<%BANNER%>

Controlled Ophthalmic Drug Delivery by Surfactant-Laden Contact Lenses

Permanent Link: http://ufdc.ufl.edu/UFE0022886/00001

Material Information

Title: Controlled Ophthalmic Drug Delivery by Surfactant-Laden Contact Lenses
Physical Description: 1 online resource (221 p.)
Language: english
Creator: Kapoor, Yash
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: brij, cyclosporine, dexamethasone, draize, liposomes, micelle, microemulsion, phema, surfactants, toxicity
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Drug delivery via eye drops has been found to be inefficient due to low bioavailability of less than 5%, and bioavailability of these drugs can be substantially improved to about 50% by delivering it via contact lenses. Commercial contact lenses are not suitable for drug delivery as the drug diffuses from these gels within a few hours to a day. Also, they have limited loading of the drug due to limited solubility of the drug in the contact lens material. Surfactants are commonly incorporated into hydrogels to increases solute loading and attenuate the drug release rates. Our study focused on development of nanostructured poly (2-hydroxyethyl methacrylate) (p-HEMA) hydrogels containing microemulsions of Brij 97 surfactant and micelles of Brij 97, 78, 98 and 700 surfactants for extended delivery of ophthalmic drugs, specifically Cyclosporine A (CyA). Also, the effect of loading drug nanoparticles inside the hydrogels is explored. The release of CyA from these nanostructured hydrogels was performed in vitro and the effects of surfactant concentration, processing conditions, shelf life stability and gel thickness on the release of CyA was studied. We focus on various Brij surfactants to investigate the effects of chain length and presence of an unsaturated group on the drug release dynamics and partitioning inside the surfactant domains inside the gel. The release of drug and surfactant from the hydrogels is found to be diffusion controlled and the duration of drug release increases with increasing surfactant loading and is relatively similar for both surfactant and microemulsion-laden gels. We also focus on understanding and modeling the mechanisms of both surfactant and drug transport in hydrogels. These models can aid in tuning the drug release rates from hydrogels by controlling the surfactant concentration. These studies also show that these hydrogels retain their effectiveness as drug delivery vehicles even after exposure to the relevant processing conditions needed in contact lens manufacturing including unreacted monomer extraction, autoclaving and packaging. The gels were imaged by Cryogenic Scanning Electron Microscopy (Cryo-SEM) to obtain direct evidence of the presence of surfactant-aggregates in the gel and to investigate the detailed microstructure for different surfactants. The images show a distribution of nano pores inside the surfactant laden hydrogels that we speculate are regions of surfactant aggregates, possibly vesicles that have a high affinity for the hydrophobic drug molecule. The gels are further characterized by studying their mechanical and physical properties such as transparency, surface contact angle, and equilibrium water content to determine their suitability as extended wear contact lenses. Results show that Brij surfactant-laden p-HEMA gels provide extended release of CyA, and have suitable mechanical and optical properties for contact lens applications. Surfactants can also cause potential toxicity if they diffuse from the contact lenses to the ocular surface. Hence, we also designed an in vitro liposomal study to predict ocular toxicity of the surfactants utilized in the study for which ocular toxicity data is currently unavailable.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yash Kapoor.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Chauhan, Anuj.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022886:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022886/00001

Material Information

Title: Controlled Ophthalmic Drug Delivery by Surfactant-Laden Contact Lenses
Physical Description: 1 online resource (221 p.)
Language: english
Creator: Kapoor, Yash
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: brij, cyclosporine, dexamethasone, draize, liposomes, micelle, microemulsion, phema, surfactants, toxicity
Chemical Engineering -- Dissertations, Academic -- UF
Genre: Chemical Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Drug delivery via eye drops has been found to be inefficient due to low bioavailability of less than 5%, and bioavailability of these drugs can be substantially improved to about 50% by delivering it via contact lenses. Commercial contact lenses are not suitable for drug delivery as the drug diffuses from these gels within a few hours to a day. Also, they have limited loading of the drug due to limited solubility of the drug in the contact lens material. Surfactants are commonly incorporated into hydrogels to increases solute loading and attenuate the drug release rates. Our study focused on development of nanostructured poly (2-hydroxyethyl methacrylate) (p-HEMA) hydrogels containing microemulsions of Brij 97 surfactant and micelles of Brij 97, 78, 98 and 700 surfactants for extended delivery of ophthalmic drugs, specifically Cyclosporine A (CyA). Also, the effect of loading drug nanoparticles inside the hydrogels is explored. The release of CyA from these nanostructured hydrogels was performed in vitro and the effects of surfactant concentration, processing conditions, shelf life stability and gel thickness on the release of CyA was studied. We focus on various Brij surfactants to investigate the effects of chain length and presence of an unsaturated group on the drug release dynamics and partitioning inside the surfactant domains inside the gel. The release of drug and surfactant from the hydrogels is found to be diffusion controlled and the duration of drug release increases with increasing surfactant loading and is relatively similar for both surfactant and microemulsion-laden gels. We also focus on understanding and modeling the mechanisms of both surfactant and drug transport in hydrogels. These models can aid in tuning the drug release rates from hydrogels by controlling the surfactant concentration. These studies also show that these hydrogels retain their effectiveness as drug delivery vehicles even after exposure to the relevant processing conditions needed in contact lens manufacturing including unreacted monomer extraction, autoclaving and packaging. The gels were imaged by Cryogenic Scanning Electron Microscopy (Cryo-SEM) to obtain direct evidence of the presence of surfactant-aggregates in the gel and to investigate the detailed microstructure for different surfactants. The images show a distribution of nano pores inside the surfactant laden hydrogels that we speculate are regions of surfactant aggregates, possibly vesicles that have a high affinity for the hydrophobic drug molecule. The gels are further characterized by studying their mechanical and physical properties such as transparency, surface contact angle, and equilibrium water content to determine their suitability as extended wear contact lenses. Results show that Brij surfactant-laden p-HEMA gels provide extended release of CyA, and have suitable mechanical and optical properties for contact lens applications. Surfactants can also cause potential toxicity if they diffuse from the contact lenses to the ocular surface. Hence, we also designed an in vitro liposomal study to predict ocular toxicity of the surfactants utilized in the study for which ocular toxicity data is currently unavailable.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Yash Kapoor.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Chauhan, Anuj.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022886:00001


This item has the following downloads:


Full Text

PAGE 1

1 CONTROLLED OPHTHALMIC DRUG DELIVERY BY SURFACTANT-LADEN CONTACT LENSES By YASH KAPOOR A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Yash Kapoor

PAGE 3

3 To my parents and my lovely wife, Ananya

PAGE 4

4 ACKNOWLEDGMENTS I thank m y advisor, Dr. Anuj Chauhan for his excellent guidance and help during my research work. He has been an extraordinary mentor to me, never letting me down. He has also been a great friend and I have pe rsonally learnt more from him than anyone else I have had a chance to interact. He has been a constant motiv ational force behind me a nd this work would not have been possible if it was not for his conti nuous presence throughout my graduate life. I would also like to thank Professor Vijay John and his student Dr. Grace Tan from Tulane University who helped me getting the Cryo SEM imag es for my samples. I am also thankful to Professor Dinesh Shah, Dr. Jason Butler, Dr. Sergey Vasenkov, Professor Christopher Batich, Dr. Yiider Tseng and Dr. Tanmay Lele from Chem ical Engineering department at University of Florida for their insightful discus sions regarding my research, progr ess and my future endeavors. I have been extremely lucky to have had an excellent advisor and an amazing set of group members, who have been very understanding and friendly throughout my stay in the department. First and foremost I would like to thank Dr. Jina h Kim, whose mere presence in the lab made everybody jovial and energetic. She helped me designing some of my experiments and will always be a great friend. I ha ve had some stimulating interactions with Dr. Heng Zhu and Dr. Chi-Chung Li in our group. They helped me in se ttling in our lab when I joined and were always there to help. I also thank Cheng Chun Peng and Hyun Jung Jung for their support in the lab. Chavvi Gupta has been an excellent friend to me and he has helped me with some of my lab work, especially through some difficult times. I will always remain indebted to him. Lastly, I would like to extend my thanks to Brett Howell w ith whom I have had a chance to work on the Liposome permeability assay for in vitro toxicit y. He is and will always be an exceptional researcher. His dedication to his work and his zeal to learn and succeed are unparalleled. Chapter 6 in my thesis would not ha ve been possible without his help.

PAGE 5

5 I would also like to thank so me wonderful undergraduate students whom I had a chance to work with. Justin Thomas, Anthony Conwa y, Andrew Cohen and Maung Zaw. All these undergraduate students were exceptional in their work and were excellent learners. I would also like to acknowledge the technical help provided by James Hinnant and Dennis Vince. I would like to thank the staff memb ers, specially, Shirley Kelly, Deborah Aldrich, Cynthia Sain, Deborah Sandoval and Janice Harris for their support and comfort during my stay in the Chemical Engineering Department. Help provided by Sean Poole for computer support is also greatly appreciated. Finally, I would like to thank my mother and father, Meena Rani Kapoor and Kailash Kapoor, for their continuous suppor t in all my endeavors, I am here because of them and hope that I can extend the same love, care and support to them as they have bestowed on me. I love them and they will always be my heroes. I thank my wife, Ananya, from the bottom of my heart. She is one of the most understanding and caring persons I have ever met. Her sacrifices during my extended work hours in the lab cannot be evaluate d. I just know that she is the most precious thing in my life and I cannot thank he r enough for being there to support me.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ........10 LIST OF FIGURES.......................................................................................................................12 LIST OF ABBREVIATIONS........................................................................................................ 15 ABSTRACT...................................................................................................................................17 CHAP TER 1 INTRODUCTION..................................................................................................................19 2 CYCLOSPORINE A RELEASE FROM POLY(HYDOXYETHYL METHACRYLATE) HYDROGELS .....................................................................................37 2.1 Introduction............................................................................................................... ........37 2.2 Materials and Methods.....................................................................................................37 2.2.1 Materials.................................................................................................................37 2.2.2 Methods..................................................................................................................37 2.2.2.1 Synthesis of drug laden and pure p-HEMA gels.......................................... 37 2.2.2.2 Drug detection: HPLC assay........................................................................ 38 2.2.2.3 Drug release: Equilibrium experiments........................................................ 38 2.2.2.4 Drug release: PBS change experiments.......................................................39 2.2.2.5 Transmittance measurements....................................................................... 39 2.2.2.6 Statistical analysis........................................................................................ 39 2.3 Results and Discussion..................................................................................................... 40 2.3.1 Drug Release: Equilibrium Experiments................................................................ 40 2.3.2 Model for Drug Release......................................................................................... 42 2.3.3 Drug Release: PBS Replacement...........................................................................44 2.3.4 Theoretical Model..................................................................................................45 2.3.5 Effect of Drug Concentration on Transparency.....................................................46 2.4 Conclusion........................................................................................................................47 3 CYCLOSPORINE A RELEASE FROM BRIJ 97 MICROEMULSION AND SURFACTANT LADEN HYDROGELS .............................................................................. 53 3.1 Introduction............................................................................................................... ........53 3.2 Materials and Methods.....................................................................................................53 3.2.1 Materials.................................................................................................................53 3.2.2 Microemulsion Formulation................................................................................... 53 3.2.3 Particle Size Analysis............................................................................................. 54 3.2.4 Preparation of Microemulsion Laden Gels.............................................................54

PAGE 7

7 3.2.5 Preparation of Surfactant Laden Gels..................................................................... 55 3.2.6 CyA Detection by HPLC........................................................................................ 55 3.2.7 Drug Release Kinetics from Gels Load ed with CyA by Drug Addition to the Monom er................................................................................................................55 3.2.8 Drug Uptake and Release Kinetics from Gels Loaded with CyA after Polymerization....................................................................................................... 56 3.2.9 Packaging Solution for Drug Release.....................................................................56 3.2.10 Processing Conditions in Contact Lens Manufacturing.......................................57 3.2.11 Surfactant Release................................................................................................58 3.2.12 Statistical Analysis............................................................................................... 58 3.3 Results and Discussion..................................................................................................... 58 3.3.1 Particle Size Analysis of Microem ulsions and Drug Release from Microemulsion-Laden Gels...................................................................................58 3.3.2 Release of Drug after Soaking Microe m ulsion-Laden Gels in Drug Solution....... 60 3.3.3 Effect of Packaging Conditions on Drug Release.................................................. 60 3.3.4 Drug Release from Micelle Laden Hydrogels........................................................ 61 3.3.5 Mechanism of Drug Release.................................................................................. 62 3.3.6 Processing Conditions............................................................................................63 3.3.7 Brij 97 Release from p-HEMA Hydrogels............................................................. 64 3.4 Conclusion........................................................................................................................66 4 MODEL FOR SURFACTANT AND DRUG TRANSPORT FROM P-HEMA HYDROGEL ..........................................................................................................................77 4.1 Introduction............................................................................................................... ........77 4.2 Materials and Methods.....................................................................................................77 4.2.1 Materials.................................................................................................................77 4.2.2 Synthesis of Surfactant Laden Gels........................................................................ 77 4.2.3 Drug Release Experiments..................................................................................... 78 4.2.4 Surfactant Release.................................................................................................. 79 4.3 Results and Discussion..................................................................................................... 79 4.3.1 Drug Release from Pure p-HEMA Gels with Daily PBS Replacement................. 79 4.3.2 Surfactant Release from the Hydrogels.................................................................. 81 4.3.2.1 Model...........................................................................................................81 4.3.2.2 Experimental results.....................................................................................85 4.3.3 Drug Release from Surfactant Laden Gels............................................................. 87 4.3.3.1 Model...........................................................................................................87 4.3.3.2 Experimental results.....................................................................................91 4.3.4 Model Comparison with Published Data................................................................ 93 4.4 Conclusion........................................................................................................................95 5 SURFACTANT LADEN HYDROGELS FO R OPHT HALMIC DRUG DELIVERY WITH INCREASED WETTABILI TY AND WATER CONTENT.................................... 107 5.1 Introduction............................................................................................................... ......107 5.2 Materials and Methods...................................................................................................107 5.2.1 Materials...............................................................................................................107

PAGE 8

8 5.2.2 Preparation of Surfactant Laden Gels................................................................... 108 5.2.3 Drug Release Experiments................................................................................... 108 5.2.4 Drug Detection.....................................................................................................109 5.2.5 Surfactant Release Experiments........................................................................... 109 5.2.6 Lysozyme Sorption............................................................................................... 110 5.2.7 Preparation and Cryo-SEM of Hydrogels............................................................ 110 5.2.8 Dynamic Mechanical Analysis.............................................................................111 5.2.9 Surface Contact Angle Measurements................................................................. 111 5.2.10 Transmittance Measurements............................................................................. 112 5.2.11 Equilibrium Water Content................................................................................ 112 5.2.12 Statistical Analysis............................................................................................. 112 5.3 Results and Discussion................................................................................................... 112 5.3.1 Surfactant Release from the Hydrogels................................................................ 112 5.3.2 CyA Release: Equilibrium Experiments.............................................................. 115 5.3.3 Effect of Surfactant Dissolved in the Release Medium........................................ 117 5.3.4 CyA Release: PBS Change Experiments.............................................................118 5.3.5 DMS and DMSA Release: Equilibrium Experiments.......................................... 121 5.3.6 Uptake of Lysozyme in the Hydrogels................................................................. 122 5.3.7 Microstructure of Hydr ogels: Cryo-SEM Im aging..............................................123 5.3.8 Physical Properties...............................................................................................124 5.3.8.1 Mechanical properties................................................................................124 5.3.8.2 Transparency, equilibrium water c onten t, and surface contact angle of gels..............................................................................................................126 5.4 Conclusion......................................................................................................................127 6 LIPOSOME ASSAY FOR EVALUATING OCULAR TOXICITY OF SURFACTANTS .................................................................................................................. 158 6.1 Introduction............................................................................................................... ......158 6.2. Materials and Methods..................................................................................................159 6.2.1 Materials...............................................................................................................159 6.2.2 Liposome Preparation for Calcein Leakage Studies ............................................ 160 6.2.3 Liposome Leakage Studies................................................................................... 160 6.2.4 Draize Scores........................................................................................................161 6.2.5 Data Analysis........................................................................................................ 162 6.3 Results and Discussion................................................................................................... 162 6.3.1 Draize Score / Leakage Correlations at a C onstant Test Concentration.............. 162 6.3.2 Draize Score / Leakage Correlati ons at Adjusted Concentrations ....................... 163 6.3.3 Mechanism of Surfactant Toxicity....................................................................... 164 6.3.4 Comparison of Liposome Assay with other in Vitro Assays ............................... 169 6.3.5 Prediction of Ocular Toxici ty for Non-ionic Surfactants ..................................... 171 6.3.6 Model for Micelle Depletion from the Ocular Surface........................................ 172 6.4 Conclusion......................................................................................................................174 7 ASSESING CRITICAL AGGREGATION CONCENTRATION FOR SURFACTANTS IN HYDROGELS ................................................................................... 186

PAGE 9

9 7.1 Introduction............................................................................................................... ......186 7.2 Materials and Methods...................................................................................................186 7.2.1 Materials...............................................................................................................186 7.2.2 Synthesis of Surfactant and Drug Laden Gels......................................................187 7.2.3 Drug Detection: HPLC Assay..............................................................................187 7.2.4 Drug Release........................................................................................................188 7.2.5 Equilibrium Water Content.................................................................................. 188 7.3 Results and Discussion................................................................................................... 188 7.3.1 Method I: Drug Release........................................................................................188 7.3.2 Method II: Water Uptake...................................................................................... 192 7.3.3 Surfactant Diffusivity........................................................................................... 193 7.4 Conclusion......................................................................................................................193 8 CONCLUSION..................................................................................................................... 201 9 FUTURE WORK.................................................................................................................. 206 9.1 Gels with Higher Surfactant Loading............................................................................. 206 9.2 Oxygen Permeability...................................................................................................... 206 9.3 Release of Bio-active Agents like Vitam in E from Contact Lenses............................... 207 9.4 Surfactant laden Silicone Contact Lenses....................................................................... 208 9.5 Polymerizable Surfactants..............................................................................................208 9.6 In-vivo Experiments....................................................................................................... 209 LIST OF REFERENCES.............................................................................................................210 BIOGRAPHICAL SKETCH.......................................................................................................221

PAGE 10

10 LIST OF TABLES Table page 2-1 Physical properties of CyA at room temperature............................................................... 52 3-1 Diffusion coefficients of the drug for the m icroemulsion laden systems.......................... 75 3-2 Drug uptake by microemulsion laden ge l M2 after soaking in drug solution ................... 75 3-4 Drug loading and release from surfact ant-laden and m icroemulsion-laden gels subjected to processing conditions. The e rror bars represent half the difference between the data from two repeat runs.............................................................................. 76 5-1 Physical properties of the su rfactants explored in this study ........................................... 155 5-2 DSC* for all systems obtained from fitt ing of the surfactant release data....................... 155 5-3 Concentration of surfactants dissolved in the release m edium........................................ 155 5-4 Partition coefficient of CyA for all the surfactant system s.............................................. 156 5-5 Partition coefficient of DMS and DMSA in p-HEMA and Brij 78 surfactant laden hydrogels ...................................................................................................................... ....156 5-6 Physical properties of the surf actant laden and pure p-HEMA hydrogels ....................... 157 5-7 Parameters obtained by fitting Standard Line a r Solid Model to the experimental data.. 157 6-1 Draize scores used for in vitro/in vivo correlations.........................................................181 6-2 Correlation comparisons for Draize scores and leakage experim ents performed at surfactant concentrations of 1 g/mL and CMC/200......................................................182 6-3 Critical micelle concentrations for su rfactants studied and subsequent test concentrations for liposom e leakage................................................................................183 6-4 Surface area to volume ratio comparisons for lipo somes and epithelial cells................. 184 6-5 Correlation comparisons between the lipos om e leakage method of assessing toxicity and other published methods........................................................................................... 184 6-6 Predicted Draize scores for 10% stock solutions of selected Brij surfactants ................. 185 6-7 Predicted Draize scores for 1% stock solutions of selected Brij surfactants ...................185 7-1 Physical properties of the surfactant u tilized in solving th e drug release model ............. 200

PAGE 11

11 7-2 Critical aggregation concentration for each surfactant system evaluated from two different techniques..........................................................................................................200

PAGE 12

12 LIST OF FIGURES Figure page 2-1 Percentage drug release from 200 m thick p-H EMA hydrogel. Amount of drug loaded in the hydrogel is indicated.................................................................................... 48 2-2 Partition coefficient plotted against concentration of drug in the PBS at equilibrium...... 48 2-3 A schematic of drug interaction with the p-HEM A matrix as drug loading is increased............................................................................................................................49 2-4 Effect of drug loading on cum ulative drug release from 200 m thick gels in PBS change experiments............................................................................................................ 50 2-5 Cumulative percentage release of drug from 200 m thick surfactant-laden gels in PBS change experiments after rescaling the time..............................................................50 2-6 Transmittance of the gel with increasing drug concentrations.......................................... 51 3-1 Size distribution of microemulsions with three different surfactant loadings ................... 68 3-2 Cumulative percentage rel ease of drug from microemulsi on laden gels with varying surfactant loading and pure p-HEMA gels........................................................................68 3-3 Linear fits for the short time release da ta to obtain the effective diffusivity for m icroemulsion and pure p-HEMA gels............................................................................. 69 3-4 Cumulative percentage rel ease of drug from microemuls ion gels after loading the drug into gels by soaking in a drug solution for 5, 10 and 15 days................................... 69 3-5 Cumulative percentage release of dr ug from microemulsion laden gels after packaging in three different salt soluti ons for different durations of time........................70 3-6 Cumulative percentage release of drug from Brij 97 surfactant laden, microemulsion laden and pure p-HEMA gels............................................................................................ 71 3-7 Linear fits for the short time release da ta to obtain the effective diffusivity for surfactant laden gels .......................................................................................................... .72 3-8 Effect of thickness on percentage release for p-H EMA gels and surfactant laden gels for equilibrium experiments............................................................................................... 72 3-9 Effect of surfactant loading and proc essing conditions on cumulative percentage release from pure p-HEMA, the m icroemulsion laden and surfactant laden gels............. 73 3-10 Dependence of surface tension on the bul k surfactant concen tration. for Brij 97 surfactant............................................................................................................................74

PAGE 13

13 3-11 Cumulative percentage release of surfactant from surfactant laden gels ........................... 74 4-1 Transport of drug in the p-HEMA hydrogel...................................................................... 97 4-2 Percentage release of drug from pure p-HEMA gels......................................................... 98 4-3 Transport from surfactant laden hydrogel.......................................................................... 99 4-4 Dependence of surface tension on the bul k surfactant concen tration for Brij 98 surfactant..........................................................................................................................100 4-5 Cumulative percentage releas e of surfactant from hydrogels .......................................... 100 4-6 Cumulative percentage rel ease of surfactant from the hydrogels after rescaling the tim e........................................................................................................................... .......101 4-7 Effect of surfactant loading on cumula tive percentage release of the drug for surfactan t lade n thick (200 m) gels during PBS change experiments...........................102 4-8 Effect of surfactant loading on cumula tive percentage release of the drug for surf actant laden thin (100 m) gels during PBS change experiments............................. 102 4-9 Cumulative percentage release of drug for surfactant laden thick (200 m ) gels during PBS change experiments afte r rescaling the time of release................................ 103 4-10 Cumulative percentage release of drug for surfactant laden thin (100 m ) gels during PBS change experiments after re scaling the time of release........................................... 104 4-11 Drug release from surfactant laden gels for equilibrium (no PBS change) experiments.................................................................................................................... ..105 4-12 Drug release from agarose hydrogels cont aining S DS surfactant obtained from Liu et al. [94]..............................................................................................................................106 4-13 Drug release from agarose hydrogels contai ning DTAB surfa ctant obtained from Liu et al. [95]..........................................................................................................................106 5-1 Dependence of surface tension on th e bulk surfactant concen tration.............................. 130 5-2 Percentage release of surfactan t during water change experim ents................................. 131 5-3 Cumulative percentage rel ease of surfactant from the hydrogels after rescaling the tim e........................................................................................................................... .......133 5-4 Effect of thickness on percentage releas e of drug during equilibrium experiments........ 135 5-5 Microstructure of th e surfactant-laden gel .......................................................................136

PAGE 14

14 5-6 Effect of surfactant disso lved in the release m edium on cumulative percentage release of the drug during equilibrium experiments........................................................ 137 5-7 Effect of surfactant loading on cumulative drug release from surfactant-laden gels in PBS change experiments.................................................................................................. 139 5-8 Cumulative percentage release of drug from 200 m thick surfactant-laden gels in PBS change experiments after rescaling the time............................................................141 5-9 Percentage release of DMS from 100 m thick p-HEMA and Brij 78 laden in equilibrium experiments.................................................................................................. 143 5-10 Percentage release of DMSA from 100 m thick p-HEMA and Brij 78 laden gels in equilibrium experiments.................................................................................................. 143 5-11 Lysozyme sorption in surfactant laden and pure p-HEMA hydrogels ............................ 144 5-12 Cryo-SEM image for 200 m thick gels.......................................................................... 145 5-13 Frequency dependence of moduli for 800 m thick surfactant laden and pure pHEMA gels......................................................................................................................153 5-14 Standard Linear Solid Model used fo r f itting the viscoelasticity data of the surfactant-laden gels........................................................................................................154 5-15 Effect of thickness on the storage and loss m oduli of pure p-HEMA gels...................... 154 6-1 Draize scores versus liposome leakage after 10 m inutes induced by surfactants at concentrations of 1 g/mL and logarithmic correlations................................................. 176 6-2 Draize scores versus liposome leakage after 10 m inutes induced by surfactants at CMC/200 and logarithmic correlations........................................................................... 178 6-3 The 95% confidence intervals for mean Draize scores at 10% ocular loading for surfactan ts based on logarithmic corrlea tions from percent dye leakage from liposomes after ten minutes at surfactant CMC/200........................................................180 6-4 Surfactant induced toxic ity on the corneal surface .......................................................... 180 7-1 Error between theoretical and experime ntally determ ined slope for drug release experiments from hydrogels containing va rying surfactant loading against CAC.......... 194 7-2 Plot of slope vs inverse of initia l surf actant loading inside the hydrogels....................... 196 7-3 Equilibrium water content of surfactan t lad en hydrogels with varying initial surfactant loading inside the p-HEMA matrix................................................................. 198

PAGE 15

15 LIST OF ABBREVIATIONS BKC Benzalkonium Chloride Brij 56 Poly Oxy Ethylene (10) Cetyl Ether Brij 58 Poly Oxy Ethylene (20) Cetyl Ether Brij 76 Poly Oxy Ethyl ene (10) Stearyl Ether Brij 78 Poly Oxy Ethyl ene (20) Stearyl Ether Brij 700 Poly Oxy Ethylene (100) Stearyl Ether Brij 97 Poly Oxy Ethylene (10) Oleyl Ether Brij 98 Poly Oxy Ethylene (20) Oleyl Ether C* Critical Aggregation Concentration CH Cholestrol CMC Critical Micelle Concentration CPB Cetyl Pyridinium Bromide CPC Cetyl Pyridinium Chloride CTAB Hexadecyl Trimethylammonium Bromide CyA Cyclosporine A D Diffusivity of the Drug DS Diffusivity of Surfactant DI Deionized DMPC 1,2 Dimyristyl-sn-Glycero 3 Phosphocholine DMPG 1,2 Dimyristyl-sn-Gly cero 3 Phospho-rac-glycerol DMS Dexamethasone DMSA Dexamethasone Acetate DOPE 1,2 Dioleoyl-sn-Glycero 3 Phosphoethanolamine DTAB Dodecyl Trimethylammonium Bromide

PAGE 16

16 EGDMA Ethylene Glycol Dimethacrylate EWC Equilibrium Water Content HEMA Hydroxy Ethyl Methacrylate HLB Hydrophilic Lipophilic Balance MTAB Myristyl Trimethylammonium Bromide OTAB Octadecyl Trimethylammonium Bromide PBS Phosphate Buffered Saline PLTF Pre Lens Tear Film POLTF Post Lens Tear Film SDS Sodium Dodecyl Sulfate Tween 20 Polyoxyethylene Sorbitan Monolaurate Tween 40 Polyoxyethylene Sorbitan Monopalmitate Tween 80 Polyoxyethylene Sorbitan Monooleate

PAGE 17

17 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy CONTROLLED OPHTHALMIC DRUG DELIVERY BY SURFACTANT-LADEN CONTACT LENSES By Yash Kapoor December 2008 Chair: Anuj Chauhan Major: Chemical Engineering Drug delivery via eye drops has been found to be inefficient due to low bioavailability of less than 5%, and bioavailability of these drugs can be s ubstantially improved to about 50% by delivering it via contact lenses. Commercial contact lenses are not suitable for drug delivery as the drug diffuses from these gels within a few hour s to a day. Also, they have limited loading of the drug due to limited solubility of the drug in the contact lens material. Surfactants are commonly incorporated into hydrog els to increases solute loadi ng and attenuate the drug release rates. Our study focused on development of nanostructured poly (2-hydroxyethyl methacrylate) (p-HEMA) hydrogels containing microemulsions of Brij 97 surfactant and micelles of Brij 97, 78, 98 and 700 surfactants for exte nded delivery of ophthalmic drugs, specifically Cyclosporine A (CyA). Also, the effect of loading drug nanoparticles inside the hydrog els is explored. The release of CyA from these nanostructured hydrogels was performed in vitro and the effects of surfactant concentration, pro cessing conditions, shelf life stab ility and gel th ickness on the release of CyA was studied. We focus on various Brij surfactants to investigate the effects of chain length and presence of an unsaturated gr oup on the drug release dyna mics and partitioning

PAGE 18

18 inside the surfactant domains inside the gel. The release of drug a nd surfactant from the hydrogels is found to be diffusion controlled and the duration of drug release increases with increasing surfactant loading and is relatively similar for both surfactant and microemulsionladen gels. We also focus on understanding and modeling the mechanisms of both surfactant and drug transport in hydrogels. These models can aid in tuning the drug release rates from hydrogels by controlling the surfactant concentr ation. These studies also show that these hydrogels retain their effectivene ss as drug delivery vehicles even after exposure to the relevant processing conditions needed in contact lens manufacturing including unreacted monomer extraction, autoclaving and packaging. The ge ls were imaged by Cryogenic Scanning Electron Microscopy (Cryo-SEM) to obtain di rect evidence of the presence of surfactant-aggregates in the gel and to investigate the detailed microstructure for different surfactants. The images show a distribution of nano pores inside the surfactant laden hydr ogels that we speculate are regions of surfactant aggregates, possibly vesicles that have a high affinity for the hydrophobic drug molecule. The gels are further characteri zed by studying their mechanical and physical properties such as transparency, surface contac t angle, and equilibrium water content to determine their suitability as extended wear contact lenses. Results show that Brij surfactantladen p-HEMA gels provide extended release of Cy A, and have suitable mechanical and optical properties for contact lens applications. Surfactants can also cause potentia l toxicity if they diffuse from the contact lenses to the ocular surface. Hence, we also designed an in vitro liposomal study to predic t ocular toxicity of the surfactants utilized in the study for which oc ular toxicity data is currently unavailable.

PAGE 19

19 CHAPTER 1 INTRODUCTION Millions of people around the world suffer from various eye diseases and the effective treatment of these diseases ha s become a focus of researchers around the world. The most convenient way of instilling drugs to ocular tissues is through topical administration and approximately 90% of all ophthalmic drugs formulat ions used to treat these diseases are applied as eye-drops [1]. Drugs applied through topica l administration can reach the ocular tissue by either penetrating through the co rnea or penetrating through the c onjunctiva. Penetration of the drug inside the cornea depends on two important factors that include pe rmeability of the drug across corneal epithelium and the residence time of the drug on the corneal surface. Permeability of most ocular drugs is very small and there is small residence time of the drug is small when applied through eye drops [2-7]. Thus, the bioava ilability of the drug when using eye drops is minimal and only about 1-5% of the drug applied via eye drops eventually reaches the target tissue while the remaining 95-99% enters the systemic circulation through conjunctival uptake or drainage into the nasal cavity [8]. The low bioavailability leads to drug wastage and, more importantly, the systemic uptake of ophthalm ic drugs can lead to side effects. One way of increasing residence times of the drug on the corneal surface is by increasing the viscosity of the applied solu tion. Various viscosity enhancers have been used to increase the viscosity of the applied drops to increase the resulting residence time and thus the bioavailability [9,10]. Apart from viscosity enha ncers, shear thinning polymers ha ve also been explored which conform under shear generated by blinking to reduce stress on the corneal surface and also increase the residence time of the drug. Also, mucoadhesive polymers have been explored in literature which leads to higher contact time of the drug containing solu tion due to its physical

PAGE 20

20 binding with the mucin layer [11]. Major disadv antages of these polymer based delivery systems is blurring of vision and uncomfortable feeling due to the sticky nature of these polymers. Several solid particle suspensions have also been explored for instilling various drugs. These systems mainly include liposomes, niosom es, microsperes, and microemulsion particles [12,13]. Major issues concerni ng these systems are the stabili ty and sterilization for mass production making them quite ineffective for use. Various soluble, biodegradable, and nonsoluble ocular inserts have also been explored in literature. These systems can lead to programmed delivery of the drug to the ocular surface, but the difficulty involved in handling these systems along with bad patient complia nce has led to limited applications. To avoid all these issues, contac t lenses have been widely stud ied due to the high degree of comfort and biocompatibility. On insertion of a medicated contact lens in the eye, drug diffuses through the lens matrix into the thin tear f ilm named post-lens tear film (POLTF) trapped between the lens and the cornea, and the drug has a residence time about 30 min in the eye [14,15]. An increase in the residence time leads to a significant increase in the bioavailability. Both mathematical models and c linical data suggest that the bioavailabil ity for ophthalmic drug delivery using contact lenses can be as large as 50% [16]. This work addresses the issue of drug de livery using microemulsion and micelles laden contact lenses. An in-depth understanding of drug interaction with contact lens material followed by changes in drug interaction in the presence of microemulsions and micelles inside the contact lenses is presented by rigorous e xperiments and by proposing some new models to understand these interactions. Th e focus of this work was to tackle the most common ocular disease, dry eyes, by using the currently approve d drug called Cyclosporin A (CyA). Chapter 1

PAGE 21

21 focuses on drug release from contact lenses load ed with drug nanoparticle s and key factors in polymer drug interactions are evaluated in this chapter. Dry eye disease is mainly treated by instilling artificial tears into the eyes. Dry eyes syndrome, one of the most commonly occurring pr oblems in the world, can be broadly classified into mild, moderate, and severe depending upon the intensity of the disease. Mild damages to cornea and conjunctiva can be treated readily usin g tear substitutes available in the market. More frequent instillation of artificial tears is needed in the case of moderate damage and this can again be treated readily by using commercial products. Severe dry eye syndrome on the other hand can lead to far greater damage to the ocular surface, both in the cornea and conjunctiva, and simple treatments like artificial tears and oint ments can be of minimal use. Also, mimicking the complex composition of real tears has been a ma jor challenge in the industry. Significant advances have been made to develop systems which resemble the real ocular physiology and current artific ial tears available to the patients come with various additives which have specific roles such as providing co mfort, improving retention time on the corneal epithelium and maintaining pH le vels at physiological values. Natural tears include salts, proteins, lipids, and hydrocarbons and the tear film is made up of three layers: a mucin layer in contact with the ocular surface, an aqueous layer above this mucin layer with a lipid bi-layer in contact with the environment. Artificial tears cannot generate th e three layer structure of tears which is necessary for its effective function [17-19]. Another proposed way of treating this disease is by blocking the lacrim al drainage route so that the residence time of tears on the ocular surf ace increases [20]. This treatment can lead to a build-up of inflamed tears on the ocular surface that can further enhance the severity of the condition by influencing the infla mmation cascade [21]. Loss of corneal sensation has also been

PAGE 22

22 reported as an associated problem with this technique [22]. Overal l, significant problems associated with this method can lead to furthe r discomfort on the patients eye and may not be well accepted. Elimination of tears from the ocular surf ace can also be controlled by controlling the evaporation rate of the tears fr om the ocular surface. This can be achieved by wearing goggles or by maintaining some humid environment arou nd the eyes, but this can lead to minimal comfort as the evaporation rate cannot be altere d significantly and thus cannot help in treating the disease. Increasing tear production by stimulating the lacr imal gland by drugs such as pilocarpine has also been studied [23]. Excessi ve production of inflamed tears from already suffering eyes can lead to enhanced severity be cause of increased inflammatory response of the immune system. All the above listed methods can give relief to the patients during the disease and may cure the disease if it is not very severe, but none of them address treating the underlying cause of the disease. Though still under scru tiny, physicians believe that dr y eyes is caused by an antiinflammatory immune response of the body which itself leads to inflammation. Dry eye syndrome can be triggered by common reasons such as living in a very dry environment or due to some infection in the eye. This results in inflammation on the ocular surface causing an immune response of the body whic h activates the T-cel ls which in turn produce cytokines to fight the inflammation. The ensuing inflammation, of both the ocular surface and the lacrimal gland results in production of inflamed tears. Si nce the tears are not normal, the irritated eye is not properly nourished or lubricat ed, encouraging the cycle of inflammation to repeat. CyA prevents T-cell activation, breaking the inflammato ry cycle and hence treating the cause of the disease.

