<%BANNER%>

Ophthalmic Drug Delivery by Soft Contact Lenses

Permanent Link: http://ufdc.ufl.edu/UFE0021964/00001

Material Information

Title: Ophthalmic Drug Delivery by Soft Contact Lenses
Physical Description: 1 online resource (171 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: angle, barrier, bulk, coefficient, contact, contacts, content, controlled, delivery, dexamethasone, diffusion, diffusivity, dma, drug, equilibrium, extended, eye, hema, hydrogel, ion, lens, loading, ocular, ophthalmic, partition, permeability, release, silicone, surface, timolol, transport, tris, uptake, vitamin
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 contact lenses is known to significantly increase the bioavailability of ophthalmic drugs compared to that via eye drops. Our study focused on investigating the feasibility of delivery ophthalmic drug for extended period of time by contact lenses. To have the fundamental insight of drug transport in hydrogel contact lenses, we explored ocular delivery of dexamethasone via poly(hydroxylethyl methacrylate) contact lenses, which are common daily disposable contact lenses. Three derivatives of dexamethasone (dexamethasone 21- disodium phosphate (DXP), dexamethasone (DX), and dexamethasone 21-acetate (DXA)) are explored. These drugs are loaded in the gels by soaking in aqueous or ethanol solutions, and also by direct addition of the drug to the polymerizing mixture. Dynamic drug concentrations in the aqueous phase are monitored both in loading and release experiments. The data is utilized to determine the partition coefficients and the mean diffusivity, which includes contributions from both bulk and surface diffusion. Finally we utilize the transport model to predict the bioavailability of the three forms of dexamethasone for drug delivery via contact lenses. While hydrogel contact lenses lead to higher bioavailability, these cannot be used for extended drug delivery because these cannot be worn overnight. Silicone hydrogel contact lenses can be worn for extended periods but commercial silicone hydrogel lenses are not suitable for extended drug delivery. So we developed new extended wear silicone hydrogel soft contact lenses that deliver ophthalmic drugs for an extended period of time ranging from weeks to months. Silicone hydrogels comprising of N,N-dimethylacrylamide, 3-methacryloxypropyltris (trimethylsiloxy) silane, bis-alpha,omega-(methacryloxypropyl) polydimethylsiloxane, 1-vinyl-2-pyrrolidone, and ethylene glycol dimethacrylate were prepared with varying ratios of monomers and transport of three different ophthalmic drugs, timolol, dexamethasone, and dexamethasone 21-acetate was explored. All the silicone hydrogels of 0.1 mm thickness exhibit diffusion limited transport and extended release varying 20 days up to more than three months depending on the compositions of hydrophobic and hydrophilic components of silicone hydrogels. Also, there are multiple time scales in transport of at least certain molecules, which is perhaps due to the complex microstructure of these gels. The mechanical and physical properties of lenses such as ion permeability, equilibrium water content, transparency, and surface contact angles of some of the gels are suitable for contact lens application. We also focused on the concept of increasing the release duration of drugs from commercial silicone contact lenses by creation of transport barriers. As diffusion barriers, we loaded the vitamin E, which is considered as an important ocular nutraceutical, into five different commercially available silicone contact lenses and transport of dexamethasone and timolol were measured. The results conclusively show that vitamin E loading can substantially increase the release duration of drugs, particularly for hydrophilic drugs. The effect of vitamin E loading on the release rates is qualitatively similar for ACUVUE OASYS, NIGHT & DAY, and O2OPTIX. The effect of vitamin E loading in dynamics of both DX and timolol is minimal for PureVision.
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.
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 2009-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021964:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021964/00001

Material Information

Title: Ophthalmic Drug Delivery by Soft Contact Lenses
Physical Description: 1 online resource (171 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: angle, barrier, bulk, coefficient, contact, contacts, content, controlled, delivery, dexamethasone, diffusion, diffusivity, dma, drug, equilibrium, extended, eye, hema, hydrogel, ion, lens, loading, ocular, ophthalmic, partition, permeability, release, silicone, surface, timolol, transport, tris, uptake, vitamin
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 contact lenses is known to significantly increase the bioavailability of ophthalmic drugs compared to that via eye drops. Our study focused on investigating the feasibility of delivery ophthalmic drug for extended period of time by contact lenses. To have the fundamental insight of drug transport in hydrogel contact lenses, we explored ocular delivery of dexamethasone via poly(hydroxylethyl methacrylate) contact lenses, which are common daily disposable contact lenses. Three derivatives of dexamethasone (dexamethasone 21- disodium phosphate (DXP), dexamethasone (DX), and dexamethasone 21-acetate (DXA)) are explored. These drugs are loaded in the gels by soaking in aqueous or ethanol solutions, and also by direct addition of the drug to the polymerizing mixture. Dynamic drug concentrations in the aqueous phase are monitored both in loading and release experiments. The data is utilized to determine the partition coefficients and the mean diffusivity, which includes contributions from both bulk and surface diffusion. Finally we utilize the transport model to predict the bioavailability of the three forms of dexamethasone for drug delivery via contact lenses. While hydrogel contact lenses lead to higher bioavailability, these cannot be used for extended drug delivery because these cannot be worn overnight. Silicone hydrogel contact lenses can be worn for extended periods but commercial silicone hydrogel lenses are not suitable for extended drug delivery. So we developed new extended wear silicone hydrogel soft contact lenses that deliver ophthalmic drugs for an extended period of time ranging from weeks to months. Silicone hydrogels comprising of N,N-dimethylacrylamide, 3-methacryloxypropyltris (trimethylsiloxy) silane, bis-alpha,omega-(methacryloxypropyl) polydimethylsiloxane, 1-vinyl-2-pyrrolidone, and ethylene glycol dimethacrylate were prepared with varying ratios of monomers and transport of three different ophthalmic drugs, timolol, dexamethasone, and dexamethasone 21-acetate was explored. All the silicone hydrogels of 0.1 mm thickness exhibit diffusion limited transport and extended release varying 20 days up to more than three months depending on the compositions of hydrophobic and hydrophilic components of silicone hydrogels. Also, there are multiple time scales in transport of at least certain molecules, which is perhaps due to the complex microstructure of these gels. The mechanical and physical properties of lenses such as ion permeability, equilibrium water content, transparency, and surface contact angles of some of the gels are suitable for contact lens application. We also focused on the concept of increasing the release duration of drugs from commercial silicone contact lenses by creation of transport barriers. As diffusion barriers, we loaded the vitamin E, which is considered as an important ocular nutraceutical, into five different commercially available silicone contact lenses and transport of dexamethasone and timolol were measured. The results conclusively show that vitamin E loading can substantially increase the release duration of drugs, particularly for hydrophilic drugs. The effect of vitamin E loading on the release rates is qualitatively similar for ACUVUE OASYS, NIGHT & DAY, and O2OPTIX. The effect of vitamin E loading in dynamics of both DX and timolol is minimal for PureVision.
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.
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 2009-05-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021964:00001


This item has the following downloads:


Full Text

PAGE 1

1 OPHTHALMIC DRUG DELIVERY BY SOFT CONTACT LENSES By JINAH KIM 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 Jinah Kim

