, Marco Fabrizio. \Progress and Problems in Ophthalmic Drug Delivery." Technical report, Business Brie ng: Pharmatech, 2002. [116] Sakurai, Eiji, Miho Nozaki, Komei Okabe, Noriyuki Kunou, Hideya Kimura, and Yuichiro Ogura. \Scleral plug of biodegradable polymers containing tacrolimus (FK506) for experi- mental uveitis." Investigative Ophthalmology and Visual Science 44, 11: (2003) 4845{4852. [117] Santini, John T., Amy C. Richards, Rebecca Scheidt, Michael J. Cima, and Robert Langer. \Microchips as controlled drug-delivery devices." Angewandte Chemie International Edition 39, 14: (2000) 2396{2407. [118] Sasaki, H, K Yamamura, T Mukai, K Nishida, J Nakamura, M Nakashima, and M Ichikawa. \Modi cation of ocular permeability of peptide drugs by absorption promoters." Biological & Pharmaceutical Bulletin 23, 12: (2000) 1524{1527. [119] Schmidt-Erfurth, Ursula M, and Christian Pruente. \Management of neovascular age- related macular degeneration." Progress in Retinal and Eye Research 26, 4: (2007) 437{451. [120] Schoenwald, Ronald D. Textbook of Ocular Pharmacology, Lippincott Williams & Wilkins, 1997, chapter Ocular Pharmacokinetics, 119{138. 3rev ed edition. 115 [121] Sharma, Ashutosh, and Eli Ruckenstein. \Mechanism of tear lm rupture and its implica- tions for contact lens tolerance." American Journal of Optometry and Physiological Optics 62, 4: (1985) 246{253. [122] Sklubalova, Z, and Z Zatloukal. \Systematic study of factors a ecting eye drop size and dosing variability." Pharmazie 60, 12: (2005) 917{921. [123] Srinivas, S P. \In situ measurement of uorescein release by collagen shields in human eyes." Current Eye Research 13, 4: (1994) 281{288. Comparative Study. [124] Stuart, J.C., and J.G. Linn. \Dilute sodium hyaluronate (Healon) in the treatment of ocular surface disorders." Annals of Ophthalmology 17, 3: (1985) 190{2. [125] Sultana, Yasmin, M Aqil, and Asgar Ali. \Ion-activated, Gelrite-based in situ ophthalmic gels of pe oxacin mesylate: comparison with conventional eye drops." Drug Delivery 13, 3: (2006) 215{219. Comparative Study. [126] Sultana, Yasmin, M Aqil, Asgar Ali, and Shadaab Zafar. \Evaluation of carbopol-methyl cellulose based sustained-release ocular delivery system for pe oxacin mesylate using rabbit eye model." Pharmaceutical Development and Technology 11, 3: (2006) 313{319. Compar- ative Study. [127] Sultana, Yasmin, R Jain, M Aqil, and Asgar Ali. \Review of ocular drug delivery." Current Drug Delivery 3, 2: (2006) 207{217. [128] Tamesis, R R, A Rodriguez, W G Christen, Y A Akova, E Messmer, and C S Foster. \Systemic drug toxicity trends in immunosuppressive therapy of immune and in ammatory ocular disease." Ophthalmology 103, 5: (1996) 768{775. [129] Thompson, John T. \Cataract formation and other complications of intravitreal triamci- nolone for macular edema." American Journal of Ophthalmology 141, 4: (2006) 629{637. [130] Toda, Ikuko. \LASIK and dry eye." Comprehensive ophthalmology update 8, 2: (2007) 79{85. [131] Tsubota, K, and M Yamada. \Tear evaporation from the ocular surface." Investigative Ophthalmology & Visual Science 33: (1992) 2942{2950. [132] Uchida, Rei, Takao Sato, Haruyasu Tanigawa, and Kenji Uno. \Azulene incorporation and release by hydrogel containing methacrylamide propyltrimenthylammonium chloride, and its application to soft contact lens." Journal of Controlled Release 92, 3: (2003) 259{264. [133] Urquhart, J. Ophthalmic delivery systems, Washington D.C.: American Pharmaceuti- cal Association, Academy of Pharmaceutical Science, 1980, chapter Development of the OCUSERT pilocarpine ocular therapeutic systems a case history, 105118. [134] Urtti, Arto. \Challenges and obstacles of ocular pharmacokinetics and drug delivery." Advanced Drug Delivery Reviews 58, 11: (2006) 1131{1135. 116 [135] Vandamme, Th F. \Microemulsions as ocular drug delivery systems: recent developments and future challenges." Progress in Retinal Eye Research 21, 1: (2002) 15{34. [136] Vasantha, R, PK Sehgal, and KP Rao. \Collagen ophthalmic inserts for pilocarpine drug delivery system." International Journal of Pharmaceutics 47, 1-3: (1988) 95{102. [137] Venkatesh, Siddarth, Stephen P. Sizemore, and Mark E. Byrne. \Biomimetic hydrogels for enhanced loading and extended release of ocular therapeutics." Biomaterials 28, 4: (2007) 717{724. [138] Wei, Gang, Ping-Tian Ding, Jun-Min Zheng, and Wei-Yue Lu. \Pharmacokinetics of timolol in aqueous humor sampled by microdialysis after topical administration of thermosetting gels." Biomedical Chromatography 20, 1: (2006) 67{71. [139] Whikehart, David R. Biochemistry of the Eye, Philadelphia: Butterworth-Heinemann, 2003, chapter Water and Ocular Fluids, 1{14. 