PAGE 23

23 CyA is a cyclic polypeptide consisting of 11 amino acids. It is the most commonly used immunosuppressant, and it is prescribed for a numb er of ophthalmic applications such as dry eyes [24], uveitis in children and adolescents [25], vernal keratoconjunctivitis [26], and peripheral ulcerative keratitis [27]. Due to its li mited solubility in water, it is not useful to formulate an aqueous solution of CyA, and thus a variety of systems such as solutions of drug in castor oil [28] and olive o il [29] have been explored in litera ture for controlled topical delivery of this drug in eyes. Since CyA is a lipophilic compound, it seems r easonable to use the oils to instill the drug but these studies show that CyA does not penetrate the cornea and there is a small bioavailability of CyA at the corneal surface. Furthermore, these oils do not have a good compatibility with eyes and cause irritation, bl urred vision and toxic effects. Some of the problems associated with oily solutions of dr ug can be eliminated by formulating drug loaded emulsions in water. Emulsions of castor oil have been investigated for delivery of CyA in albino rabbits and beagle dogs. These systems showed promising results but were only effective for a period of about 12 hours [30]. Other types of emulsions, particularly the positively charged emulsions such as triglyceride emulsion stabilized by -tecopherol, which show extended interaction with negatively charged epithelial co rneal cells, have also been extensively studied for ophthalmic delivery of CyA [31]. These system s seem to enhance the retention time of the drug but are not well tolerated in the eyes. Cyclodextrins, which have an internal lipophilic and outer hydrophilic region, can trap the hydrophobic drug in th e internal core and also be dispersed in the aqueous tear fluid. The studies on loading of CyA in the hydrophobic core of the cyclodextrins increase the pene tration by a factor of about 10 in comparison to the oil based systems [32]. The major problem with the CyA release by cyclodextri n is the lack of a continuous delivery which makes the drug delive ry ineffective. Penetration enhancers like

PAGE 24

24 benzalkonium chloride, dimethylsulfoxide and Cremophor-EL have been used to increase the permeation of CyA [33]. These substances increased the permeability of CyA across the cornea, but only marginally compared to pure CyA and the release were observed for only a few hours. Also, these enhancers are poorly tolerated by eyes and add to th e toxicity level in blood. Nonionic surfactants such as polyoxy l 40 stearate have been used to form micelles and increase CyA solubility in aqueous medium. These susp ension show a 60 fold increase in the corneal uptake compared to oil water emulsions and other oil based delivery system s [34]. However, the stability of micelles and their low shelf life is one major disadvantage of this system. Liposome particles formed by phospholipids namely phosph atidylcholine and phosphatidylserine, have been used as a vehicle to deliver CyA [35]. There is large upload efficiency in these particles but there short retention times make them of limite d use. Nanoparticles such as chitosan particles have also been researched as poten tial delivery vehicles for CyA. They are of significant interest because of their good tolerance and higher cornea l permeability [36]. These systems release drugs for as long as 48 hours, but a major part of the drug has been shown to diffuse in the initial hour making sustained delivery difficult [37]. A nother major problem associated with chitosan is the lack of reproducibility in drug release, which occurs due to the heterogeneities in the structure of this naturally occurring polymer. Later studies of positively charged chitosan particles show only a slight improvement in bioa vailability and retention time as compared to uncharged chitosan particles [38]. Collagen shie lds have been another popular vehicle to study CyA to enhance contact time and continuous delive ry to ocular tissues [39]. These shields can be effective for a period of about eight hours but patient discomfort due to blurring of vision and difficulty in self administrati on by patients leads to a lack of acceptability. Disaggregating collagen matrix in the eyes during implant is a nother major concern. Lastly, prodrugs have been

PAGE 25

25 studied to enhance penetration of CyA in eyes [ 40]. The only advantage of this system seems to be better tolerance in the system, but the re tention time of the drug is low making the method less affective. A controlled way of deliver ing CyA for an extended period of time using vinylpyrrolidone and HEMA polymer has been discussed in the literature [41]. Authors looked into the affect of different copolymer compositions on drug release data but no further investigation is presente d on the release rate dependency on di fferent drug loading. They have also looked in to the in-vivo studies for a particular composition [42]. As discussed earlier, contact le nses can increase the bioavailability of the ocular drugs by increasing the residence time on the ocular surf ace. Also, easy avai lability and ease of application make them a suitable vehicle for drug delivery. Soft contact lenses as a drug delivery vehicle were first used in 1965 [ 43]. The major problem with comme rcial contact lenses is that most of the drug diffuses from these systems within a few hours [44]. Drugs like cromolyn sodium, ketotifen fumarate, ketorolac tromet hamine, dexamethasone sodium phosphate [45], timolol [46], pilocarpin e [47], and fluoroquinolon es [48] have been studied for uptake and release by soft contact lenses. None of these drugs seem to release drugs for more than 6 h. A number of studies have been conducted for uptake of the drug by soaking the lens in concentrated drug solution followed by in vitro or in vivo release studies [49-57]. The major problem of loading drug by this method is that in most cases the loading capacity of the soaked contact lenses is inadequate. The drug loading capacity could be increased by 2-3 times by designing a molecularly imprinted soft contact lens [58,59]. A nother commonly used method of entrapping drugs in gels is direct addition of drug in the polymerizing medium [60-62]. None of these methods seem to be effective in controlling the drug release and effectively designing a system which can be tweaked according to the patients needs.

PAGE 26

26 This work represents a thorough overview of how microemulsion and micelles can be used to control delivery of CyA. Release kinetics from hydrogels is a st rong function of hydrogel properties such as pore size and swelling of th e hydrogels. Larger po re sizes can affect permeability of the drug or alternatively the di ffusion of the drug from the hydrogels and both these properties can be significantly altered by incorporati on of surfactants inside the gels. Thus, addition of surfactants inside conventional hyd rogels can not only induce structural and rheological changes but also effect the drug re lease kinetics and drug in teraction within the hydrogel, having considerable implications on drug delivery mechanisms. The drug, which was previously interacting with the polymer, can now associate with the surfactants inside the hydrogels through hydrogen bonding, electrostatic or hydrophobic in teractions. To comprehend polymer-drug-surfactant interactions it is impera tive to first clearly understand how surfactantdrug, drug-polymer and polymer-surfactant inter act with each other and what factors govern these interactions. Surfactants, which are amphiphillic molecules ca n self assemble inside a solution after a critical concentration referred to as the critical micelle c oncentration (CMC) is reached. Surfactants as drug delivery vehicles have been explored in the past mainly because of their property of self assembly. Surfactants can be used to a) enhance the permeability of drug through lipid bi-layers, b) contro l drug delivery, c) increase solubil ity of poorly soluble drugs, d) decrease toxicity inside the body due to excess dosage and e) increase drug bioavailability. Park et al explained how surfactants can enhance perm eability of drugs across the lipid bi-layer on the skin and did a thorough study of hydrophilic chain length, hydrophobic chain length and hydrophilic-lipophilic balance (HLB) on permeation of ibuprofen through outer most layer of the skin, stratum corneum [63]. There have been nu merous studies in the similar area [64-66] and

PAGE 27

27 researches have also studied use of surfactants to enhance ocular bioavailability of certain hydrophobic and hydrophilic drugs [67-69]. It is believed that penetration enhancers increase permeability of hydrophilic drug more significantly than the hydrophilic drugs. Polymeric micelles have been also shown to be effective as drug carriers and controlled delivery vehicles especially for delivery of dr ugs inside the body. Worm -like micelles prepared from degradable polymer (polyethylene oxi de (PEO)-polylactic acid) and inert (PEOpolyethylethylene and PEO-polybutadiene) we re studied for uptake of a hydrophobic drug Triameterene and release of two hydrophobic dyes, showing that these systems can be used for controlled delivery of hydrophobic molecules [70]. Authors further showed that the size of worm micelles was ideal to deliver drug into por ous tissues by conducting permeation studies of the worm micelles through agarose gels wherea s 100nm size vesicles we re unable to permeate through the same gels making vesi cles less effective than micelle s to deliver drug inside the tissues. For in vivo applications, interacti on of the micelles with th e blood plasma and other lipids and proteins inside the body determines its stability an d further usefulness as a drug carrier. Polymeric micelles are ge nerally 10-100 nm in size, smaller than liposomes but larger in size than conventional surfactant micelles (<5 nm in size) making them more stable inside the blood plasma. Due to their larger size, polymeric micelles if injected intravenously have high residence time due to reduced elimination from the body. Major advantage of this drug delivery system is the reduced toxicity of the drug as they can be used to target specific organs inside the body. This not only increases drug bioavailability, but also reduces loss of drug activity due to drug interaction with other proteins and lipids inside the body [71]. Drug loading inside a micelle ty pically depends on a) its solub ility in the continuous phase, b) initial drug loading, c) dr ug loading procedure (chemical c onjugation, physical adsorption or

PAGE 28

28 electrostatic interaction), d) drug affinity to the micellar interi or and e) pH and temperature of the system. Drug loading has also been shown to be a strong function of the aggregation number of the surfactant inside the solution [72]. Larger the hydrophobic part of the surfactant, larger would be the hydrophobic interior of the micelle resulting in increased solubility of a hydrophobic drug. On the other hand, as the hydr ophilic chain length is increased, CMC of surfactant increases resulting in lesser number of surfactant micelles for a given concentration of surfactant resulting in lesser drug solubilization. A review by Torc hillin discussed the usefulness of micelles in pharmaceutical industry as drug carri ers and other potential applications pertaining to properties and nature of surf actant-drug interactions [73]. Th e author has also discussed the possible mechanism for targeted drug delivery us ing micelles as drug delivery vehicles [74]. Polymer-surfactant interaction can be classifi ed into three categories: ionic, hydrophobic, and through hydrogen bonding. For non-ionic surfactants the interaction is predominantly hydrophobic in nature whereas in the presence of anionic and catio nic surfactants the electrostatic interaction dominates. These intera ctions can be significantly altered by changes in pH, ionic strength of the solution, presence of solutes and physiological changes such as temperature. Strong interacti ons between polymer and surfac tant once mixed together can significantly affect the polymeriza tion process, affecting the microstructure of chemical hydrogel during polymerization. Vl achou et. al. demonstrated that inco rporation of surfactants inside polymeric tablets increases wettability and wate r content on incorporatio n of non-ionic surfactant Tegobatain, though a clear understanding of polymer surfactant interaction was not presented [75]. Surfactants can alter the swelling propertie s of chemical hydrogels which can be directly correlated to the pore size change inside the hyd rogel. In some cases, the presence of oppositely charged surfactant in a polymeric matrix can also lead to collapse of the gel so that the water

PAGE 29

29 content inside the gel decreases due to ion pairing between the gel and the surf actant [76]. A more detailed study of interactions between charged and uncharged polymers with cationic (Cetylpyrridium chloride) and anionic (sodium dodecylbenzene sulfonate) surfactants has been shown by Philippova et al [77]. Th ey showed that a charged gel co llapses in presence of cationic and anionic surfactants due to el ectrostatic and hydrophobic interactions respect ively. They also discussed the charge specific interaction of uncharged polym er with cationic and anionic surfactants showing that poly acrylic acid (PAA) gels and hyd rophobically modified PAA gels swelled in presence of anionic surfactant while they collapsed in presence of cationic surfactants. Surfactants have been shown to have signi ficant effect on viscosity of the polymer mixtures especially at concentra tion above their critical aggrega tion concentration (CAC) as they start associating with the hydrophobic polymer chai ns. Furthermore, mixed micelles seem to interact more with polymers consisting of hydrophobi c units than single su rfactants since change in viscosity is more pronounced in case of mixe d micelles [78]. A significant effect of ions on surfactant interaction with polymer matrix has al so been observed. Researchers have observed that carbopol gels interact differently with i onic and non-ionic surfactants in water, 0.9% NaCl solution, and lacrimal fluid [79] So, interaction between surfa ctant-polymer changes not only due to presence of salt but also due to nature of ions. Many drugs also have amphiphillic character, interacting with polymers in si milar way as does a surfactant molecule. Using hydrogels to control drug release has been explored extensively in literature [80]. The rate of release strongly de pends on the interacti on of the solutes such as drugs with the hydrogels. Various rheological and structural changes can be introduced in hydrogels by changing physiological conditions such as pH [81] and temper ature [82] or providing some stimuli like magnetic field [83] a nd ultrasound [84]. Even presence of ions or specific chemicals

PAGE 30

30 can bring significant changes in th e gels [85]. All th ese parameters or co nditions can then be controlled to deliver drug s at specific location in the body. If some lipophilic modifications are done to the polymer, they can further lower the release rates of the drug molecule [86]. Drug, which can be bound to the polymer due to hydr ophobic attraction, hydrogen bonding or ionic pairing, can either diffuse along the polymer, i. e., by surface diffusion or by dissolving in the solution surrounding the polymer. Mathematical models to describe diffusion from nonswellable polymeric slab have been explored previously [87]. Many theoretical mechanisms have been proposed to describe the diffusion of solutes from hydrogels and these can be divided into three basic theori es: a) Free volume theory, b) Obst ruction theory and c) hydrodynamic theory [88]. These theories predict the diffusion coefficient of the solute when incorporated in various gels depending on various parameters such as volume fraction of polymer in gel, area of solute, radius of solute, length of polymer ch ain, molecular weight of the polymer, hydraulic permeability of the medium etc. These models and experiments are limited to a low solute and polymer concentration inside th e hydrogel. Drug loaded abov e a critical con centration can precipitate and form aggregates inside the hydrog els. Overall diffusion of drug molecules will then governed by diffusion of free drug molecule s and dissolution of drug aggregates. The earliest model describing drug release from a me dium in which dug concentration is above the solubility limit was done by Higuchi [89]. Properties of polymers and surfact ants and their interaction w ith the drug molecules can be combined to affect the drug release mechanism and rates. Incorporation of surfactants inside the gel matrix can enhance gel properties and fu rther enhance the drug loading capacity of hydrogels, especially if the concentration of surfactant is above CAC inside the hydrogel. Hydrophobic drugs can partition into these aggregates leading to enhanced loading, and the drug-

PAGE 31

31 laden micelles can act as depots of drug leading to extended drug release. Also, surfactants present in the release medium can enhance drug diffusion from polymeric gels to the release medium by a) increasing the solu bility of the drug in the oute r fluid and b) by lowering the interfacial tension between the gel and the releas e medium. In the gel matrix, surfactants can exist in three forms: a) free form, not interacti ng with other surfactants or the polymer b) as micellar aggregates or c) interact ing with the polymer matrix. Sim ilarly, drug also exists in three different forms: a) free form, b) inside micellar aggregates, c) adsorbed on the polymer. In most cases, drug interacting with surfac tant aggregates controls the drug release rates from the gel matrix. If drug has substantial affinity for the mi cellar aggregates inside the hydrogel, i.e., it has a very high partition coefficient favoring its adsorption inside th e micelles, then the free drug concentration would be less and the lower free drug concentration may lead to slower drug release rates and longer duration of release. Paulsson and Edsman explored diffusion of hydrophobic drugs in carbopol gels loaded with Brij 58 and sodium dodecyl su lfate (SDS) and showed that as the hydrophobic nature of the drugs was increased, there was a significant decrea se in the diffusion rates [90]. They concluded that the reduction in diffusion ra te can be attributed to the lipophilic interactions between the drug and the surfactant micelles. They also showed that the interacti ons between charged drug and oppositely charged surfactant can further de crease the diffusion of the drug [91,92]. The polymer content in all their formul ations was less than 2%. Lin et al also explored carbopol gels and showed that pluronic F-127 surfactant can be used to control the release of the drug especially if the gel and surfactan t are mixed in a particular ratio [93]. Liu et al used a cationic surfactant (dodecyl trimethyl ammonium bromide) a nd an anionic surfactant (SDS) to solubilize a hydrophobic drug Camtothecin (CPT), and the surfactant-drug mixture was loaded in agarose

PAGE 32

32 gels [94,95], They showed that the releas e of CPT was slowed down with increasing concentration of surfactant. In a later study, authors reported that if a mixture of a cationic surfactant (cetyltrimethylamm onium bromide) and an ani onic hydrophilic polysaccharide (kcarrageenan) were introduced in agarose gels, they further affected the re lease of the drug [96]. Concheiro et al explored the changes in microviscos ity of mixtures due to presence of surfactants and suggested that these systems could be used in ophthalmic applications to increase the retention time of eye drops and thus prolong the release of the drug to the oc ular tissues [97]. The gel-surfactant-drug interactio ns and the consequences on the drug release rates have also been reviewed in detail by Alvarez-Lorenzo and Concheiro [98]. Wu et al did a study on interaction of the drug, lidocaine hydrochloride w ith silica based xerogels in presence of a nonionic surfactant Igepal CO 720 [9 9]. They found that drug re lease was slowed down due to hydrophobic interactions with surf actant micelles but more inte restingly, surfactant release increased as drug concentration was increased insi de the xerogel due to reduced interaction of the surfactant with the gel. Thermosensitive polym ers as controlled delivery vehicles have also been explored. Above a certain temperature, call ed the cloud point, the pol ymer is insoluble in a solution thereby not releasing any drug. Wher eas below the cloud point, polymer starts dissolving, gradually eluding th e drug molecules. This phe nomenon can be used to design systems with polymers having a cloud point less than the physiological temperature of 37 oC. Cloud point of polymers can be increased by incor porating surfactants insi de the polymer matrix and delay in drug release can be controlled by changing surfactant con centration inside the polymers [100]. Despite a significant focus on incorporating su rfactants in hydrogels for impeding drug release rates, there has been very less work on drug interaction with surfactant containing

PAGE 33

33 chemical hydrogels where polymer content is cons iderably higher, for example, hydrogel contact lenses which have 60% polymer w/ w when swollen. In such system s, there is a competing affect of drug interaction with the polymer matrix and the surfactant micelles. It is thus essential that for the surfactants to retard drug transport in thes e systems, the surfactant aggregates must have a very high partitioning for the drug compared to the hydrogel. Incorporat ion of microemulsions inside polyhydroxy ethyl me thacrylate was studied to deliver li docaine at therapeutic dosages to the ocular surface [101]. It was reported that the drug release from these gels was combination of diffusion from the hydrogel and from the microemu lsion particle. In ch apter 3 we try to use microemulsions to impede CyA release from th e gels. It is shown in this chapter that microemulsion laden gels and surfactant laden gels utilizing same surfactant, behave similarly and that oil phase has minimal role in impe ding drug release. A t horough investigation of surfactant loading, storage conditions and processing conditions are studied in this chapter. Chapter 4 focuses on modeling surfactant a nd drug release from these hydrogels. Drug release model can be used to predict the partitio n coefficient of the drug and it is shown that we can control specific parameters during formulatio n to control drug release from the hydrogels. Experimental data agrees well with the proposed model and alr eady published data from other authors is also shown to agree well with the model. In Chapter 5 we extend our understanding of these systems by studying various surfactants and elucidate factors governing the drug release rates based on su rfactant structure. In this chapter we also evaluate physical and mechanical properties of su rfactant laden gels to determine the suitability of these surfactant laden systems as contact lenses. It is shown that presence of surfactants enhances water uptake and surface properties of the hydrogels whereas there is no effect on the mechanical propertie s and transparency of the gels We also propose a model for

PAGE 34

34 viscoelastic response of the gels in this chapter. Also, a thorough understanding of microstructure change inside th e hydrogels is investigated in this chapter by performing Cryo SEM imaging on the surfactant laden gels. In troduction of surfactants can change the microstructure of the gels though not compromising the mechanical properties. After application of these hydr ogels on the ocular surface, surfactants can diffuse out and they can cause potential toxicity. Surfactants us ed in this study are non-ionic surfactants which are expected to cause minimal toxicity on the ocul ar surface. Since litera ture had scarce data on the ocular toxicity of th e surfactants used in this study, we used an in vitr o assay to evaluate their toxicity in chapter 6. The popular method to eval uate toxicity of substances is by performing a Draize eye test. The Draize eye test has been critic ized in the past for its lack of reproducibility and the cruelty associated with harsh testing c onditions for animals [102,1 03]. Alternatives to this test have been proposed, but a widely accepted model to a ssess toxicity in vivo has not yet been found. Varied levels of success have been obtained by each newly proposed method, with some researchers showing excellent correlations to in vivo data and others showing insufficient ones, sometimes even with the same method. Vian et al. showed relatively poor correlations for neutral red uptake (NRU), the MTT tetrazolium sa lt assay, and cell prolifer ation via total protein content measurements [104]. Matsukawa et al. reported mixed results for the EYTEX test, clearly showing several weaknesses of the test [105]. The use of red blood cells to measure ocular toxicity has been tout ed as quick, inexpensive, and effective [106]. The correlation between this test and the Drai ze eye score has been found to be poor, though the authors have claimed that it could effectively verify the toxicity of chemicals with Draize scores greater than 50 [107]. Okahata and Ebato us ed a lipid-coated quartz microba lance to correlate partition coefficients of surfactants between lipid bilaye rs and distilled water with Draize scores and

PAGE 35

35 found excellent correlations [108]. Perhaps the most successful type of test developed thus far has been those utilizing cell cultures to a ssess the cell permeability of the compounds in question. Cottin and Zanvit reporte d successful correlation results fr om such a test [109]. The fact remains, however, that no test has been pr oven robust enough to completely replace in vivo testing [110]. One in vitro method of assessing ocular toxicity is the utili zation of liposomes to mimic cell permeation by the test substance. The advant ages of using liposome leakage to assess ocular toxicity include low cost and the ability to assess many compounds rapi dly. Additionally, the test is quantitative and so it lack s the unpredictability that can be associated with using live cells and requires no specialized equipment or expertise to conduct. This test is based on the idea that the permeation of a test substan ce through lipid bilayers is the root cause of inducing ocular toxicity, with toxicity being caused due to leakage of cellular com ponents which increases substantially after bindin g of the test substance to the bilayer. The liposome based assay is designed so that the lipid composition of the bilayers imitates the composition of corneal epithelial cells. The te st measures the leakag e of fluorescent dye from the liposome core upon interaction with a test substance. The maximum score of th e Draize eye test is 110, with 80 out of 110 coming from the cornea alone, suggesting th at the assessment of co rneal toxicity should be the main focus of an in vitro alternative. This fact first inspired researchers to test liposomes as a possible means of assessing th e ocular toxicity of surfactan ts [111,112]. Since that time, a few others have examined liposome leakage as well [108,113]. Good correlations to in vivo data were obtained in some cases for some surfactants with gross outliers sometimes present. We show that the lack of good corre lation in some studies between the liposome based assay and the

PAGE 36

36 Draize test was due to neglect of mechanistic issu es, and that a better correlation can be obtained by designing the liposome assay after mechanistic considerations.

PAGE 37

37 CHAPTER 2 CYCLOSPORINE A RELEASE FROM PO LY(HYDOXYETHYL METHACRYLATE) HYDROGE LS 2.1 Introduction In this Chap ter we investigate the uptake a nd release of CyA from p-HEMA gels focusing on the drug polymer interaction and the dynamics of drug transport both below and above the solubility limit of the drug in the p-HEMA matrix. This is a first attempt to show that contact lenses loaded with drug above th e solubility limit as drug-nanopa rticles can be potentially used to deliver drugs like CyA at therapeutic dosages for extended periods. 2.2 Materials and Methods 2.2.1 Materials 2-Hydroxyethyl m ethacrylate (HEMA) mono mer, ethylene glycol dimethacrylate (EGDMA), Dulbeccos phosphate buffered saline (P BS) acetonitrile and HPLC grade water were purchased from Sigma-Aldrich Chemicals (St Louis, MO). 2,4,6-trimethylbenzoyl-diphenylphophineoxide (TPO) was kindly provided by Ciba (Tarrytown, NY). CyA was purchased from LC Laboratories (Woburg, MA). All the chemicals were reagent grade. Acetonitrile was filtered before use and all the other chemicals were used without further purification. 2.2.2 Methods 2.2.2.1 Synthesis of drug laden and pure p-HEMA gels p-HEMA hydrogels were synthesized by free radical solution polym erization of the monomer with chemical initiation. Drug was load ed in the p-HEMA gels by dissolving the drug in the monomer mixture before polymerization and the drug concentratio n was varied from 0.125 % to 5.25 % (w/dry gel w) for all the experime nts. Briefly, 2.7 ml of drug loaded HEMA monomer was mixed with 15 l of the crosslinker (EGDMA) and 2ml of deionized (DI) water. The solution was then degassed by bubbling nitroge n for 10 minutes. Next, 6 mg of the initiator

PAGE 38

38 (TPO) was added, and the solution was stirred at 300 rpm for 10 minutes to ensure complete solubililization of the initiator. The solution was then poured into a mold comprising two glass plates separated by a polyeste r spacer having a thickness of 200 m. The mold was then placed on Ultraviolet transilluminiator UVB-10 (Ultra Lum, Inc.) and the gel was cured by irradiating UVB light (305 nm) for 40 min. After polymeriz ation, each gel was removed from the glass mold and was cut into smaller square pieces wei ghing about 40 mg and these gels were dried at room temperature for two days before drug rel ease was initiated. p-HEMA gels without any drug were synthesized in a similar manner as de scribed above except that the drug was not mixed in the monomer solution before polymerization. 2.2.2.2 Drug detection: HPLC assay CyA concentration was m easured using a HPLC (Waters, Alliance System) equipped with a C18 reverse phase column and a UV detector [114]. The mobile phase composition was 70% acetonitrile and 30% DI water, and the column was maintained at 60C. The flow rate was fixed at 1.2 ml/min and the detection wavelength was set at 210 nm. The retention time for CyA under these conditions was 4.55 minutes, and the ca libration curve for area under the peak vs. concentration was linear (R2 = 0.995). 2.2.2.3 Drug release: Equilibrium experiments The inte raction of the p-HEMA matrix w ith the drug can be characterized by the equilibrium partition coefficient, which is the ratio of the drug c oncentration in the gel and that in PBS at equilibrium. The partition coefficien t was obtained by soaking the square gel pieces about 40 mg in weight with known drug amounts in 3.5 ml PBS and measuring dynamic drug concentrations. The drug concentrations after equilibrium were used to obtain the partition coefficient. In another set of experiments, the square gel pieces without drug were soaked in 3.5 ml drug solutions, and the dynamic drug concentrati ons in the loading solutions were measured.

PAGE 39

39 Again, the concentration after equilibrium was us ed to obtain the parti tion coefficient. Again, dynamic uptake was recorded for these systems til l there was no further upt ake of the drug by the gel. The injection volume in HPLC was set as 20 l which was significantly lower than the total fluid volume (3.5 ml) to ensure negligible volume changes during the equilibrium experiments. 2.2.2.4 Drug release: PBS change experiments Drug release experim ents were also conducte d under perfect sink c onditions by periodic replacement of the PBS. As before, square ge l pieces about 1.5X1.5 cm in size and 40 mg in weight were utilized for drug re lease in PBS change experiments. Drug release kinetics was measured by soaking the gel in 3.5 ml PBS buf fer which was replaced every 24 hours. The volume of the release medium was chosen to be 3.5 ml to approximately match the in vivo conditions of tear turnover. These experime nts were conducted till a majority of the drug diffused from the gel matrix. 2.2.2.5 Transmittance measurements Transparency of drug containing hydrogels was q uantified by measuring the transmittance of 200 m thick hydrated gels at 600 nm using a UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). Gels were soaked in 3 ml of PBS for a day before the transmittance measurements were made. 2.2.2.6 Statistical analysis Linear regression analysis to determ ine slope s, correlation coefficients and confidence intervals was done in JMP which was developed by SAS (Cary, North Carolina). Slopes were compared by determining the confidence interval for the respective systems.

PAGE 40

40 2.3 Results and Discussion 2.3.1 Drug Release: Equilibrium Experiments Figure 2-1 shows the drug releas e profiles from gels loaded with varying drug amounts through direct drug addition to the polymerizing mixture. In the same figure, uptake studies are also plotted for different initial drug PBS solutio ns. In these profiles, the percentage of drug absorbed (for loading studies) or released (for release studies) is plotte d as a function of time. The results show that all the upt ake profiles overlap w ithin 95% confidence interval (CI) and the percentage drug release profiles also overlap within 95% CI if the initial drug loading is less than or equal to 0.4% w/w in the dry gel. The rele ase profiles begin to differ if the drug loading is beyond 0.4% with the total percenta ge release decreasing with incr eased initial loading. This effect likely arises due to the fact that at high drug loadings the drug precip itates inside the gel, and so even after equilibrium is established a fraction of the dr ug is present in the gel as drug particles. This issue is further explored by calculating the partition coefficients. The partition coefficient is defined as the ratio of the concentration in the gel and the concentration in the aqueous phase at equilibrium, w gC C K (2-1) The amount of CyA in the gel phase was calcu lated by subtracting the content of CyA in PBS phase from the initial loaded content of the drug in the gel matrix at equilibrium. Similarly, for experiments in which uptake of the drug wa s measured, the amount of CyA inside the gel phase was calculated by subtracting the amount left in the solution from the initial drug dissolved. Figure 2-2 shows the dependence of partition coefficient on the equilibrium drug concentration in the release medi um. Since the partition coefficien t of the drug is very large in all the cases, it can be safely concluded that th ere is significant bindi ng of the drug to the p-

PAGE 41

41 HEMA matrix. The partition coefficient of th e drug between the polyHEMA matrix and the PBS solution is relatively constant at about 148.16 till a critical concentration Ccr of about 0.02 mM, which equals 24.05 g/ml. Beyond this critical concen tration, the partition coefficient increases very sharply. The critical value of 24.05 g/ml is close to the so lubility limit of CyA in water at room temperature, whic h has been reported to be about 27.67 g/ml [115]. We propose that the sharp increase in the partition coe fficient is evidence of the fact that for gels with sufficiently large initial drug loading even after equilibrium is reached, a fraction of the drug inside the gel is present as particles. In these systems, the drug concentration of the unaggregated form is fixed by equilibrium to be K(Cr)Cr, and the remaining amount is present as precipitates. Thus the rapid increase in the partiti on coefficient is only an ar tifact of the presence of the aggregates in the gel. In these systems, the drug in the gel is present in three possible forms: free drug dissolved in th e aqueous phase in the gel, drug bound to the polymer, and the drug particles. At equilibrium the chemical potential of the drug molecules in the aqueous phase in the gel must equal the chemical potential of the drug molecules in the release medium. If the drug molecules in the aqueous phase in the gel are unaffected by the constraining effects of the gel, one may expect that the e quilibrium concentrati on of the drug in the aqueous phase of the gel will equal the concentration in the release me dium. At sufficiently high drug loadings in the gel, the concentration of the drug in the release medium is close to the solubility limit, and this suggests that the concentration of the drug in the aqueous phase in the gel is indeed equal to the concentration in the release medium, and the high value of the partition co efficient is due to the drug bound to the p-HEMA polymer.