PAGE 3

3 To my loving husband Byoung Sam and my precious daughter Dahyoung.

PAGE 4

4 ACKNOWLEDGMENTS I thank my advisor and supervisory committee chair, Dr. Anuj Cha uhan, for his continuous guidance and encouragement on my research as well as great understandings and support on my life. He has been willing to give me lots of heartwarming advice professionally and also personally. He is the best advisor that I have ever had. I also thank my committee members, Dr. Spyros A. Svoronos, Dr. Kirk J. Ziegler, and Dr. Gregory S. Schultz for their time and support for this work. I would like to thank all our group members: Yash Kapoor who yielded lots of time and research subjects to me all the time, Brett Howell, Chhavi Gupta, Hyun Jung Jung, Chen-Chun Peng, and former members, Dr. Chi-Chung Li, Dr Heng Zhu, Dr. Chen Zhi, and Dr. Marissa Fallon. Their friendship, help, and encourag ement are unforgettable. I also thank the undergraduate students: Anthony B. Conway helped me with synthe sizing transparent silicone hydrogel lenses; and Joshua R. McCarty and Andy Cohen with different experiments. In addition, I would like to gr atefully acknowledge the techni cal support for this research by James Hinnant and Dennis Vince. I also thank all the other staffs in the department of Chemical Engineering at University of Florid a. The help with characterizing hydrogels and particles by the people at Particle Engineering Research Center (P ERC) and at Major Analytical Instrument Center (MAIC) at UF is appreciated. I thank my parents and my brother who are al ways with me for their unchanging love and support with pray throughout my life. Their enorm ous love that they have shown to me is not describable. I would like to give additional thanks to my other family members for their love. Most of all, I thank my dear husband, Byoung Sa m, who totally supports me to pursue PhD. His love and understandings as well as his scarificat ion of his lots opportuniti es in his career for the

PAGE 5

5 support are invaluable. I am also indebted to my precious daughter, Dahyoung. I am thankful for her existence. Finally, I am grateful to God (who is alwa ys with me) for His sincere love and for guarding me.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES................................................................................................................ .......10 LIST OF ABBREVIATIONS........................................................................................................14 ABSTRACT....................................................................................................................... ............15 CHAPTER 1 INTRODUCTION................................................................................................................. .17 2 DEXAMETHASONE TRANSPORT AND OCULAR DELIVERY FROM POLY(HYDROXYETHYL METHACRYLATE) GELS.....................................................22 2.1 Introduction.............................................................................................................. ........22 2.2 Materials and Methods....................................................................................................2 2 2.2.1 Materials............................................................................................................... .22 2.2.2 Synthesis of Poly(Hydroxyethyl Methacrylate) (PHEMA) Gels.........................23 2.2.3 Drug Loading.........................................................................................................23 2.2.4 Drug Release Experiments....................................................................................24 2.2.5 Effect of Gel Thickness on Loading and Release of Drug....................................25 2.2.6 Effect of Ionic Strength in Releas e Solution on Loading and Release of DXP....25 2.2.7 Conversion of UV-VIS Absorbance to the Corresponding Concentration of Drug........................................................................................................................... ..25 2.2.8 Determinations of Critical Mice lle Concentration (CMC) of DXP......................26 2.3 Results and Discussion.................................................................................................... 26 2.3.1 Dexamethasone (DX) Loaded Gel........................................................................26 2.3.1.1 Loading (by soaking in aqueous solution) and release studies of DX........26 2.3.1.2 Release of DX loaded in the gel by direct entrapment................................31 2.3.1.3 Effect of gel thickness on DX loading........................................................33 2.3.1.4 Effect of crosslink density of gel on DX loading........................................34 2.3.1.5 Release of DX into PBS..............................................................................34 2.3.2 Dexamethasone 21-acetate (DXA) Loaded Gel....................................................35 2.3.2.1 Release of DXA loaded by direct entrapment.............................................35 2.3.2.2 Release of DXA loaded by DXA-ethanol presoaking................................36 2.3.3 Dexamethasone 21-disodium phosphate (DXP) Loaded Gel................................37 2.3.3.1 Partition coefficient and dynamics of DXP loading and release.................37 2.3.3.2 Effect of gel thickness on DXP loading......................................................39 2.3.3.3 Effect of ionic strength of outer solution on DXP loading and release.......39

PAGE 7

7 2.3.4 Mathematical Model for Drug Transport to the Cornea........................................42 2.3.4.1 Concentration profiles in the POLTF..........................................................42 2.3.4.2 Fraction of drug that enters cornea..............................................................43 2.4 Conclusions................................................................................................................ .......44 3 EXTENDED DELIVERY OF OPHTHALM IC DRUGS BY SILICONE HYDROGEL CONTACT LENSES..............................................................................................................70 3.1 Introduction.............................................................................................................. ........70 3.2 Materials and Methods....................................................................................................7 0 3.2.1 Materials............................................................................................................... .70 3.2.2 Preparation of Silicone Hydrogels.........................................................................71 3.2.3 Drug Loading.........................................................................................................72 3.2.4 Drug Release Experiments....................................................................................73 3.2.5 Packaging Tests.....................................................................................................73 3.2.6 Gel Characterization..............................................................................................73 3.2.6.1 Dynamic mechanical analysis.....................................................................73 3.2.6.2 Ion permeability measurements..................................................................74 3.2.6.3 Surface contact angle measurements...........................................................74 3.2.6.4 Transmittance measurements......................................................................74 3.3 Results and Discussion.................................................................................................... 75 3.3.1 Timolol Loaded Gel..............................................................................................75 3.3.1.2 Loading and release of timolol by the GELS 1 and 4 soaked in PBS solution.................................................................................................................75 3.3.1.2 Release of timolol from GELS 1-6 after loading by soaking in drugethanol solution.....................................................................................................76 3.3.1.3 Rate limiting mechanism for drug transport...............................................78 3.3.1.4 Impact of concentration in drugethanol solution on drug loading and release...................................................................................................................81 3.3.2 DX or DXA Loaded Gel........................................................................................82 3.3.2.1 Loading and release of DX by th e gel soaked in PBS solution...................82 3.3.2.2 Release of DX(or DXA) loaded in the gel by soaking in drug-ethanol solution.................................................................................................................83 3.3.3 Packaging Effects..................................................................................................84 3.3.4 Mechanical Properties of Gels...............................................................................85 3.3.5 Ion Permeability, Equilibrium Water C ontent, and Surface Contact Angle of Gels........................................................................................................................... ...86 3.3.6 Transparency of Gels.............................................................................................88 3.4 Conclusions............................................................................................................... .......88 4 EXTENDED DRUG DELIVERY FROM SILICONE-HYDROGEL CONTACT LENSES CONTAINING DIFFUSION BARRIERS...........................................................105 4.1 Introduction.............................................................................................................. ......105