2nd edition. [140] Whitesides, George M. \The origins and the future of micro uidics." Nature 442, 7101: (2006) 368{373. [141] Wichterle, Otto. \Cross-linked Hydrophilic Polymers and Articles made therefrom.", 1965. U.S. Patent 3,220,960. [142] Winterton, Lynn C., John M. Lally, Karen B. Sentell, and L. Lawrence Chapoy. \The elution of poly (vinyl alcohol) from a contact lens: The realization of a time release mois- turizing agent/arti cial tear." Journal of Biomedical Materials Research Part B: Applied Biomaterials 80B, 2: (2006) 424{432. [143] Wolko , P, J K Nojgaard, C Franck, and P Skov. \The modern o ce environment desiccates the eyes?" Indoor Air 16, 4: (2006) 258{265. [144] Wu, Chunjie, Hongyi Qi, Wenwen Chen, Chunyan Huang, Cheng Su, Wenmin Li, and Shixiang Hou. \Preparation and evaluation of a Carbopol/HPMC-based in situ gelling ophthalmic system for puerarin." Yakugaku Zasshi 127, 1: (2007) 183{191. Evaluation Studies. [145] Young, Graeme, Jane Veys, Nicola Pritchard, and Sarah Coleman. \A multi-centre study of lapsed contact lens wearers." Ophthalmic and Physiological Optics 22, 6: (2002) 516{527. Clinical Trial. [146] Zderic, Vesna, Shahram Vaezy, Roy W Martin, and John I Clark. \Ocular drug delivery using 20-kHz ultrasound." Ultrasound in Medicine and Biology 28, 6: (2002) 823{829. [147] Zeimer, R, and M F Goldberg. \Novel ophthalmic therapeutic modalities based on noninva- sive light-targeted drug delivery to the posterior pole of the eye." Advanced Drug Delivery Reviews 52, 1: (2001) 49{61. [148] Zimmer, Andreas, and Jorg Kreuter. \Microspheres and nanoparticles used in ocular deliv- ery systems." Advanced Drug Delivery Reviews 16: (1995) 61{73. 117 [149] Zurowska-Pryczkowska, K, M Sznitowska, and S Janicki. \Studies on the e ect of pilocarpine incorporation into a submicron emulsion on the stability of the drug and the vehicle." Euro- pean Journal of Pharmaceutics and Biopharmaceutics 47, 3: (1999) 255{260. Comparative Study. 118 Appendices 119 Appendix A Dynamic release of HA in various concentrations The rst objective is to ensure that the lens is capable of delivering a therapeutic dosage of HA over an acceptably long time period. Too low a dosage would be insu cient for treatment, while an excessive dosage would be economically unfeasible. We selected three concentrations of HA in Nel lcon, made lenses, and conducted release studies to help us benchmark the appropriate amount of HA to add to the formulation. Figure A.1 shows the release curves obtained from lenses made with formulations containing 2 mg/g, 6.5 mg/g and 40 mg/g of HA in Nel lcon. Release was measured over 24 hours. The 2 mg/g formulation released 15 g over the rst 6 hours, and then ceased to release any more. The 6.5 mg/g formulation released 41 g over the rst six hours, and then decreased its release rate. The 40 mg/g formulation released 581 g for the rst 6 hours, with irregular release thereafter. The 6.5 mg/g and 40 mg/g lenses released HA for at least 24 hours. In contrast, lenses with the formulation 2 mg/g released the maximum amount of HA within the rst 6 hours. Hydrogel Di usion coe cient Std dev. R2 Order Std. dev. R2 Nel lcon with 6.5 mg/mL HA 5.689 10 10 0.005 10 10 0.99 0.609 0.018 0.99 Nel lcon with 40 mg/mL HA 6.370 10 10 0.054 10 10 0.91 0.599 0.109 0.87 Table A.1: Di usion and release order of Nel lcon hydrogel with 6.5 and 40 mg HA/g Nel lcon It appears that the di usion coe cient does not vary signi cantly with the concentra- tion of HA in the Nel lcon lens, and the released amount of HA depends on the concen- tration incorporated into the hydrogel. The 6.5 mg/g formulation gave the closest match to the desired HA delivery rate, thus we selected it as the basis for all future experiments. 