PAGE 42

42 2.3.2 Model for Drug Release Below the solubility lim it of the drug in the gel, the drug rele ase under perfect sink conditions at short times is given by [87], 100 4 Release Drug%2 h Dt (2-2) Here D is the effective diffusivity of the drug that includes contri bution from both bulk and surface diffusion of the drug, h is the half-thickness of the gel and t is the time of release. A detailed model for evaluating diffusivity of Cy A from p-HEMA gels has been discussed in Section 4.3.1 in chapter 4, and the value of diffusi vity determined by fitti ng the release data to the above equation is 1.44x10-14 m2/s. The drug release profiles cannot be described by the above equation if the drug loading in the gel is increased above a thres hold corresponding to the solubility limit inside the gel matrix. In such situations, as mentioned earlier and illu strated in Figure 2-3, the drug is present in three different forms: free form dissolved in the aque ous phase, bound to the polymer, and particles. When the drug concentration is below a threshold limit inside the gel matrix, most of the drug is bind to the polymer (Figure 2-3A). As the polymer matrix reaches saturation, drug aggregates start forming inside the hydrogel (Figure 2-3B ) and after further lo ading of the drug, the concentration of drug present as aggregates far exceeds the drug concentration bound to the polymer matrix (Figure 2-3C). On soaking of this gel in PBS solution, free drug and the adsorbed drug would diffuse out from the gel in to the release medium. This transport would then reduce the free drug concentration in th e gel matrix leading to breakup of the drug aggregates to compensate for the drug loss. Th is mechanism results in creation of a depletion zone near the surface which does not contain drug aggregates because these are already dissolved, and the thickn ess of this zone ( ) increases with time. Th e free concentration of the

PAGE 43

43 drug in the zone that still contai ns the aggregates must be equal to the solubility concentration (CS), and the drug contained in the aggr egates should be at concentration CA = CTOTAL CS, where CTOTAL is the initial drug concentration. Thus we get the following model to describe drug transport from gel systems when the loading is above the solubility li mit of the drug in the gel matrix, )(0 from 2 2ty y C D t C (2-3) )( from hyt CCS (2-4) With the following initial and boundary conditions, 0)0,( ytC (2-5) C)yC(t,S (2-6) yfor t C DA y C (2-7) 0),0( yt (2-8) In the above equations C is the concentrati on of the un-aggregated drug molecules, which includes both free and polymer bound drug. The first boundary condition (Equation 2-5) assumes perfect sink conditions, the second bounda ry condition (Equation 2-6) arises from continuity of concentration and the third condition (Equation 2-7) st ates that the drug flux at the intersection of the zone with drug aggregates and the one wit hout drug aggregates is equal to the amount released by the dissolution of the aggr egates. The above m odel is only valid till is less than h, and after that the transport is purely diffusive. A similar model is solved for surfactant diffu sion from p-HEMA hydrogels in Chapter 4, Section 4.3.2. The amount of drug release under the limit 1 S AC C is given by

PAGE 44

44 ASCDtCAN 22 (2-9) Where A is the surface area of the polymer gel and t is the time of release. If the above limit is satisfied, the value of CA would be equal to the total drug con centration initially loaded into the gel (CTOTAL). Thus, the percentage drug release (%) from the system can than be given by, 100 2 Release Drug% TOTAL SC t h DC (2-10) Based on the above equation, the plots of percentage drug release for systems with various CTOTAL vs. TOTALC t should be a straight lines with the same slope of h DC1002* as long as the drug concentration inside the hydrogel is above the solubility limit. This result is identical to an earlier proposed equation by Hi guchi for describing a drug re lease from ointments that contain drug as suspension [89]. We conducted release experiment s under perfect sink conditions from gels with various drug loadings to explore the validity of the model proposed above, which are described below. 2.3.3 Drug Release: PBS Replacement To validate the model developed above, it was de cided to explore the e ffect of initial drug loading on release rates of the drug from the p-HEMA gels. The release experiments were conducted in 3.5 ml PBS with PBS replaced ever y day, and the drug loading was varied from 0.125% to 5% inside the gel matrix. The th ickness of the gel was ke pt constant at 200 m. The data from all the experiments is plotted in Figure 2-4 as percentage release vs. time. As expected the percentage release profiles do not overlap, as would be the case for gels with initial drug loadings below the solubility limit. In Figur e 2-5, the percentage release is plotted vs. TOTALC t The release curves overlap for initial dr ug loadings larger than or equal to 0.77%

PAGE 45

45 but the profile for 0.125% is clearly different. This suggests that the solubility limit in the gels is above 0.125%. The slope from the data with higher drug loadings represents h DCS1002 and has a value of 17.68.47. Utilizing the value of diffusivity of the drug evaluated from the pHEMA matrix (1.44x10-14 m2/s), we can determine the solubility, CS, of the drug in the p-HEMA matrix to be 3mM, which equals 0.37% w/w in the dry gel. The results shown above prove the validity of the developed model. Also, the release duration of the drug can be significantly altere d once the drug forms aggregates inside the hydrogel as can be seen in Figure 2-5. Thus, the desired release rate s of the drug can be achieved by loading varying amount of drug inside hydrogels, e.g., contact lenses. 2.3.4 Theoretical Model Brinkman combined Darcys law and Navier-S tokes equation to describe diffusion of solutes through a porous medium su ch as a hydrogel [143]. Phillips et al. confirmed the validity of using Brinkmans equation for solute diffu sion through a porous medium by both experiments and Stokesian dynamics [144]. The following eq uation can then be used for evaluating diffusivity of a molecule from a hydrogel, 1 2 2/1 23 1 1 6 k r k r r Tk DB TH (2-11) Where, DTH is the theoretically determined diffusiv ity of the molecule, r is the solute radius, k is the hydraulic permeability of the medium, is the viscosity of water, T is the temperature and kB is the Boltzmanns constant. Later it was suggested that Equation 2-11 is valid only if there is a pressure driven flow and in its absence, the coefficient of r2/k should be 1/9 instead of 1/3 [145]. Due to lack of pressure driven flow in our syst ems, we use a modified Equation 2-11 with 1/3 replaced by 1/9. The values of various parameters needed for evaluating

PAGE 46

46 the theoretical diffusivity are listed in Table 2-1 and the value of the diffusivity for CyA is determined to be, 5.6x10-12 m2/s. This value is significantly different from the experimentally determined value of 1.44x10-14 m2/s. To reconcile this difference, we reiterative that below the solubility limit, the drug mol ecules in the gel are present both in free and polymer-bound form. The effective diffusivity obtained from the macros copic transport model is a combination of both surface and bulk diffusion and can be defined as, K fKDfD DSUf)( (2-12) Where, Df and DSU are the diffusivities of the free a nd the bound drug, respectively, f is the fraction of water in the gel which is 0.4 fo r p-HEMA gels, and K is the partition coefficient determined in section 2.3.1 as 148. It is imp licitly assumed in the a bove equation that the free and the bound form are always in equilibrium, i,e., the binding-unbinding events occur on a time scale much faster than diffusion. The surface diffusion is expected to be smaller than the bulk diffusion due to friction with the polymer and so as an extreme case we assume DSU to be 0, and evaluate Df from Equation 2-12 to be 5.33x10-12 m2/s. This value is similar to that evaluated from the Brinkman model for solute diffusivity in porous medium suggesting that hydrodynamic interactions of the polymer matrix with the dr ug molecule have a sign ificant contribution in molecular transport from p-HEMA gels. This also suggests that surface diffusivity for CyA inside the gel matrix is negligible. 2.3.5 Effect of Drug Concentration on Transparency The presence of drug aggregates at loadings beyond the solubility li mit could potentially cause scattering leading to a loss of transparen cy, which is a critical requirement for contact lenses. To explore this issue, the transmittance of the gels was measured, and these are reported in Figure 2-6. The gels are transparent with tr ansmittance >99% for drug loadings below 0.3%.

PAGE 47

47 On increasing the drug loading transparency decrea ses, the gels become hazy, and at loadings beyond 5.25%, the gels are almost opaque. Thus the transition point for loss of transparency coincides with the solubility limit that was determin ed to be 3mM (0.37%). This suggests that while gels with drug particles may not be suitable for contact lens appli cations. It is however noted that the size of the drug particles will li kely depend on the polymerization dynamics and if conditions are found in which the particle aggr egates are nanosized, the gels may retain transparency even above the solubility limit. 2.4 Conclusion The drug CyA can be released from the 200 m thick pHEMA gels for about a week, which is consistent with prior reports from VP -HEMA copolymer gels [41]. Here the authors did not mention the volume of the container in which drug releas e was conducted and so a direct comparison of their results with our results is no t possible. Since the dr ug transport is diffusion controlled, 100 m thick gels will release drug only for about 1.5 days. The drug CyA exhibits strong interaction with the p-HEMA matrix as evident by the high and co ncentration independent partition coefficient of 148.16. The releas e duration of the drug from the gel can be increased by loading drug above the solubility limit. This however l eads to a re duction of transparency which is undesirable for contact lens applications. Th e release of drug from systems with drug particles can be described by the same model as proposed earlier for release from drug suspension in ointments. The diffusive transport of the drug mainly occurs due to diffusion of the free drug. The polymer bound dr ug does not diffuse along the polymer chains, but it always in equilibri um with the free drug, and thus su rface bound drug can desorb and then diffuse. The diffusivity of the drug CyA in the p-HEMA gels can be described by the Brinkman model with some modifications to account for the absence of pressure driven flow.

PAGE 48

48 0 10 20 30 40 50 60 70 80 90 100 0246 Time (Days)% Drug Release Uptake 1 Uptake 2 Uptake 3 Uptake 4 1st Release, CyA = 0.125% 1st Release, CyA = 0.26% 1st Release, CyA = 0.35% 1st Release, CyA = 0.4% 1st Release, CyA = 0.7% 1st Release, CyA = 0.85% 1st Release, CyA = 1.98% 1st Release, CyA = 5.25% Figure 2-1. Percentage drug release from 200 m thick p-HEMA hydrogel. Amount of drug loaded in the hydrogel is indicated. Upta ke experiments were conducted from a drug solution where starting concentr ation of CyA in PBS was 11 g/ml for Uptake 1 & Uptake 2 and it was 14 g/ml for Uptake 3 &Uptake 4. 0 200 400 600 800 1000 1200 1400 1600 00.0050.010.0150.020.025 Cw (mM)K Figure 2-2. Partition coefficient plotted against c oncentration of drug in the PBS at equilibrium. There is a significant jump in the partition coefficient around 0.02 mM which is the solubility limit of Cy A in the PBS buffer.

PAGE 49

49 Figure 2-3. A schematic of drug interaction w ith the p-HEMA matrix as drug loading is increased. A) Drug loading is below the solubility limit and most of the drug is adsorbed on the polymer matrix. B) As the drug loading is increased beyond the solubility limit of the drug in the hydrogel matrix, drug starts to precipitates and forms aggregates inside the hydrogel. C) At a very high drug loading, concentration of drug aggregates is much higher than the concentration of polymer bound drug. Increasing Drug Concentration A B C

PAGE 50

50 0 10 20 30 40 50 60 70 80 90 100 0102030 Time (Days)% Drug Release CyA = 0.125% CyA = 0.77% CyA = 0.85% CyA = 1.98% CyA = 3.92% CyA = 5.25% Figure 2-4. Effect of drug loadi ng on cumulative drug release from 200 m thick gels in PBS change experiments Slope = 17.68 R2 = 0.9952 0 10 20 30 40 50 60 70 80 90 100 05101520% Drug Release CyA = 0.125% CyA = 0.77% CyA = 0.85% CyA = 1.98% CyA = 3.92% CyA = 5.25% Model Figure 2-5. Cumulative percenta ge release of drug from 200 m thick surfactant-laden gels in PBS change experiments after rescaling the time. represents t/CTOTAL where t is time (h) and CTOTAL (mM) is the total concentration of drug present inside the gel matrix.

PAGE 51

51 0 10 20 30 40 50 60 70 80 90 100 0.00%1.00%2.00%3.00%4.00%5.00%6.00% % CyA LoadingTransmittance (%) Figure 2-6. Transmittance of the gel with increasing drug concentrations. Transmittance values start to go down as the drug loading inside the hydrogel reaches the solubility limit.

PAGE 52

52 Table 2-1. Physical properties of CyA at room temperature Properties of CyA Values Diffusivity 1.44x10-14 m2/s Solubility in p-HEMA matrix 3 mM Solubility in PBS 0.02 mM Solubility un Water 0.023 mM

PAGE 53

53 CHAPTER 3 CYCLOSPORINE A RELEASE FROM BRIJ 97 MICROEMULSION AND SURFACTANT LADE N HYDROGELS 3.1 Introduction In this chapter we show that contact lenses made from microemulsion and surfactant-laden hydrogels can be used for extended delivery of Cyclosporine A (CyA) at therapeutic dosages. Also, we show that surfactant-laden hydrogels can go through all the pr ocessing steps that a typical contact lens goes through including monome r extraction, autoclaving and packaging, and still provide extended dr ug release at therapeuti c dosages. The results of this study provide strong evidence that microemulsion and/or surf actant-laden contact lenses can be used for extended delivery of various ophthalmic drugs including CyA. 3.2 Materials and Methods 3.2.1 Materials 2-Hydroxyethyl methacrylate (HEMA) monomer, ethylene glycol dimethacrylate (EGDMA), ethyl butyrate, Dulbeccos phosphate buffered saline (PBS), acetonitrile and polyoxyethylene (10) oleyl ether (Brij 97) were purchased from Sigma-Aldrich Chemicals (St Louis, MO). 2,4,6-trimethylbenzoyl-diphenyl -phophineoxide (TPO) was kindly provided by Ciba (Tarrytown, NY). CyA wa s purchased from LC Laborator ies (Woburg, MA). All the chemicals were reagent grade. Acetonitrile was filtered before use and all the other chemicals were used without further purification. 3.2.2 Microemulsion Formulation The surfactant solution was prepared by adding the Brij 97 surfactant to de-ionized (DI) water in the required ratio and then stirring the mixture at 600 rpm and at room temperature for a period of about 10 hours. Specifically, 1, 1.5, a nd 2 g of Brij 97 was dissolved in 10 ml DI water to prepare three different surfactant solutions (named M1, M2, M3, respectively). Separately,

PAGE 54

54 0.4 g of CyA was dissolved in 5 ml of ethyl butyr ate to prepare the drug loaded oil phase. Next, 100 l of the drug loaded oil was added to 5 ml of the surfactant soluti on, and the mixture was then stirred at 600 rpm (70oC) for 20 minutes. The soluti on was then cooled to room temperature, resulting in forma tion of a clear microemulsion. Microemulsions without the drug were synthesized by eliminating CyA in the formulation described above. 3.2.3 Particle Size Analysis The particle sizes for microemulsions were measured using a Precision Detectors PDDLS/CoolBatch+90T instrument. The data was analyzed with th e Precision Deconvolve32 Program. The measurements were obtained at 20 C and 90 scattering angle, using a 683 nm laser source. 3.2.4 Preparation of Microemulsion Laden Gels The microemulsion-laden p-HEMA hydrogels were prepared by free radical solution polymerization with UV initiation. Specifically 2.7 ml of HEMA mono mer was mixed with 15 l of the crosslinker (EGDMA) a nd 2ml of the CyA containing microemulsion. The solution was then degassed by bubbling nitrogen for 10 minut es. Next, 6 mg of the initiator (TPO) was added, and the solution was stirred at 300 rpm for 10 minutes to ensure co mplete solubililization of the initiator. The solution was then poured in a mold that comprised two glass plates separated by a 200 m (thick gels) or 100 m (thin gels) thick spacer. The mold was placed on Ultraviolet transilluminiator UVB-10 (Ultra Lum, Inc.) and the gel wa s cured by irradiating UVB light (305 nm) for 40 min. The gels load ed with microemulsions M1, M2 and M3 were named M1, M2 and M3 gels, respectively. Microemulsions without any dr ug were incorporated in the polymerizing mixture for synthesizing gels with no drug and the synthesis protocol was same as described above.

PAGE 55

55 3.2.5 Preparation of Surfactant Laden Gels The surfactant solution was prepared as describe d earlier. Specifically, 0.2, 0.6, 1.5 g of Brij 97 surfactant was dissolved in 10 ml DI wa ter to make three different surfactant solutions (named S1, S2, S3, respectively). Separately, 3. 5 mg of CyA was dissolved in 2.7 ml of HEMA monomer and stirred at 600 rpm fo r a period of 5 hours. Next 15 l of the crosslinker and 2ml of surfactant solution we re added to the 2.7 ml of drug loaded monomer. The hydrogels were then prepared by adding the mixture to the molds followed by UV curing, as described above. Control, drug loaded p-HEMA gels without surfactants (D1) were prepared by following procedures identical to those described above except that the 2 ml surfactant solution was replaced by 2 ml DI water. Also, surfactant la den gels without any drug were synthesized in a similar manner as above by not incorporating drug in the monomer mixture. 3.2.6 CyA Detection by HPLC CyA concentration was measured using a HPLC (Waters, Alliance System) equipped with a C18 reverse phase column and a UV detector [114]. The mobile phase composition was 70% acetonitrile and 30% DI water, and the column was maintained at 60C. The flow rate was fixed at 1.2 ml/min and the detection wavelength was set at 210 nm. The retention time for CyA under these conditions was 4.55 minutes, and the ca libration curve for area under the peak vs. concentration was linear (R2 = 0.995). 3.2.7 Drug Release Kinetics from Gels Loaded with CyA by Drug Addition to the Monomer After polymerization, each gel was removed from the glass mold and was cut into smaller pieces that were about 1.5X1.5 cm (for thick gels) and 1.5X3 cm (for thin gels) in size and about 40 mg in weight. Drug release kinetics was m easured by soaking the gel in 3.5 ml PBS buffer which was replaced every 24 hours a nd all the measurements were done at room temperature. The volume of the release medium was chosen to be 3.5 ml to approximately match the in vivo

PAGE 56

56 conditions of tear turnover. Additionally, some experime nts were conducted without PBS replacement till the gel and the release medium equilibrated. These experiments were conducted to explore the rate limiting step in the drug tr ansport by conducting drug release from gels with two different thicknesses. The dynamic drug c oncentrations in the release medium were measured for both sets of the drug release e xperiments by the HPLC method described above. The injection volume in HPLC was set as 20 l which was significantly lower than the total fluid volume (3.5 ml) to ensure negligible volume changes during the equilibrium experiments. 3.2.8 Drug Uptake and Release Kinetics from Ge ls Loaded with CyA after Polymerization In the release protocol described above, Cy A was loaded in the hydrogels by dissolving it into the oil phase of the microemulsion. It is conceivable that the proc ess of gel formation may lead to partial loss of drug activity and some irreversible entrapment of the drug. To eliminate the possible loss of activity due to the polym erization process, it was decided to conduct experiments in which the microemulsions (without drug) were entrapped in the gel, and the drug was loaded by soaking the gels into aqueous dr ug solutions. Specifically, drug was loaded by soaking the gels, about 40 mg in weight, in 4 ml of 11.5 g/ml drug solution. To determine the time needed for uptake of drug by the microemuls ion-laden gels, the duration of soaking period was varied between 5 and 15 days. The rele ase kinetics was subs equently measured by following the same protocols as described in the previous section and these results are discussed in section 3.3.2. 3.2.9 Packaging Solution for Drug Release To explore the effect of packaging solution on drug release from hydr ogels containing drug incorporated inside the microemulsions, it was decided to soak the drug containing microemulsion laden gels (M2) in 1.5 ml of a packaging medium for specific durations and then

PAGE 57

57 measure drug release kinetics from these gels. The duration of soaking in packaging solutions was varied from 10 100 days to evaluate the effect of storage on thes e hydrogels. Also, three different compositions of packag ing solutions were explored. The first packaging medium was simply DI water, and the second and the thir d were 0.85% and 4.25% w/w salt solutions, respectively. The salt concentr ation of 0.85% (0.14 M) was chosen to match the typical concentration in commercial p ackaging solution [116], and higher (4.25%) and lower (DI water) salt concentration was used to observe the effect of salt on equilibrium amount of CyA released in the packaging solution. Drug release from th ese gels after packaging period were carried out in 3.5 ml of PBS with daily PBS replacement as described earlier, and release kinetics from gels used for packaging studies are discussed in section 3.3.3. 3.2.10 Processing Conditions in Contact Lens Manufacturing To evaluate the suitability of the Brij 97 microemulsion and surfactant laden gels as contact lenses, it was decided to subject these ge ls through processing conditions similar to those used in contact lens manufacturing. The gels we re first subjected to an extraction stage in which the un-reacted monomer was extracted from the gels by soaking gels in 10 ml of DI water at 50oC. The DI water was replaced every 5 minutes, and this step was repeated 5 times. After extraction, each gel was soaked in 4 ml of CyA-water (12 g/ml) for a period of 12 days. Each gel was then soaked in 1.5 ml of DI water and autoclaved for 15 min at 121oC followed by storage at room temperature for a period of 10 days. In the fina l step, each gel was submerged in 3.5 ml of PBS, which was replaced every 24 hours, and the concentr ation of the drug was measured to determine the release kinetics. The results from this study are discussed in section 3.3.6.

PAGE 58

58 3.2.11 Surfactant Release To measure the surfacta nt release from the gels, surfactan t laden gel samples were soaked in 3.5 ml of DI water. The DI water was repl aced after regular intervals and the surface tension of the solution was measured by the Wilhelmy pl ate method to determine the concentration of Brij 97. We used a sand blasted platinum plate attached to a Scaime France Microbalance which was further connected to a Stathan Universal tran sducer (SC001). The transducer was calibrated by using DI water ( = 72 mN/m) and acetone ( = 23 mN/m) as standards. To ensure complete removal of impurities, the platinum plate was rinsed with DI water and acetone, followed by annealing till red hot using a propane burner before each measurement. For each measurement, the solution was kept still for a period of 1 hour to ensure an equilibrium surface coverage of surfactant at the ai r-liquid interface. 3.2.12 Statistical Analysis Linear regression analysis to determine sl opes, correlation coefficients and confidence intervals was performed in JMP developed by SA S (Cary, North Carolina). The 95% confidence intervals (CI) were obtained to compare release kinetics. 3.3 Results and Discussion 3.3.1 Particle Size Analysis of Microemulsio ns and Drug Release from MicroemulsionLaden Gels Figure 3-1 plots the particle size distributions for the thr ee microemulsions that were explored in this study. These microemulsions have mean particle sizes ranging from 10-13 nm, which is typical for microemulsions. The mean particle size increased with a reduction in surfactant loading, which was expected. The dr ug release profiles (with PBS change every 24 hours) from a control p-HEMA gels and gels lo aded with microemulsions are compared in Figure 3-2. The amount of surfactant in the th ree systems was 5.6%, 8% and 9.4% w/dry gel w

PAGE 59

59 for M1, M2 and M3, respectively. The CyA release from p-HEMA gels last only about 6-7 days but the microemulsion-laden gels release drug for about 20 days. The results in Figure 2 clearly demonstrate a significant reduction in delivery rate and a concurrent increase in the duration of release on addition of microemulsions to the gels. We speculate that since CyA is a hydrophobic molecule, it preferentially partitions into the oil phase of the microemulsions, leading to a reduction in the free drug concentra tion, and thus a reduction in the drug flux. It is also possible that during the gel preparation and subsequent h ydration, some surfactant molecules desorb from the oil drops and then aggregate in the gel pores to form micelles. The presence of micelles is also expected to retard drug transport from the hydrogels. The short time release from a hydrogel can be described by the following equation [87], 100 Dt 4Release Drug%2 h (3-1) Equation 3-1 is valid for short time scales when the released drug percentage is less than 60%. In Equation 3-1, D is the effective diffusiv ity of the drug, h is the thickness of the gel, and t is the release time. The release data shown in Figure 3-2 we re fitted to the above equation to determine the effective diffusivity of the drug for the control p-HEMA gel and the microemulsion-laden gels. Based on Equati on 3-1, a plot of cumulative release vs. t must be a straight line. The fit between the data and the model is shown in Figure 3-3, and the values of the slopes along with the 95% CI are listed in Ta ble 3-1. For clarity, on ly average values for each system are plotted while th e fitted line is evaluated by us ing all the data points for each system. The slopes were then ut ilized to determine the effective diffusivities, which are also listed in Table 3-1. We observe that the dr ug release from the microemulsion laden systems depends on surfactant loading, a nd effective diffusivity decreas es with increasing surfactant loading. It is noted that the cumulative rel ease profiles for the microemulsion-laden gels are

PAGE 60

60 linear at short times but they intercept the x-axis at about t = 2.9, implying t ~ 8.4 h. This suggests that at very short times the drug transport rates are much smaller than thos e predicted by diffusion mechanism, leading to a delay in releas e. The delay in drug release could potentially be caused due to the time needed to hydrate the interfacial region of the microemulsions. A similar phenomenon is observed with surfactant-lade n gels (see section 3.3. 4) which supports the hypothesis of delay caused by hydration of micelles or microemulsion interface. 3.3.2 Release of Drug after Soaking Microem ulsion-Laden Gels in Drug Solution As explained in section 3.2.8, in some expe riments, drug was loaded into the gels after polymerization by soaking them in drug solution. After the soaking phase, the gels were withdrawn and the concentration of drug in the aqueous phase wa s measured. The mass of drug taken up by the gels was determined by subtracting the mass of drug left in the solution from the initial mass of drug in the soaking solution. The systems explored here had 8% surfactant in the dry gel (w/dry gel w), and Table 3-2 lists the amounts of drug th at were taken up by these gels for the different soaking durations The results in Table 3-2 show that the mass of the drug taken up by the gels is relatively similar for all the gels This shows that 5 days of soaking time is sufficient to establish equilibrium. The drug rele ase profiles shown in Figure 3-4 are within 95% CI of each other (CI not shown in the plot), whic h is expected because each gel absorbed similar amount of drug. These results also show that the duration of drug release for the systems in which the drug is loaded by soaking in the drug so lution is similar to the systems in which the drug is entrapped in the microemulsions before polymerization, wh ich suggests that the surfactant loss during the drug loading step was negligible. 3.3.3 Effect of Packaging Co nditions on Drug Release At the end of the packaging phase described in section 3.2.9, the ge ls were withdrawn and the concentration of drug in the liquid was measured to determine the amount of drug that

PAGE 61

61 diffused out during this step (Table 3-3). The amount of drug that diffuses out of the gel during storage is less in salt solutions because CyA is a hydr ophobic drug and so increasing ionic strength reduces drug solubility. Also, the amount of drug released into the packaging solution is largest for the 100 day soak whic h shows that the equilibration time for drug release from these gels is at least more than 30 days. After the end of the storage phase, the gels were withdrawn and dr ug release experiments were conducted as described earlier. At short ti mes, all the drug release profiles in Figure 3-5 are within 95% CI except the 100 day soak in 0. 85% salt solution. Also the release duration from these systems is comparable to that from ge ls that were not subjected to packaging (Figure 3-4). These results demonstrate that the dr ug release profiles are relatively unaffected by soaking in packaging solution, and also by th e composition of the p ackaging solution. These results are encouraging but the micr oemulsion-laden gels also have several drawbacks. Firstly, preparation of microemulsion-laden gels is a two step process, which renders it cumbersome. Secondly, although th e oil phase of the microemulsi on is only slightly soluble in tears, it will likely elude at a slow rate, and thus could potentially cause ocular t oxicity. Ethyl butyrate is food grade oil suitable for in vivo ap plications [117,118], but to our knowledge ocular toxicity of ethyl butyrate has not been investigated. To avoid pote ntial ocular toxicity due to oil and to simplify the gel preparation, it is desi rable to replace the microemulsions with micelles which may also impede drug release leading to ex tended drug delivery. To test this hypothesis, surfactant laden gels were fabricated, and drug re lease studies from these systems are described in section 3.3.4. 3.3.4 Drug Release from Micelle Laden Hydrogels Figure 3-6 shows the drug release profile for S3 gels, i.e., gels loaded with 8% surfactant (w/dry gel w) with daily PBS change. The drug release profiles from microemulsion-laden gels

PAGE 62

62 (M2) with similar surfactant loading are included in the figure for comparison. The results show that surfactant-laden gels also provide extended drug release la sting more than 20 days. The release rates of the microemulsion-laden gels are less than that for the surfactant-laden gels with 8% surfactant loading sugge sting that the presence of oil further slows down drug transport. The effective diffusivity of the drug from the S3 gels was obtained by fitting the short time data to Equation 3-1 (Figure 3-7). The fitted diffusivity is 4.34x10-15 m2/s, which is more than that for the M2 gels within 95% CI. We speculate that the surfactant-laden gels contain micelles and the drug preferentially partitions into the hydrophobic core of these micelles. The reported values of the critical micelle concentr ation (CMC) of Brij 97 range fr om 0.217 mM [119] to 0.94 mM [120]. The hydrated p-HEMA gels absorb about 39% water w/w. Based on these values, if the surfactant loading in a gel exce eds 0.061% 0.27% (w/dry gel w), its concentration in the hydrated state is expected to be above the CMC. It is noted that this estimation neglects binding of surfactant to the gel, which is likely because the gel has some hydrophobic sites to which the surfactants are expected to adsorb. Also, th e shapes of these micellar aggregates may be complex due to the confining effects of the gel. The cumulative releas e profile for surfactant system intercepts the x-axis at t= 2.5, which lies within 95% CI of the intercepts for the microemulsion-laden gels (2.9.6), supporting th e hypothesis that the initial delay in the drug release is caused due to hydrat ion of surfactant aggregates. 3.3.5 Mechanism of Drug Release The drug release from the surfactant-laden gels could be controlled by two potential mechanisms: transport of the drug from inside the micelles trapped in the gel to the bulk gel, or diffusion through the gel. The linear rela tionship between cumulative release and t suggested that the transport is controlled by diffusion thro ugh the gel. If the release is controlled by

PAGE 63

63 diffusion, the release time scales as (thickness)2, and if the release is controlled by transport from inside the micelles to outside, the release time should be independent of thickness. To determine the rate limiting st ep, drug release profiles from 100 m thick gels were compared with those from the 200 m thick gels. These were equi librium studies in which PBS was not replaced and the system was allowed to equ ilibrate. It is noted that the weights of both the thick and thin gels were about same because the cross sectional ar ea of the thin gel was double that of the thick gel. The results from these studies are shown in Figure 3-8, where the percentage release of drug is plotted as a function of scaled time, where Scaled Time = 2microns)(in h 100 x Time (3-2) In Equation 3-2, h represen ts the thickness of the hydrogel. To compare the profiles shown in Figure 3-8, we computed the relative e rror defined as the ratio of the difference in cumulative percentage release between the thin a nd the thick gels and the cumulative release for the thick gels. We observe that the percentage re lease vs. scaled time profiles for the thin and the thick pure p-HEMA gels and also for the thick and the thin S3 gels are similar with root mean square of the relative error being 13.5% and 5.7% for the p-HEMA and the S3 gels, proving that the diffusion through the gel matrix is the rate controlling step. 3.3.6 Processing Conditions S1, S2, S3, M2 and D1 gels were prepared by following procedures described earlier, and the drug was not loaded in these gels. These gels were about 100 m thick and weighed about 40 mg. After monomer extraction described earlier, drug was loaded in these gels by soaking in drug solution and the amount of drug loaded in to the gel was determined by calculating the

PAGE 64

64 difference between the initial and the final concen trations in the drug solution. The results for the amount of drug loaded into thes e gels are shown in Table 3-4. After autoclaving and 10 day storage in 1.5 ml DI water, the concentration in the aqueous phase was measured to determine the amount of drug that was released from the gel during the autoclaving and storage period. By subtracting this amount from the amount of drug taken up by the gel, amount of drug retained by the gel was determined. These results are also shown in Table 3-4. Each gel released about 25% of the entrapped dr ug into the solution during autoclaving and packaging. The drug release profiles for the cumulative release as a function of time are shown in Figure 3-9 and the drug amounts released are shown in Table 3-4. We observe that each gel releases almost 100% of the entrapped drug. A 100% release along with the fact that the elusion spectra of the drug from the HPLC column (abs orbance vs. time) is not altered by autoclaving suggests that the drug does not de grade during processing. The duration of drug release from the surfactant and microemulsion-lade n gels is much longer than that for the pure p-HEMA gels, which shows that the processing steps, particular ly autoclaving do not ca use significant changes in the gel structure. We also believe that a significant amount of surf actant diffused out from these systems during processing and hence the release duration decrea ses due to processing conditions. Nevertheless, the sy stems with the higher surfactant loading release drug at the slower rate compared to p-HEMA gels, which su ggests that even after processing steps there is enough surfactant in these systems to attenuate the drug release rates. 3.3.7 Brij 97 Release from p-HEMA Hydrogels Surfactant is likely to diffuse into the tear fi lm after a surfactant-l aden contact lens is placed on the eye. It was thus important to measur e the rate of surfactant release from the gels. The rate of surfactant released was measured from surfactant laden gels which contained 8%

PAGE 65

65 surfactant by weight in dry gel (S3). By perf orming control experiments, it was verified that other components in the gels were not surface ac tive. During the surface tension measurement, the surface area created was small and so the ch ange in bulk concentration due to surfactant adsorption at the surface was negligible. Firstly relationship between th e surface tension and the bulk concentration of Brij 97 was measured, and th is was used as a calibration curve to later relate the measured surface tension to the bulk concentration of the su rfactant in the release experiments (Figure 3-10). To maximize the sensitivity of the measurements, the 3.5 ml solution was diluted by trial and error to surface tensions above 40 mN/m at which the surface tension is most sensitive to concentration. The percentage releases of the surfact ant from both thick and thin gels are plotted against 2 2100 h t in Figure 3-11 where h is the thickness of the hydrogel and t is time in hours. These curves when fitted with a straig ht line had a slope of 1.47.08 (R2 = 0.99) for the thick gels and 1.4.23 (R2 = 0.98) for thin gels, matching within a 95% CI. The thin gels, which were about the sa me thickness as typical c ontact lenses, released about 48% of the surfactant in a period of 65 days. This corresponds to around 1500 g of surfactant released in 65 days, or equivalently an average of 25 g/day. Brij 97 surfactant has been previously explored as oral delivery vehicle but ocular toxicity of th is surfactant has not been evaluated in literature [117,121,122]. However, similar surfactants fr om the series of Brij surfactants (Brij 35, Brij 78, Br ij 98) have been shown to be nontoxic at high concentrations on the corneal surface and have also been shown to be useful as cornea permeability enhancers [67,68,123]. Thus a slow release of surfactants from the lens could have the beneficial effect of increase in corneal permeability of the dr ug leading to increased bioavailability.

PAGE 66

66 3.4 Conclusion This chapter focused on exploring microemu lsion and surfactant-laden hydrogels for extended delivery of CyA. It was shown that by using Brij 97 surfactant, both surfactant and microemulsion-laden gels exhibit slow and extend ed drug release lasting for about 20 days. This is a significant improvement compared to the co ntrol (pure p-HEMA gels), which releases drug for less than 5 days. The duration of drug rel ease depends on the surfac tant loading, and the rates of drug release are slightly smaller for microemulsion-laden gels compared to surfactantladen gels with same surfactant loading. The transport of CyA in the surfactant-laden gels is controlled by diffusion. The hydrated gels are expected to contain surfactant aggregates and Cy A, which is a hydrophobic drug, partitions into the hydrophobic domains of these a ggregates leading to an increase in partition coefficient leading to a slow down in transport rates from the gel. The drug release profiles are unaffected by the method of drug loading. The gels which had CyA loaded by soaking in solutions had simila r results compared to those gels in which the drug was added before polymerization. The duration of drug release was longest for highest surfactant containing gels after pr ocessing conditions which include d autoclaving and packaging. These results are very encouragi ng and it seems that su rfactant or microemulsion-laden gels may be suitable for delivering CyA to eyes. In addition to treating ocular disorders, CyA has also shown promise in treating contact lens mediated dry eyes, and so these systems could also be very useful for a large population that is unable to wear contact lenses due to discomfort [124]. Furthermore, the surfactant-laden gels are expected to have better wettability that could also lead to improved comfort. While these systems are promising, it is not ed that p-HEMA lenses cannot be used for extended wear because of low oxygen permeability. Thus the surfactant-laden p-HEMA contact

PAGE 67

67 lenses will need to be taken off at night and cleaned to remove the protein and lipid deposits. The impact of these steps on CyA release needs to be assessed.

PAGE 68

68 0 0.05 0.1 0.15 0.2 0.25 010203040 Size (nm)Value (counts/s) M1 M2 M3 Figure 3-1. Size distribu tion of microemulsions with three different surfactant loadings 0 10 20 30 40 50 60 70 80 90 100 0510152025 Time (Days)% Drug Release M1 M2 M3 D1 Figure 3-2. Cumulative percentage release of drug from microemu lsion laden gels with varying surfactant loading and pure p-HEMA gels. All the gels were 200 m thick in dry state and gels M1, M2, M3 and D1 contained 48.5, 52.2, 53.4 and 53.3 g of drug, respectively. Data are plotted as mean SD fo r M2, D1 gels (n = 3). The error bars for M1 and M3 systems represent half the diffe rence between the data from two repeat runs.

PAGE 69

69 0 10 20 30 40 50 60 70 80 05101520 t0.5 (h0.5)% Drug Release M1 M2 M3 D1 Figure 3-3. Linear fits for the short time release data to obta in the effective diffusivity for microemulsion and pure p-HEMA gels 0 10 20 30 40 50 60 70 80 90 100 0510152025 Time (Days)% Drug Release 5 Day 10 Day 15 Day Figure 3-4. Cumulative percentage release of drug from microemu lsion gels after loading the drug into gels by soaking in a drug solution for 5, 10 and 15 days

PAGE 70

70 0 10 20 30 40 50 60 70 80 0510152025 Time (Days)% Drug Release Solution I_10 Solution II_10 Solution III_10 Solution I_30 Solution II_30 Solution III_30 Solution I_100 Solution II_100 Solution III_100 Figure 3-5. Cumulative percenta ge release of drug from microemulsion laden gels after packaging in three different salt solutions fo r different durations of time. Solution 1 DI water; Solution II 0.85% salt solution; Solution III 4.25% slat solution. All the gels were 200 m thick in dry state.

PAGE 71

71 0 10 20 30 40 50 60 70 80 90 100 0510152025 Time (Days)% Drug Release D1 S3 M2 Figure 3-6. Cumulative percentage release of drug from Brij 97 surfactant laden, microemulsion laden and pure p-HEMA gels All the gels were 200 m thick in dry state and gels S3, M2 and D1 contained 49.2, 52.2 and 53.3 g of drug, respectiv ely. Data are plotted as mean SD (n = 3).