PAGE 8

8 4.2 Materials and Methods..................................................................................................105 4.2.1 Materials..............................................................................................................1 05 4.2.2 Drug Loading into Pure Lenses...........................................................................106 4.2.3 Vitamin E Loading into Pure Lenses...................................................................106 4.2.4 Drug Loading into Vitamin E Loaded Lenses.....................................................107 4.2.5 Drug Release Experiments..................................................................................107 4.3 Results and Discussion..................................................................................................10 8 4.3.1 Dynamics of Drug By Pure Contact Lenses........................................................108 4.3.1.1 DX loaded lenses.......................................................................................108 4.3.1.2 Timolol loaded lenses................................................................................109 4.3.2 Effect of Vitamin E Loadi ng in Water Content of Lenses..................................110 4.3.3 Effect of Vitamin E Loading in Size of Contact Lenses.....................................112 4.3.4 Effect of Vitamin E Loading on Dynamic of Drug.............................................112 4.3.4.1 DX-vitamin E loaded lenses......................................................................112 4.3.4.2 Timolol-vitamin E loaded lenses...............................................................114 4.3.4.3 Scaling theory............................................................................................117 4.3.4.4 Diffusivities of DX or timolol in vitamin E loaded lenses........................118 4.4 Conclusions............................................................................................................... .....122 5 CONCLUSIONS.................................................................................................................. 153 6 FUTURE WORK.................................................................................................................. 158 6.1 Drug Transport in PHEMA Copolymer........................................................................158 6.2 Silicone Contact Lenses for Extend Delivery of Ophthalmic Drug..............................158 6.3 Drug Transport in Commerci al Silicone Contact Lenses..............................................159 APPENDIX MODEL DESCRIPTION FOR DRUG DELIVERY BY SOAKED CONTACT LENSES..................................................................................................................162 LIST OF REFERENCES............................................................................................................. 167 BIOGRAPHICAL SKETCH.......................................................................................................171

PAGE 9

9 LIST OF TABLES Table page 2-1 Concentration and partition coefficients ( K ) of DX in a soaking method.........................65 2-2 Apparent partition coefficients ( Kapp), intrinsic partition coefficients ( K ), and permanently entrapped drug amount ( Mp) for DX release from the PHEMA gel synthesized in a direct entrapment method........................................................................66 2-3 Effect of crosslink density on partition coefficient and diffusivities of DX in the gel synthesized in a direct drug entrapping method................................................................66 2-4 Apparent partition coefficients ( Kapp), intrinsic partition coefficients ( K ), and permanently entrapped drug amount ( Mp) for DXA release from the PHEMA gel synthesized in a direct entrapment method........................................................................67 2-5 Concentration and partition coefficients ( K ) of DXP in a soaking method.......................67 2-6 Effect of ionic strength ( I ) on DXP loading and release....................................................68 2-7 Effect of ionic strength on tim olol maleate load ing and release........................................68 2-8 Various fractions ( Fc, Fs, and Fp) in the eye for three deri vatives of dexamethasone.......69 3-1 Compositions of the monomer mixtures (in mL) for various gels (0.18 mL of NVP and 15 L of EGDMA was added to each compos ition before preparing the gels)........103 3-2 Drug concentration and partition coefficients ( K ) of drug...............................................103 3-3 Diffusivities of timolol in gels.........................................................................................10 4 3-4 Physical properties of gels...............................................................................................1 04 4-9 List of silicone containing commercial c ontact lens (dipoter -6.50) used in this study. (n=6).......................................................................................................................... .......149 4-10 Changes in diameter and water content of vitamin E loaded contact lenses before and after loading vitamin E.....................................................................................................150 4-11 Products of partition coefficient of DX and lens volume ( KVlens) for lenses soaked in DX-PBS solution.............................................................................................................151 4-12 Parameters of scaling theory for timolol..........................................................................151 4-13 Ratio of diffusivities of drug in vi tamin E loaded lens to pure lens ( D / D0)....................152 A-1 Model parameters for three derivatives of dexamethasone.............................................166

PAGE 10

10 LIST OF FIGURES Figure page 2-1 Molecular structures of model drugs A) DX B) DXA C) DXP......................................48 2-2 The model geometry of the gel..........................................................................................48 2-3 Comparison of the model prediction and experimental data for DX loading into PHEMA gel soaked in drug solution.................................................................................49 2-4 Comparison of the model prediction and e xperimental data for DX release into fresh water from a PHEMA gels which have been soaked in drug solution..............................50 2-5 Comparison of the model prediction and e xperimental data for DX release into the fresh water from PHEMA gels synthesi zed in a direct entrapment method......................51 2-6 Effect of gel thickness on DX loading a nd release of a PHEMA gel soaked in drug solution....................................................................................................................... ........52 2-7 Comparison of DX release profiles in water and PBS from a PHEMA gel synthesized in a direct entrapment method........................................................................53 2-8 Comparison of the model prediction and e xperimental data for DXA release into the fresh water from PHEMA gels synthesi zed in a direct entrapment method......................54 2-9 Effect of gel thickness on DXA release of a PHEMA gel which has been soaked in DXA-ethanol solution........................................................................................................55 2-10 Comparison of DXA release profiles in water and PBS from a PHEMA which has been soaked in DXA-ethanol solution...............................................................................56 2-11 Comparison of the model prediction a nd experimental data for DXP loading by PHEMA gel soaked in DXP solution.................................................................................57 2-12 Comparison of the model prediction and e xperimental data for DXP release into the PBS from a PHEMA gels which have been soaked in drug solution................................58 2-13 Effect of gel thickness on DXP loading to a PHEMA soaked in DXP solution................59 2-14 Effect of ionic strength on DXP lo ading and release in a soaking method.......................60 2-15 Plot of diffusivities of DXP for loading and release in a soaking method as a function of ionic strength.............................................................................................................. ...61 2-16 Effect of ionic strength on timolol mal eate loading and release in a soaking method......62 2-17 Plot of diffusivities of timolol maleate for loading and re lease in a soak ing method as a function of ionic strength. Cw,i = 0.086 mg/mL...............................................................63

PAGE 11

11 2-18 Plot of surface tension of DXP as a func tion of concentration for determination of CMC............................................................................................................................ .......63 2-19 Concentration transients of DX in the PO LTF at different axial locations for case 1, i.e., no flux to the PLTF.....................................................................................................6 4 3-1 Profiles of timolol uptake and release fr om 0.1 mm thick A) GEL1 B) GEL4 in PBS solution................................................................................................................... ....91 3-2 Profiles of timolol releas e from 0.1 mm thick GELS 1-6..................................................92 3-3 Profiles of timolol release from 0.2 mm thick A) GEL7 B) GEL8..................................93 3-4 Continued.................................................................................................................. .........95 3-5 Profiles of timolol release from three diffe rent gels (0.4 mm thick) vs. scaled time.........96 3-6 Plot of mass of timolol releas ed verses square root of time..............................................96 3-7 Profiles of DX uptake by 0.1 mm thick A) GEL1 B) GEL4 in PBS solution..................97 3-8 Profiles of DX release from four diff erent composition gels (0.1 mm thick). DX was loaded in the gel by soaking in drug-ethanol solution of 4.99 mg/mL..............................98 3-9 Effect of gel thickness on DX release from (a) GEL1 (b) GEL4......................................99 3-10 Plot of mass of DX released verses square root of time (0.1mm thick gels)...................100 3-11 Profile of DXA release from a 0.4 mm thick GEL1........................................................100 3-12 Profiles of timolol release from th e GEL1 packaged for 1.3 and 2 month......................101 3-13 Profiles of DXA release from the GEL1 packaged for 1.5 and 2 month.........................101 3-14 Dependence of storage modu li of the gels on frequency.................................................102 3-15 Dependence of loss moduli of the gels on frequency......................................................102 4-1 Effect of DX loading method on profile of DX release by A) ACUVUE ADVANCE™ B) ACUVUE OASYS™ C) NIGHT&DAY™ D) O2OPTIX™ E) PureVision™ contact lenses. F) The plot of (DX release time)-1 versus water content of contact lenses.............................................................................................................. .126 4-1 Continued.................................................................................................................. .......127 4-1 Continued.................................................................................................................. .......128