120 121 0 200 400 600 800 1000 1200 1400 1600 0 5 10 15 20 25 30 Time (hours) Cumulative Mass Release d (micr ograms) Figure A.1: Cumulative release of different concentrations of HA from Nelfilcon hydrogels Release studies were done on Nelfilcon lenses containing 2 mg/g (?), 6.5 mg/g (?) and 40 mg/g (?) of HA in the prepolymer mixture. No functional monomers were added. Appendix B Dynamic release of HA of various sizes As HA is a long-chain molecule, we predicted that it would encounter di culty in di using from the hydrogel.A small molecule would encounter relatively little hindrance from the polymer network as it di uses out of the hydrogel. The di usion coe cient would therefore be relatively high. As the molecular weight of the drug increases and the e ective radius becomes comparable to the polymer mesh size, it faces more steric hindrance and the di usion coe cient decreases. We conducted dynamic release studies to determine the e ect of the size of HA on the di usion coe cient using HA of di erent molecular weights. We synthesized HA-Nel lcon lenses with 6.5 mg of HA per g of Nel lcon, using HA of three di erent sizes: 1 million Da, 100,000 Da and 50,000 Da. The fractional release pro les are shown in Figure B.1. Figure B.1 is a plot of the release pro les of HA of di erent sizes normalized to the maximum amount of HA released. It is evident that the smaller HA molecules di used out of the hydrogel lenses rapidly, compared to the largest HA molecule which was di using out of the lens for up to 3 days. 122 123 0 0.2 0.4 0.6 0.8 1 1.2 0 20 40 60 80 100 120 140 Time (hours) Fr actional mas s of HA r e lea sed Figure B.1: Fractional release of various sizes of HA from Nelfilcon hydrogels Release studies were done on Nelfilcon lenses containing 50 kDa (?), 100 kDa (?) and 1 MDa (?) of HA in the prepolymer mixture. No functional monomers were added. Appendix C Tensile testing of Hydrogels To eveluate the tensile strength of the Nel lcon based hydrogels, the synthesized hy- drogels were cut into strips and extended uniaxially. The hydrogel sample made with Nel lcon without added HA or functional monomers was tested in duplicate, while the 6.5 mg/g HA in Nel lcon sample, the 0.25% functional monomers in Nel lcon sample, and the 0.25% functional monomers and 6.5 mg/g HA in Nel lcon sample were all measured in triplicate as shown in Figures C.1 to C.11. 124 125 y = 193120x - 3905.8 R 2 = 0.9955 0 2000 4000 6000 8000 10000 12000 14000 16000 18000 20000 0 0.02 0.04 0.06 0.08 0.1 0.12 ? ? 1/? 2 Str ess (Pa) Figure C.1: Tensile test Nelfilcon? sample 1 y = 209335x - 7547.9 R 2 = 0.9997 0 10000 20000 30000 40000 50000 60000 70000 80000 0 0.05 0.1 0.15 0.2 0.25 ? ? 1/ ? 2 Stress (Pa) Figure C.2: Tensile test Nelfilcon? sample 2 126 y = 176949x + 4028.1 R 2 = 0.9993 0 10000 20000 30000 40000 50000 60000 70000 0 0.05 0.1 0.15 0.2 0.25 0.3 ? ? 1/? 2 Str ess (Pa) Figure C.3: Tensile test Nelfilcon with HA? sample 1 y = 149189x + 5732 R 2 = 0.9999 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 50000 0 0.05 0.1 0.15 0.2 0.25 0.3 ? ? 1/? 2 Str ess (Pa) Figure C.4: Tensile test Nelfilcon with HA? sample 2 127 y = 132718x + 3891.1 R 2 = 0.9976 0 5000 10000 15000 20000 25000 30000 35000 40000 0 0.05 0.1 0.15 0.2 0.25 ? ? 1/? 2 Stress (Pa ) Figure C.5: Tensile test Nelfilcon with HA? sample 3 y = 212018x + 1858.5 R 2 = 0.9994 0 10000 20000 30000 40000 50000 60000 70000 80000 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 ? ? 1/? 2 Str e ss (Pa) Figure C.6: Tensile test Nelfilcon with functional monomers? sample 1 128 y = 189636x + 4208.6 R 2 = 0.9994 0 10000 20000 30000 40000 50000 60000 70000 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 ? ? 1/? 2 Str e ss (Pa) Figure C.7: Tensile test Nelfilcon with functional monomers? sample 2 y = 183089x + 7230.4 R 2 = 0.9995 0 5000 10000 15000 20000 25000 30000 35000 40000 45000 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2 ??1/? 2 St re ss ( Pa ) Figure C.8: Tensile test Nelfilcon with functional monomers? sample 3 129 y = 192832x - 1585.4 R 2 = 0.9976 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 100000 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 ? ? 1/? 2 St ress (Pa) Figure C.9: Tensile test Nelfilcon with functional monomers and HA ? sample 1 y = 206752x - 4173.3 R 2 = 0.9951 0 20000 40000 60000 80000 100000 120000 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 ? ? 1/? 2 Stress (Pa) Figure C.10: Tensile test Nelfilcon with functional monomers and HA ? sample 2 130 y = 209111x + 5314.9 R 2 = 0.9994 0 10000 20000 30000 40000 50000 60000 70000 80000 90000 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 ? ? 1/? 2 Stress (Pa) Figure C.11: Tensile test Nelfilcon with functional monomers and HA ? sample 3