PAGE 72

72 0 10 20 30 40 50 60 70 80 0 5 10 15 20 t h% Drug Release M2 S3 Figure 3-7. Linear fits for the short time release data to obta in the effective diffusivity for surfactant laden gels 0 5 10 15 20 25 30 35 40 012345 Scaled Time (Days)% Drug Release S3_Thin S3_Thick D1_Thin D1_Thick Figure 3-8. Effect of thickness on percentage rel ease for p-HEMA gels and surfactant laden gels for equilibrium experiments. S3_Thin, S 3_Thick, D1_Thin and D1_Thick contained 49.2, 52.4, 56.2 and 57.7 g of drug, respectively. Data are plotted as mean SD (n = 3).

PAGE 73

73 0 10 20 30 40 50 60 70 80 90 100 0246810 Time (Days)% Drug Release D1 M2 S2 S1 S3 Figure 3-9. Effect of surfactan t loading and processing condi tions on cumulative percentage release from pure p-HEMA, the microemulsi on laden and surfactant laden gels. All the gels were 100 m in thickness. The error bars represent half the difference between the data from two repeat runs.

PAGE 74

74 y = -13.191Ln(x) + 64.913 R2 = 0.993520 30 40 50 60 70 80 0.11101001000Concentration ( g/ml)Surface Tension (mN/m) Figure 3-10. Dependence of surface tension on th e bulk surfactant concentration. for Brij 97 surfactant 0 5 10 15 20 25 30 35 40 05101520 % Surfactant Release S3_Thin S3_Thick Figure 3-11. Cumulative percentage release of surfactant from su rfactant laden gels. Thin gels contained 3245 g of surfactant and thic k gels contained 3344 g of surfactant. Data are plotted as mean SD (n = 3).

PAGE 75

75 Table 3-1. Diffusion coefficients of the drug for the microemulsion laden systems System Slope 95% CI for slope Dx1015 m2/s R2 M1 4.47 (4.30, 4.64) 4.36 0.99 M2 4.12 (3.98, 4.26) 3.70 0.98 M3 3.52 (3.41, 3.62) 2.70 0.99 D1 8.12 (7.46, 8.78) 14.4 1.00 Table 3-2. Drug uptake by microemulsion la den gel M2 after soaking in drug solution X Days Amount of CyA initially in solution ( g) Amount of CyA left in solution after X Days( g) Amount of CyA loaded in the gel ( g) 5 46 16.2 29.8 10 46 15.1 30.9 15 46 15.2 30.8

PAGE 76

76 Table 3-3. Drug released in the packaging medium from microemulsion-laden gels M2 X Days Drug inside the gel before packaging ( g) Drug released in packaging ( g) Drug left in the gel after packaging ( g) 10 44.2 7.1 37.1 30 49.8 9.7 40.1 Solution I (DI Water) 100 46.9 17.6 29.3 10 49.8 6.7 43.1 30 47.7 10.3 37.4 Solution II (0.85% w/w Salt Solution) 100 47.9 16.6 31.3 10 52.3 3.3 48.0 30 47.9 5.6 42.3 Solution III (4.25% w/w Salt Solution) 100 52.6 9.6 43.0 Table 3-4. Drug loading and release from su rfactant-laden and microemulsion-laden gels subjected to processing conditions. The e rror bars represent half the difference between the data from two repeat runs. System Drug in solution initially ( g) Drug remaining in the solution after 12 days of soaking ( g) Amount of drug inside the gel ( g) Amount of drug released during storage ( g) Amount of drug retained in the gel ( g) Amount of drug released during experiments ( g) S3 48 18.2 1.1 29.8 1.1 5.5 0.4 24.4 1.5 22.1 0.0 S2 48 25.0 2.9 23.0 2.9 4.5 0.1 18.6 2.9 20.0 1.1 S1 48 23.2 0.6 24.9 0.6 6.9 0.1 17.9 0.6 14.6 1.0 M2 48 22.5 0.6 25.6 0.6 6.5 0.1 19.1 0.5 19.5 0.1

PAGE 77

77 CHAPTER 4 MODEL FOR SURFACTANT AND DRUG TR ANSPORT FROM P-HEMA HYDROGEL 4.1 Introduction Goals of this chapter are to understand and model the transport of both surfactant and drug from the hydrogels. In this chapter we specifi cally focus on p-HEMA hydr ogels that are loaded with Brij 98 surfactant and drug CyA. Experiments are done to validate the model and data from publications of other authors are also shown to approve with th e developed model. While this work focuses on a specific set of drug, surfactant and polymer, the model developed here is expected to be valid for a wide variety of systems. 4.2 Materials and Methods 4.2.1 Materials HEMA monomer, ethylene glycol dimeth acrylate (EGDMA), Dulbeccos phosphate buffered saline (PBS), Acetonitr ile, HPLC grade water and Brij 98 were purchased from SigmaAldrich Chemicals (St Louis, MO). Darocur TPO was kindly provided by Ciba (Tarrytown, NY). CyA was purchased from LC Laboratories (Woburg, MA). A ll the chemicals were reagent grade. Acetonitrile was filtered before use and al l the other chemicals were used without further purification. 4.2.2 Synthesis of Surfactant Laden Gels Surfactant solutions of three different concen trations were prepared by adding 0.25, 0.6, 1.5 g of the surfactant to 10 ml water and then stirring th e mixture at 600 rpm and at room temperature for a period of about 10 hours. Separa tely, 3.8 mg of CyA was dissolved in 2.7 ml of HEMA monomer and stirred at 60 0 rpm for a period of 5 hours. 15 l of the crosslinker (EGDMA) and 2ml of surfactant so lution were added to the 2.7 ml of drug-HEMA mixture. The solution was then degassed by bubbling nitrogen for 10 minutes. Next, 6 mg of the initiator

PAGE 78

78 (TPO) was added and the solution was stirred at 300 rpm for 10 minutes to completely solubilize the initiator. The solution was then poured in be tween two glass plates that were separated either by a 200 m (thick gels) or 100 m (thin gels) thick spacer. The mold was then placed on Ultraviolet transilluminiator UVB-10 (Ultra Lum, Inc.) and the gel wa s cured by irradiating UVB light (305 nm) for 40 min. To synthesize HEMA gels without surfactants, 2 ml of the surfactant solution was replaced by 2 ml DI water. These gels are referred to as pure gels in the following sections. 4.2.3 Drug Release Experiments After polymerization, each gel was removed fr om the glass mold and was cut into smaller pieces that were about 1.5 cm X 1.5 cm X 200 m for the thick gels and 1.5 cm X 3 cm X 100 m for the thin gels, with each gel weighed nearly 40 mg. Two sets of experiments were performed for the drug release studie s. In the first set of experiments, gel was soaked in 3.5 ml of PBS and the drug concentrations in the releas e medium were measured periodically until the drug flux approached zero. In the second set of e xperiments, we attempted to create perfect sink conditions in the release medium of 3.5 ml PBS by replacing th e medium every 24 hours. CyA concentration was measured using an HPLC (W aters, Alliance System) equipped with a C18 reverse phase column and a UV detector. The mobile phase composition was 70% acetonitrile and 30% DI water, and the column was maintained at 60C. The flow rate was fixed at 1.2 ml/min and the detection wavelength was set at 210 nm [114]. The retention time for CyA under these conditions was 5.3 minutes, and the calibration curve for area under the peak vs. concentration was linear (R2 = 0.98).

PAGE 79

79 4.2.4 Surfactant Release The rates of surfactant release from the gels we re measured by soaking them in 3.5 ml DI water and replacing the release medium at regular intervals. The surfactant concentration in the release medium was determined by measuring surface tensions ( ) which was then related to the concentration through a (C) calibration curve. The surface tension was measured by using a Wilhelmy plate (sand blasted platinum plate) attached to a Scaime France Microbalance which was further connected to a Stathan Universal tran sducer (SC001). The transducer was calibrated by using DI water ( = 72 mN/m) and acetone ( = 23 mN/m) as standards. The Wilhelmy plate was rinsed with DI water and acetone followed by annealing till red hot using a propane burner. The annealing process was done to remove impuriti es which rinsing was not able take away from the platinum surface. This process was repeated before every measurement and the plate was left to cool for one minute before taking the measurem ent. The solution was allowed to equilibrate for an hour prior to the measurement to ensure th at the surface coverage was in equilibrium with the bulk concentration. It was also ensured th at the surface area to volume ratio was sufficiently small to cause a negligible cha nge in bulk concentration due to adsorption of the surfactant on the surface. 4.3 Results and Discussion 4.3.1 Drug Release from Pure p-HEMA Gels w ith Daily PBS Replacement The drug loaded into pure p-HEMA gels is pr esent either as free drug in the water phase (C) inside the gel or as drug bound to the polymer ( ). The mean drug concentration in the gel (Cg), which is essentially the sum of the bound and the free drug concentration, is given by (S/V)gel + fC, where (S/V)gel is the surface area per volume of the gel available for the drug to adsorb and f is the volume fraction of water in hydrated gel (Figure 4-1). The value of f for

PAGE 80

80 p-HEMA gels was determined to be 0.39 from the swelling experiments. The free and the bound drug are expected to be in e quilibrium and so the total gel concentration and the free drug concentration can be related th rough a partition coefficient Kd = Cg/C, which is assumed to be concentration independent. The tr ansport of the drug in the hydrogel is expected to occur by a combination of bulk and surface diffusion, and thus it can be described by the modified diffusion equation, i.e., 2 2 2 2y A P D y C fD t CSU f g (4-1) where Df and DSU are the diffusivities of the drug in solution and on the surface, respectively, and (P/A) is the perimeter of the gel fibers per unit cross-sectional area, which can be approximated as (S/V). For a number of drugs including CyA, the partition coefficient Kd is much larger than 1, and so most of the drug can be safely assumed to be bound to the polymer matrix. Utilizing the definition of parti tion coefficient in the above equation gives 2 2 2 2)( y C D y C K fKDfD t Cd dSUf (4.2) In Equation 4-2, d dSUfK fKDfD D )( is the effective diffusivity of drug in the gel. Equation 4-2 is subjected to the following boundary conditions, 0)0,( yt y C (4-3) 0),( hytC (4-4) where h is the half-thickness of the gel. The first boundary condition (Equation 4-3) assumes symmetry at the center of the gel and the s econd boundary condition (Equation 4-4) assumes

PAGE 81

81 infinite sink conditions in th e release medium. The initial condition for the drug release experiments is: iCtyC )0,( (4-5) Where, Ci is the initial concentratio n of drug inside the hydrogel. Equations 4-2 to 4-5 can be solved to give 0 4 )12(2 22) 2 )12( cos( )12( 4)1(n Dt h n i ney h n n C C (4-6) The fraction released can be computed from Equation 4-6 to give 100) 1( 12 18 (%)2 224 )12( 0 2 2 Dt h n n De n R (4-7) In the short time limit, Equation 4-7 simplifies to the following form [87], 100 4 (%)2h Dt RD (4-8) Figure 4-2 plots the percentage release of CyA from pure p-HEMA gels vs. t and the slope of the plot is used to obtai n D, which has a value of 1.44x10-14 m2/s. This value is much lower than the free solution diffusivity of CyA due to the small pore size of the p-HEMA gels. 4.3.2 Surfactant Release from the Hydrogels 4.3.2.1 Model The hydrated surfactant laden hydrogels are exp ected to contain the surfactant in three different forms: free surfactant, surfactant ad sorbed on the polymer and surfactant present in aggregates. On soaking of this gel in aque ous solution, free surfactant and the adsorbed surfactant diffuse and elude from the gel into the release medium. This transport reduces the free concentration of the surfactant below the critical aggregation concentration (CAC) leading to breakup of the aggregates. This mechanism resu lts in creation of a depletion zone near the

PAGE 82

82 surface which does not contain aggregate particle s because these are already dissolved, and the thickness of this zone ( ) increases with time (Figure 4-3). The free concentration of surfactant in the zone that still contains the aggregates must be equal to the CAC (C*), and the surfactant contained in the aggregates should be at concentration Cp = CiS C*, where CiS is the initial surfactant concentration. Thus, we get the following model to describe surfactant transport from gel systems, )(0for 2 2ty y C D t CS S S (4-9) )(for *hyt CCS (4-10) With the following initial and boundary conditions, 0)0,( ytCS (4-11) SC)y(t, C (4-12) yfor t C Dp S y CS (4-13) 0),0( yt (4-14) Where, h is the half thickness of the gel, and CS is the concentration of the un-aggregated surfactant, which includes both free and polymer bound surfactants. Also DS is the average surfactant diffusivity that incl udes contributions from both free and surface diffusion of the surfactant. The first boundary condition (Equati on 4-11) assumes perfect sink conditions, the second boundary condition (Equation 412) arises from continuity of concentration and the third condition (Equation 4-13) states th at the surfactant flux at the in tersection of the zone with aggregates and the one without ag gregates is equal to the amount released by the dissolution of the aggregates. The above model is only valid till is less than h, and after that the transport is

PAGE 83

83 purely diffusive. It is noted th at the above model is very simplistic because it neglects drugsurfactant interaction. The above partial differential equation ad mits a similarity solution of the form CS(y,t) = CS( ), where = y/ 4DSt and = 4DSt. With these substituti ons, the equations and the boundary conditions reduce to the following form: 0 22 2 d Cd d dCS S (4-15) 0)0( SC (4-16) *)( C CS (4-17) 2p SC d dC (4-18) Equation 4-15 is a ordinary differential eq uation and it can be solved to yield the concentration profile de de CCS 0 0 *2 2 (4-19) Where can be obtained by solving the following implicit equation: de e C Cp 0 *2 22 (4-20) For case of 1* C Cp, Equation 4-20 can be simplified to yield pC C 2* (4-21)

PAGE 84

84 The flux from the gel to the outside fluid medium can be calculated as de C t D y C DjS y S S 0 024 (4-22) The total amount of surf actant released from a gel can then be calculated by integrating the flux in time and then multiplying by the surface area 2A to give de C tDANS S0 *22 (4-23) Again in the limit 1* C Cp the above expression can be simplified to yield pS SCtCDAN*22 (4-24) Physically, the limit *CCP represents the pseudo-steady solution to the above equations as in this limit the time scale of changes in due to micellar breakup far exceed the diffusive time scales in region I and so the conc entration profile of the surfactant in region I is linear. Furthermore, in this limit, the values of Cp are close to the total surfactant concentration initially loaded into the gel (CiS). The amount of drug initiall y loaded into the gel can be approximated as 2AhCp, and thus the percentage surfactant release RS(%) from the system is given by 100 2(%)2 hC t CD Rp S S (4-25) Thus plots of percentage release vs. 2hC tp should be a straight line with slope 100 2* CDS. It is noted that this result sharply contrasts with the result for pure Fickian

PAGE 85

85 diffusion without any aggregate formation for wh ich case the release rates scale linearly with loading and so the percentage release is independent of the initial surfactant loading. A similar equation has been proposed in the past for drug diffusion from suspension inside an ointment [89]. So it seems that both the systems are si milar, though this method is more rigorous and helps in understanding of the structural details due to incorporation of surfactants inside the hydrogels. Our model can be solv ed numerically to de termine the complete release profiles of surfactants whereas Higuchi model was solved assuming pseudo steady state and is valid for a small time of release. 4.3.2.2 Experimental results The concentration of the surfactant in the release medium was determined by measuring the surface tension of the fluid and relating it to the surfactant concentration through the measured (C) relationship shown in Figure 4-4. It s hould be noted that dur ing these studies, 3.5 ml of water was used as the release medium. To maximize the sensitivity of the measurements, the 3.5 ml solution was diluted by trial and error to surfactant concentrations below its critical micelle concentration (CMC) at which the surface te nsion is most sensitive to concentration. We also performed control experiments to check fo r surface activity of the un-reacted monomer from the gels and found no change in surface tension of water. To validate the model developed above, it wa s decided to explore the effects of gel thickness and initial surfactant loading on release rates of the surfactant. Surfactant-laden gels with 8% surfactant by weight in dry state were prepared having two different thicknesses (100 and 200 m). Also 200 m thick gels were prepared with 4% surfactant loading by weight in dry state. The weights of both the thick and the thin gels were about the same because the cross sectional area of the thin gel was double that of th e thick gel. Surfactant release data from all the

PAGE 86

86 three sets of hydrogels is plot ted in Figure 4-5 and we observe that the percentage release decreases with an increase in surf actant loading, which agrees with the model proposed above. Also, the release rates decrease with an increasing thickness, which is expected. Equation 4-25 predicts that a plot of cumulative percentage vs. 2hC tpshould be a straight line with slope 100 2* CDS. To quantitatively validate the model, in Figure 4-6, the percentage release is plotted against All the release curves match and are line ar at short times validating the model. The slope indicated in the figure ha s a value of 1.1 0.03. The m odel predicts that the plot in Figure 4-6 must stay linear ti ll the boundary layer thickness becomes equal to h. At this time the gel does not contain any micelles and so the concentration in the entire gel is less than or equal to C*. For cases in which C* is much less than Cp, the amount of surfact ant in the gel after all the micelles are broken is a very small fraction of the initial surfactant loading. This is clearly evident in Figure 6 as the plot deviates from lin ear behavior after the percentage release exceeds 95%. The slope of the line in Figur e 4-6 depends only on the parameter (DSC*) and so we cannot determine both DS and C*. The non-linear part of th e curve depends on these parameters individually, but since the non-linear part of the curve contributes to a very small fraction of the release, it is not prudent to use that limited data to determine DS and C* individually. The value of surfactant diffusivity can be obtained by techniques such as nuclear magnetic resonance or by measuring surfactant release rates with low surfacta nt loading, and then the value of C* can be obtained from the slope of the data in Figure 4-6. Later in chapter 7 we will also discuss two model techniques to determine the CAC of surfacta nts. However it is shown later that the drug transport from the gels does not depend on DS and C* individually, but only on the product.

PAGE 87

87 4.3.3 Drug Release from Surfactant Laden Gels 4.3.3.1 Model As described in the previous section, diffusion and breakup of micelles leads to creation of two zones: the first zone contains free surfacta nt below the critical a ggregation concentration, C* and the second zone contains micelles at concentration Cp and free surfactant at concentration C*. The drug in region I can exist either as fr ee or bound to the polymer and both of these forms can diffuse with respective diffusivities. Accordingly, the mass bala nce in region I can be described by the following equation: (t)y0for 2 2 y C D t CI I (4-26) Where CI is the total drug concentration in regi on I, which includes both free and polymer bound drug, and D is the average diffusivity that includes contributions from both free and surface diffusion. Furthermore, it is assumed that surfactant and drug transport are uncoupled in this region, which may be a reasonable assumption due to small drug concentrations explored here. It is noted that the drug diffusiv ity in the above equation is assu med to be the same as obtained from the experiments of drug release from pure pHEMA gels in section 4.3.1. In region II, a fraction of the drug partitions insi de the micelles and this fraction is not available for diffusion. Assuming that the time scale of drug transport from inside the micelle to outside is rapid, the concentration of drug inside the micelle is denoted by KmCII where Km is the partition coefficient of the drug. The average concentration of micelle bound drug can then be given by hSpIImfMWCCK where Cp is the concentration of surf actant present as micelles, MWS is the molecular weight of the surfactant, is the density of micellar core inside the gel volume and fh is the fraction of the surfactant length which is hydrophobic in nature. For Brij 98

PAGE 88

88 surfactant, fh can be approximated as 0.3. This expression can then be written as KCpCII where hSmfMWK K Thus, the total concentration of the drug in region II isIIpCKC)1( However, only the drug present outside the micel les can diffuse towards region I. Thus, mass balance in region II can be described by hy(t)for )1(2 2 y C D t CKCII IIP (4-27) The average diffusivity in the above equation is assumed to be the same as the average drug diffusivity in region I. The boundary a nd the initial conditions for the mass balance equation are, 0)0,( ytCI (4-28) 0),( hyt y CII (4-29) ),(),( ytCytCII I (4-30) yat dt d CKC y C D y C DIIp II I (4-31) ),0(f IICytC (4-32) Where, Cf is the initial free drug concentra tion in the gel, and it equals )1/(p iKCC where Ci is the initial total drug concen tration. The first boundary condition (Equation 4-28) assumes perfect sink conditions in the release medium and the second boundary condition (Equation 429) assumes symmetry at the gel center. Th e third (Equation 4-30) and the fourth boundary conditions (Equation 4-31) arise fr om assumptions of continuity of free concentr ation and total drug flux, with the second term on the RHS of the last boundary c ondition (Equation 4-31) accounting for the drug flux due to breakup of micelles.

PAGE 89

89 To facilitate analytical solutions at short ti mes, i.e., times in which the thickness of the drug concentration boundary layer is less than the half gel thickness, we replace Equation 4-29 by the following equivalent equation f IIChytC ), ( (4-33) The above set of equations admits a simila rity solution at short times of the form )(' 'I f I IC C C C and )(' 'II f II IIC C C C where tDyS4/ and as shown in the previous section, tDS4 This transformation leads to th e following equations and the boundary conditions, 0 22 '2' d Cd d dC D DI I s (4-34) 0 )1( 22 '2' d Cd d dC D DKCII II sp (4-35) 0)0('IC (4-36) 1)('IIC (4-37) )()(' 'II IC C (4-38) for 2' D DCKC d dC d dCsIIp II I (4-39) Solving the above set of equations utilizing the first three boundary conditions (Equations 4-36,4-37,4-38) yields the following solutions, 0 '2 deBCD D Is (4-40)

PAGE 90

90 1 12 1 2 21 0 de de deB Cp Kc D s D p s sKc D D D D II (4-41) As shown above, for case of 1* C Cp, pC C 2*. In this limit the expression for the concentration simplify to yield 0 '2deBCD D Is (4-42) p S p S IIKcD D de KcD D B Cp Kc D s D14 14 1 10 '2 1 (4-43) Applying the flux equality boundary condition (Equation 4-39) at = gives the following expression for B P s p S p SC C D D KC KCD D KC KCD D B 2 14 14 1* (4-44) Assuming that a major fraction of the dr ug in Region II is inside the micelles KCp>>1, and thus the expression for B simplifies to P s P S PSC C D D KC DC KD C KCD D B 2 4 4 1* (4-45) The flux from the gel to the flui d can than be calculated as tD BDC y C Djs f y I40 (4-46)

PAGE 91

91 The total amount of drug releas ed from a gel can be calcula ted by integrating the flux in time and then multiplying by the to tal surface area A to give P s P S PS s fC C D D KC DC KD C KCD D D AtDC N 2 4 4 1 2* (4-47) The amount of drug initially loaded into the gel is iAhC 2 and thus the percentage drug release RD(%) from the system is given by 100 (%)2 hC t Rp D (4-48) Where, 2 2/3* 2/3 *2 )( 4 )( 4 1 D K DC D K DC D KS S (4-49) 4.3.3.2 Experimental results 4.3.3.2.1 Experiments with Daily PBS Replacement To validate the model developed above, it was decided to explore the effects of initial surfactant loading on release rates of the drug. In these experiments, the release medium was replaced daily to simulate perfect sink condi tions. Experiments were performed with three different surfactant loadings and thickness of gel was also varied for all the systems to validate the model. Release profiles from the thick and th e thin gels are shown in Figures 4-7 and 4-8. As expected, surfactant loading inside the gel pha se significantly affects the release rates of the drug. Based on the above model, a plot of percentage release vs. 2hC tp should be a straight line with slope100 The plots are shown in Figures 4-9 and 4-10 after rescaling the time

PAGE 92

92 scale for both, the thin and the thick gels, respec tively. For each case, the release curves match for the 2, 4 and 8% surfactant loadings. The sl opes indicated in each fi gure represent the values of100 The values of slope are 1.47.018 and 1.1.12 for the thick and the thin gels, respectively. The small difference in these values for the gels of two different thicknesses implies that the data does not satisfy the thickne ss scaling as predicted by the model. This difference likely arises from the fact that th e PBS in the release medi um was replaced every 24 hours for both the thick and the thin gels. Since the amount of drug released from the thin gels in a 24 hour period is double the amount released by the thick gels, the perfect-sink condition is perhaps not satisfied for the case of release from the thin gels. For the experiments from the thick and the thin gels to be equivalent, the PB S has to be replaced every 6 hours for the case of thin gels, which is a more difficult schedule to maintain. So it was decided to conduct release experiments without PBS replacement to validat e the thickness scaling predicted by the model (data shown in Section 4.3.3.2.2). The release from thick gels was in perfect sink conditions and so the model developed above is valid. Accordingly, the slope value of 1.47.018 2s mMobtained from the thick gel data can be equated to 100x Since both D and *CDS are known from the fits in the pr evious sections, we can now obtain the value of K. By using values of 1.44 x 10-14 m2/s and 0.0078 2s mM for D and *CDS, respectively, and a value of 1.47 2s mM for 100 we determine K to be 142.9 M-1. The molecular weight of the surfactant is 1149.5 and if we approximate to be 1000 kg/m3, we can determine the approximate partition coefficient Km of the drug from the expression hS mfMW K K to be 414.4.

PAGE 93

93 The partition coefficient Km is the ratio of the drug concentration inside the micelle and the mean concentration outside, which includes both the fr ee concentration in the aqueous phase and the bound concentration. A high value of partition coefficient is re sponsible for the significant increase in release duration. 4.3.3.2.2 Experiments without Replacing the Re lease Medium (Equilibrium Release) The model developed above can be extended to apply to experiments in which the release medium is not replaced, and eventually equilib rium is achieved. However that extension requires the partition coefficients between gel an d PBS for both the surfactant and the drug. While the percentage release prediction of Equation 4-48 is not valid for equilibrium experiments, the thickness scaling predicted by the model is still valid if the volume of the fluid to the gel is kept the same for both the thin and the thick gels. The thickness scalings are verified by the data in Figure 4-11 in which the drug release rates are plotted against where 2h t The results overlap for the thick and thin gels validating the thickness sc alings predicted by the model. It is also observed that the time requi red to reach equilibrium is much greater for surfactant laden gels then for pure p-HEMA gels. 4.3.4 Model Comparison with Published Data To further validate the model it was decided to utilize the data from the work of Liu et al [94, 95] and compare it to the model developed for drug release from surfactant laden hydrogels. In [94], Liu et al explored the effect of changing surfactant (S DS) concentration on release of a hydrophobic drug CPT from agarose hydrogels. Speci fically, they measured drug release from gels with 0.2, 0.4, 0.6, 0.8 and 1 wt% surfactan t. The drug release data was obtained from their published work by a digitalization program and was re-plotted in Figur e 4-12 as fraction drug

PAGE 94

94 release vs. SDSC t where CSDS represents the concentration of surfactant (%wt) inside the hydrogel. The release profiles match for 0.4, 0.6, 0.8 and 1 wt% surfacta nt, whereas the profile for 0.2% loading deviates. It is noted that the analytical solu tion developed for the drug release is based on the assumption that the concentration of surfactant present as micelles (Cp) is substantially larger than the CAC (C*). The value of CMC of SDS is about 0.24 wt% in the presence of the drug, and since the gels prepared by the authors had only about 3% polymer by weight, it can be assumed that the critical aggreg ation concentration inside the gel (C*) is close to the CMC. A surfactant loading of 0.2% would th en be very close in value to C*, and thus the profiles for the 0.2% loading are in fact not e xpected to overlap those for the higher surfactant loadings. Liu et al also explor ed the effect of surf actant (DTAB) concentrations on release of CPT from agarose hydrogels [95]. They measur ed drug release rates from gels with 10mM, 30mM, 40mM and 50mM DTAB. The data obtai ned by Liu et al. for the four surfactant loadings is re-plotted in Figure 4-13 as fraction drug release vs. DTABC t where CDTAB represents the concentration of surfactant inside the hydrogel. Again, th e results in Figure 4-13 show that the profiles for 30mM, 40mM and 50mM DTAB ar e very similar while the profile for 10mM DTAB is significantly different. The value of CMC of DTAB is about 7mM in the presence of the drug and therefore the profiles for the 10mM lo ading are expected to not match those for the higher loadings. It should be not ed that even though the analytical solution developed here is not valid in conditions where the total surfactant con centration is close to th e critical aggregation concentration, the general model developed here is still valid and it can be solved numerically to fit the data.

PAGE 95

95 4.4 Conclusion We have explored the mechanisms of tran sport of drugs and surfactants in hydrogels loaded with CyA and Brij 98. Transport models were developed for both the surfactant and the drug, and the results of the model were verified by measuring release rate s of both surfactant and drug from hydrogels. The experi mental results are in good agreement with the model. The model of drug release from surfactant laden hydroge ls seems to be in good agreement with the published work of Liu et al [94, 95]. The transport of both the drug and the surf actant is controlled by diffusion through the gel. At concentrations above the critical aggregation concentration, excess surfactant forms micellar aggregates, into which hydrophobic drugs can partition preferentially. The diffusion of surfactant leads to breakup of micelles causing form ation of a depletion z one near the surface. The plots of percentage release vs. 2hC tp should be a straight line with slope 100 2* CDS. This result sharply contrasts with the result for pure Fickian diffusion without any aggregate formation for which case the release rates scale linearly with loading and so the percentage release is independent of the initial surfactant loading. The drug transport is strongly coupled to the su rfactant transport. As the micelles break, the drug is released into the gel, and becomes av ailable for diffusion. The model predicts that the percentage drug release is linear with 2hC tp, and thus a four fold increase in surfactant loading leads to a two fold reduction in percenta ge release for drug at a given time. The model can be fitted to the experimental data to dete rmine important physical parameters such as the partition coefficient between the hydrophobic core of th e micelle and the hydrogel.

PAGE 96

96 The model developed here neglect a number of issues including interactions between the drug and the micelle, concentration dependen ce of drug and surfactan t adsorption on the polymer, etc. Also, the surfactant laden gel ma trix is assumed to be homogenous on the length scales relevant to transport. A good agreemen t between the model and the experiments suggest that the assumptions are perhaps valid. The ge ometry for the gel is considered to be twodimensional, which is a good assumption. The models can be applied in ge neral to any arbitrary geometry by including diffusive flux in othe r directions in the mass balance equations. The transport models developed here can be ve ry helpful in tuning the drug release rates from hydrogels by controlling the surfactant concentr ation. The results also show that Brij 98 loaded p-HEMA exhibit an exte nded release of CyA and so contact lenses made with this material can be used for extended ocular delivery of CyA.

PAGE 97

97 Figure 4-1. Transport of drug in the p-HEMA hydrogel. A large fraction of the drug (denoted by circles) is bound to the polymer ( ) and a small fraction is present in the water phase of the hydrogel (C). The free and the bound drug are in equilibrium. h C

PAGE 98

98 Slope = 8.120 10 20 30 40 50 60 70 80 0246810Time0.5 (h0.5)% Drug Release Pure p-HEMA_Thick Figure 4-2. Percentage release of drug from pure p-HEMA gels. All the gels contained about 50 g of drug. Data are plotted as mean SD (n = 3).

PAGE 99

99 Region I Region II y = 0 y = y = h Figure 4-3. Transport from surfactant laden hydrogel. Region I represent the depletion zone with no micelles because the surfact ant concentration is below the critical aggregation concentration. Region II contains surfactant aggregates along with free surfactant.

PAGE 100

100 y = -8.6657Ln(x) + 64.578 R2 = 0.999930 35 40 45 50 55 60 65 70 75 0.1 1 10 1001000Concentration ( g/ml)Surface Tension (mN/m) Figure 4-4. Dependence of surface tension on th e bulk surfactant concentration for Brij 98 surfactant 0 10 20 30 40 50 60 70 80 90 100 02004006008001000Time (h % Surfactant Release Thick_8% Thin_8% Thick_4% Figure 4-5. Cumulative percentage release of surfactant from hydrogels. Data are plotted as mean SD (n = 2).

PAGE 101

101 Slope = 1.1 0 10 20 30 40 50 60 70 80 90 100 020406080100% Surfactant Release Thick_8% Thin_8% Thick_4% Figure 4-6. Cumulative percentage release of surfactant from th e hydrogels after rescaling the time. represents 2/ hCtP, where t is time in seconds, Cp is surfactant concentration in M, and h is half thickness of the gel in m. Data are plotted as mean SD (n = 2).

PAGE 102

102 0 10 20 30 40 50 60 70 80 90 100 020040060080010001200Time (h)% Drug Release Thick_2% Thick_4% Thick_8% Figure 4-7. Effect of surfactant loading on cu mulative percentage release of the drug for surfactant laden thick (200 m) gels during PBS change experiments. All the gels contained nearly 50 g of drug. Data is plotted as mean SD (n = 2). 0 10 20 30 40 50 60 70 80 90 100 0200400600800Time (h)% Drug Release Thin_2% Thin_4% Thin_8% Figure 4-8. Effect of surfactant loading on cu mulative percentage release of the drug for surfactant laden thin (100 m) gels during PBS change experiments. All the gels contained nearly 50 g of drug. Data is plotted as mean SD (n = 2).

PAGE 103

103 Slope = 1.47 0 10 20 30 40 50 60 70 80 90 100 020406080% Drug Release Thick_2% Thick_4% Thick_8% Figure 4-9. Cumulative percentage releas e of drug for surfactant laden thick (200 m) gels during PBS change experiments afte r rescaling the time of release. represents 2/ hCtP, where t is time in seconds, Cp is surfactant concentration in M, and h is half thickness of the gel in m. Data is plotted as mean SD (n = 2).

PAGE 104

104 Slope = 1.18 0 10 20 30 40 50 60 70 80 90 100 0306090120% Drug Release Thin_2% Thin_4% Thin_8% Figure 4-10. Cumulative percen tage release of drug for surfactant laden thin (100 m) gels during PBS change experiments afte r rescaling the time of release. represents 2/ hCtP, where t is time in seconds, Cp is surfactant concentration in M, and h is half thickness of the gel in m. Data is plotted as mean SD (n = 2).

PAGE 105

105 0 10 20 30 40 50 60 70 80 90 010203040% Drug Release Thin_8% Thick_8% Figure 4-11. Drug release from surfactant la den gels for equilibrium (no PBS change) experiments. represents 2/ ht, where t is time in seconds and h is half thickness of the gel in m. All the gels contained about 50 g of drug. Data are plotted as mean SD (n = 2).

PAGE 106

106 0 10 20 30 40 50 60 70 80 90 100 0510152025Fraction Drug Release 0.2% SDS 0.4% SDS 0.6% SDS 0.8% SDS 1% SDS Figure 4-12. Drug release from agarose hydrogels containing SDS surfact ant obtained from Liu et al. [94]. represents SDSCt /, where t is time in minutes and CSDS is percentage of surfactant. 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 00.511.522.5Fraction Drug Release 50mM DTAB 40mM DTAB 30mM DTAB 10mM DTAB Figure 4-13. Drug release from agarose hydrogels containing DTAB surf actant obtained from Liu et al. [95]. represents DTABCt /, where t is time in minutes and CDTAB is surfactant concentration in mM.