PAGE 12

12 4-2 Effect of timolol loading method on pr ofile of timolol release by A) ACUVUE ADVANCE™ B) ACUVUE OASYS™ C) NIGHT&DAY™ D) O2OPTIX™ E) PureVision™ contact lenses............................................................................................129 4-2 Continued.................................................................................................................. .......130 4-2 Continued.................................................................................................................. .......131 4-3 Correlation of vitamin E loading and con centration of soaking solution for different lenses......................................................................................................................... .......131 4-4 Plot of A) water content ( Q ) B) EW of vitamin E loaded lenses versus vitamin E loading........................................................................................................................ ......132 4-5 Percent increase in diameter of A) dry lens es B) wet lenses before and after loading vitamin E. Lines are best fit strai ght lines passing zer o to the data.................................133 4-6 Profiles of DX uptake and release by vitamin E loaded contact lenses A) ACUVUE OASYS™ B) NIGHT&DAY™ C) O2OPTIX™ D) PureVision™. Solid legends represent uptake and the hollow legends release......................................134 4-6 Continued.................................................................................................................. .......135 4-7 Profiles of timolol release by vitamin E lo aded contact lenses. Timolol and vitamin E were loaded together by soaking A) ACUVUE OASYS™ B) NIGHT&DAY™ C) O2OPTIX™ D) PureVision™ c ontact lens in timolol/v itamin E-ethanol solution (0.8 mg timolol in 1 mL of vitamin E-etha nol solution of various concentrations) for 24 hours....................................................................................................................... .....136 4-7 Continued.................................................................................................................. .......137 4-8 Profiles of timolol release by vitamin E loaded contact lens es A) NIGHT&DAY™ B) O2OPTIX™.................................................................................................................138 4-9 Duration increase of A) DX loading B) DX release C) and D) timolol release by vitamin E loaded contact lenses.......................................................................................139 4-9 Continued.................................................................................................................. .......140 4-10 Profile of timolol release by TRIS loaded NIGHT&DAY™ lens...................................141 4-11 Plot of % drug release by vitamin E loaded lenses versus square root of time...............142 4-12 Slope of best fit straight line to the pl ot of A) % DX release B) % timolol release versus square root of time for short time by vitamin E loaded lenses (Figure 4-11).......146 4-13 Plot of D0 / D versus vitamin E loading for timolol release..............................................147

PAGE 13

13 4-14. Plot of % drug release by vita min loaded lenses versus (Dh0 2/D0h2)t. A) ACUVUE OASYS™ B) NIGHT& DAY™ contact lens.................................................................148 A-1 Geometry of lens utilized in the model............................................................................166

PAGE 14

14 LIST OF ABBREVIATIONS CMC Critical micelle concentration DI Deionized DMA N,N-dimethylacrylamide DX Dexamethasone DXA Dexamethasone 21-actate DXP Dexamethasone 21-disodium phosphate EGDMA Ethylene glyc ol dimethacrylate EW Equilibrium water content FEP A polymer of tetrafluoroe thylene and hexafloropropylene HEMA 2-Hydroxyethyl methacrylate NVP 1-Vinyl-2-pyrrolidone PBS Dulbecco’s phosphate buffered saline PHEMA Poly(hydroxyethyl methacrylate) PLTF Pre-lens tear film POLTF Post-lens tear film TRIS 3-Methacryloxypropyltris (trimethylsiloxy)silane

PAGE 15

15 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 OPHTHALMIC DRUG DELIVERY BY SOFT CONTACT LENSES By Jinah Kim May 2008 Chair: Anuj Chauhan Major: Chemical Engineering Drug delivery via contact lenses is known to significantly in crease the bioavailability of ophthalmic drugs compared to that via eye drops. Our study focused on investigating the feasibility of delivery ophthalmic drug fo r extended period of time by contact lenses. To have the fundamental insight of drug trans port in hydrogel contact lenses, we explored ocular delivery of dexamethasone via poly(hydroxyl ethyl methacrylate) cont act lenses, which are common daily disposable contact lenses. Three derivatives of dexamethasone (dexamethasone 21disodium phosphate (DXP), dexamethasone (DX), and dexamethasone 21-acetate (DXA)) are explored. These drugs are loaded in the gels by soaking in aqueous or ethanol solutions, and also by direct addition of the drug to the polymer izing mixture. Dynamic drug concentrations in the aqueous phase are monitored bot h in loading and release experime nts. The data is utilized to determine the partition coefficients and the mean diffusivity, which includ es contributions from both bulk and surface diffusion. Finally we utilize the transport model to predict the bioavailability of the three forms of dexame thasone for drug delivery via contact lenses. While hydrogel contact lenses lead to highe r bioavailability, these cannot be used for extended drug delivery because these cannot be wo rn overnight. Silicone hydrogel contact lenses can be worn for extended periods but commercial silicone hydrogel lens es are not suitable for

PAGE 16

16 extended drug delivery. So we developed new ex tended wear silicone hydrogel soft contact lenses that deliver ophthalmic drugs for an extended period of time ranging from weeks to months. Silicone hydrogels comprising of N, N-dimethylacrylamide, 3-methacryloxypropyltris (trimethylsiloxy) silane, bi s-alpha,omega-(methacryloxypropyl) polydimethylsiloxane, 1-vinyl2-pyrrolidone, and ethylene gl ycol dimethacrylate were prep ared with varying ratios of monomers and transport of three different ophthalmic drugs, timolol, dexamethasone, and dexamethasone 21-acetate was explored. All the s ilicone hydrogels of 0.1 mm thickness exhibit diffusion limited transport and ex tended release varying 20 days up to more than three months depending on the compositions of hydrophobic and hydrophilic components of silicone hydrogels. Also, there are multiple time scales in tr ansport of at least certain molecules, which is perhaps due to the complex microstructure of th ese gels. The mechanical and physical properties of lenses such as ion permeability, equilibrium water content, transparency, and surface contact angles of some of the gels are su itable for contact lens application. We also focused on the concept of increas ing the release duration of drugs from commercial silicone contact lenses by creation of transport barriers. As diffusion barriers, we loaded the vitamin E, which is considered as an important ocular nutraceutic al, into five different commercially available silicone contact lenses an d transport of dexamethasone and timolol were measured. The results conclusively show that v itamin E loading can substantially increase the release duration of drugs, partic ularly for hydrophilic drugs. The e ffect of vitamin E loading on the release rates is qualitatively simi lar for ACUVUE OASYS™, NIGHT&DAY™, and O2OPTIX™. The effect of vitamin E loading in d ynamics of both DX and timolol is minimal for PureVision™.

PAGE 17

17 CHAPTER 1 INTRODUCTION In the last few decades, a number of novel a pproaches have been developed for controlled drug delivery of ophthalmic drugs. However, t opical application of ey e drops is still the dominant treatment methodology desp ite its inefficiency [1]. Afte r application of an eye drop, the drug solution mixes with tear fluid, and then within about 5 minutes, a majority of drug is eliminated by tear drainage and conjunctival upt ake. Due to the short residence time, only about 15% of the applied drug penetrates the cornea and reaches the intraocular tissues. To maintain therapeutic levels of drug concentration, frequent instillation of drops with large drug loadings are required, which is inconveni ent for patients. Moreover, the drug that gets absorbed in conjunctiva or in the nasal cavity eventually re aches other organs thr ough the blood circulation leading to side effects [2, 3]. To achieve in crease ocular bioavailabi lity, researchers have explored a variety of vehicles including suspension of nanopa rticles, nanocapsules, liposomes and niosomes, ocular inserts like collagen shield s and Ocusert, and therapeutic contact lenses. Among theses, contact lenses have been widely studied due to the high degree of comfort and biocompatibility. On instillation of medicated cont act lens in the eye, drug diffuses through the lens matrix into the thin tear film named postlens tear film (POLTF) trapped between the lens and the cornea, and the drug has a residence time a bout 30 min in the eye [4, 5]. An increase in the residence time leads to a significant increas e in the bioavailability. Both mathematical models and clinical data suggest that the bioavailability for o phthalmic drug delivery can be as large as 50%, which is an order of magn itude larger than that for drops [6]. There have been a number of attempts in the past to use contact le nses for ophthalmic drug delivery; however most of these focused on soaking hydrophilic lenses in a drug solution followed by insertion into the eye [7 13]. The major problem of lo ading drug by this method is

PAGE 18

18 that in most cases the loading capacity of the soaked contact lenses is inadequate. Another commonly used method of entrapping drugs in gels is direct addition of drug in the polymerizing medium [14 16]. Although such direct loadi ng of drug into the lenses can allow higher loadings of the drugs, it can result in an activity loss dur ing polymerization. Furthermore, a majority of the drug can diffuse from the lenses into the pack aging medium and the drug retained in the lens can diffuse from the lens rapidly after insertion into the eye. In chapter 2, we explored ocular deli very of dexamethasone via poly(hydroxylethyl methacrylate) contact lenses to understand the drug transport by the contact lens in the fundamental point of view. Dexamethasone is a glucocorticoid steroid, which is similar to the natural steroid hormone made by the adrenal glands in the body. It relieves eye inflammation and swelling, heat, redness, and pain cau sed by chemicals, infection, and/or severe allergies. It is also used to treat persistent macular oe dema in retina, which is a major cause of visual disabilities and blindness among individuals with di abetes [17]. Prolonged systemic administration of steroid can cause serious side effects such as diabetes, hemorrhagic ulcers, skin atrophy, myopathies, osteoporosis and psychosis [18]. Fu rthermore, it has been reported that continuous application of eye drops of 0.1% dexamethasone for extended pe riods of time (varying from 3 weeks to a year) can cause glaucoma accompanied by optic nerve damage, defects in visual acuity and fields of vision, and posterior subcapsular ca taract formation and thinning of the cornea or sclera [19, 20]. Controlled drug delivery of dexamethasone to the eye through daily we ar contact lens is expected to be safer than delivery via drops because of reduction in the amount of drug that reaches other body tissues through systemic circul ation [6]. Several forms of dexamethasone with significantly different aque ous solubilities have been util ized in ocular studies including dexamethasone 21disodium phosphate, de xamethasone, and dexamethasone 21-acetate

PAGE 19

19 (arranged in the order of incr easing partition coefficients meas ured in octanol-phosphate buffer solution) [21, 22]. Molecular structures of thes e three dexamethasone de rivatives are shown in Figure 2-1. Dexamethasone and dexamethas one 21acetate are hydrophobic drug while dexamethasone 21disodium phosphate is ionic a nd thus freely water soluble. Civiale et al. screened the ocular permeability of dexa methasone derivatives through cell culture ( in vitro ) and excised rabbit cornea ( ex vivo ) and studied in vivo concentration of dexamethasone in rabbit aqueous humor as well. They reported that the permeability rates of dexamethasone derivates through cornea generally increase as octanol/water partitioning coefficients (log P) increase [22]. Weijtens et al. determined the dexamethasone concentration in aqueous humor, vitreous and serum after dexamethasone administ ration through various routes such as topical application of eye drop, subconjunctival injection, a peribu lbar injection, and an oral dose [23 27]. They showed that the dexamethasone concentration in the aqueous humor is far lower for eye drops of 0.1% dexamethasone disodium phosphate compared to a subconjunctival injection even if an eye drop is instilled every 1.5 hour. However the conj unctival injection is al so not the optimal drug delivery vehicle because it needs to be applied daily to have sufficiently high dexamethasone concentrations in the aqueous humor. It is thus de sirable to have other drug delivery vehicles that can deliver dexamethasone in a non invasive manner, and yet achieve sufficiently high concentrations in aqueous humor. To address the issues of insufficient drug lo ading by soaking and rapid diffusion of drug by direct loading in the polymeri zing medium Gulsen and Chauhan have proposed the development of nanoparticle laden gels that can load subs tantial amount of drug to the gel which can be released at a controlled rate from the nanopartic les [28, 29]. Also, a numbe r of researchers have focused on developing biomimetic and ‘imprinted’ contact lenses [30 35]. It has been

PAGE 20

20 demonstrated both by in vivo and in vitro studies that soft contact lenses fabricated by the molecular imprinting method have a drug loading capaci ty 2to 3-fold greater than that of the contact lenses made by a conve ntional method and leads to an increase in the partition coefficients and slower release of drugs. While the approaches listed above are effec tive at increasing the release duration from contact lenses, the lenses are st ill not suitable for extended rel ease lasting a week or longer. Additionally, the studie s cited above focused on hydrophilic hydrogel based contact lenses, which are not suitable for extended wear due to limited oxygen permeability. Due to high oxygen permeability silicone hydrogel contact lenses can be prescribed for extended wear lasting several weeks, and so these lenses are most suitable fo r development of extended drug delivery vehicles. Recently Karlgard et. al. measured the uptake an d release of a number of ophthalmic drugs by commercially available HEMA base d and silicone contact lenses in vitro studies [36]. A number of drugs including cromolyn sodium, ketotife n fumarate, ketorolac tromethamine, and dexamethasone sodium phosphate which are all hydrophilic were loaded into these lenses by soaking these in drug solutions fo r a limited period of time. The release studies showed that a majority of the drug taken up by the gels was re leased in a short period of time, proving that commercial silicone hydrogel cont act lenses do not have the appr opriate composition to provide extended drug release. Also they sh owed that the type of drug and the type of material affect the drug uptake and release In the chapter 3, we developed new extended wear silicone hydrogel soft contact lenses that deliver ophthalmic drugs for an extended pe riod of time ranging from weeks to months. The important properties for contact le ns applications were explored and furthermore we focused on

PAGE 21

21 correlation between composition and transport mechanism of drug such as timolol and dexamethasone in the silicone hydrogels. A number of approaches have been explored in the past to increase the duration of drug release from contact lenses. These include disper sing colloidal particles in the gels, molecular imprinting and biomimetic recogni tion. All these approaches focu s on increasing the affinity of the gel for the drug. In chapter 4, we propose a n ovel approach of creatin g diffusion barriers in the contact lenses by loading the lenses with exci pients that have a very small affinity for the drug. If a drug has negligible affinity for the ex cipients, drug molecules must diffuse around the regions containing the excipients leading to longer diffusive path lengths, which must reduce the mean diffusivity. Essentially, the proposed role of the excipients is to increase the tortuosity of the gel for drug transport. It is important to choose excipients with a very small aqueous solubility so that it does not diffu se into the tear film in large amounts. It is also important to choose an excipient that does not have ant toxic effects on the cornea. We focus on using vitamin E as the excipient because of very low solubility of hydrophilic drugs in v itamin E, and also its high viscosity, low aqueous solubility and a high degree of compatibility with ocular tissue.