PAGE 107

107 CHAPTER 5 SURFACTANT LADEN HYDROGELS FOR OPHT HALMIC DRUG DELIVERY WITH INCREASED WETTABILITY AND WATER CONTENT 5.1 Introduction Hydrogel contact lenses typically have ab out 60% polymer in the swollen state which competes with the surfactant-aggregat es for drug binding. It is thus essential that in order for the surfactants to retard drug transport in contact le nses, the surfactant aggregates must have a very high affinity for the drug compared to the hydrogel. Accordingly, to develop a contact lens suitable for extended delivery of a given drug, it is important to investigate the microstructure of the gel with particular focus on the micellar-aggr egates, and also investigate the mechanisms that impact the partitioning of the drugs in the aggregates It is also equally important to investigate the effect of surfactant loading on gel physical pr operties relevant to contact lenses such as transparency, modulus, protein binding, wettability, and water content. This chapter is an exhaustive study that focuses on each of the issues listed above. The results of this study will be helpful in deliverin g CyA to eyes through contact lenses, and also in designing suitable contact lenses for delivering other ophthalmic drugs. This is the first study that reports the microstructure and physical characterization of the surfactant-laden hydrogels with a large polymer fraction as large as 60%. 5.2 Materials and Methods 5.2.1 Materials Hydroxy ethyl methacrylate (HEMA) mono mer, ethylene glycol dimethacrylate (EGDMA), Dulbeccos phosphate buffered saline (PBS), dexamethasone (DMS), dexamethasone acetate (DMSA), Acetonitrile, ly sozyme from chicken egg white, HPLC grade water, Brij 97, Brij 98, Brij 78 and Brij 700 were purchased from SigmaAldrich Chemicals (St Louis, MO). 2,4,6-Trimethylbenzoyl-diphenyl-phophineoxide (Daroc ur TPO) was kindly

PAGE 108

108 provided by Ciba (Tarrytown, NY). Cyclos porine A (CyA) was purchased from LC Laboratories (Woburg, MA). All the chemicals were reagent grade. Acetonitrile was filtered before use and all the other chemicals we re used without further purification. 5.2.2 Preparation of Surfactant Laden Gels Surfactant laden gels were prepared by polymerizing the monomer solution containing surfactant and drug mixed in specific ratio. Brie fly, 0.25, 0.6, 1.5 g of surfactant was dissolved in 10 ml DI water to make three different surfactant solutions (corresponding to 2%, 4%, 8%, surfactant in dry gel respectively). Separately, 3. 5 mg of drug was dissolved in 2.7 ml of HEMA monomer and stirred at 600 rpm for a period of 5 hours. Next 15 l of the crosslinker and 2ml of surfactant solution were added to the 2.7 ml of drug loaded monomer. The solution was degassed by bubbling nitrogen gas through it for 10 minutes followed by a ddition of 6 mg of UV initiator (TPO) and stirring the solution for 10 minutes. The so lution was then poured between two glass plates separated by a spacer and the ge l was cured by irradiating UVB light (305 nm) for 40 min from an Ultraviolet transilluminiato r UVB-10 (Ultra Lum, Inc.). Four different spacers, 100, 200, 400 and 800 m in thickness were utilized to synthesize gels of various thicknesses. Control, drug loaded p-HEMA gels without surfactants were prepared by following procedures identical to those described abov e except that the 2 ml surfactant solution was replaced by 2 ml DI water. 5.2.3 Drug Release Experiments After polymerization, each gel was removed fr om the glass mold and was cut into smaller pieces that weighed about 40 mg in the dry state. These 40 mg gels were used in all experiments described in this chapter. As the thickness of the gel was varied, the size of the gel piece was adjusted to maintain similar weig ht for all the gels used in the study. Two sets of experiments were performed for the drug release studies. In the first set of experiments, gel was soaked in 3.5

PAGE 109

109 ml of PBS and measurements were taken until equilibrium was reached for the drug. In the second set, gel was soaked in 3.5 ml of PB S and PBS was replaced every 24 hours, mimicking perfect sink conditions for the release experiments. Equilibrium experi ments were conducted for all the three drugs explored in this study (C yA, DMS, and DMSA), whereas, PBS replacement experiments were performed for CyA only. 5.2.4 Drug Detection CyA concentration was measured using a HPLC (Waters, Alliance System) equipped with a C18 reverse phase column and UV detector. The mobile phase composition was 70% acetonitrile and 30% DI water, and the column was maintained at 60C. The flow rate was fixed at 1.2 ml/min and the detection wavelength was se t at 210 nm[114]. The retention time for CyA under these conditions was 4.5 minutes, and the calibration curve for the area under the peak vs. concentration was linear (R2 = 0.995). DMS and DMSA were detected using a UV-Vis spectrometer (Thermospectronic Genesys 10 UV) by measuring the absorbance spectra over a range of 190-290 nm. The absorbance data for the release experiments of DMS and DMSA were converted to the respective concen tration value by a de-convolution technique as reported earlier [125]. 5.2.5 Surfactant Release Experiments The rates of surfactant release were measured in 3.5 ml of DI water with water replacement after each measuremen t to maintain perfect sink conditions. The surfactant concentration in the release medium was determ ined by measuring surface tension, which was then related to the concentration through a calib ration curve. The surface tension was measured by using a Wilhelmy plate (sand blasted platin um plate) attached to a Scaime France Microbalance which was further connected to a Stathan Universal transducer (SC001). A

PAGE 110

110 detailed description of the process for measur ing surfactant concentrat ion by surface tension measurements has also been reported earlier in chapter 3 (Section 3.2.11). 5.2.6 Lysozyme Sorption A lysozyme solution was prepared by adding 40 mg of lysozyme to 40 ml of PBS. The 8% surfactant laden gels (about 40 mg in weight ) were soaked in 3.5 ml of lysozyme solution and the amount of lysozyme that was taken up by the hydrogels was monitored by UV detection in the wavelength range 240-340 nm. The conc entration of lysozyme was evaluated following a similar protocol as reported above for DMS and DMSA. 5.2.7 Preparation and Cryo-SEM of Hydrogels All samples were soaked in 1x PBS buffer for at least 24 hours. The hydrogel samples were trimmed down to approximately 1 cm x 1 cm in size and mounted vertically on the cryoSEM sample holder with a small amount of Ti ssue-Tek adhesive (Sakura). The samples were rapidly plunged into liquid nitr ogen at a temperature below oC (Gatan, Alto 2500), withdrawn into a vacuum transfer device under the protection of high vacuum, and transferred into the cryo-preparation chamber where the temperature was maintained at -130 oC and the anticontaminator at around oC. The hydrogel samples were freeze fractured using the flat edge of a cold knife maintained at -130 oC and sublimated for 5 minutes at -95 oC to etch away surface water and expose the internal structural f eatures. After sublimation, the temperature of the stage was adjusted back to oC and the samples were sputter coated with platinum at 11 mA for 100 seconds. The samples were then tr ansferred into the main chamber of the Field Emission SEM (Hitachi S-4800) via an interloc ked airlock and mounted onto a cold stage module ( oC) fitted to the SEM stage. Images were acquired at a voltage of 2 kV.

PAGE 111

111 5.2.8 Dynamic Mechanical Analysis A dynamic mechanical analyzer (DMA Q800, TA instruments) was used to determine the mechanical properties of different surfactant la den systems synthesized above. For this study 400 m and 800 m thick gels were utilized to avoid br eaking of the gel duri ng the experiment. A hydrated gel was mounted on the clamp and the ge l was kept submerged in DI water at room temperature during the experiment. Gel response in the form of storage and loss modulus of the gel was determined by applying tensile force in the longitudinal direction while keeping the gel tightly screwed between the clamps by applying a preload force of 0.01 N. To determine the linear viscoelastic range, strain test were firs t conducted at a frequency of 1 Hz followed by frequency sweep (1-35 Hz) measurements performed for all the samples at 20 m strain. 5.2.9 Surface Contact Angle Measurements Surface contact angles were measured for all the surfactant laden systems with 8% loading to investigate the effect of surf actant release on wettability. Th e contact angles were measured by captive bubble technique with a Drop Shape Analyzer (DSA100, KRSS). This technique was preferred over the sessile drop technique to eliminate contact angle change due to sample drying during measurements. A 200 m thick gel was mounted on a glass slide which was then placed on a water filled cuvette with the lens su bmerged in water. An air bubble was created by an inverted syringe inside the cuvette, and allowed to detach and ri se till it came in contact with the gel, and then the contact angle ( ) was measured. The gels were presoaked in a PBS buffer solution for one day before the experiment.

PAGE 112

112 5.2.10 Transmittance Measurements The transparency of all the surfactant la den hydrogels was quantified by measuring the transmittance of 100 m thick hydrated gels at 600 nm using a UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). 5.2.11 Equilibrium Water Content The gels of known weight were soaked in 3.5 ml of DI water, and the dynamic weight was measured as a function of time. The excess water from the gel surface was removed before each measurement by dabbing with Kimwipes (F ischer Scientific). The equilibrium water content (EWC) of the surfactant laden gels was calculated by determining the amount of water uptake per dry gel weight, i.e., 100 % DRY DRY WETW WW EWC (5-1) 5.2.12 Statistical Analysis Linear regression analysis to determine sl opes, correlation coefficients and confidence intervals was performed in JMP developed by SA S (Cary, North Carolina). The 95% confidence intervals (CI) were utilized to compare release rates. 5.3 Results and Discussion 5.3.1 Surfactant Release from the Hydrogels Table 1 lists the surfactants utilized in this study and their releva nt physical properties obtained from the literature. We also list the value of fh which is defined as the fraction of hydrophobic chain length of the surfactant and is calculated by taking th e ratio of number of carbons in the hydrophobic tail of the surfactant to the total numb er of carbons present in the surfactant.

PAGE 113

113 The rate of surfactant release from surfactant-laden contact le nses needs to be measured because an excessive release could lead to toxici ty. Additionally, the rate of surfactant release impacts drug release. As described in the previous section, the surfactant concentration in the release medium was determined by measuring th e surface tension of the solution, which is related to the bulk concentration. Figure 5-1A -B shows the calibration curves for Brij 78 and Brij 700 surfactants. Calibration curves for Brij 98 and Brij 97 have been shown earlier (Figure 3-10 and Figure 4-4). A model for surfactant release from hydrogels laden with surfactant aggregates has been proposed earlier (Chapter 4), and it predicts the following equation to describe surfactant release at short times, 100 2 Release Surfactant%2 hC t CDP S (5-2) where DS is the surfactant diffusivity, C* is the cri tical aggregation concentration (CAC), i.e., the concentration beyond which the surfactant forms aggregates inside the gel, t is time, Cp is the concentration of the surfactant present as aggreg ates inside the hydrogel, and h is the halfthickness of the hydrogel. The above equation is valid for diffusion of any solute that is loaded in the gel above saturation limit and so a fraction of the solute precipitates into aggregates. This equation is the equivalent of the Higuchi equa tion that is commonly utilized to model drug release from ointments when drug is present as a suspension [89]. To validate the model and to understand the m echanism of surfactant transport, surfactant release studies were conducted from gels of di fferent surfactant loadings (approximately 2%, 4% and 8% w/w in drug gel) and diffe rent gel thicknesses (~ 100 and 200 m in the dry state). Figure 5-2A-C shows the su rfactant release from 100 m thick and 200 m thick gels with three different surfactant concentr ations explored for 200 m thick gels. The data is re-plotted in

PAGE 114

114 Figure 5-3A-C with the time axis rescaled to 2hC tp, where t is time in seconds, Cp is the concentration in moles per liter and h is half-thickness of the gel in m. The model (Equation 52) predicts that the rescaled plots should over lap for all thicknesses and surfactant loadings, and the plots should be linear with the slope 100 2*CDS which agrees with all the experimental results for Brij 97 and Brij 78 laden hydrogel s within 95% CI. However, for the Brij 700 surfactant, the data matches the model only for 4% and 8% surfactant loading, while thickness scaling and surfactant release from gels containi ng 2% surfactant in dry gel do not follow the predicted behavior. We speculate that associa tion of Brij 700 with the gel matrix is strongly dependent on the gel thickness maybe due to larg e number of ethylene oxide (EO) units in the surfactant. Also, the CAC values for the Brij 700 laden system might be closer to 2% surfactant loading where the model assumption that C*<
PAGE 115

115 surfactants inside the gels to estimate both the parameters individually. We discuss some novel methods to evaluate CAC for surfactant laden hydr ogels in chapter 7. As mentioned above, the critical aggregation concentra tion for Brij 700 in p-HEMA gels is more than 1%, and so the release data from gels loaded with 1% (w/w) can be fitted to the following equation to obtain the surfactant diffusivity [87], % Release = 100 42h tDS (5-3) Here DS is the diffusivity of the surfactant and h is the half-thickness of the gel. The diffusivity of Brij 700 from these hydrogels was determined to be 7.85x10-17 m2/s and subsequently the value of C* was evaluated fr om the already determined parameter DsC* (Table 5-2) to be 2.4.1 mM, which is equivalent to 1.12.51% surfactant in a dry gel. This value is in reasonable agreement with our prior hypothesis that the C* for Brij 700 is approximately 2%, which was based on the overla pping percentage releas e data for 1% and 2% Brij 700 loaded gels. 5.3.2 CyA Release: Equilibrium Experiments A large fraction of hydrophobic drugs such as CyA are expected to partition inside the surfactant aggregates. During the drug release pr ocess, the drug molecules have to first diffuse through the surfactant h ead region into the p-HEMA gel, and subsequently diffuse through the gel. The head group of the surfactants may offe r resistance to transport of the molecules, and this transport could potentially be rate limiting. However if the resistance to transport from the surfactant aggregates to the p-HEMA gel is small, the concentrations insi de the aggregates will be in equilibrium with that in the p-HEMA gel, and in this case, diffusion through the p-HEMA gel will be rate controlling. If transport through the gel is rate controlling, the time scale for drug release scales as the square of th e gel thickness, and if the transport across the aggregates is rate

PAGE 116

116 controlling, the time scale for release should be in dependent of the gel thickness. To investigate the rate limiting step, drug release studies were conducted from 100 m and 200 m thick gels with 8% surfactant loading. It is noted that th e gel weight and the fluid volume in the release medium was kept the same for gels of both thicknesses. To determine the rate limiting process, we plot percentage release of the drug against where, 2Time Release i i = 1 for 100 m thick gels and 2 for 200 m thick gels in Figures 5-4A-B. It has been shown earlier that the rate limiting step in drug diffusion from p-HEMA, Brij 97 and Brij 98 laden gels is also diffusion controlled (Figure 3-8 and Figure 4-11). The data for different thicknesses for all the systems overlaps proving that the transport of drug in all cases is controlled by diffusion through the gel, and that the drug concentrations inside the aggregates and in the p-HEMA gel are in equilibrium. The data in Figures 5A-B, Figure 4-11 and Figure 3-8 also show that the time required to reach equilibrium is much greater for surfactant la den gels then for pure p-HEMA gels and Brij 78 systems take the longest time to equilibrate. Al so, the percentage of drug that diffuses out, till equilibrium is attained, is different for each syst em, with the relative order Brij 97 laden gels < Pure p-HEMA < Brij 78
PAGE 117

117 surfactant chain length and the packing in the ag gregates. Also, due to surfactant diffusing out during equilibrium experiments, the drug solubility can also increase in the release medium. It is thus probable that the presence of surfactants in the release medium can affect the amount of drug diffusing out. This interaction between th e drug and the surfactant in the release medium should be dominant only above CMC values and it is possible that for Brij 97 system, the surfactant concentration in the bulk phase does not reach CMC, while for other systems it does reach above CMC value, leading to smaller perc entage drug release for Brij 97 laden gels than pure p-HEMA systems. Essentially, the combin ed interactions of drug molecules with the surfactants present inside the gel matrix and outsi de in the release medium would then determine the equilibrium percentage release from the hydrogel. 5.3.3 Effect of Surfactant Dissolved in the Release Medium To understand the effect of surfactants (prese nt in the release medium) on the release rates and amount of drug released from the gel sy stem, we conducted experiments in which pure pHEMA gels (loaded with drug) were soaked in release mediums containing varying surfactant concentration. Three different surfactant concen trations were explored for each surfactant based on the theoretical value of their CMC as shown in Table 5-3. Specifically, concentration of surfactant was chosen as C1 (CMC/3), C2 (CMC ) and C3 (3*CMC) for all the systems studied and the results from all the experiments are shown in Figure 5-6 A-D. It was observed in all the experiments that as we increase the surfactant concentration in the re lease medium, amount of drug and its release rates also increase. We also observe that at short time scales, drug release rates from all the surfactant containing system s are slower than from systems containing no surfactants. This observation can be explained by taking into consideration the initial uptake of surfactants by the gel matrix which then can gi ve rise to drug-surfactant interaction at the periphery of the gel resulting in slower release rate s of the drug. It is also observed that amount

PAGE 118

118 of drug released at equilibrium for pure HEMA gel is nearly same for Brij 78 and Brij 700 containing release medium when the concentrati on of surfactant is below its CMC value (C1). This shows that there is no effect of surfactan t concentration on the release percentage of the drug when concentration of surfactant is below its CMC value. We do not observe this when Brij 98 and Brij 97 surfactants are dissolved in the release medium at concentrations below CMC and we speculate that the theo retical value of CMC for these two surfactants might not be accurate which may lead to an erroneous assumption that C1 concentration for these surfactants is below CMC. Researchers have also show n a wide range of CMC being reported in the literature for these non-ionic surfactants [119,120 ]. As we start increasing the surfactant concentration to a value above CMC, there is more drug released at equilibrium indicating that the drug is interacting with the micelles present in the release medium. Also, we observe that there is not much change in release rates and rele ase percentage of the drug for concentration C2 and C3 for Brij 98 surfactant. This may be becau se at much higher concentrations of surfactant in the release medium, rod shaped micelles may al so start forming and so at concentrations C2 and C3 we observe similar release behavior of the drug. 5.3.4 CyA Release: PBS Change Experiments Figures 5-7A-C show the percentage release of the drug with differe nt surfactant loading for all the surfactant laden gels that are 200 m in thickness and loaded with 50 g of drug. As expected, the percentage release of the drug de creases as the surfactan t loading is increased inside the hydrogel. A model for drug release from hydrogels laden with surfactant aggregates has been proposed earlier in Chap ter 4, and it predicts the foll owing equation to describe drug release at short times, 100 Release Drug%2 *h t (5-4)

PAGE 119

119 P P S P S PCD KCK DC D KC D K DC D KC 2 )1( 4 )1( 4 )1( 12 2/3 (5-5) K= hSmfMWK (5-6) Here, D = Diffusivity of drug Km = Partition coefficient of the drug define d as ratio of drug concentration inside the micelle to drug concentration in the hydrogel MWS = Molecular weight of the surfactant = Density of micellar core To establish the validity of the model and to determine all the model parameters, we replotted the percentage release of the drug against where = 2h t where t is the time in seconds and h is half thickness of the gel in m. We have also included the re-plotted data for systems loaded with Brij 98 surfactant and the re sults are shown in Figures 5-8A-D. The slopes between various surfactant loadings for each system differ as the 95% CI for the slopes do not overlap. We fitted the release data to a straight line to obtain and then used Equations 5-5 and 5-6 to obtain Km. The parameters determined for each system are listed in Table 5-4. The values of Km are relatively independent of the surfactant loading, thus validating the model. In these calculations, the diffusivity of the drug was taken to be 1.44x10-14 m2/s and was taken to be 1000 kg/m3. The value of K for 2% Brij 700 loading in the system could not be determined which is expected since it was shown earlier that at this concentration is close to the CAC value for this surfactant. Furthermore, in Figure 5-7C we observed the percentage release of CyA from p-HEMA gels overlapping that from the 2% Brij 700 loaded gels, again suggesting that there was no significant partitioning inside the surfactant ag gregates due to an insignificant number of such aggregates inside the hydrogel.

PAGE 120

120 The partition coefficient of the drug between the surfactant aggregates and the p-HEMA gels is highest for the Brij 700 laden gels and smallest for the Brij 97 laden hydrogels. Even though the value of Km is highest for Brij 700, these systems do not attenuate drug release significantly for two reasons. First, the value of C* is large (~2%) implying that only a small amount of surfactant is available for forming ag gregates. Second, due to the large molecular weight, the volume fraction of the hydrophobic core which provides the site for drug partitioning is small. The values of the partition coefficients clearl y suggest that an increase in the hydrophilic chain length leads to an increase in the partitio n coefficient. This is most likely due to an improved shielding of the hydrophobic core from water on increasi ng the hydrophilic chain length. Also, a comparison of Brij 78 and Brij 98 systems shows that there is a significant increase in the partition coefficient of the drug if the hydrophobic ta il of the surfactant is saturated (Brij 78). This could be attributed to the fact that an unsaturated chain is more rigid than a saturated chain, and an increase in rigidity will likely lead to reduced packing resulting in a reduced shielding of the core from water, an d a consequent reduction in the partitioning of hydrophobic drugs. The results reported above show that amongst the systems explored, Brij 78 is the most suitable candidate for extended drug delivery. Th e toxicological response of this surfactant on the ocular surface has been investigated in rabbits and it is reported that administration of 0.1 ml of 2% (20000 g/ml) eye drops does not cause toxic eff ects [67]. Assuming a bioavailability of about 2%, about 40 g of the surfactant delivered in the dr op reaches the cornea without causing any toxicity. The same study also showed that exposure to 0.05% Brij 78 solution for about 5 hours does not lead to any significant increase in corneal hydration again suggesting that Brij 78

PAGE 121

121 has negligible toxicity even with extended exposu re. The surfactant release studies reported here demonstrate that about 10% of the Brij 78 loaded in the 100 m thick gel with 8% loading is released in a period of 10 da ys, which corresponds to around 300 g of surfactant or equivalently an average release of about 30 g/day. When a contact lens is pl aced on an eye, at the most half this amount, i.e., 15 g/day will be released into the post lens tear film, which is the thin tear film in between the cornea and the contact lens Ocular conditions are likely not perfect-sink conditions and so the amounts released would be le ss than this level. T hus, it might be expected that the lenses loaded with Br ij 78 will not cause any toxicity even if the lenses are worn continuously for a few days. 5.3.5 DMS and DMSA Release: Equilibrium Experiments Since Brij 78 laden hydrogels were found to be most effective in attenuating CyA release rates, these systems were explored for delivering other hydrophobic drugs such as DMS and DMSA. All the gels prepared for DMS and DMSA were 100 m in thickness and weighed about 40 mg. In Figure 5-9 and Figure 5-10 we plot the percentage release of these drugs over a period of one day from hydrogels loaded with varying Brij 78 loading. The data shows no significant difference between release times of th e drugs from the surfactant laden gels when compared to release from control p-HEMA gels. To explain the negligible effect of surfactant loading on release times, we calculated the partit ion coefficient of the drug between the gel and the release medium (Table 5-5). These release e xperiments equilibrated in less than a day, and it could be assumed that during this time there was negligible surfactant loss from the hydrogel. We can then calculate the contribution of mice lles inside the hydrogel to the overall partition coefficient by the following equation, hMicelle HEMAp AvgfK KK )1 ( (5-7)

PAGE 122

122 Where, KAvg = Calculated partition coefficient in gel with respect to PBS Kp-HEMA = Partition coefficient of the drug in control p-HEMA gels with respect to PBS KMicelle = Partition coefficient of the drug in th e cores of the surfactant aggregates with respect to PBS = Fraction of surfactant inside the hydrogel, and thus fh is the fraction of the hydrophobic cores inside the gel The values of KMicelle calculated from Equation 5-7 are list ed in Table 5-5 for both drugs. The partition coefficient of the drug in the hydrophobic cores with respect to the p-HEMA matrix is simply the ratio KMicelle/Kp-HEMA, defined earlier as Km. The values of this ratio are 7.5.8 and 18.0.0 for DMS and DMSA, respec tively, which are both much less than the value of 458.9.5 obtained earlier for CyA. This implies that the cores of the surfactant aggregates have a much larger affinity for CyA than DMS and DMSA. These results also suggest that the ratio KMicelle/Kp-HEMA, which can be determined relatively easily, is the most critical parameter for the successful attenuatio n of drug release from the surfactant-laden hydrogels. 5.3.6 Uptake of Lysozyme in the Hydrogels Binding of tear proteins to contact lenses is undesirable as it can lead to increased bacterial binding to the lenses. Lyzozyme is the main protei n present in tears, and it is frequently used as the test protein to investigate pr otein binding to contact lenses [126]. It may be speculated that the presence of hydrophobic re gions inside the surfactant-laden contact lenses can lead to increased protein binding. To te st this hypothesis, hydrogels (200 m thick) were soaked in lyzozyme solution, and the mass of lysozyme bo und to the lens was determined by assaying the free lyzozyme concentration through UV-Vis sp ectrophotometry. Lysozyme uptake by the surfactant laden gels (8% w/dry ge l w) and pure p-HEMA gels is shown in Figure 5-11. It is

PAGE 123

123 observed that the presence of surfactant in th e gels does not have any significant effect on lysozyme adsorption in the gel matrix. 5.3.7 Microstructure of Hydrogels: Cryo-SEM Imaging Figures 12 A-O show a series of SEM imag es of the cross-sections of p-HEMA and surfactant laden gels (8% w/dry gel w). The pur e p-HEMA gels have a uniform structure with no visible pores, which is expected because pores in p-HEMA gels are a few nm in size. The microstructure of surfactant laden gels is in sharp contrast to pure p-HEMA gels, as they show a uniform distribution of pores with pore sizes ranging from 40-50 nm. The sizes of these pores are much larger than the expected micelle size suggesting that the structure of the surfactant aggregates inside these pores is likely more co mplex than micelles. The volume fraction of the pores is much larger than the surfactant loading, which implies that these pores are likely water rich environments, and so the presence of these pores should lead to incr eased water content in the gels. To investigate this issue further, the area fraction of por es in the hydrogels was determined by image analysis using ImageJ (Na tional Institute of Health) software and these values are listed in Table 5-6. If we assume that after hydration these pores are filled with water, then the equilibrium water content of the syst ems can be predicted by the following equation EWCpred(%) = 100) 100 )1( ( HEMApEWC (5-8) Where, = Fraction of area occupied by pores EWCp-HEMA = Water uptake in pure p-HEMA hydrogel The values of water content (EWCPred) from Equation 5-8 are also listed in Table 5-6. A quantitative analysis of the pore structure along with the equilibrium water content (see 5.3.8.2) shows that these pores are essentially filled with water when the gels are hydrated. Together with the 50 nm size this suggests that the su rfactant structures could possibly be vesicles.

PAGE 124

124 5.3.8 Physical Properties The surfactant-laden gels were characterized by several techniques to investigate the impact of surfactant loading a nd the resulting porous structur e and also to determine the suitability of these systems as a typical contact lens It should be noted that in all the studies in this section we used gels which contained 8% surfactant per dry gel weight. 5.3.8.1 Mechanical properties The storage moduli (G) and the loss moduli (G) for all the five systems explored in this study are plotted as a function of frequency in Figure 5-13A-B. All the gels explored for studying the mechanical properties were 400 m and 800 m thick. The results show a negligible effect of surfactant loading on the me chanical properties. The elastic modulus G continuously increases with increasing frequency and the values of the modulus at frequencies approaching zero for all the systems is close to the desired value for commercial lenses [46]. The loss modulus first increases and then decreases with frequency. The mechanical response of a contact lens plays an important role in lens se ttling, lens shape, and the pressure distribution on the post-lens tear film. It may thus be useful to obtain a model that can describe the mechanical properties of the lenses so that this model can be utilized in modeling lens deformation in the eyes due to application of the forces during blinking. To mode l the frequency dependence of the storage and loss modulus, we used a three para meter Standard Linear Solid Model which is shown in Figure 5-14. In this model there is an elastic spring connected to a viscous dashpot in series like a Maxwell model, with an addition of an elastic spring in parallel. To determine the expressions for G and G, it is instructive to co nsider application of a periodic strain. The stresses and strains in the individual elements are then related by the following expressions tie0321 (5-9)

PAGE 125

125 3 22E (5-10) The complex modulus G (=G + iG) is the rati o of the stress and the strain, and it can be obtained from the following expression tie EE iGG 0 2211 ''' (5-11) Using Equations 5-9, 5-10 and 5-11 we can evaluate G and G to be, 222 2 22 12 2 21 ')( E EEEE G (5-12) 222 2 2 2 '' E E G (5-13) The experimental data averaged from all the systems was fitted to the model, and the model fit is also plotted in Figure 5-13A-B, and the parameters are listed in Table 5-7. The data fits the model reasonably for lower frequencies, while there is a clear deviation from the model at frequencies greater than 25 s-1. A more generalized Maxwell model is needed for the data at higher frequencies but physiological blink frequencies are about 10 s-1 and so the model proposed above may be adequate. The loss moduli of the gels may partially be attributed to water flow during the gel stretching. The cont ribution of water flow to the moduli could be explored by measuring the moduli for gels of different thicknesses. Since water transport depends on the gel thickness, a significant dependence of the loss modulus on the thickness will indicate that water transport is an important cont ributor to the loss modulus. The storage and the loss modulus for pure p-HEMA gels are relativel y independent of the ge l thickness as shown in Figure 5-15. This suggests that water transport during gel stretching does not contribute to the loss modulus for the p-HEMA gels. The results were similar for the surfactant-laden gels and are not shown in Figure 5-15 for clarity of presentation.

PAGE 126

126 5.3.8.2 Transparency, equilibrium water content, and surface contact angle of gels All the gels explored in this study were visibly transparent and clear, and 100 m thick hydrated gels had transmittance valu es larger than 98.5% (Table 56) at a wavelength of 600 nm, and so are suitable for contact lens application. It was also observed th at the surfactant laden gels had a higher transparency than the pure p-HE MA gel likely due to the higher water content. The EWC of contact lenses is crucial as it lik ely impacts comfort, and also an increase in the EWC leads to an increase in the oxygen permeability of p-HE MA contact lenses. The EWC values of all the gels obtained by hydrating 200 m thick gels are listed in Table 5-6. All the surfactant systems had higher water content than pu re p-HEMA gels, and thus are more suitable for contact lens application. The water content appears to be a function of length of hydrophilic part of the surfactants and it increases as the length of (EO) groups of the surfactants increase. The measured values match the EWC predictions based on analysis of SEM images. This suggests that water content inside the hydrogel increases due to formation of pores which are filled with water and since the por e size is less than 40 nm for all the surfactant laden hydrogels, these remain visibly transparent after polymerization. The contact angles were measured for all systems to determine the effect of surfactants on the wettability of the hydrogels. Thickness of th e gels utilized for meas uring the contact angle was 200 m. It has been shown previously that su rfactants can significan tly alter the contact angles of hydrogels [127,128]. The captive bub ble technique was utilized to determine the contact angle to ensure that the gel remains hyd rated during the measurements. The values of contact angles listed in Table 5-6 are all lower th an those for the p-HEMA gels likely due to the presence of surfactant on the su rface making surfactant laden gels more suitable for contact lens application.

PAGE 127

127 5.4 Conclusion We have explored Brij surfactant laden pHEMA hydrogels for ophthalmic drug delivery by contact lenses. The surfactants explored he re have the same length of the hydrophobic group but differed in chain length of the hydrophilic unit (EO) and in pr esence of an unsaturated group in the tail region of the surfactant. Ophthalmic drugs were loaded in these systems by direct addition to the polymerizing mixture. The results reported here clearly prove that the duration of CyA release can be significantly reduced by incorp oration of surfactants inside the gel matrix by as much as a factor of five compared to pure p-HEMA gels. The mechanism of reduction in release rates is through a preferential partitioni ng of the drug into the surfactant domains that form inside the gels. The rate controlling step is diffusion through the gel but its rate is reduced due to a reduction in the free drug concentration. The concentration and type of the surfactant plays an important role as an increase in the hy drophilic chain length in creases partitioning into the hydrophobic cores of the surfactant aggregates and the presence of a double bond in the hydrophobic chain reduces the partitioning. Both these effects likely occur due to the impact of these factors on shielding of the hydrophobic core from the water molecules. Amongst the surfactants explored here, Brij 78 is most promising for extended release of CyA from p-HEMA contact lenses due to a high partition coefficien t of 458.9.5 for partitioning between the hydrophobic cores and the p-HEMA matrix. The pa rtition coefficient is even higher for Brij 700 due to the large hydrophilic chain length but th e total partitioning is small due to a smaller fraction of the hydrophobic segment and also due to a relatively large value of the critical aggregation concentration in the ge l. Furthermore, Brij 78 surfactants have been used in ocular studies as cornea permeability enhancers, and have been shown to have negligible toxicity at concentrations as large as 2% (w/w) [67]. Even though Brij-78 laden gels showed promising results for the drug CyA, these could not signifi cantly attenuate release of two other hydrophobic

PAGE 128

128 drugs, DMS and DMSA, due to a lower partition coefficient of these drugs inside the Brij 78 aggregates. Freeze fracture SEM imaging provides a direct evidence of the presence of pores in the surfactant-laden hydrogels. The pores are about 50 nm in size and are filled mostly with water, which suggests that the surfactant aggregates ar e perhaps more complex than micelles, and could possibly be vesicles. The surfactant-laden gels were characterized to determine their suitability as contact lenses materials. All the systems were clear and tr ansparent, and had storage moduli suitable for contact lens applications. The water content fo r all the surfactant laden gels was much higher than that for pure p-HEMA hydrogel, which is encouraging as it will incr ease oxygen permeability and perhaps comfort. The wettability of the lenses also improved due to surfactant entrapment with maximum improvement with Brij 78 which could also have beneficial effects. While the results reported here are very encouraging, they need to be supplemented with animal studies to explore the potential toxicity du e to continuous exposure to surfactants. Also, even though incorporation of the surfactants is likely to increase the oxygen permeability, it may not be sufficient for extended wear applications, and so it will be desirable to conduct studies similar to those reported here for silicone h ydrogels. Since silicone hydrogels are a mix of silicone and hydrogel materials, it may be po ssible to create both hydrophobic and hydrophilic domains, which may be useful for extended rele ase of both hydrophobic and hydrophilic drugs. In addition to the surfactants explored here, othe r surfactants and self assembling molecules such as lipids, and block-co-p olymers could be used to create dom ains that could trap and slowly release hydrophobic and/or hydrophilic drugs. Thus while issues related to toxicity still need to be explored, this study provides conclusive evid ence that p-HEMA contact lenses loaded with

PAGE 129

129 Brij 78 surfactant have physical properties suitable for contact lens applications, and also for extended drug delivery.

PAGE 130

130 y = -10.303Ln(x) + 66.038 R2 = 0.99235 40 45 50 55 60 65 70 75 0.11101001000Concentration ( g/ml)Surface Tension (mN/m)A y = -3.7195Ln(x) + 61.902 R2 = 0.994445 50 55 60 65 70 0.11101001000Concentration ( g/ml)Surface Tension (mN/m)B Figure 5-1. Dependence of surface tension on th e bulk surfactant concentration A) Brij 78 surfactant B) Brij 700 surfactant

PAGE 131

131 0 5 10 15 20 25 30 05101520Time (Days)% Surfactant Release Brij 97_100_8% Brij 97_200_8% Brij 97_200_4% Brij 97_200_2%A 0 5 10 15 20 25 30 051015202530 Time (Days)%Surfactant Release Brij 78_100_8% Brij 78_200_8% Brij 78_200_4% Brij 78_200_2%B Figure 5-2. Percentage release of surfactant during water change experiments. The gel thicknesses in m and the percentage of surfactant lo aded in the gel are indicated in the legend. A) Brij 97 surfactant system B) Brij 78 surfactant system C) Brij 700 surfactant system.