PAGE 22

22 CHAPTER 2 DEXAMETHASONE TRANSPORT AND OCULAR DELIVERY FROM POLY(HYDROXYETHYL METHACRYLATE) GELS 2.1 Introduction The aim of this chapter was to investigate loading and release of different forms of dexamethasone in PHEMA gels, which are a co mmon contact lens material. Three different derivatives of dexamethasone, i.e., dexamethas one (DX), dexamethasone 21-actate (DXA), and dexamethasone 21-disodium phosphate (DXP) were incorporated in the ge l by soaking gels in drug solutions (soaking method) or by direct dissolution of the drug in the polymerizing mixture (direct entrapment method). The loaded drug was then released by soaking the drug-loaded gels in DI water or PBS. Dynamic drug concentrations in the aqueous phase were monitored both in loading and release experiments. The equilibriu m uptake in these experiments was utilized to determine the partition coefficients, and the dynamic data was fitted to a modified diffusion equation to determine the mean diffusivity, wh ich includes contributi ons from both bulk and surface diffusion. Finally the transport model fo r the drugs was utili zed to predict the bioavailability of the three forms of dexamethas one for drug delivery via PHEMA contact lenses. 2.2 Materials and Methods 2.2.1 Materials 2-hydroxyethyl methacrylate (HEMA, 97%) mo nomer, dexamethasone 21-acetate (DXA, 99%), dexamethasone 21-disodium phosphate (DXP, 99%), timolol maleate, ( 98%), ethanol ( 99.5%), and Dulbecco’s phosphate buffered saline (PBS) were purchased from Sigma-Aldrich Chemicals (St. Louis, MO); dexamethasone (DX, 98%) and ethylene glycol dimethacrylate (EGDMA) from Sigma-Aldrich Chemicals (Milwaukee, WI). Darocur TPO was kindly provided by Ciba Specialty Chemicals (Tarry town, NY). Nitrogen was bought from Praxair

PAGE 23

23 (Danbury, CT). All the other chemicals were of reagent grade. All the chemicals were used without further purification. 2.2.2 Synthesis of Poly(Hydroxyethyl Methacrylate) (PHEMA) Gels Poly(hydroxyethyl methacrylate) (PHEMA) gels were synthesized by free radical solution or bulk polymerization of the monomer with photoinitiation. Briefl y, 2.7 mL of monomer HEMA and 10 l of EGDMA were mixed with 2 mL of deionized (DI) water. The solution was purged by bubbling nitrogen for 10 min. 6 g of photoinitiator (Darocur TPO) was added to the monomer mixture with stirring for 5 min and the resulting solution was immediately injected into a mold composed of two 5 mm thick glass plates separated by a plastic spacer. The spacer thickness was chosen to be either 0.1 or 0.2 mm. The mold was then placed on Ultraviolet transilluminiator UVB-10 (UltraLum, Inc.) and the gel was cured by irradiating by UVB light (305 nm) for 40 min. The gel was cut in square shaped pieces (about 1.5 1.5 cm) and dried in air overnight for further use. 2.2.3 Drug Loading The drug was loaded into the gels either by di rectly dissolving the drug in the polymerizing mixture (direct entrapment) or by soaking the gel in an aqueous drug solution. Due to minimal solubility of DXA in water, it was loaded by di rect entrapment or by soaking the gel in drugethanol solution. The drug concentrations in the loading solutions and the drug loading by direct entrapment were conducted under conditions that led to drug loadings comparable to those required for therapeutic applications. Before th e soaking step, a square piece of PHEMA gel (about 1.5 1.5 cm) was boiled in 200 mL of DI water for 30 min to remove the unreacted monomer. Next the gel was soaked in 3mL of drug solution. DXA was loaded by soaking in drug-ethanol solution for a peri od of 3 hours. During the soaki ng in drug-ethanol solution, the dynamic concentration in the ethanol phase was not monitored. The total amount of drug loaded

PAGE 24

24 into the gel was estimated by determining the amount of ethanol uptake by the gel, and then multiplying it by the drug concentration in the et hanol solution. It was thus implicitly assumed that in view of the very hi gh solubility of DXA in ethanol absorption of drug on the PHEMA polymer can be neglected when the gel is soaked in ethanol. At the end of the loading stage the gel was taken out and dried in air overnight, a nd subsequently used for release experiments. During soaking of gels in aqueous DX and DXP solutions, the dynami c drug concentration in the DI water (or PBS) was monitored by measur ing the absorbance spectra of the solution over the wavelength range of 220 to 270 nm with a UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). The loading step was conducted till equilibrium was reached. The total amount of drug loaded into the gel was determined by finding the total amount of drug-loss from the aqueous solution. 2.2.4 Drug Release Experiments The drug release experiments were conducted by soaking the square shaped gels (about 1.5 1.5 cm) in 3 mL of DI water (or PBS) at room te mperature. It is noted th at the gels that were loaded with drug by soaking in aqueous solution s were directly transf erred from the loading solution to the release solution, and so these were fully hydrated at the beginning of the release experiment. However, gels that had the drug load ed by direct entrapment or by soaking in drugethanol solution were dry at the beginning of the release experiment The effect of this difference was shown to be minimal by comparing release profiles from dry and hydrated gels with the same drug loading (results not shown here). Du ring the release experiments, the dynamic drug concentration in the DI water (or PBS) was m onitored by measuring the absorbance spectra of the solution over the wavelength ra nge of 220 to 270 nm. After equi librium was reached, the gels were transferred to fresh 3 mL solution (DI water or PBS) and the process was repeated. These

PAGE 25

25 two releases are refereed as the 1st and the 2nd release. In some cases even a 3rd release was performed. 2.2.5 Effect of Gel Thickness on Loading and Release of Drug To determine whether the drug transport in ge ls is controlled by di ffusion, the procedures described above were repeated with thin gels (0.1 mm thick). The drug loading and release procedures for the thin gels were identical to those described above except that the volume of solutions was reduced by a factor of 2 to ensure that the ratio of gel to fluid volume was kept the same. 2.2.6 Effect of Ionic Strength in Release Solution on Loading and Release of DXP Since DXP is a charged molecule, its transpor t may be affected by the ionic strength. To investigate the effect of electrostatics on DXP tr ansport, the loading and release experiments for DXP were conducted at several diffe rent ionic strengths. In these experiments the ionic strength was changed by adding NaCl to the PBS buffer. DXP loading and release experiments were carried with three different NaCl-DXP solution in PBS corre sponding ionic strengths of 594, 1022, and 1202 mM. Also experiments were conducte d in PBS that has an ionic strength of 173 mM. 2.2.7 Conversion of UV-VIS Absorbance to the Corresponding Concentration of Drug The absorbencies of solution in both the loadi ng and the release steps arise mainly from the drug, but there is some additional absorbance due to small polymer chains that continue to diffuse out from the PHEMA gels. The absorban ce spectra of HEMA and dexamethasone (DX, DXA, and DXP) partially overlap, and thus to distinguish between the absorbance from HEMA and the drug, the measured absorbance spectr a was deconvoluted by expressing the measured absorbance as Abs = Drug + Control (2-1)