PAGE 132

132 0 2 4 6 8 10 12 14 16 18 20 0510152025 Time (Days)%Surfactant Release Brij700_200_1% Brij700_100_8% Brij700_200_8% Brij700_200_4% Brij700_200_2%C Figure 5-2. Continued

PAGE 133

133 Slope = 0.48020 5 10 15 20 25 30 35 40 020406080% Surfactant Release Brij 97_100_8% Brij 97_200_8% Brij 97_200_4% Brij 97_200_2%A Slope = 0.1846 0 2 4 6 8 10 12 14 16 18 20 020406080100% Surfactant Release Brij 78_100_8% Brij 78_200_8% Brij 78_200_4% Brij 78_200_2%B Figure 5-3. Cumulative percentage release of surfactant from th e hydrogels after rescaling the time. represents t/Cp/h2 where t is time in seconds, Cp is surfactant concentration in M, and h is half thickness of the gel in m. A) Brij 97 surfactant system B) Brij 78 surfactant system C) Brij 700 surfactant system.

PAGE 134

134 Slope = 0.0613 0 2 4 6 8 10 12 14 16 18 20 050100150%Surfactant Release Brij700_200_4% Brij700_100_8% Brij700_200_8% Brij700_200_2%C Figure 5-3. Continued

PAGE 135

135 0 10 20 30 40 50 60 010203040% Drug Release Brij 78_100 Brij 78_200A 0 10 20 30 40 50 60 70 80 0 5 101520% Drug Release Brij 700_100 Brij 700_200B Figure 5-4. Effect of thickness on percentage re lease of drug during equilibrium experiments. represents t/i2 where t is release time in hours an d i = 1 for thin gels and i = 2 for thick gels. All the gels contained 50 g of drug. A) Brij 78 surfactant system B) Brij 700 surfactant system.

PAGE 136

136 Figure 5-5. Microstructure of the surfactant-laden gel Free Drug Drug bound to Surfactant aggregates Free Surfactant Gel Matrix Surfactant adsorbed on the polymer Drug adsorbed on the polymer Surfactant aggregates

PAGE 137

137 0 10 20 30 40 50 60 70 0 5 10 15 20 Time (Days)% Drug Release Brij 97_C1 Brij 97_C2 Brij 97_C3 No SurfactantA 0 10 20 30 40 50 60 70 80 05101520 Time (Days)% Drug Release Brij 98_C1 Brij 98_C2 Brij 98_C3 No SurfactantB Figure 5-6. Effect of surfactan t dissolved in the release medium on cumulative percentage release of the drug during equi librium experiments. All the gels contained nearly 55 g of drug and were around 200 m thick. A) Brij 97 surfactant system B) Brij 98 surfactant system C)Brij 78 surfactant system D) Brij 700 surfactant system.

PAGE 138

138 0 10 20 30 40 50 60 05101520 Time (Days)% Drug Release Brij 78_C1 Brij 78_C2 Brij 78_C3 No surfactantC 0 10 20 30 40 50 60 70 05101520 Time (Days)% Drug Release Brij 700_C1 Brij 700_C2 Brij 700_C3 No surfactantD Figure 5-6. Continued

PAGE 139

139 0 10 20 30 40 50 60 70 80 90 100 0510152025Time (Days)% Drug Release Brij 97_200_2% Brij 97_200_4% Brij 97_200_8% Pure p-HEMAA 0 10 20 30 40 50 60 70 80 90 100 0102030405060Time (Days)% Drug Release Brij78_200_2% Brij78_200_4% Brij 78_200_8% Pure p-HEMAB Figure 5-7. Effect of surfactant loading on cumula tive drug release from surfactant-laden gels in PBS change experiments. All the gels were 200 m thick and contained 50 g of drug. A) Brij 97 surfactant system B)Brij 78 surfactant system C) Brij 700 surfactant system.

PAGE 140

140 0 10 20 30 40 50 60 70 80 90 100 0510152025Time (Days)% Drug Release Brij700_200_2% Brij700_200_4% Brij700_200_8% Pure p-HEMAC Figure 5-7. Continued

PAGE 141

141 0 10 20 30 40 50 60 70 80 90 100 051015% Drug Release Brij 97_200_2% Brij 97_200_4% Brij 97_200_8%A 0 10 20 30 40 50 60 70051 01 5% Drug Release Brij78_200_2% Brij78_200_4% Brij 78_200_8%B Figure 5-8. Cumulative percentage release of drug from 200 m thick surfactant-laden gels in PBS change experiments after rescaling the time. represents t/h2 where t is time in seconds and h is half thickness of the gel in m. A) Brij 97 surfactant system B) Brij 78 surfactant system C) Brij 700 surfact ant system D) Brij 98 surfactant system.

PAGE 142

142 0 10 20 30 40 50 60 70 80 90 100 051015% Drug Release Brij700_200_2% Brij700_200_4% Brij700_200_8%C 0 10 20 30 40 50 60 70 80 90 100 05101520% Drug Release Brij 98_200_2% Brij 98_200_4% Brij 98_200_8%D Figure 5-8. Continued

PAGE 143

143 0 10 20 30 40 50 60 70 051015202530 Time (h)% Drug Release Brij 78_2% Brij 78_4% Brij 78_8% Pure HEMA Figure 5-9. Percentage re lease of DMS from 100 m thick p-HEMA and Brij 78 laden in equilibrium experiments 0 2 4 6 8 10 12 14 16 18 20 0510152025 Time (h)% Drug Release Brij 78_2% Brij 78_4% Brij 78_8% Pure hema Figure 5-10. Percentage release of DMSA from 100 m thick p-HEMA and Brij 78 laden gels in equilibrium experiments

PAGE 144

144 700 750 800 850 900 950 1000 1050 051015202530 TIme (Days)Concnetration (mg/ml) Brij 78 Brij 97 Brij 98 Pure p-HEMA Figure 5-11. Lysozyme sorption in surfactant laden and pure p-HEMA hydrogels

PAGE 145

145 A Figure 5-12. Cryo-SEM image for 200 m thick gels. A) pure p-HEMA gels:Scale-1m B) pure p-HEMA gels:Scale-2m C) pure p-HEMA gels:Scale-5m D) Brij 98 laden gels:Scale-1m E) Brij 98 laden gels:Scale-2m F) Brij 98 laden gels:Scale-5m G) Brij 97 laden gels:Scale-1m H) Brij 97 laden gels:Scale-2m I) Brij 97 laden gels:Scale-5m J) Brij 78 laden gels:Scale-1m. K) Brij 78 laden gels:Scale-2m L) Brij 78 laden gels:Scale-5m M) Brij 700 laden gels:Scale-1m N) Brij 700 laden gels:Scale-2m O) Brij 700 laden gels:Scale-5m. 1 m

PAGE 146

146 B C Figure 5-12. Continued 2 m 5 m

PAGE 147

147 D E Figure 5-12. Continued 1 m 2 m

PAGE 148

148 F G Figure 5-12. Continued 5 m 1 m

PAGE 149

149 H I Figure 5-12. Continued 5 m 2 m

PAGE 150

150 J K Figure 5-12. Continued 1 m 2 m

PAGE 151

151 L M Figure 5-12. Continued 5 m 1 m

PAGE 152

152 N O Figure 5-12. Continued 2 m 5 m

PAGE 153

153 0 0.5 1 1.5 2 2.5 3 05101520253035Frequency (Hz)Storage Modulus (MPa) Brij 78 Brij 97 Brij 98 Brij 700 pure p-HEMA ModelA 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 05101520253035Frequency (Hz)Loss Modulus (MPa) Brij 78 Brij 97 Brij 98 Brij 700 pure p-HEMA ModelB Figure 5-13. Frequency depe ndence of moduli for 800 m thick surfactant laden and pure pHEMA gels. A) Storage Modulus B) Loss Modulus

PAGE 154

154 Figure 5-14. Standard Linear Solid Model used for fitting the viscoelasticity data of the surfactant-laden gels 0 0.5 1 1.5 2 2.5 3 05101520 Frequency (Hz)Modulus (MPa) Pure p-HEMA_Storage_800 Pure p-HEMA_Storage_400 Pure p-HEMA_Loss_800 Pure p-HEMA_Loss_400 Figure 5-15. Effect of thickness on the storage an d loss moduli of pure p-HEMA gels. The gel thicknesses in m are indicated in the legends. 1, E1 3, 2, E2

PAGE 155

155 Table 5-1. Physical properties of th e surfactants explored in this study System Name Chemical Formula Molecular weight HLB [120] fh Brij 97 POE (10) oleyl ether C18H35(OCH2CH2)10OH 709.0 12.4 0.47 Brij 98 POE (20) oleyl ether C18H35(OCH2CH2)20OH 1149.5 15.0 0.31 Brij 78 POE (20) stearyl ether C18H37(OCH2CH2)20OH 1151.5 15.0 0.31 Brij 700 POE (100) stearyl ether C18H37(OCH2CH2)100OH4670.0 18.8 0.08 Table 5-2. DSC* for all systems obtained from fitting of the surfactant release data System Slope DSC* (m)2M/s x 106 Brij 97 0.4802 0.010 11.52 Brij 98 1.1000 0.013 39 60.80 Brij 78 0.1846 0.050 1.70 Brij 700 0.0613 0.013 0.19 Table 5-3. Concentration of surfactan ts dissolved in the release medium System CMC (g/ml) C1 (g/ml) C2 (g/ml) C3 (g/ml) Brij 97 290.0 100 286 861 Brij 98 304.5 101 307 924 Brij 78 53.0 18 53 159 Brij 700 93.4 31 93 270

PAGE 156

156 Table 5-4. Partition coefficient of CyA for all the surfactant systems System Amount of surfactant Slope K (M-1) Km 2% 11.08 0.83 14.14 4% 9.22 0.20 18.35 Brij 97 8% 7.76 0.33 16.40 48.9 6.3 2% 6.37 0.43 142.20 4% 4.51 0.40 163.10 Brij 78 8% 3.19 0.33 186.10 458.9.5 2% 9.17 0.20 99.46 4% 7.46 0.16 96.46 Brij 98 8% 5.71 0.07 83.10 261.0.4 2% 13.87 2.32 4% 7.26 0.67 258.48 Brij 700 8% 5.33 0.17 252.47 675.8.4 Table 5-5. Partition coefficient of DMS and DM SA in p-HEMA and Brij 78 surfactant laden hydrogels DMS DMSA System KAvg KMicelle KAvg KMicelle Pure p-HEMA 54.5 341.3 Brij 78 (2%) 55.3 366.8 368.0 5885.4 Brij 78 (4%) 57.0 399.5 384.2 4563.6 Brij 78 (8%) 61.4 456.9 503.7 7950.4

PAGE 157

157 Table 5-6. Physical properties of the surfactant laden and pure p-HEMA hydrogels System EWCPred (%) EWC (%) Transmittance (%) Contact Angle Brij 97 (8%) 31.0 67.6 64.1 99.8 27.9.06 Brij 98 (8%) 29.1 66.7 67.2 99.2 24.9.40 Brij 78 (8%) 41.0 72.3 70.3 99.9 19.8.78 Brij 700 (8%) 46.4 74.8 70.4 99.5 27.2.59 Pure p-HEMA 53.0 98.9 30.3.18 Table 5-7. Parameters obtained by fitting Standard Linear Solid Model to the experimental data Parameter Value E1 0.55 MPa E2 0.90 MPa 0.0076 MPa.s

PAGE 158

158 CHAPTER 6 LIPOSOME ASSAY FOR EVALUATING OCULAR T OXICITY OF SURFACTANTS 6.1 Introduction One in vitro method of assessing ocular toxicity is the utilization of liposomes to mimic cell permeation by the test substance. The advant ages of using liposome leakage to assess ocular toxicity include low cost and the ability to assess many compounds rapidly. Additionally, the test is quantitative and so lacks the unpredictabil ity that can be associated with using live cells. This test is based on the idea that the permeation of a test substance through lipid bilayers is the root cause of inducing ocular toxicity, with toxicity being ca used by the leakage of cellular components, which increases substantially on binding of the test substance to the bilayer. The liposome based assay is designed so that the lipid composition of the bilayers imitates the composition of corneal epithelial cells. The test measures the leakage of fluorescent dye from the liposome core upon interaction with a test s ubstance. The maximum score of the Draize eye test is 110, with 80 out of 110 coming from the cornea alone, suggesting that the assessment of corneal toxicity should be the main focus of an in vitro alternative. This fact first inspired researchers to test liposomes as a possible means of assessing the ocular toxicity of surfactants [111,112]. Since that time, a few others have examined liposome leakage as well [108,113]. Good correlations to in vivo data were obtained in some cases for some surfactants, with gross outliers sometimes present. In this Chapter, we propose that the lack of good correlation in some studies between the liposome based assay and the Draize test was due to neglect of mechanistic issues, and that a better correlation can be obtained by design ing the liposome assay after mechanistic considerations. Specifically, while compari ng the Draize test to increases in liposome permeability upon exposure to surfactants, most researchers evaluated the liposome permeability

PAGE 159

159 at a fixed surfactant concentration or at a conc entration which induced 50% dye leakage, and the concentrations were significantly below the conc entration used in the Draize test. Herein, we first show that the liposome perm eability in the presence of su rfactants at a fixed surfactant concentration does not correlate well with the Dr aize scores. We then show that the correlations are significantly improved when the surfactant c oncentration in the liposome assay is chosen to be CMC/200, where CMC is the critical micelle co ncentration, which varies significantly across the surfactants explored in this study. We also show the rationale for th is choice of surfactant concentration in the liposome assay based on mechan istic considerations. Finally, we utilize the liposome assay developed here to determine the ocular toxicity of several Brij surfactants for which available ocular toxicity data was very limited or nonexistent and have been used previously to attenuate drug release from contact lenses. 6.2. Materials and Methods 6.2.1 Materials Methanol, chloroform, Dulbeccos Phosphat e Buffered Saline (PBS) without calcium chloride and magnesium chloride, Sephadex G-50 (fine), cholesterol (CH), sodium dodecyl sulfate (SDS), polyoxyethylene sorbitan monolaurate (Tween 20) polyoxyethylene sorbitan monooleate (Tween 80), hexadecyltrimethylam monium bromide (CTAB), cetyl pyridinium chloride (CPC), benzalkonium chloride (BKC), Brij 56, Brij 58, Brij 98, Brij 76, Brij 78, Brij 97, and Brij 700 were purchased from Sigma Aldric h. Whatman GF/B glass microfiber filters, calcein dye (fluorexon), polyoxyethylene sorbitan monopalmitate (Tween 40), myristyltrimethylammonium bromide (MTAB), and Triton X-100 were pur chased from Fisher Scientific. Octadecyltrimethylammonium bromide (OTAB) was purchased from K & K Laboratories. The lipids 1,2-Dioleoyl-sn-Glycero-3-Phosphoethanolamine (DOPE), dissolved in chloroform, 1,2-Dimyristoyl-sn-Glycero-3-[P hospho-rac-(1-glycerol)] (Sodium Salt) (DMPG),

PAGE 160

160 in powder form, and 1,2-Dimyristoyl-sn-Glycero-3-Phosphocholine (D MPC), in powder form, were purchased from Avanti Polar Lipids, Inc. A Mini-Extruder kit for liposome preparation was also purchased from Avanti Polar Lipids, Inc. 6.2.2 Liposome Preparation for Calcein Leakage Studies Liposomes composed of a molar ratio of 8:6:1.5:1.5 of DMPC:DOPE:DMPG:CH encapsulating an aqueous 100 mM calcein dye so lution were prepared via a combination of mixing, sonication, and extrusion. The ratio of lipids was chosen based on the composition of the corneal epithelium [113]. Lipids were combined in their respective molar ratios and then dissolved in a 9:1 mixture (by volume) of chloroform:methanol such that a 20 mg/mL concentration of lipids was obtain ed. The organic solvent was then evaporated under a stream of nitrogen. After an even and uniformly dried lip id film was obtained, the dried lipid layer was hydrated with 100 mM calcein dye dissolved in PB S such that the lipid concentration was 25 mg lipid/mL. The lipid suspension was then mixed using a Fisher vortex mixer for 1-2 minutes, followed by bath sonication for 20 minutes. The ve sicles were then gently stirred overnight at 30C for approximately 20 hours, due to the incr ease in entrapped aqueou s volume of liposomes with increased stir times [129]. After stirri ng, the liposome solution was extruded through a 100 nm membrane 15 times. To remove excess dye from the bulk, the liposome solution was passed through a mini-column of Sephadex G-50 (fine) us ing the centrifugation method to ensure that a large portion of the lipids added to the Sephad ex bed were recovered [130]. The resulting liposome solution was diluted by a factor of 1001 based on the observation that the calcein release at the subsequent concentration fell in the linear detection regime. 6.2.3 Liposome Leakage Studies Surfactant induced calcein leakage from li posomes was used to mimic surfactant permeation into corneal epithelial cells in vivo The 100 mM calcein solution inside the

PAGE 161

161 liposomes was in the quenched state, as is the case for highly concentrated calcein and carboxyfluorescein, so that only the diluted dye which leaked into the bulk medium gave a signal [130]. The baseline fluorescence of the liposome solution prior to leak age was first measured using a Quantech Digital Filter Fluorometer w ith excitation and emission filters at 490 and 515 nm, respectively. A concentrated surfactant solution was then added such that the final surfactant concentration was either 1 g/mL or the critical micelle concentration (CMC) of the surfactant divided by 200 (discussed later). Th e liposome solution was kept at room temperature in between subsequent fluorescent measurements. To compute the percent release, the following formula was used, %Release = 100 o total otFF FF (6-1) Where Ft was the fluorescence measurement at time t (10 minutes), Fo was the fluorescence at time zero, and Ftotal was the total calcein released, which was determined by breaking the liposomes with 100 L of 20% (v/v) Triton X-100. Corrections were made to account for the dilution upon addition of the surfactant and Trit on X-100 solutions. All release experiments were carried out at least twice. 6.2.4 Draize Scores Draize scores for ten of the surfactants te sted (SDS, Tween 20, Tween 40, Tween 80, Triton X-100, MTAB, CTAB, OTAB, CPC, BKC) were obtained from three independently published studies [105,131,132]. Kennah et al published 24 h average Draize scores for multiple concentrations of surfactants, enabling two correlations from their publication [131]. Their scores for concentrations of 10% and 1% (vol %) were used. For SDS, authors reported Draize scores for 30%, 15%, and 3% surfactant co ncentrations, and their data was fitted to a calibration curve to compute the corresponding Dr aize scores at 10% and 1%. Tachon et al.

PAGE 162

162 reported the maximum Draize score after 1 h or 24 h at 10% w/v surfactant concentrations [132]. Finally, Matsukawa et al. reported both the maxi mal average scores and the 24 h average scores for 10 % solutions, and the latter was used for consistency [105]. Table 6-1 shows the surfactants used for the in vitro/in vivo correlations and their corres ponding Draize scores from each publication. Note that not all surf actants were studied in each published work. 6.2.5 Data Analysis Draize scores and percentage of calcein re lease from liposomes after 10 minutes were correlated using Pearsons correlation coeffici ent (r) and the coefficient of determination (R2), as well as Spearmans rank correlation coefficient. Both types of measures were computed based on the variety of coefficients used in the literature and the ques tion of validity associated with using ordinal data for Pearsons coefficient. The prediction equations calculated from Draize data at 10% from Kennah et al., Tachon et al., and Matsukawa et al. [105,131,132] were compared using 95% confidence intervals for mean Draize scores as a function of percentage calcein leakage generated in JMP (SAS, Cary NC). Also, the predicted Draize scores from the correlations were compared to th e actual values qualitatively to determine the percentage of surfactants correctly categorized. One of three categories was assigned to each surfactant based on the following scale for Draize scores: mild/mod erate (0-25), irritant (>25-50), severe (>50110) 6.3 Results and Discussion 6.3.1 Draize Score / Leakage Correlations at a Constant Test Concentration In some of the prior studies, researchers meas ured the leakage of dyes from liposomes after exposure to surfactants at a give n fixed concentration and then a ttempted to correlate the leakage with the Draize score. In this section, we report results from a similar study in which the bulk concentration of surfactant in the liposome solution was fixed at 1 g/ml, which was below the

PAGE 163

163 CMC value of all the surfactants tested. It is a common belief that the Draize eye score, if reported from different labs, can have statistically significant differences [103]. To avoid this, we chose three publications with Draize eye scor es available for most of the surfactants we utilized in our study (Table 6-1) In Figure 6-1 we pl ot the Draize scores from different authors versus the percentage dye leakage after 10 minute s at fixed surfactant con centrations. We used two different concentrations from the publicatio n of Kennah et al. to obtain two correlations [131]. The correlation coefficients (Pearson and Spearman) and the coefficient of determination (R2) for the fits at fixed concentrations are report ed in Table 6-2. As expected, the correlation coefficients for all the publications do not show a promising relationship between the Draize test and the liposome study. We also did experiments at two higher concentrations (data not shown) and obtained poor correlations for the fits be tween the Draize data and the percentage dye leakage from the liposomes after 10 minutes. 6.3.2 Draize Score / Leakage Correlations at Adjusted Concentrations As shown in the previous section, the liposom e assay does not correlate well to the Draize scores for surfactants if the surfactant concentrati on in the liposome leakage studies is kept fixed. In this section, we report the results for leak age studies from liposomes with the surfactant concentration kept at CMC/200. Specifically, we correlate the Draize data with the percentage dye leakage from liposomes after 10 minutes when the concentration of the surfactant introduced in the solution was adjusted to be CMC/200. Table 6-3 shows the CMC values taken from literature and the concentrations utilized for the test. Figure 6-2 shows th e correlations between the reported Draize scores and the percentage dy e leakage after 10 minute s. Consequently, the following correlation equations for various sets of Draize scores obtained via different in vivo concentrations and from diffe rent groups were produced:

PAGE 164

164 a) Kennah et al. [131], Concentration 1% v/v: Draize score = 12.57 Ln (% Leakage) + 3.76 (6-2) b) Kennah et al. [131], Concentration 10% v/v: Draize score = 16.19 Ln (% Leakage) + 34.83 (6-3) c) Matsukawa et al. [105], Concentration 10% w/v: Draize score = 18.18 Ln (% Leakage) + 21.64 (6-4) d) Tachon et al. [132], Concentration 10% w/v: Draize score = 13.75 Ln (% Leakage) + 22.35 (6-5) Correlation coefficients in Table 6-2 show a si gnificant improvement over the correlations obtained by comparing the dye release at fixed surfact ant concentrations. This is true of both the Spearman and Pearson coefficients. The Spearman co efficient is perhaps more significant in this case, despite the common practice of reporting the Pearson value, due to the ordinal nature of the data. Equally important for application purposes is the percent of surfactants categorized correctly, based on the system described in Sec tion 6.2.5. Clearly, the number of surfactants correctly identified into one of three irritanc y categories has improved in most cases. These observations are confirmation that the correction factor introduced does indeed provide a clearer understanding of the relationship between Draize scores and the percentage dye leakage from liposomes. We also evaluated the 95% confidence intervals for the mean values for all the correlations at a fixed concentration of 10% for the Draize scores and the resu lts are shown in Figure 6-3. The three confidence intervals overlap and this suggest that the correlations obtained from different sources with different surfactants may be used simultaneously in this case for Draize score prediction purposes. 6.3.3 Mechanism of Surfactant Toxicity Surfactants have been shown to cause toxicity by penetrating the epithelial cell membrane, causing irreparable damage to the ocular tissue. Jester et al. suggested that the degree of ocular

PAGE 165

165 irritation caused by surfactants depends on the initi al area and depth of inju ry, suggesting that the area and volume both determine the toxicity levels of surfactants [133]. Tachon et al. pointed out that damaged cells release lysosomal enzymes, histamine, and inflammatory mediators and suggested that the ability to permeate cells was a major sign of toxicity [132]. Okahata and Ebato claimed that eye irritancy is due to the pe netration of surfactant molecules into the lipid bilayer as they correlated Draize scores to the partition coefficients of surfactants [108]. Authors did not found a significant correlation between the hydrophilic-lipophilic balance (HLB) and the Draize score and concluded that mere lipophilicit y arguments cannot determine the irritancy of a surfactant. Other properties such as electrostatic interactions, steric effects, and charge density significantly contribute to the interaction between the surfactant and the corneal epithelium. The liposomes used in our study have been designe d to mimic the lipid composition of the corneal epithelium [113]. Based on this, an increase in the permeability of liposomes in the presence of surfactants should directly correlate with eye irri tancy. The mechanism of surfactant toxicity is shown in Figure 6-4. Release of substances from the corneal epithelium is shown to be directly related to monomer surfactant penetrating inside the cells. Liposome leakage as a toxicity assay to meas ure surfactant toxicity has been explored in the past with little understanding of the underlying criteria needed to corr ectly test the potential irritant [108,111-113]. The liposome composition is chosen to mimic the cornea lipid composition, and so the key parameters that need to be chosen for the liposome assay are the lipid loading, liposome size, and the surfactant concentration. Among these, the surfactant concentration is the most important parameter and must be chosen based on mechanistic considerations. Surfactant molecules will form micelles on the ocular surface if introduced at concentrations above their critical micelle c oncentration (CMC) and th ese micelles should have

PAGE 166

166 little or no interaction with the epithelial layer due to thermodynamic considerations. Essentially, only the surfactant monomers should dire ctly bind to the lipid bilayers of the corneal epithelium. This point was specifically addres sed by Okahata and Ebato, who measured the absorption of surfactants onto a lipid-coated quartz microbalance and showed good correlation with Draize scores [108]. Absorption became saturated for nonionic and cationic surfactants above the CMC, pointing to a lack of micelle interaction with th e membrane. Furthermore, this point was supported by the work of Hall-Manning et al., who found skin irrita ncy to be related to the amount of surfactant monomer on the surface (CMC), rather th an the absolute concentration tested [134]. Thus, only the free surfactant, wh ich is present at the CMC, should interact with the corneal epithelium. As the surfactant molecu le starts to penetrate the epithelial bilayer, micelles on the ocular surface break and the concen tration of free surfactan t on the ocular surface is constantly maintained at the CMC until all the surfactant micelles have either drained or have been broken. Thus, if a surfactant is introduced at varying concentrations on the ocular surface, the drainage of surfactant micelles should affect the extent of ocular toxicity. As the concentration of surfactant introduced on the ocul ar surface increases, the total time for drainage of the surfactant from the ocular surface will also vary, leading to an increase in toxicity. Once the concentration of the surfactan t introduced exceeds a critical va lue where the time of drainage is longer than the time at which the toxicity is assessed, ocular toxicity should saturate. These points are supported by Draize da ta given by Kennah et al. and Matsukawa et al., which show increases in Draize scores well above the CMC values of surfactants and then subsequent leveling off of those Draize scores [105,131]. Thus, residence time is a key factor to consider when comparing one Draize score to another for va rying surfactant concentra tion. This point is quantitatively addressed in a later section. For this reason, we have used Draize data generated

PAGE 167

167 from the same stock surfactant concentration for each of our correlations, rather than using any Draize data available at any con centration above the surfactant CMC. Draize data also varies widely from study to study, and so we have generated separate correlations for each study, rather than intermingling all of the available data. The degree of surfactant penetration inside the ocular surface should also be proportional to the surface area availa ble for the surfactant to diffuse in. Similarly, when liposomes are used as an alternative, the amount of surfactant penetrating the lipid bilayer should be directly proportional to the surface area of the liposomes pr esent. This clearly suggests that experiments have to be designed with differences between liposomes and the corneal epithelium taken into consideration. To illustrate the importance of differences between liposome and corneal geometry, it is instructive to consider a mass balance on species such as lysosomal enzymes, histamine, and inflammatory mediators that begin to leak from inside the corneal cells due to the toxic effects of surfactant penetration into the bilayer of the corneal epithelium. The mass balance yields S Cornea Perm S CorneaCAK dt dC V (6-6) where VCornea is the cellular volume of the corneal epithelium, KPerm is the permeability of the corneal epithelium to the species that leaks out, ACornea is the corneal area available for penetration, and CS is the concentration of the species of interest inside the corneal cells. The above equation treats the cornea epithelium as well-mixed, which is perhaps not a precise assumption. However this simpler treatment is su fficient to illustrate the approach for obtaining the surfactant concentration that should be used in the liposome assay. Similarly, a mass balance on a test component such as a dye present inside the liposomes can be given by

PAGE 168

168 Dye Liposome LipoPerm Dye LiposomeCAK dt dC V, (6-7) Where VLiposome is the volume of the liposomes, KPerm,Lipo is the permeability of the liposomes, ALiposome is the liposome surface area available for penetration, and CDye is the concentration of the species that leaks out. The presence of su rfactants is manifested in increased permeabilities for both the cornea and the liposomes. Based on the above equations, the time scale for the leakage of the molecules is V KA. The surface area to volume ratio (A/V) is much larger for liposomes due to their small size. Thus, if the surfactant concentration for the liposome assay is chosen to be the same as that in the Draize te st, or even the CMC, the time scale for dye leakage will be extremely small, and so the percentage leakage will be very large unless the measurements are done at extremely short times. Since short-time measurements are prone to artifacts due to issues such as mixing, it is more appropriate to ensure that the time scale for the leakage in the liposomes is comparable to that in the corneal cells. Since the time scale is V KA, the higher values of A/V can be compensate d by a lower K for the liposome assay. The permeability K is related to the amount of surfactant that binds to the liposomes, and so the value of K in the liposome assay can be controlled by controlling the surfactant concentration in the assay. Based on these arguments, Equation 6-8 can be used to evaluate the effective concentration that should be tested in the liposome assay to correctly predict the irritancy of surfactants in vivo, CMC AV VA CLiposome Cornea Liposome Cornea TEST (6-8) Equation 6-8 implicitly assumes that permeab ility is linearly related to the surfactant concentration, and also that the appropriate conc entration that controls binding in the Draize test

PAGE 169

169 is not the total concentration but the surfactant monomer concentration, which equals the CMC. In Table 6-4, we show the calculated values for the surface area to volume ratios for the corneal epithelium and the liposome system. From these values we can evaluate the test concentration for different surfactant systems based on their CMC s. Since the surface area to volume ratio is a rough calculation and the reported CMC values for surfactants can vary, we rounded the correction factor up to 200. Thus, we divi ded each CMC value by 200 to obtain the most physiologically relevant test concentration possibl e for our liposomal system. This surface area to volume ratio correction has not been done in previous reports and is crucial to ensuring that the two systems allow for a correct comparison. Our correction factor of 200 depends heavily on our liposomes having mean diameters of around 110 nm and new correction factors must be computed when working with liposomes of dramatically different sizes [135]. It is also noted that since the value of this ratio was obtained from several qualitative or partially quantitative arguments, there is likely a range of the value of this ratio, perhap s from about 100 to 1000, that could be used in the assay. The central hypothesis is that the test concentrations for the liposome assays for several different surfactants should be chosen such that the ratio of the test concentrations and the respective CMCs are the sa me for each surfactant, an d that this ratio is around 200. 6.3.4 Comparison of Liposome Assay with other in Vitro Assays Matsukawa et al. [105] used the EYTEX test as an in vitro model to predict ocular toxicity. They argued that since protein denaturation is one of the most important factors in determining the extent of ocular irritation, it could be used as an alternative. The overall correlation coefficient between Draize scores and the EYTEX test was reported as 0.313, which is very poor when compared to our values (Table 6-2). Also, using their technique had a major limitation of not being able to predict toxi city for cationic surfactants, whereas with the

PAGE 170

170 liposome leakage technique we can accurately predic t the irritancy of cationic surfactants. Vian et al. compared three different in vitro techniques to determine the ocular toxicity of various surfactants [104]. They utilized neutral red uptake assay, the MTT tetrazolium salt assay, and the total protein content assay fo r correlating in vivo Draize da ta and concluded that among the three techniques, neutral red uptak e assay was best correlated with the Draize score though even this assay had insufficient correl ation. Tachon et al. used cell mortality and inhibition of cell growth as assays to assess surfactant toxicity [132]. They got reasonable agreement with the Draize test score and concluded that penetration of surfactant in the cell lines was responsible for cell damage. Cottin et al. correlated toxicity to cell leakage using a fluorescent dye [109]. They measured the amount of surfactan t needed to induce 20% leakage and related it to Draize scores via a non-linear relationship. They found strong correlations between in vivo and in vitro assays and suggested that this method co uld be another addition for an in vitro alternative to the Draize eye test. Kennah et al. have suggested in the past that there is a need for another in vivo alternative to the Draize test due to poor re producibility and the subjective assessment which differs from one researcher to other [103,131]. They sought to accomplish this by measuring the corneal thickness of rabbit eyes before and afte r exposure. They found a linear relationship between their test and the Draize score with a reasonable correl ation, but failed to comment on why a linear fit should be observed even though the Draize score saturates at a maximum value of 110. A reasonable logarithmic fit can also be obtained from their data and it remains unclear as to why the authors chose a linear relationship. All the relevant corr elations, including our work, are compared in Table 6-5. As can be seen from the results, liposome leakage studies have potential to be used as one of the in vitro alternatives for early predictions of the irritancy of surfactants and possibly other substances as well.

PAGE 171

171 6.3.5 Prediction of Ocular Toxici ty for Non-ionic Surfactants Surfactants have been actively explored in lit erature as potential permeability enhancers to increase the permeability of some common ocular drugs [67]. We have also shown previously that surfactants can be used to attenuate drug rele ase from contact lenses and they can potentially diffuse from the lenses to the ocul ar surface and lead to potential to xicity. This is a first attempt at evaluating the potential ocular toxicity of similar surfactants for which Draize eye scores are currently unavailable. We predic ted the ocular irritancy levels of these surfactants by performing similar experiments as before, where the test co ncentration of surfactant was adjusted according to Equation 6-8. Leakage from liposomes was eval uated as discussed in th e previous section and Draize scores were predicted from the correla tions (Equations 6-2 to 6-5) obtained by fitting Draize scores from various sources. In Table 6-6, Draize score predictions for six non-ionic surfactants are presented. Since all the predictio ns were made for a surfactant concentration of 10% (w/v) on the ocular surface, the predicted Draize scores sh ould correspond to the same tested concentration. Similarly, predicted valu es for ocular irritancy at 1% surfactant loading were obtained by using the correlation from the da ta of Kennah et al. and is shown in Table 6-7 [131]. As expected, the predicted Draize scores ar e higher for the higher concentration with Brij 78, Brij 700, Brij 56 and Brij 58 showing neg ligible toxicity for the 1% concentration correlation. The mechanism of toxicity should be identical for a particular class of surfactants. For non-ionic surfactants the determining factor in toxicity should be the hydrophobic interaction of the surfactant with the lipid bilayer. Hydrophobic interac tion of stearyl and cetyl chains should be stronger than oleyl chains because of the double bond of the oleyl chain, which makes oleyl surfactants more hydrophilic. This would suggest that Brij 78 (C18H37(OCH2OCH2)20OH), Brij 700 (C18H37(OCH2OCH2)100OH), Brij 56 (C16H33(OCH2OCH2)10OH), and Brij 58

PAGE 172

172 (C16H33(OCH2OCH2)20OH) should be more toxic compared to Brij 97 (C18H35(OCH2OCH2)10OH) and Brij 98 (C18H35(OCH2OCH2)20OH). On the other hand, the CMC values of Brij 97 and Brij 98 surfactants are much higher than the other surfactants (Table 6-3). Thus, when present at concentrations higher than the CMC, the number of monomers for Brij 98 and Brij 97 surfactants would be much larg er than the other surfact ants, leading to more toxicity. This explains why we get a higher Draize score for Brij 97 and Brij 98 compared to other surfactants at both 10% and 1% surfactant lo ading. Since no Draize data is available for these systems, we can at best speculate that the oleyl series of surfactants should be more toxic than the stearyl and cetyl groups of surfactants if administered on the ocular surface above their respective CMCs due solely to the larger num ber of monomers available for epithelial cell penetration for the oleyl series. Draize scores for Brij 78 have been reported previously and the au thors found that the Draize score at 1% w/v surfactant loading was 2 [67] This value is in agreement with what we observe (Table 6-7) for this surfactant at similar co ncentrations. It is noted here that we try to predict the overall Draize score from liposomes mimicking only the corneal epithelium whereas there is a contribution from both the iris and conjunctiva in the overall Draize score, which can result in prediction errors. Our correlations on the other hand, suggest that reasonable predictions can be derived by correlating the overall Draize da ta with percentage liposome leakage. 6.3.6 Model for Micelle Depletion from the Ocular Surface In the previous sections we have proposed that the interaction of surfactants and the epithelial layer occurs at a constant concentration which corresponds to the CMC of the surfactant. The surfactant micelles on the ocular surface have little intera ction with the ocular tissues and either break due to surfactant adsorption inside the corneal epithelium and

PAGE 173

173 conjunctiva to maintain the free surfactant concen tration at the CMC, or drain from the ocular surface through the nasal route. Researchers have shown that as the conc entration of the test material on the ocular surface is increased to concentrations above CMC, the corresponding Draize scores also increase, whereas at really higher concentrations th ere is a saturation of Draize scores [131]. At first glance, this ap pears to be a conundrum in that though the interaction between surfactants and the ocular surface occurs at a maximum concentration corresponding to the surfactant CMC, the toxici ty still increases with increasing surfactant concentration. This seems to be a direct effect of the residence time of surfactant micelles on the ocular surface, which should vary with incr easing surfactant concentration. This can be explained more clearly by looking closely at the mass balances on the tears and surfactant on the ocular surface. A mass balance on the surf actant on the ocular surface can be given by, CMCAkCMCAkCMC Cq dt VCdConj Conj Cornea Cornea Micelle Drainage Micelle ) ( )( (6-9) And the tear balance on the ocular surface can be given by, nEvaporatio aConjunctiv Drainage Secretionqqqq dt dV (6-10) Where CMicelle = Concentration of mice lles on the ocular surface CMC = Critical micelle concentration ACornea = Area of the cornea AConj = Area of the conjunctiva kCornea = Permeability of the surfactant in the cornea kConj = Permeability of surfact ant in the conjunctiva qSecretion = Rate of tear production qDrainage = Rate of tear drainage qConjunctiva = Rate of tear penetra tion inside the conjunctiva qEvaporation = Rate of tear evapora tion from the ocular surface V = Volume of tears on the ocular surface These coupled equations show that the residence time of the surf actant micelles on the ocular surface should be a functi on of surfactant concentration.