PAGE 26

26 where Abs is the measured absorbance spectra from 220 to 270 nm, Drug is the absorbance spectra of the drug in the same wavelengt h range at some arbi trary concentration, Control is the absorbance spectra of the solution in which PHEMA gel without any drug was soaked, and and are constants. These constants were obtained by finding best fit values minimizing the error between measured absorbance and calcul ated absorbance according to Eq.2-1 using the function ‘fminsearch’ in MATLAB. The concentration of drug was finally obtained by multiplying to the concentrati on corresponding to Drug The above procedure assumes that the absorbencies of these components are simply additive and linear in concentration, and this was verified by conducting several experiments. Also the accura cy of the procedure described above was established by determining drug concentratio ns in solutions of known composition. The difference between the fitted and the measured absorbance spectra was typically less than 1%. 2.2.8 Determinations of Critical Mi celle Concentration (CMC) of DXP Surface tension isotherm of DXP was measur ed at room temperature (about 23C) by creating a pendant drop, digitizing the shape and then fitting it to the Young-Laplace equation by using the Drop Shape Analysis System DSA100 (KRSS). A concentrated solution of the DXP (40 mg/mL) in PBS was prepared and then di luted successively to 0.021 mg/mL which is far below the expected CMC. 2.3 Results and Discussion 2.3.1 Dexamethasone (DX) Loaded Gel 2.3.1.1 Loading (by soaking in aqueous solution) and release studies of DX 2.3.1.1.1 Partition coefficient The partition coefficient K is defined as the ratio of the drug concentration in the gel and the concentration in the aqueous phase at equilibrium. The values of partition coefficient can be

PAGE 27

27 obtained from both the loading and the release experiments. For the loading experiment the partition coefficient can be calculated by w,f g w,f w,i w w,f g,fC V ) C (C V C C K (2-2) where Vw and Vg are the volumes of the aqueous phase and the fully hydrated gel, respectively, and Cg,f Cw,i and Cw,f are the equilibrium concentr ations of the drug in the gel, and the initial and equilibrium concentrations in the aqueous phase, respectively, in the load ing experiment. For the 1st release, the partition coefficient is given by r,f g f r w,f w,i w w,f g,fC V ) C C (C V C C K, (2-3) where Cr,f is the equilibrium concentr ation of the drug in the a queous phase gel in the 1st release, and the other variables correspond to the values obtained from the loading phase for the same gel. Similar expressions can be written for the 2nd and the 3rd releases. Table 2-1 shows the equilibrium concentra tions in the water phase and corresponding partition coefficients. The partition coefficien t of DX seems to be constant in the whole concentration range that was explored in these experiments. Moreover, the K values are similar for the loading, 1st release, and 2nd release. This suggests that th e process of load ing and release of DX is reversible and that K is not function of concentratio n in the explored concentration range. The K values for loading and release were 39.00 3.14 and 37.00 3.41, respectively and mean K was 38.00. 2.3.1.1.2 Dexamethasone (DX) loading dynamics A schematic of the geometry of the gel used in the drug loading and release experiments is shown in Figure 2-2. The experimental data sh own above demonstrates that the partition coefficient for dexamethasone is much larger than 1, which implies that a large fraction of the

PAGE 28

28 drug is bound to the PHEMA polymer. The binding of the drug to the gel can be modeled as a Langmuir adsorption isotherm, which relates the adsorbed concentration of the drug on the gel () to the free concentration in th e aqueous phase inside the gel (C) by the following equation C k C (2-4) where is the surface concentration at the maximum packing on the surface and k is the ratio of the rate constants for desorption and adsorpti on of the drug on the HEMA surface. The mean concentration in the gel (Cg), which is essentially the sum of the bound concentration and the free concentration, is given by fC V S Cgel g (2-5) where gelV S is the surface area per volume avai lable for the drug to adsorb and f is the volume fraction of water in hydra ted gel. The value of f for PHEMA gels was determined to be 0.42 from the swelling experiments. At equilibrium the free drug concentration in th e gel is expected to equal to the concentration in th e PBS, and thus the partition coefficients determined above are simply the ratio of the tota l drug and the free drug, i.e., f C k a C C Kg (2-6) where gelV S a. The values of K shown above are independent of C, which suggests that the value of k for DX adsorption on HEMA is much larger than 0.045 mg/mL which is the highest equilibrium concentration investigated in this study. Thus K is treated as independent of concentration in the model developed below. The transport of the drug in the hydrogel is

PAGE 29

29 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 2) (y A P D y C fD t Cs f g (2-7) where Df and Ds are the diffusivities of the drug in so lution and on the surface, respectively, and P/A is the perimeter of the gel fibers per unit cros s-sectional area, which can be approximated as S/V. Utilizing Eq.2-5 and Eq.2-6 in the a bove equation and noting the fact that K is independent of C gives 2 2 2 2) (y C D y C f K D fD t C Ks f (2-8) where ) ( f K D fD Ds f is the effective diffusivity of drug in the gel. The above differential equation is subjected to the following boundary conditions 0 ) (0 yy t y C (2-9a) wC H y t C ) ( (2-9b) where H is the half-thickness of the gel. The bounda ry condition Eq.2-9a arises due to symmetry at the center of the gel and th e boundary condition Eq.2-9b assume s that the drug concentration in the water phase in the gel phase at the inte rface of gel and solution is the same as the concentration in the outer water phase. From a mass balance of drug in the aqueous phase, we get H y gy C D A dt dCw Vw 2 (2-10) where Ag is a cross-sectional area of the gel. The initial conditions for the drug loading are

PAGE 30

30 i w wC t C,) 0 ( 0 ) 0 ( y t C (2-11a,b) The above set of equations was solved by imp licit finite difference method with 21 spatial nodes and a dimensionless time step ( Dt / KH2) of 0.0025. The DX loading experiments were conducted with different initial concentrati ons and the dynamic drug concentration in the aqueous phase was measured. These data was fitte d to the model described above and diffusivity of the drug in the gel was determined. The e rror between the experime ntal data and model prediction was defined as ex w ex w wC C C, 2 ,/ ) (, where Cw and Cw,ex are the predicted concentration in the aqueous phase by model and the experimental concentration, respectively. The diffusivity for the each set of experimental dynamic concentration da ta was evaluated using the function ‘fminsearch’ in MATLAB. The value of D for DX was 1.00 10-11 6.32 10-13 m2/sec (n = 10). The theoretical profiles predicte d by the model and the experimental data in drug loading experiments are given in Figure 2-3. Two different theoretical profiles for the same initial concentration co rrespond to the two different gel volu mes. The model predictions match the experimental data very well with the error ranging from 1.86 to 3.88%. 2.3.1.1.3. Dexamethasone (DX) release dynamics The dynamics of drug release from a gel into fresh solution can be described by Eq.2-8 with boundary conditions Eq.2-9, Eq.2-10, and the following initial conditions 0 ) 0 ( t Cw, iC y t C ) 0 ( (2-11c, d) where Ci is an initial free drug concentration in the gel. If the gel properties do not change in the drug loading experiments, the values of K and D for drug release are expected to be the same as during the loading step. The numer ical procedure described above was repeated with the release data to determine the drug diffusivity in the gel. The experimental data and the fitted profiles are plotted in Figure 2-4. The amount of drug loading in the gel and the rms errors in the fit are also