PAGE 174

174 It is instructive to consider an extrem e case in which the dominant mechanism for surfactant loss from the tear film is canalicular dr ainage. This is a likely scenario at surfactant concentrations much above CMC. In this case the time for the surfactant concentration in the tear film to go from the initial concentration Ci to CMC is equal to CMC C q Vi Drainageln. Thus, a higher initial Ci will lead to a larger duration in which the surfactant concentration in the tear volume is larger than CMC, and the larger time w ill lead to a larger influx of the surfactant into the cornea, causing higher toxicity. This clear ly indicates that the Draize eye score, which should be indicative of the amount of surfactant monomer diffusing inside the cornel epithelium, should increase with increasing surfactant concentra tions due to increased residence times. 6.4 Conclusion Damage to the corneal epithelium can be attributed to the disruption of membrane fluidity due to the penetration of external agents such as surfactants, and the subsequent release of lysosomal enzymes, histamine, and inflammatory mediators. Interactions between surfactants and the corneal surface are govern ed by their respective CMCs, as micelles are not expected to interact with lipid bilayers. This has been indi cated indirectly in the past by researchers who have tried to correlate Draize scores with specif ic surfactant concentrations resulting in 20% to 50% leakage of fluorescent dyes from liposom es [111,113]. Thus, it has been shown that different surfactants interact differently with the corneal epithelium, and the interaction is dependant on, but not limited to, the concentrati on of the surfactants on the ocular surface. We propose that this concentr ation is the CMC of different surfactants, and that to successfully develop an in vitro alternative to the Draize eye test using liposomes, it is imperative to account for CMC differences. Moreover, liposomes mimicking the corneal epithelium can be successfully utilized to assess th e toxicity of various su rfactants if a correction

PAGE 175

175 factor is introduced to account for the increa sed surface area to volume ratio of liposomes compared to the corneal epithelium. Once this factor was introduced, the correlations between dye leakage from liposomes and Draize scores impr oved significantly. This method can be used to evaluate the initial toxicity of various surfactants, and coul d thus become a key method to assess ocular toxicity in vitro Accordingly, the correlations between Draize eye scores and liposome leakage produced were used to predict th e ocular toxicity of six non-ionic surfactants for which ocular toxicity data was non-existent. We predict that Brij 78, Brij 700, Brij 56, and Brij 58 are mildly/moderately comfortable when placed in the eye at concentrations of 10% (w/v), while Brij 97 and Brij 98 appear to be irritati ng at similar concentrations. At 1% (w/v), all of the surfactants examined ar e most likely in the mild/modera te category, causing little to no discomfort.

PAGE 176

176 0 20 40 60 80 100 0246810 Liposome leakage ( % )Draize scoreA 0 10 20 30 40 50 60 0246810 Liposome leakage ( % )Draize scoreB Figure 6-1. Draize scores versus liposome leakage after 10 minut es induced by surfactants at concentrations of 1 g/mL and logarithmic correlations for A) Draize scores from Kennah et al. [131] evaluated at 10% (v/v) with rp = 0.38, rs = 0.49, B) Draize scores from Kennah et al. [131] eval uated at 1% (v/v) with rp = 0.74, rs = 0.59, C) Draize scores from Tachon et al.[132] evaluated at 10% (w/v) with rp = 0.26, rs = 0.43, d) Draize scores from Matsukawa et al.[105] evaluated at 10% (w/v) with rp = 0.74, rs = 0.63.

PAGE 177

177 0 10 20 30 40 50 60 0246810 Liposome leakage ( % )Draize scoreC 0 20 40 60 80 100 0246810 Liposome leakage ( % )Draize scoreD Figure 6-1. Continued

PAGE 178

178 0 20 40 60 80 100 120 010203040506070 Liposome leakage ( % )Draize scoreA 0 10 20 30 40 50 60 010203040506070 Liposome leakage ( % )Draize scoreB Figure 6-2. Draize scores versus liposome leakage after 10 minut es induced by surfactants at CMC/200 and logarithmic correlations for A) Draize scores from Kennah et al. [131] evaluated at 10% (v/v) with rp = 0.82, rs = 0.79, B) Draize scores from Kennah et al.[131] evaluated at 1% (v/v) with rp = 0.99, rs = 0.94, C) Draize scores from Tachon et al. [132] evaluated at 10% (w/v) with rp = 0.74, rs = 0.85, D) Draize scores from Matsukawa et al.[105] evaluated at 10% (w/v) with rp = 0.78, rs = 0.79.

PAGE 179

179 0 10 20 30 40 50 60 024681012 Liposome leakage ( % )Draize scoreC 0 20 40 60 80 100 010203040506070 Liposome leakage ( % )Draize scoreD Figure 6-2. Continued

PAGE 180

180 0 20 40 60 80 100 03691215 Liposome leakage ( % )Draize score Figure 6-3. The 95% confidence intervals for mean Draize scores at 10% ocular loading for surfactants based on logarithmic corrlea tions from percent dye leakage from liposomes after ten minutes at surfact ant CMC/200. Kennah et al. [131] ( ); Tachon et al. [132] ( ); Matsukawa et al. [105] ( ). Figure 6-4. Surfactant induced t oxicity on the corneal surface Tear Film Corneal Epithelium Release of components from the cells Monomer diffusion inside the cells Surfactant micelles do not interact with the corneal epithelium a) Break due to lowering of monomer concentration below CMC b) Drain from the ocular surface through tear drainage Micelle Surfactant monomer Lysosomal enzymes, Histamine, Inflammatory mediators

PAGE 181

181 Table 6-1. Draize scores used for in vitro/in vivo correlations Surfactant Tachon et al. Draize Scores (10%) [132] Matsukawa et al. Draize Scores (10%) [105] Kennah et al. Draize Scores (10%) [131] Kennah et al. Draize Scores (1%) [131] SDS 37.34 14.70 40.33 b 5.83 b Tween 20 5.67 0.01 a 1.00 0.01 Tween 80 3.83 0.01 Triton X 100 40.33 59.00 2.00 CPC c 52.67 93.00 84.00 36.00 Tween 40 1.50 MTAB 42.66 CTAB 44.00 39.00 BKC 78.00 98.00 56.00 OCTAB 56.30 aValues of 0.01 were used in place of 0 for nonlinear correlation purposes. bInterpolated from a linear relationship for concentr ations of 0, 3, and 15%. cCetyl pyridinium bromide was tested in Kennah et al. [131] and Tachon et al. [132] and assumed to behave similarly to cetyl pyridinium chloride which was tested here.

PAGE 182

182 Table 6-2. Correlation comparisons for Draize scores and leakage experiments performed at surfactant concentrations of 1 g/mL and CMC/200 Source Pearson Spearman R2 Percent Categorized Correctly Kennah et al. a (10% 1 g/mL) [131] 0.38 0.49 0.14 50.0 Kennah et al. b (10% CMC/200) [131] 0.82 0.79 0.68 60.0 Kennah et al. c (1% 1 g/mL) [131] 0.74 0.59 0.54 100.0 Kennah et al. d (1% CMC/200) [131] 0.99 0.94 0.99 100.0 Matsukawa et al. a (10% 1 g/mL) [105] 0.74 0.63 0.55 71.4 Matsukawa et al. b (10% CMC/200) [105] 0.78 0.79 0.61 71.4 Tachon et al. a (10% 1 g/mL) [132] 0.26 0.43 0.07 62.5 Tachon et al. b (10% CMC/200) [132] 0.74 0.85 0.55 87.5 aDraize scores were evaluated at stock surf actant concentrations of 10%, while leakage experiments were evaluated at a final concentration of 1 g/mL. bDraize scores were evaluated at stock surfactant concentrations of 10%, while leakage experime nts were evaluated at a final concentration of CMC/200. cDraize scores were evaluated at stock surfactant concentrations of 1%, while leakage experiments were eval uated at a final concentration of 1 g/mL. dDraize scores were evaluated at stock surfactant concentrations of 1% while leakage experiments were evaluated at a final c oncentration of CMC/200.

PAGE 183

183 Table 6-3. Critical micelle concentrations for surfactants studied and subsequent test concentrations for liposome leakage Surfactant MW CMC (mM) CMC ( g/mL) Test Concentratio n CMC/200 ( g/mL) Source Triton X-100 647.0 0.200 129.40 0.647 Roche Applied Science SDS 288.4 8.200 2364.72 11.824 Rosen [136] Tween 20 1226.0 0.050 61.30 0.307 Hait and Moulik [120] Tween 40 1283.6 0.023 29.52 0.148 Hait and Moulik [120] Tween 80 1309.7 0.010 13.10 0.065 Hait and Moulik [120] BKC 340.0 8.800 2992.00 14.960 Rosen [136] OTAB 391.9 0.310 121.49 0.607 Rosen [136] CTAB 364.5 0.980 357.19 1.786 Rosen [136] MTAB 308.0 3.600 1108.80 5.544 Rosen [136] CPC 340.0 0.900 305.99 1.530 Rosen [136] Brij 58 1120.0 0.007 7.84 0.039 Hait and Moulik [120] Brij 56 682.0 0.002 1.36 0.007 Hait and Moulik [120] Brij 97 709.0 0.400 283.60 1.418 Hait and Moulik [120] Brij 98 1153.5 0.265 305.69 1.528 Hait and Moulik [120] Brij 78 1151.5 0.006 6.56 0.033 Hait and Moulik [120] Brij 700 4670.0 0.020 93.40 0.467 Hait and Moulik [120]

PAGE 184

184 Table 6-4. Surface area to volume ratio comp arisons for liposomes and epithelial cells Liposome (SA / V) Ratio (m2/L) Epithileal Cell (SA / V) Ratio (m2/L) 54545 a 314 d 66439 b 375 e 68412 c Mean (SA / V) ratios 63132 345 Ratio of (SA / V) 183 aCalculated from the av erage surface area, 62 2, occupied by a single lipid [130,137,138] and the entrapped volume [139] based on a liposome radius of 55 nm.[135]. bCalculated based on a unilamellar bilayer thickness of 35 and a liposome radius of 55 nm [135,138]. cCalculated based on a unilamellar bilayer thickness of 40 and a liposome radius of 55 nm [140]. dTaken from Maric et al. for epithelial cells [141]. eTaken from Farinas and Ve rkman for epithelial cells [142]. Table 6-5. Correlation comparisons between the liposome leakage method of assessing toxicity and other published methods Source Study Pearson Spearman % Correct Matsukawa et al. [105] EYETEXTM 0.20 0.40 N / A 61 70 Vian et al. [104] Cell Studies 0.48 0.62 0.53 0.64 75 Kennah et al. [131] Corneal Thickness 0.86 N / A 76 Tachon et al. [132] Cell Growth Inhibition N / A 0.65 0.85 N / A Cottin and Zanvit [109] Cell Leakage 0.94 0.92 93 Current Study Liposome Leakage 0.74 0.99 0.79 0.94 60 100 a aBased on the three class system described in section 6.2.5.

PAGE 185

185 Table 6-6. Predicted Draize scores for 10% st ock solutions of selected Brij surfactants Surfactant Kennah et al. (10%) a [131] Matsukawa et al. (10%) a [105] Tachon et al. (10%) a [132] Average Standard Deviation Irritation Class b Brij 97 45.1 34.5 31.1 36.9 7.3 Irritant Brij 98 42.9 32.1 29.2 34.7 7.2 Irritant Brij 78 29.8 18.3 18.1 22.1 6.7 Mild / Moderate Brij 700 32.2 20.9 20.1 24.4 6.8 Mild / Moderate Brij 56 29.3 17.8 17.6 21.6 6.7 Mild / Moderate Brij 58 34.3 23.1 21.9 26.4 6.8 Mild / Moderate aDraize scores were evaluated at surfactant conc entrations of 10%, while liposome leakage was evaluated at CMC/200 for each surfactant. bBased on the scale presented in section 6.2.5. Table 6-7. Predicted Draize scores for 1% st ock solutions of selected Brij surfactants Surfactant Kennah et al.(1%) a [131] Irritation Class b Brij 97 11.0 Mild / Moderate Brij 98 9.2 Mild / Moderate Brij 78 0.0 c,d Mild / Moderate Brij 700 0.8 Mild / Moderate Brij 56 0.0 c Mild / Moderate Brij 58 2.5 Mild / Moderate aDraize scores were evaluated at surfactant conc entrations of 1%, while liposome leakage was evaluated at CMC/200 for each surfactant. bBased on the scale presented in section 2.5. cIn cases where predictions were slightly negativ e, the Draize score was taken to be zero. dCompares well with the work of Saettone et al ., who reported a score of 2 for Brij 78 at a concentration of 1% [67]. They also showed an increase at 2%, which supports our finding of a higher score at 10% (Table 6-6).

PAGE 186

186 CHAPTER 7 ASSESING CRITICAL AGGREGATION CONC ENT RATION FOR SURFACTANTS IN HYDROGELS 7.1 Introduction It is important to determine the CAC values for surfactants in hydroge ls to design and tune drug release. In this chapter we discuss two methods of determining the critical aggregation concentration (CAC) inside p-HEMA matrix for the surfactants used in this work. In chapter 5 we determined the CAC of Brij 700 and here we obtain the CAC values for Brij 97, Brij 98 and Brij 78. The first method for determining CAC relies on the fact that the formation of the surfactant aggregates leads to a slowdown of drug transport due to drug partitioning into the surfactant aggregates. Thus gels with fixed drug loading and different surfactant loadings were prepared and drug diffusion from these gels was measured. The second method relies on the fact that the water content of the ge ls increase with increasing surf actant amount. However the rate of increase will likely be discontinuous at the critical aggregation concentration signaling a first order phase transition. Thus ge ls with varying surfactant loadi ng were prepared, and their water content was measured. 7.2 Materials and Methods 7.2.1 Materials 2-Hydroxyethyl methacrylate (HEMA) mono mer, ethylene glycol dimethacrylate (EGDMA), Dulbeccos phosphate buffered saline (PBS), acetonitrile, polyoxyethylene(20) stearyl ether (Brij 78), polyoxyethylene(10) oleyl ether (Brij 97), polyoxyethylene(20) oleyl ether (Brij 98) and HPLC grade water were purch ased from Sigma-Aldrich Chemicals (St Louis, MO). 2,4,6-trimethylbenzoyl-diphenyl-phophi neoxide (TPO) was ki ndly provided by Ciba (Tarrytown, NY). CyA was purchased from LC Laboratories (Woburg, MA). All the chemicals

PAGE 187

187 were reagent grade. Acetonitrile was filtered be fore use and all the other chemicals were used without further purification. 7.2.2 Synthesis of Surfactant and Drug Laden Gels The surfactant laden gels were prepared by polymerizing the monomer solution containing surfactant and drug mixed in spec ific ratio. Briefly, specifi c amounts of surfactants were dissolved in DI water to make surfactant solu tions of compositions such that the surfactant loadings in the dry gel ranged from 0.05% to 8% Separately, 3.5 mg of drug was dissolved in 2.7 ml of HEMA monomer and stirred at 600 rpm for a period of 5 hours. Next 15 l of the crosslinker and 2ml of su rfactant solution were ad ded to 2.7 ml of drug loaded monomer. The solution was degassed by bubbling nitrogen gas through it for 10 minutes followed by addition of 6 mg of UV initiator (TPO) and stirring for 10 minutes. The solution was then poured between two glass plates separa ted by a spacer and the gel was cured by irradiating UVB light (305 nm) for 40 min from an Ultraviolet transill uminiator UVB-10 (Ultra Lum, Inc.). Control drug loaded p-HEMA gels without surfactants were prepared by following procedures identical to those described above except that the 2 ml su rfactant solution was replaced by 2 ml DI water. Control gels without any drug were synthesized in a similar manner as described above except that the drug was not mixed in the monome r solution before polymerization. After polymerization, each gel was removed from the glass mold, and was cut into smaller square pieces that were dried at room temperature for two days before being used for any experiments. 7.2.3 Drug Detection: HPLC Assay CyA concentration was measured using a HPLC (Waters, Alliance System) equipped with a C18 reverse phase column and a UV detector The mobile phase composition was 70% acetonitrile and 30% DI water, and the column was maintained at 60C. The flow rate was fixed at 1.2 ml/min and the detection wavelength was set at 210 nm. The retention time for CyA under

PAGE 188

188 these conditions was 4.55 minutes, and the ca libration curve for area under the peak vs. concentration was linear (R2 = 0.995). 7.2.4 Drug Release Square gel pieces about 1.5X1.5 cm in size a nd 40 mg in weight were utilized for drug release experiments. Drug release kinetics was measured by soaking the gel in 3.5 ml PBS buffer, which was replaced every 24 hours and all the measurements were done at room temperature. These experiments were conducte d till 60% of the drug diffused from the gel matrix. 7.2.5 Equilibrium Water Content The gels of known weight with varying surfactant loading but no drug were soaked in 3.5 ml of PBS, and the dynamic weight was measured as a function of time. The excess water from the gel surface was removed before each meas urement by dabbing with Kimwipes (Fischer Scientific). The equilibrium water content (EWC ) of the surfactant laden gels was calculated by determining amount of water uptake per dry gel weight, i.e, 100 % DRY DRY WETW WW EWC (7-1) For PBS uptake experiments, surfactants laden gels without the drug were utilized. 7.3 Results and Discussion 7.3.1 Method I: Drug Release We have earlier developed a model for drug transport from surfactant laden hydrogels (Chapter 4, Section 4.3.3.1). Here we propose a new method to determine CAC for various surfactants by performing drug release from th e surfactant lade n hydrogels by vary ing surfactant concentration and using the model. To understand th is method, we first need to clearly decipher the role of surfactant in drug release from hydrogels. When su rfactants are introduced inside the

PAGE 189

189 gel matrix containing a drug, their interaction w ith the drug solute depend s strongly on surfactant concentration. Below their respective CAC, surf actants have only minor interactions with the drug solute but as the concentr ation of the surfactant is in creased beyond the CAC, surfactant starts aggregating, and th e hydrophobic interact ion between the drug and the hydrophobic interior of the surfactant aggregates becomes pronounced. Hydrophobic drug molecules such as CyA partition inside the hydrophobic interior of the surfactant aggr egates, and this leads to a slowdown in the drug release from the hydrogel. This phenomenon can be utilized to determine the CAC for various surfactants, especially if th e presence of surfactants can significantly slow down the drug release above their CAC. Firstly, we conducted drug release from gels containing varying amount of surfactant loading. The rate of release scales as t1/2, and so we determined the slope (SlopeEXP) of the linear fit between % Drug Release and 2)2( h twheret is time in hours and h is half-thickness of the gel. This slope is relatively unaffected by surfactant addition till the surfactants begin to form aggr egates, and thus the surfactant concentration at which the slope begins to decrease equals the critical aggregatio n concentration. A better and more quantitative value of the CAC can be determined by utiliz ing the model for drug release from surfactantladen gels proposed earlier and briefly described below. Drug release from the surfactant laden hydroge ls was discussed in chapter 4 and the following concentration profiles were determined for solute diffusion from surfactant laden hydrogels. 0 '2deBCD D Is (7-2)

PAGE 190

190 de d e deB Cp KC D s D p s sKC D D D D II2 1 2 21 0 '1 1 (7-3) In Equation 7-2 and 7-3, D is th e diffusivity of the drug (1.44x10-14 m2/s), DS is the diffusivity of the surfactant, Cp is the concentration of surfactant present as aggregates inside the hydrogel defined as Ci-C* (Ci = Initial surfactant concen tration inside the hydrogel, C* = CAC) and K is a constant related to the partition coefficient of th e drug when partition coeffi cient is defined as the ratio of drug concentration inside the micelles and drug concentration in side the gel matrix. Equation 7-2 is the solution for drug diffusion from region inside the gel matrix which does not contain any surfactant aggregates whereas Equation 7-3 is the concen tration profile of the drug in region II, i.e., the region with surfactant aggregates (See Figure 4-3). is given by the solution to following equation, de e C Cp 0 *2 22 (7-4) In Equation 7-2 and 7-3, is defined as y/ 4DSt and the thickness of region I by 4DSt (= ). The unknown B in the above equations can be dete rmined by using the following flux balance at y = (or = ), D DCKC d dC d dCsIIp II I '2 (7-5) Utilizing Equation 7-2, 7-3 and 7-5 we can evaluate the constant B as,

PAGE 191

191 0 )1( )1(2 214 1 d e KCD D e Bp D p DKC D D p S KC D D (7-6) Where, D deDKC d e KCD D de e eD D Sp KC D D p S D D KC D D D DD p D D p D D 0 0 )1( 0 )1(2 2 2 2 22 14 (7-7) The percentage release of drug fr om the gel can be determined by calculating the flux from the gel and integrating it over time (See Section 4.3.3.1) to give the following relation, 100 )2( )1( 2 (%)2 h t DKC DB RSp (7-8) And slope for a plot of % Drug Release vs. 2)2( h tshould be, 100 )1( 2 Sp THDKC DB Slope (7-9) For a given value of CAC, we can determine th e theoretical slope for any value of initial surfactant loading, sin ce all the other parameters in Equa tion 7-9 are known, and are listed in Table 7-1. The value of CAC can thus be dete rmined by minimizing the differences between the measured and the theoretical slopes. This was accomplished by minimizing the Error between the model and the experiments defined as, iC TH EXP THSlope Slope Slope Error2 (7-10)

PAGE 192

192 In Figure 7-1A-C we plot the Error vs. CAC for the three surfactants and the minimum value for error in each graph corresponds to the CAC for different surfactant systems. The values of CAC for various surfactant systems are listed in Table 7-2. For the CAC corresponding to the minimum error, we also plot the theoretical a nd experimentally determined slopes for various initial surfactant loadings in Figure 7-2A-C where the x-axis represents1/Ci where Ci is in mM. As can be seen from the plots, there is a good agreement between the theoretical and experimentally determined slopes for all the surfactant systems. 7.3.2 Method II: Water Uptake It was observed earlier that pr esence of surfactants inside the hydrogel can significantly alter the microstructure of the gel (Chapter 5, Sect ion 5.3.7). It was also shown that this change in microstructure can alter the water uptake proper ties of the gels. This change in water uptake properties of the hydrogels could be used to de termine the CAC for these surfactants. This method assumes a direct correlation between th e equilibrium water uptake (EWC) of hydrogels with the surfactant loading. We believe that as the concentrati on of the surfactant is increased beyond the CAC of the hydrogel, ther e is a significant alteration of the microstructure of the hydrogel, which should result in changes in the rate of increase in water content with increase in surfactant concentration. Figure 7-3A-C shows th e water uptake by gels at varying surfactant loadings for the three surfactant of interest. As can be seen in all the systems, the water content increases with surfactant loading but there is a sudden change in the slope of EWC increase with increasing surfactant loading at critical surf actant loading, which likely corresponds to aggregation of surfactants into aggregates. Th e increase in water content is linear in both regimes (surfactant loading less or greater than CAC). Accordingly, the EWC vs. surfactant loading data (Ci) was fitted to two strai ght line relationships with discontinuous slope at the CAC. The CAC values calculated from these syst ems are listed in Table 7-2 for comparison with

PAGE 193

193 the CAC determined from the previous method, a nd we see that both the methods lead to similar values except for Brij 98 surfactant system. 7.3.3 Surfactant Diffusivity We had earlier fitted the su rfactant release data to a model to determine the product DSC* for all surfactants of interest. Now that we have also determined C*, which is the critical aggregation concentration, we can determine the surfactant diffusivities. The values of DSC* are listed in Table 7-1 and the diffusi vities obtained by using the C* (=CAC) determined above are reported in Table 7-2. It is observed that the diffu sivities of the surf actant are not a strong function of their molecular weight as the diffusivity for Brij 97 is smaller than that for Brij 98. This clearly shows that surfactant diffusivity, un like solute diffusivity from theses hydrogels depends strongly on the surfact ant-polymer interactions. 7.4 Conclusion Drug release from hydrogels can be controlled only when surf actant concentration is above the CAC inside the hydrogel and hence it is im portant to know the value of CAC for various surfactant-polymer systems. Two methods for de termining the CAC for Brij series surfactants are suggested in this chapter. Both the met hods are based on the hypothesis that there is a significant change in gel propertie s as the surfactant concentrati on is increased beyond the CAC. Both methods yield relatively similar CAC values.

PAGE 194

194 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 345678 CAC (mM)Error x 103A 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 4681012 CAC (mM)Error x 104B Figure 7-1. Error between theoretical and expe rimentally determined slope for drug release experiments from hydrogels containing vary ing surfactant loading against CAC. A) Brij 78 System B) Brij 97 system C) Brij 98 system

PAGE 195

195 0 2 4 6 8 10 12 66.577.588.59 CAC (mM)Error x 105 Figure 7-1. Continued

PAGE 196

196 0.2 0.6 1 1.4 1.8 2.2 00.511.522.5 1/Ci (mM-1)SlopeA 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3 00.511.522.53 1/Ci (mM-1)SlopeB Figure 7-2. Plot of slope vs inverse of initial surf actant loading inside the hydrogels. Solid line is the theoretically determined slope from the model determined at the CAC with minimum error in fitting. A) Brij 78 system B) Brij 97 system C) Brij 98 system

PAGE 197

197 0.5 0.8 1.1 1.4 1.7 2 00.511.522.5 1/Ci (mM-1)SlopeC Figure 7-2. Continued

PAGE 198

198 R2 = 0.6563 R2 = 0.9264 72 73 74 75 76 77 78 79 80 81 82 01020304050 Ci (mM)% EWCA R2 = 0.8105 R2 = 0.9525 70 72 74 76 78 80 82 84 86 88 020406080100120 Ci (mM)% EWCB Figure 7-3. Equilibrium water content of surf actant laden hydrogels with varying initial surfactant loading inside the p-HEMA matrix. Change in slope in the plot should correspond to the CAC for different surfactan t systems. A) Brij 78 system B) Brij 97 system C) Brij 98 system

PAGE 199

199 R2 = 0.9452 R2 = 0.6124 70 75 80 85 90 95 020406080 Ci (mM)% EWCC Figure 7-3. Continued

PAGE 200

200 Table 7-1. Physical properties of the surfact ant utilized in solving the drug release model System K (M-1) DSxCAC (m2mM/s) MW Brij 97 16.3 11.52x10-15 709.0 Brij 98 93.0 60.80x10-15 1149.5 Brij 78 163.8 1.70x10-15 1151.5 Table 7-2. Critical aggregation concentration for each surfactant syst em evaluated from two different techniques CAC (mM) System Method I Method II Average CAC DS (m2/s) Brij 97 7.35 7.35 7.35 1.57x10-15 Brij 98 7.95 4.48 6.22 9.77x10-15 Brij 78 4.45 4.46 4.45 3.82x10-16

PAGE 201

201 CHAPTER 8 CONCLUSION Ophthalm ic drug delivery via contact lenses is beneficial for increas ing the bioavailability of the drug and also reducing the side effects of ophthalmic drugs due to reduction in systemic uptake. Most commercially available contact lens es can be utilized to deliver various drugs to the ocular surface though the major drawback of thes e systems is a fast release from these lenses and most of the drug diffuses out within few hours of application. This work has focused on designing novel contact lenses whic h can increase the therapeutic e fficiency of contact lenses by controlling drug release rates of common ophthalmic drugs. This is achieved by introduction of microemulsion and micelles inside the contact lenses. In chapter 2 we show that by using drug nanopa rticles inside the c ontact lenses, release rates of the drug CyA can be cont rolled and sustained delivery is possible for a week. Major advantage of this system is its easy availability and ease of application. It is also pointed out that as the concentration of drug is increased above th e solubility limits of the p-HEMA gel network, the gel starts to loose its transparency. This means that though release rates of the drug can be controlled by increasing drug concen tration, there is a significant compromise of gel properties such as transparency, making these systems redundant. Chapter 3 focuses on developing microemulsion and surfactant-laden contact lenses that can deliver CyA for an extended period of time. A nonionic surfactant Brij 97 is used in these studies. Both surfactant and mi croemulsion-laden gels exhibit slow and extended drug release lasting for about 20 days. This is a significan t improvement compared to the control (pure pHEMA gels), which releases drug for less than 5 days at similar drug loadings. The duration of drug release depends on the surf actant loading, but it is similar for both surfactant and microemulsion-laden gels for similar surfactant load ings, particularly early in the release. There

PAGE 202

202 are some differences in the release profiles at longer times and so further investigations are needed to conclusively determine the role of oil in these studies. The surfactant loadings in these studies are higher than the CMC and so the hydrated gels are e xpected to contain surfactant aggregates. Since CyA is a hydrophobic drug, it preferentially partitions into the hydrophobic domains leading to an increase in partition coeffi cient of the gel. The interaction of the drug with the hydrophobic domains and th e resistance offered by the surf actants to transport of drug from inside the domains to the bulk gel leads to a slow down in transport rates from the gel. The gels release on an average about 2 g of drug each day, which is su fficient to achieve therapeutic concentrations in cornea. As shown in chapter 3, drug release profiles ar e unaffected by the method of drug loading. The gels which had CyA loaded by soaking in solu tions behave similarly to those in which the drug is added before polymerization. It is how ever interesting to note that when the drug is loaded by soaking, the process equilibrates in le ss than 5 days but these gels release drug for more than 20 days. This also suggests that the drug has a large affinity for hydrophobic domains in the gel, which leads to a rapid uptake. The duration of drug release is unaltere d by processing conditions which include autoclaving and packaging. A dditionally, the drug is not damaged by the autoclaving. These results are very encouraging and so it seems that surfactant or microemulsion-laden gels may be suitable for delivering CyA to eyes. In addition to treating ocular disorders, CyA has also shown promise in treating contact lens mediated dry ey es, and so these systems could also be very useful for a large population that is unable to wear contact lenses due to discomfort [124]. In chapter 4 we increased our understanding of this system by proposing models for both drug and surfactant transport from p-HEMA hydrogels and the results of the model were verified

PAGE 203

203 by measuring release rates of both surfactant and drug from h ydrogels. We explored Brij 98 surfactant laden p-HEMA hydrogels in this chapter. The expe rimental results were in good agreement with the models. The transport models developed here can be very helpful in tuning the drug release rates from hydrogels by c ontrolling the surfacta nt concentration. To further verify the model developed in chap ter 4, we studied more surfactants and drugs in chapter 5. Similar classes of surfactant were utilized to comment on qualitative behavior of surfactants based on their molecular structure which was further verified by experiments. These surfactants had unsaturated a nd saturated carbon chains as hydrophobic segments with three different lengths of hydrophilic (EO) segment. Experiments were conducted to study the effects of gel thickness, surfactant type and concentration on release prof iles. Mechanism of release of the drug from all the surfactant laden systems wa s diffusion controlled as was inferred from the study of two different thicknesses for each surfactant -laden system. We were able to determine partition coefficient of the drug for all the syst ems and quantitatively determine which system would provide maximum barrier to CyA transpor t from the hydrogels. Brij 78 systems seem to be the most promising because these systems rel ease CyA for longer periods of time compared to other Brij systems and also seem to have the least amount of surfactant diffusing out in the release medium. The partition coefficient of th e drug in these systems was determined to be 458.9.5 which shows that CyA had a very high affinity to Brij 78 aggregates present inside the hydrogel. Furthermore, Brij 78 surfactants have been used in ocular studies as cornea permeability enhancers, and so these are not expect ed to cause significant toxic response in the eyes [67]. Though Brij-78 laden gels showed promising results for the drug CyA, they could not be utilized to attenuate rel ease of two other hydrophobic drugs, DMS and DMSA. This failure

PAGE 204

204 was attributed to lower partition coefficient of th ese drugs inside Brij 78 micelles. Furthermore, these drugs are much smaller in size than CyA, leading to fa ster release from the hydrogel. Chapter 5 also explored various properties of the hydrogels to determine their suitability as commercial contact lenses. All the systems showed very high transparency values along with a storage modulus suitable for cont act lens applications. Water content for all the surfactant laden gels was much higher than that for pure pHEMA hydrogel which can further be related to increased oxygen permeability with better patient compliance. Brij 78 laden gels showed the best wetting properties as was determined by the contact angle meas urements on all the systems. While the examples reported here were conduc ted with CyA, other drugs could also be dissolved in the HEMA monomer, but the release rates may not be as slow as those for CyA, particularly if the drug molecules are much sm aller than CyA or if they are less hydrophobic in nature. It is also noted that in addition to surfactants, other self asse mbling molecules such as lipids, and block-co-polymers coul d be used to create domains that could trap and slowly release hydrophobic drugs. After clearly establishing the concept and proving that micelle a nd microemulsion laden gels could potentially be used for delivering CyA at therapeutic dosages to the ocular surface, we also evaluated the potential damage of the studi ed non-ionic surfactants on the ocular surface by designing an in vitro toxicity assa y as the data for potential ocular toxicity of these substances was scarce. This was discussed in chapter 6. Da mage to the corneal epithelium can be attributed to the disruption of membrane fluidity due to the penetration of external agents such as surfactants, and the subsequent release of lysosomal enzymes, histamine, and inflammatory mediators. Interactions between surfactants and the corneal surface are governed by their respective CMCs, as micelles are not expected to interact with lipid bilayers. We successfully

PAGE 205

205 developed an in vitro alternative to the popularly used ey e irritation assay known as Draize eye test by using liposomes mimicking corneal ep ithelium and determining the release of a hydrophilic dye from liposome interior due to interaction with surfact ants. We found that liposomes could be successfully utilized to as sess the toxicity of various surfactants if a correction factor is introduced to account fo r the increased surface area to volume ratio of liposomes compared to the corneal epithelium. Once this factor was introduced, the correlations between dye leakage from liposomes and Draize scores improved significan tly. This method can be used to evaluate the initial toxicity of various surfactants, and c ould thus become a key method to assess ocular toxicity in vitro. We predicted that Brij 78, Brij 700, Brij 56, and Brij 58 are mildly/moderately comfortable when placed in the eye at concentrations of 10% (w/v), while Brij 97 and Brij 98 appear to be ir ritating at similar concentrations. At 1% (w/v), all of the surfactants examined are most likely in the mild/moderate category, causing little to no discomfort.

PAGE 206

206 CHAPTER 9 FUTURE WORK 9.1 Gels with Higher Surfactant Loading This work has focused on maximum surfactant c oncentration of 8% w/dry gel w inside the hydrogel. It was observed that in troduction of surfactants inside the hydrogel can significantly alter the microstructure of the hydrogel and this can lead to an improvement of water uptake properties of these hydrogels. To further investig ate the effect of surfact ant concentration on the gel properties it will be interes ting to develop systems with mu ch higher surfactant loading and determine the effect of high surfactant loading on the microstructure of the gel. We have successfully synthesized gels with 40% surfactan t loading for Brij 97 and 25% surfactant loading for Brij 98 and Brij 78 systems in p-HEMA matrix. These gels remain transparent after introduction of such high surfact ant concentration which suggest s that the pores inside the hydrogels are not larger than 100 nm. We specula te that the pores insi de these hydrogels will likely form an interconnected ne twork, which could be verified by SEM imaging. Also, it may be possible to soak these gels in ethanol solution to remove all the surfactant from the gel matrix and then back-fill these gels with a silicone monomer. This will have a significant impact in terms of enhancing oxygen permeability of p-HEMA gels which can make these contact lenses suitable for extended wear. The wettability of these systems will also increase significantly due to high surfactant concentration, which may lead to better patient compliance. Surfactants have also been used as penetration enhancers through skin and high surfactant containing systems can be used to increase skin permeability in some other therapeutic applications. 9.2 Oxygen Permeability We were unable to measure the oxygen permeability of surfactant laden gels in our lab due to lack of required setup. This work is currentl y in progress in the lab. In future, it might be

PAGE 207

207 interesting to explore the effect of microstructure change due to surfactants inside the hydrogel on the oxygen permeability of these gels. The surfact ant presence will certainly lead to increased oxygen transport due to increased water content but there may be other mechanisms related to the microstructure that impact oxygen transport. The oxygen permeability is defined in terms of Dk or barrers and it is suggested that the oxygen permeability of an extended wear contact lens should be larger than 110 barrers. Pure pHEMA gels have an oxygen permeability of only 10 barrers and water has a maximum permeability of 80 barrers. This clearly suggests that some other component such as a silicone monomer, which has excessively large oxygen permeability (~600 barrers), has to be introduced along with su rfactants to increase the permeability to desired levels. Surfactants inside the hydrogels can also act as an emulsifier in maintaining a cocontinuous phase of HEMA and sili cone in the hydrogels. Thus, ther e is an extensive possibility of changing oxygen permeability by introduction of surfactants inside the hydrogels and these need to be explored in detail in future studies. 9.3 Release of Bio-active Agents li ke Vitamin E from Contact Lenses It was observed that release of surfactants from the hydrogels increases the partition coefficient of hydrophobic drugs and this can be potentially used to release highly insoluble agents such as a Vitamin E from the contact lens es, which can provide therapeutic advantages to commercially available contact lenses. Vitamin E can be loaded inside the gels using ethanol as the solvent as Vitamin E has negligib le solubility inside water. It can also be loaded in the gels by mixing Vitamin E in the HEMA monomer befo re polymerization. In some preliminary results, we have shown that solubility of Vita min E in water can be increased in presence of surfactants, and this suggests th at surfactant laden hydrogels can potentially be used to deliver Vitamin E to the ocular surface. Also, we have successfully synthesized surfactant and Vitamin E loaded gels in our lab and the gels retain tr ansparency on introduction of Vitamin E inside the

PAGE 208

208 gel matrix. More rigorous experiments need to be conducted to explore surfactant laden gels to increase delivery of such highly hydrophobic substances. 9.4 Surfactant laden Silicone Contact Lenses As seen previously, surfactants can enhan ce the surface properties of the hydrogels by increasing surface wettability and water content. Silicone contact lenses have very poor wettability and water content and it might be beneficial to study the property changes in these hydrogels on introduction of surfactants. We have done some preliminary studies on introduction of surfactants in silicone gels and the gels retained transparency if the surfactant concentration was below 4% surfactant loading insi de the hydrogel. We also tried to measure the surfactant release from these silicone gels but realized that surface tension measurements are not suitable for measuring surfact ant release from these gels as the control silicone gels should significant surface activity due to un-reacted monomer diffusion from these gels. Other surfactants which could be iden tified using techniques such as HPLC, UV-Vis spectrometry, Florescence, FTIR, NMR or refractive index measur ements, need to be utilized inside these silicone gels if surfactant rele ase needs to be measured. 9.5 Polymerizable Surfactants Surfactants can potentially diffuse from the c ontact lenses and penetrate the ocular surface which might lead to toxicity. Thus, it might be beneficial to utilize surfactants which can polymerize inside the hydrogel matrix, and thus elimin ating the risk of surfactant toxicity. It will also be very interesting if th e surfactants could polymerize to form nanoparticles inside the hydrogel which could help reducing the transport of hydrophobic drugs and also eliminate the toxicity of an extra excipient lik e surfactant from these hydrogels.

PAGE 209

209 9.6 In-vivo Experiments In this work we have done rigorous in vitr o analysis of surfactan t laden hydrogels for controlled drug delivery of CyA, and we have s hown that these systems could be potentially used for therapeutic use. In vivo studies still need to be performed to conclusively demonstrate the biocompatibility and efficacy of these systems fo r controlled delivery vehicle with increased bioavailability.

PAGE 210

210 LIST OF REFERENCES [1] Le Bourlais C, Acar L, Zia H, Sado PA Needham T, Leverge R. Ophthalmic drug delivery systems Recent advances. Prog Retin Eye Res 1998;17(1):33-58. [2] Prausnitz MR, Noonan JS. Permeability of corn ea, sclera and conjunctiva: A literature analysis for drug delivery to th e eye. J Pharm Sci 1998;87(12):1479-1488. [3] Snibson GR, Greaves JL, Soper NDW, Pryda ls JI, Wilson CG, Bron AJ. Precorneal residence time of sodium hyaluronate so lutions studied by quantitative gamma scintigraphy. Eye 1990;4(4):594-602. [4] Wilson CG, Zhu YP, Frier M, Rao LS, Gilchris t P, Perkins AC. Ocular contact time of a carbomer gel (GelTears) in humans. Brit J Ophthalmol 1998;82(10):1131-1134. [5] Meadows DL, Paugh JR, Joshi A, Mordaunt J. A novel method to evaluate residence time in humans using a nonpenetrating fluorescen t tracer. Invest Ophthalmol Vis Sci 2002;43(4):1032-1039. [6] Ishibashi T, Yokoi N, Bron AJ, Tiffany JM, Komuro A, Kinoshita S. Retention of reversibly thermo-gelling timolol on the human ocular surface studied by video meniscometry. Curr Eye Res 2003;27(2):117-122. [7] Yokoi N, Komuro A. Non-inva sive methods of assessing the tear film. Exp Eye Res 2004;78(3):399-407. [8] Lang JC. Ocular drug-delivery conventional ocular formulations. Adv Drug Deliver Rev 1995;16(1):39-43. [9] Felt O, Einmahl S, Gurny R, Furrer P, B aeyens V. Polymeric systems for ophthalmic drug delivery. In: Dumitriu S, editor. Polymeric Biomaterials, New York, Marcel Dekker Inc., 2002. p. 377-421. [10] Sintzel MB, Bernatchez SF, Tabatabay C, Gurny R. Biomaterials in ophthalmic drug delivery. Eur J Pharm Biopharm 1996;42(6):358-374. [11] Ludwig A. The use of mucoadhesive polymer s in ocular drug delivery. Adv Drug Deliver Rev 2005;57(11):1595-1639. [12] Calvo P, Vila-Jato JL, Alonso MJ. Comparativ e in vitro evaluation of several colloidal systems,nanoparticles, nanocapsules, and nanoe mulsions as ocular drug carriers. J Pharm Sci 1996; 85(5):530-536. [13] Gavini E, Chetoni P, Cossu M, et al. PL GA microspheres for the ocular delivery of a peptide drug, vancomycin using emulsificatio n/spray-drying as the preparation method: in vitro/in vivo studies. Eur J Pharm Biopharm 2004; 57(2): 207-212.

PAGE 211

211 [14] McNamara NA, Polse KA, Brand RJ, Grah am AD, Chan JS, McKenney CD. Tear mixing under a soft contact lens: Effects of lens diameter. Am J Ophthalmol 1999;127(6):659-665. [15] Creech JL, Chauhan A, Radke CJ. Dispersive mixing in the posterior tear film under a soft contact lens. Ind En g Chem Res 2001;40(14):3015-3026. [16] Li CC, Chauhan A. Modeling ophthalmic drug delivery by soaked contact lenses. Ind Eng Chem Res 2006;45(10):3718-3734. [17] Lemp MA. New strategies in the treatmen t of dryeye states. Cornea 1999;18:625. [18] Murube J, Paterson A, Murube E. Classification of artificial tears. I: Composition and properties. Adv Exp Med Biol 1998;438:693. [19] Pflugfelder SC. Advances in the diagnosis a nd management of keratoconjunctivitis sicca. Curr Opin Ophthalmol 1998;9(4):50. [20] Murube J, Murube E: Treatme nt of dry eye by blocking the lacrimal canaliculi. Surv Ophthalmol 1996;40(6):463. [21] Pflugfelder SC, Jones D, Ji Z, Afonso A, MOnroy D. Altered cytokine balance in the tear fluid and conjunctiva of patients with Sjogr ens syndrome keratoconjunctivitis sicca. Curr Eye Res 1999;19(3):201. [22] Yen MT, Monroy D, Pflugfelder SC. Punctal occlusion decreases tear production, clearance, and ocular surface sensation. Invest Ophthalmol Vis Sci 1999;40:S980. [23] Mathers WD, Dolney AM. Objective demonstr ation of tear stimulation with oral pilocarpine in dry eye patients. Inve st Ophthalmol Vis Sci 2000;41:S60. [24] Calonge M. The Treatment of Dry Eye. Surv Ophthalmol 2001;45(2):S227-S239. [25] Walton RC, Nussenblatt RB, Whitcup SM. Cy closporine therapy for severe sightthreatening uveitis in children and adolescents. Ophthalmology 1998;105(11):2028-2034. [26] Gupta V, Sahu PK. Topical cyclospori n A in the management of vernal keratoconjunctivitis. Eye 2001;15(1):39-41. [27] Georganas C, Ioakimidis D, Iatrou C, Vidalaki B, Iliadou K, Athanassiou P, Kontomerkos T. Relapsing Wegeners gran ulomatosis: successful treatment with cyclosporin-A. Clin Rheu matol 1996;15(2):189. [28] Wiederholt M, Kossendrup D, Schulz W, Ho ffman F. Pharmacokinetic of topical Cyclosporin A in rabbit eye. Invest Ophthalmol Vis Sci 1986;27(4):519-524. [29] BenEzra D, Maftzir G, Courten C, Timonen P. Ocular penetration of cyclosporin A. III: the human eye. Brit J Ophthalmol 1990;74(6):350-352.

PAGE 212

212 [30] Acheampong AA., Shackleton M, Tang-Liu D, Di ng S, Stern M, Decker R. Distribution of cyclosporin A in ocular tissues after topi cal administration to al bino rabbits and beagle dogs. Curr Eye Res 1999;18(2):91. [31] Klang S, Abdulrazik M, Benita S. Influe nce of emulsion droplet surface charge on indomethacin ocular tissue distributi on. Pharm Dev Technol 2000;5(4):521. [32] Kanai A, Alba RM, Takano T, Kabayashi C, Nakajima A, Kurihara K, Yokoyama T, Fukami M. The effect on the cornea of al pha cyclodextrin vehicl e for cyclosporin eye drops. Transplant P 1989;21(1):3150-3152. [33] Bijl P, Eyk AD, Meyer D. Effects of thr ee penetration enhancers on transcorneal permeation of cyclosporine. Cornea 2001;20(5):505-508. [34] Kuwano M, Ibuki H, Morikawa N, Ota A, Kawashima Y. Cyclosporine A formulation affects its ocular distribution in rabbits. Pharm Res 2002;19(1):108-111. [35] Milani JK, Pleyer U, Dukes A, Chou HJ, Lutz S, Ruckert D, Schmidt KH, Mondino BJ. Prolongation of corneal allograft survival wi th liposome encapsulated cyclosporine in the rat eye. Ophthalmology 1993;100(6):890-896. [36] Felt O, Furrer P, Mayer JM, Plazonnet B, Buri P, Gurny R. Topical use of chitosan in ophthalmology: tolerance assessment and evalua tion of precorneal retention. Int J Pharm 1999;180(2):185-193. [37] De Campos AM, Sanchez A, Alonso MJ. Chito san nanoparticles: a new vehicle for the improvement of the delivery of drugs to the ocular surface. Applica tion to cyclosporin A. Int J Pharm 2001;224(1-2):159-168. [38] El-Shabouri MH. Positively charged nanoparticle s for improving the oral bioavailability of cyclosporin-A. Int J Pharm 2002;249(1-2):101-108. [39] Reidy JJ, Gebhardt BM, Kaufman HE. The colla gen shield: a new vehicle for delivery of cyclosporin A to the eye. Cornea 1990;9(3):196-199. [40] Lallemand F, Furrer P, Felt-Baeyens O, GexFabry M, Dumont JM, Besseghir K, Gurny R. A novel water-soluble cyclosporine A prodrug : Ocular tolerance and in vivo kinetics. Int J of Pharm 2005;295(1-2):7-14. [41] Gallardo A, Fernandez F, Bermejo P, Rebuelta M, Cifuentes A, Diez-Masa JC, Roman J. Controlled release of cyclosporine from VP -HEMA copolymer systems of adjustable resorption monitorized by MEKC. Biomaterials 2000;21(9):915-921. [42] Gallardo A, Fernandez F, Cifuentes A, DiezMasa JC, Bermejo P, Rebuelta M, Roman J, Lopez-Bravo A.. Modulated release of cy closporine from soluble vinyl pyrrolidonehydroxyethyl methacrylate copolymer hydrogels A correlation of 'in vitro' and 'in vivo' experiments. J Control Release 2001;72(1-3):1-11.

PAGE 213

213 [43] Sedlavek J. Possibilities of application of ophthalmic drugs with the aid of gel contact lens. Cesk Oftalmol 1965;21:509-514. [44] Jain MR. Drug delivery through soft contac t lenses. Brit J Ophthalmol 1988;72(2):150154. [45] Karlgard CCS, Wong NS, Jones LW, Moresoli C. In vitro uptake and release studies of ocular pharmaceutical agents by silicon-containing and p-HEMA hydrogel contact lens materials. Int J Pharm 2003;257(1-2):141-151. [46] Alvarez-Lorenzo C, Hiratani H, Gomez-Amoza JL, Martinez-Pacheco R, Souto C, Concheiro A. Soft contact lenses capable of sustained delivery of timolol. J Pharm Sci 2002;91(10):2182-2192. [47] Hsiue GH, Guu JA, Cheng CC. Poly(2-hydr oxyethyl methacrylate) film as a drug delivery system for pilocarpine. Biomaterials 2001;22(13):1763-1769. [48] Tian X, Iwatsu M, Sado K, Kanai A. Studies on the uptake and release of fluoroquinolones by disposable contact lenses. CLAO J 2001;27(4):216-220. [49] Hillman JS. Management of acute glaucoma with pilocarpine-soaked hydrophilic lens. Brit J Ophthalmol 1974;58(7):674-679. [50] Ramer RM, Gasset AR. Ocular penetration of pilocarpine effect on concentration on ocular penetration of pilocarpine. Ann Ophthalmol 1974;6(11):1160-1162. [51] Ruben M, Watkins R. Pilocarpine dispensatio n for soft hydrophilic contact-lens. Brit J Ophthalmol 1975;59(8):455-458. [52] Rosenwald PL. Ocular device. US Patent No. 4,484,922, 1981. [53] Arthur BW, Hay GJ, Wasan SM, Willis WE. U ltrastructural effects of topical timolol on the rabbit cornea outcome alone and in conj unction with a gas permeable contact lens. Arch Ophthalmol 1983;101(10):1607-1610. [54] Wilson MC, Shields MB. A comparison of the cl inical variations of the iridocorneal endothelial syndrome. Arch Ophthalmol 1989;107(10):1465-1468. [55] Schultz CL, Nunez IM, Silor DL, Neil ML. Cont act lens containing a leachable absorbed material. US Patent No. 5,723,131, 1995. [56] Fristrom B. A 6-month, randomized, doublemasked comparison of latanoprost with timolol in patients with open angle glaucoma or ocular hypertension. Acta Ophthalmol Scan 1996;74(2):140-144. [57] Schultz CL, Mint JM. Drug delivery system fo r antiglaucomatous medication. US Patent No. 6,410,045, 2000.

PAGE 214

214 [58] Hiratani H, Alvarez-Lorenzo C. The nature of backbone monomers determines the performance of imprinted soft contact le nses as timolol drug delivery systems. Biomaterials 2004;25(6):1105-1113. [59] Hiratani H, Fujiwara A, Tamiya Y, Mizutani Y, Alvarez-Lorenzo C. Ocular release of timolol from molecularly imprinted soft c ontact lenses. Biomater ials 2005;26(11):12931298. [60] Ende MTA, Peppas NA. Transport of ionizab le drugs and proteins in crosslinked poly(acrylic acid) and poly(acrylic acid-co-2-hydroxyethyl methacrylate) hydrogels .2. Diffusion and release studies. J Control Release 1997;48(1):47-56. [61] Colombo P, Bettini R, Peppas NA. Observatio n of swelling process and diffusion front position during swelling in hydroxypropyl methyl cellulose (HPMC) matrices containing a soluble drug. J Control Release 1999;61(1-2):83-91. [62] Ward JH, Peppas NA. Preparation of cont rolled release systems by free-radical UV polymerizations in the presence of a drug. J Control Release 2001;71(2):183-192. [63] Park ES, Chang SY, Hahn M., Chi SC. Enhanc ing effect of polyoxyeth ylene alkyl ethers on the skin permeation of ibuprofen. Int J Pharm 2000;209(1-2):109-119. [64] Aungst BJ, Rogers NJ, Shefter E. Enhancem ent of naloxane penetration through human skin in vitro using fatty acids, fatty alc ohols, surfactans, sulfoxides and amides. Int J Pharm 1986;33(1-3):225-234. [65] Walters KA, Walker M, Olejni k O. Non-ionic surfactant ef fects on hairless mouse skin permeability characteristics. J Pharm Pharmacol 1988;40(8):525-529. [66] Sarpotdar P, Zatz JL. Percutaneous abso rption enhancement by nonionic surfactants. Drug Dev Ind Pharm 1988;12:1625-1647. [67] Saettone MF, Chetoni P, Cerbai R, Mazzan ti G, Braghiroli L. Evaluation of ocular permeation enhancers: in vitro eff ects on corneal transport of four -blockers, and in vitro/in vivo toxic activity. Int J Pharm 1996;142(1):103-113. [68] Furrer P, Mayer JM, Plazonnet B, Gurny R. Oc ular tolerance of ab sorption enhancers in ophthalmic preparations. AA PS PharmSci 2002;4(1):1-5. [69] Chetoni P, Burgalassi S, Monti D, Saetto ne MF. Ocular toxicity of some corneal penetration enhancers evaluated by electr ophysiology measurements on isolated rabbit cornea. Toxicol In Vitro 2003;17(4):497-504. [70] Kim Y, Dalhaimer P, Christian DA, Disc her DE. Polymeric worm micelles as nanocarriers for drug delivery. Nanotechnology 2005;16:S484-S491. [71] Jones MC, Leroux JC. Polymeric micelles a new generation of colloidal drug carriers. Eur J Pharm Biopharm 1999;48(2):101-111.

PAGE 215

215 [72] Hagan SA, Coombes GA, Garnett MC, Dunn SE, Davies MC, Illum L, Davis SS, Harding SE, Purkiss S, Gilbert PR. Polylac tide-poly(ethylene glycol) copolymers as drug delivery systems, 1.Characterization of wa ter dispersible micelle-forming systems. Langmuir 1996;12(9):2153-2161. [73] Torchilin VP. Structure and design of polymer ic surfactant-based drug delivery systems. J Control Release 2001 ;73(2-3):137-172. [74] Torchilin VP. Micellar nanocarriers: Ph aramceutical perspectives. Pharm Res 2007;24(1):1-15. [75] Vlachou M, Hani N, Efentakis M. Polymers fo r use in controlled release systems: The effect of surfactants on th eir swelling properties. J Bi omater Appl 2000;15(1):65-77. [76] Hansson P, Schneider S, Lindman B. Macros copic phase separation in a polyelectrolyte gel interacting with opposite ly charged surfactant: corre lation between anomalous deswelling and microstructure. Prog Coll Pol Sci 2000;115:342-346. [77] Philippova OE, Hourdet D, Audebert R, Khokhlov AR. Interaction of hydrophobically modified poly(acrylic acid) hydrogels w ith ionic surfactants. Macromolecules 1996;29(8):2822-2830. [78] Nilsson S, Thuresson K, Hansson P, Lindman Bjorn. Mixed solutions of surfactant and hydrophobically modified polymer Controlli ng viscosity and micellar size. J Phys Chem B 1998;102(37):7099-7105. [79] Barreiro-Iglesias R, Alvarez-Lorenzo C, C oncheiro A. Microcalorimetric evidence and rheological consequences of the salt effect on carbopol-surfa ctant interactions. Prog Coll Pol Sci 2003;122:95-102. [80] Peppas NA, Bures P, Leobandung W, Ichikawa H. Hydrogels in pharmaceutical formulations. Eur J Pharm and Biopharm 2000;50(1):27-46. [81] Brannon-Peppas L, Peppas NA. Equilibrium swelling behavior of pH-sensitive hydrogels. Chem Eng Sci 1991;46:715-722. [82] Beltran S, Baker JP, Hooper HH, Blanch HW, Prausnitz JM. Swelling equilibria for weakly ionizable, temperat ure sensitive hydrogels. M acromolecules 1991;24:549-551. [83] Zrinyi M, Szabo D, Kilian HG. Kinetics of the shape change of ma gnetic field sensitive polymer gels. Polym Gels Netw 1998;6(6):441-454. [84] Lavon I, Kost J. Mass transport enhancement by ultrasound in non-degradable polymeric controlled release systems. J Control Release 1998;54(1):1-7. [85] Goldbart R, Kost J. Calcium responsive bi oerodible drug delivery system. Pharm Res 1999;16(9):1483-1486.

PAGE 216

216 [86] Paulsson M, Edsman K. Controlled drug releas e from gels using lipophi lic interactions of charged substances with surf actants and polymers. J Collo id Interf Sci 2002;248(1):194200. [87] Ritger PL, Peppas NA. A simple equation for de scription of solute re lease I. Fickian and non-Fickian release from non-swella ble devices in the form of slabs, spheres, cylinders or discs. J Control Release 1987;5(1):23-36. [88] Amsden B. Solute diffusion within hydroge ls-mechanisms and models. Macromolecules 1998;31(23):8382-8395. [89] Higuchi T. Rate of release of medicaments from ointment bases containing drugs in suspension. J Pharm Res 1961;50(10):874-876. [90] Paulsson M, Edsman K. Controlled drug releas e from gels using surfactant aggregates: I. Effect of lipophilic interactions for a seri es of uncharged substances. J Pharm Sci 2001;90(9):1216-1225. [91] Paulsson M, Edsman K. Controlled drug releas e from gels using surfactant aggregates. II. Vesicles formed from mixtures of amphiphilic drugs and oppositely charged surfactants. Pharm Res 2001;18(11):1586-1592. [92] Bramer T, Bew N, Edsman K. Cationic mi xtures involving a drug: a rather general concept that can be utilized for prolonge d drug release from the gels. J Pharm Sci 2006;95(4):769-780. [93] Lin HR, Sung KC. Carbopol/pluronic phase change solutions for ophthalmic drug delivery. J Control Release 2000;69(3):379-388. [94] Liu JH, Lin L. SDS-aided immobilization and controlled release of camptothecin from agarose hydrogel. Eur J Ph arm Sci 2005;25(2-3):237-244. [95] Liu JH, Lin L, Cai YY. Immobilization of camptothecin with surfactant into hydrogel for controlled drug release. Eur Polym J 2006;42(8):1767-1774. [96] Liu JH, Lin L. Diffusion of camptothecin im mobilized with cationic surfactant into agarose hydrogel containing anionic carrageenan. J Biomed Mater Res A 2007;83(4):1103-1109. [97] Barreiro-Iglesias R, Alvarez-Lorenzo C, Conc heiro A. Incorporation of small quantities of surfactants as a way to improve the rh eological and diffusiona l behavior of carbopol gels. J Control Release 2001;77(1-2):59-75. [98] Alvarez-Lorenzo C, Concheiro A. Effect s of surfactants on gel behavior: Design implications for drug delivery systems. Am J of Drug Deliver 2003;1(2):77-101. [99] Wu Z, Joo H, Lee TG, Lee K. Controlled release of lidocaine hydrochloride from the surfactant-doped hybrid xerogels. J Control Release 2005; 104(3):497-505.

PAGE 217

217 [100] Eeckman F, Moes AJ, Amighi K. Surfactant induced drug delivery based on the use of thermosensitive polymers. Journal of Controlled Release 2003;88(1):105-116. [101] Gulsen D, Chauhan A. Dispersion of microemulsion drops in HEMA hydrogel: a potential ophthalmic drug delivery ve hicle. Int J Pharm 2005;292(1-2):95-117. [102] Wilhelmus KR. The Draize eye test. Surv Ophthalmol 2001;45(6):493-515. [103] Weil CS, Scala RA. Study of Intraand Interlab oratory Variability in the results of rabbit eye and skin irritation tests. Toxicol Appl Pharm 1971;19:276-360. [104] Vian L, Vincent J, Maurin J, Fabre I, Gi roux J, Cano JP. Comparison of three in vitro cytotoxicity assays for estimating surfact ant ocular irritation. Toxic in Vitro 1995;9(2):185-190. [105] Matsukawa K, Masuda K, Kakishima H, Suzuki K, Nakagawa Y, Matsushige C, Imanishi Y, Nakamura T, Mizutani A, Watanabe R, Shingai T, Kaneko T, Hirose A, Onno Y. Interlaboratory Validation of the in vitro eye i rritation tests for cosme tic ingredients. (11) EYTEX. Toxic in Vitro 1999;13:209-217. [106] Vinardell MP, Mitjans M. Alternative methods for eye and skin irritation tests: An Overview J Pharm Sci 2008;97(1):46-59. [107] Okamoto Y, Ohkoshi K, Itagaki H, et al. In terlaboratory Validation of the in vitro eye irritation tests for cosmetic ingredients. (3) Evaluation of the Haemol ysis Test. Toxic in Vitro 1999;13:115-124. [108] Okahata Y, Ebato H. Absorption behaviors of surfactant molecules on a lipid-coated quartz-crystal microbalance. An Alternativ e to Eye-Irritant Tests. Anal Chem 1991;63:203-207. [109] Cottin M, Zanvit A. Fluorescein leakage lest: a useful tool in ocular safety assessment. Toxic in Vitro 1997;11:399-405. [110] Curren RD, Harbell JW. In Vitro Alternatives for ocular irritation. Environ Health Persp 1998;106(S2):485-492. [111] Kato S, Itagaki H, Chiyoda I, et al. Liposomes as an in vitro model for predicting the eye irritancy of chemicals. Toxic in Vitro 1988;2(2):125-130. [112] Sunamoto J, Iwamoto K, Imokawa G, Tsuchi ya S. Possible eye-irritant test using polysaccharide-coated liposomes as a corn eal epithelium model. Chem Pharm Bull 1987;35(7):2958-2965. [113] Saarinen-Savolainen P, Jarvinen T, S uhonen P, Urtti A. Amphiphilic properties of pilocarpine prodrugs. In t J Pharm 1996:133:171-178.

PAGE 218

218 [114] Kim C, Ryuu S, Park K, Lim S, Hwang S. Preparation and physicochemical characterization of phase inverted water/oil microemulsion containing cyclosporin A. Int J Pharm 1997;147(1):131-134. [115] Ran Y, Zhao L, Xu Q, Yalkowsky SH. Solubilization of Cyclosporin A. AAPS PharmSciTech 2001;2(1):23-26. [116] Lum E, Perera I, Ho A. Osmolality and buffe ring agents in soft contact lens packaging solution. CLAE 2004;27(1):21-26. [117] Warisnoicharoen W, Lansley AB, Lawrence MJ Nonionic oil-in-water microemulsions: the effect of oil type on phase behaviour. Int J Pharm 2004;198(1):7. [118] Morey TE, Varshney M, Flint JA, Seubert CN, Smith WB, Bjoraker DG, Shah DO, Dennis DM. Activity of microemulsion-ba sed nanoparticles at the human bio-nano interface: concentration-dependa nt effects on thrombosis and hemolysis in whole blood. J Nanopart Res 2004;6:159-170. [119] Klammt C, Schwarz D, Fendler K, Haase W, Dotsch V, Bernhard F. Evaluation of detergents for the soluble e xpression of alpha-helical a nd beta-barrel-type integral membrane proteins by a prepar ative scale individual cell-free expression system. Febs J 2005;272(23):6024-6038. [120] Hait SK, Moulik SP. Determination of criti cal micelle concentration (CMC) of nonionic surfactants by donor-acceptor interaction w ith iodine and correlation of CMC with hydrophile-lipophile balance and other parameters of the surfactants. J Surfactants Deterg 2001;4(3):303-309. [121] Warisnoicharoen W, Lansley AB, Lawrence MJ. Toxicological evaluation of mixtures of nonionic surfactants, alone and in combination with oil. J Pharm Sci 2003;92(4):859-868. [122] Agatonovic-Kustrin S, Glass BD, Wisch MH Alany RG. Prediction of a stable microemulsion formulation for the oral delivery of a combination of antitubercular drugs using ANN methodology. Phar m Res 2003;20(11):1760-1765. [123] Lee Y, Simamor P, Yalkowsky SH. Effect of Brij-78 on systemic delivery of insulin from an ocular device. J Pharm Sci 1997; 86(4):430-433. [124] Hom HH. Use of Cyclosporine 0.05% ophthalm ic emulsion for contact lens-intolerant patients. Eye Contact Lens 2006;32(2):109-111. [125] Kim J, Chauhan A. Dexamethasone transport and ocular delivery from poly(hydroxyethyl methacrylate) gels. Int J Pharm 2008;353(1-2):205-222. [126] Lord MS, Stenzel MH, Simmons A, Milthrope BK. Lysozyme interaction with poly(HEMA)-based hydrogel. Biom aterials 2006;27(8):1341-1345.

PAGE 219

219 [127] Tonge S, Jones L, Godall S, Tighe B. The ex vivo wettability of soft contact lenses. Curr Eye Res 2001;23(1):51-59. [128] Ketelson HA, Meadows DL, Stone RP. Dynamic wettability properties of a soft contact lens hydrogel. Colloid Surface B 2005;40(1):1-9. [129] Szoka F, Papahadjopoulos D. Comparative properties and methods of preparation of lipid vesicles (Liposomes). Ann Rev Biophys Bioeng 1980;9:467-508. [130] Zuidam NJ, Vrueh R, Crommelin DJA. Characterization of Liposomes. In: Torchilin VP, Weissig V. Liposomes. 2nd Editi on New York, Oxford University Press, 2003. p. 31-78. [131] Kennah HE, Hignet S, Laux PE, Dorko JD, Barrow CS. An objective procedure for quantitating eye irritation based upon ch anges of corneal thickness. Fundam Appl Toxicol 1989;12(2):258-268. [132] Tachon P, Cotovio J, Dossou KG, Prunieras M. Assessment of surf actant cytotoxicity: comparison with the Draize eye test. In t J Cosmetic Sci 1989;11(5):233-243. [133] Jester JV, Li HF, Petroll WM, Parker RD, Cavanagh HD, Carr GJ, Smith B, Maurer JK. Area and depth of surfactant-induced corneal in jury correlates with cell death. Invest Ophth Vis Sci 1998;39(6):922-936. [134] Hall-Manning TJ, Holland GH, Rennie G, REvell R, Hines J, Barratt MD, Basketter DA. Skin Irritation Potential of Mixed Su rfactant Systems. Food Chem Toxicol 1998;36(3):233-238. [135] Howell B, Chauhan A. Amitriptyline overdose treatment by pegylated anionic liposomes. J Colloid Interf Sci 2008;324(1-2):61-70. [136] Rosen MJ. Surfactants and interfacial phe nomena. Hoboken NJ, Wiley-Interscience. 2004. p. 105-177. [137] Kim JH, Kim MW. Temperature effect on the transport dynamics of a small molecule through a liposome bilayer. Eu r Phys J E 2007;23(3):313-317. [138] Lis LJ, McAlister M, Fuller N, Rand RP, Parsegian VA. Interactions between neutral phospholipid bilayer membranes. Biophys J 1982;37(3):657-666. [139] New RRC. Liposomes. New Yo rk, NY, Oxford University Press. 1990. p. 137. [140] Balgavy P, Dubnickova M, Kucerka N, Kisele v MA, Yaradaikin SP, Uhrikova D. Bilayer thickness and lipid interface area in unilamellar extruded 1,2-diacylphosphatidylcholine liposomes: a small-angle neutron scattering study. Biochim Biophys Acta 2001;1512(1):40-52.

PAGE 220

220 [141] Maric K, Wiesner B, Lorenz D, Klussma nn E, Betz T, Rosenthal W. Cell Volume kinetics of adherent epithe lial cells measured by laser s canning reflection microscopy: Determination of water permeability changes of renal principal cells. Biophys J 2001;80(4):1783-1790. [142] Farinas J, Verkman AS. Cell volume and plasma membrane osmotic water permeability in epithelial cell layers measured by interferometry. Biophys J 1996;71:3511-3522. [143] Brinkman HC. A calculation of the viscous force exerted by a flow ing fluid on a dense swarm of particles. Appl Sci Res 1947;A(1):27-34. [144] Phillips RJ, Deen WM, Brady JF. Hindered tr ansport of spherica l macromolecules in fibrous membranes and gels AICHE J 1989;35(11):1761-1769. [145] Solomentsev YE, Anderson JL. Rotation of a sp here in Brinkman fluids. Phys Fluids 1996;8(4):1119-1121.

PAGE 221

221 BIOGRAPHICAL SKETCH Yash Kapoor was born in1981 in Lucknow, U ttar Pradesh, India. He did his undergradu ate work in chemical engineering at the Indian Institute of Technology Kanpur and graduated in May 2004. Thereafter, he joined the Chemical Engineering Department at the University of Florida, Gainesville, Florida as a doctorate candidate. He joined Dr. Anuj Chauhans group in January 2005 and has been pursuing his research on ophthalmic drug delivery since then.