PAGE 31

31 indicated in the figure. The model predictions ma tch the experimental re lease data well with error range between 1.60 and 3.29 %. The errors ar e slightly larger than those in drug loading, perhaps due to error accumulation to the previous drug loading expe riments. From the fitting of the release data, the value of K and D for DX were determined to be 37.00 3.41 and 1.06 10-11 1.48 10-12 m2/sec, respectively. These values are in excellent agreemen t with the values obtained from the loading data. 2.3.1.2 Release of DX loaded in the gel by direct entrapment 2.3.1.2.1 Partition coefficient Drugs can be loaded to the hydrogel contact lens es by soaking the gels in drug solutions, or by directly adding the drug to the polymerizing mixture (direct entrapment). While direct entrapment of drug may be more convenient, there is a possibility that a fraction of the drug may get irreversibly trapped in the gel. To inve stigate the feasibility of loading DX by direct entrapment, gels with different drug loading we re prepared and drug release experiments were conducted from these gels. The release experiments were conducted with four different initial drug loadings ( M0 /Mg = 1.37, 2.74, 4.78, and 6.82 mg/g). Each drug-laden gel was soaked in 3mL of DI water (1st release). After equilibrium was achieve d, the aqueous solution was replaced with 3mL fresh DI water (2nd release). The 3rd release was conducted in the same manner. The equilibrium concentration in the water phase can be used to determine the gel concentration by a using a mass balance, and then these can be used to determine the partition coefficients. In Table 2-2, the equilibrium concentrations in the water phase and the corresponding partition coefficients are listed. The values of K in Table 2-2 are significantly higher than the K values obtained above. Moreover, the K values for the 2nd release are much higher than those for the 1st release. Both of these effect s could possibly be caused due to

PAGE 32

32 irreversible entrapment of a fraction of the dr ug. Based on this hypothesis, the mass balance in the drug-water system can be modified as ) (, 0 p f w g f w w f w g f w wM C V K C V C KV C V M (2-12) where M0 is the total mass of drug loaded in the gel, Mp is the mass of drug that is irreversibly trapped in the gel, K is the true partition coefficient of drug, and the other variables have the same definitions as above. The K values that were calculated by neglecting any irreversible entrapment are now called appa rent partition coefficients ( Kapp) to differentiate them from the intrinsic partition coefficients K Based on the data obtained from the loading and release experiments from gels that had the drug loaded by soaking, it is reasonable to assume that the intrinsic partition coefficient is independent of concentration, and so for a given gel the K values and also the Mp should be the same for the 1st and the 2nd releases. Using Eq.12 for both the 1st and the 2nd releases, Mp and K can be computed for each gel, and these are listed in Table 2-2. The values of K (=34.30 2.82) are in reasonable agreement with the values obtained from gels that were loaded by soaking in drug solutions. To further validate the results for K and Mp, the apparent partition coefficient was determin ed by using the above equation for the 3rd release and the calculated values were in good agreement with the measured values. The results also show that about 17.4% of total drug loaded (Table 2-2) is permanently entrapped in the gel matrix and not available for the drug delivery. The perm anent entrapment could be either physical entrapment in tightly bound regions or chemical entrapment due to reactions. 2.3.1.2.2 Dexamethasone (DX) release dynamics The model for the release of drug directly entrapped in the gel into the water is the same as that for the release of drug loaded by soaking. Th e initial concentration in the gel, which is a parameter required for the fit is set equal to the concentration of the drug that is available to

PAGE 33

33 diffuse. In Figure 2-5, the experimental data for 1st and 2nd release are plotted along with the model predictions. The mean values of D for 1st and the 2nd releases are 1.00 10-11 6.32 1013 m2/sec, and 1.23 10-11 7.05 10-13 m2/sec, respectively. The value of D obtained by fitting the 1st release data is in excellent agreement w ith the values reported above but the value increases by about 20% in the 2nd release. This is an unexpected result because the drug release in the first release step is not sufficiently large to impact the structure of the gel. 2.3.1.3 Effect of gel thickness on DX loading A major fraction of the drug present in the gel is bound to the polymer and so the drug transport may occur by a combination of surface a nd bulk diffusion. If surface diffusion rates are small, the bound drug could desorb and then diffu se via bulk diffusion. To determine whether transport of drug is controlled by diffusion or by the process of adsorption-desorption of drug on the polymer, it was decided to fabricate gels w ith different thicknesses (0.1 mm and 0.2 mm thick in dry state). If the rate limiting step is diffusion, an increase in gel thickness by a factor of 2, while keeping fluid to gel ratio fixed, will increase the time scale of release by a factor of 4. A change in time scale by a factor of less than 4 signifies that the adsorption-desorption process occurs on a time scale comparable to the diffusion. As described in methods section, both thick and thin gels were cu t into square shaped pieces (about 1.5 1.5 cm) and unreacted monomer was removed by boiling. The gels were then dried and weighed and then soaked in water ove rnight for hydration. The hydrated thick gel ( H = 0.13 mm) was put in 3 mL of DX-water solution and the hydrated thin gel ( H = 0.06 mm) was put in 1.5 mL of DX-water solu tion for the drug loading studies. The drug loading and release profiles from the thick and the thin gels are shown in Figure 2-6. As mentioned above, if the drug transport is diffusion limited, the drug loading rate should

PAGE 34

34 be proportional to the square of the gel thickne ss. The ratio of gel to fluid volumes are not exactly equal for the two gels, and so to scale out the effect of the une qual volume ratio and the effect of the gel thickness, the dyna mic drug concentration in water ( Cw) was divided by final equilibrium drug concentration ( Cw,f), and then this Cw/ Cw,f was plotted as a function of time/ H2. As seen in this figure, the profiles for the thic k and the thin gels overlap both for loading and release experiments, which means that the drug re lease rate from the gel is exactly proportional to square of gel thickness, and thus the tr ansport is diffusion limited. The values of K and D obtained by fitting the loading and the release data fo r the thick and the thin gels are noted in the figure caption. 2.3.1.4 Effect of crosslink density of gel on DX loading To determine the effect of crosslink de nsity on the DX transport, DX was directly entrapped into gels of four di fferent crosslink densities. The cr osslink density is defined as 100 (mol) added HEMA of mole (mol) added EGDMA of mole (%) density crosslink (2-13) Since the gels were synthesized in a direct dr ug entrapment method, Eq.12 can be applied to determine the intrinsic partition coefficient K and the mass of drug perm anently entrapped in the gel Mp, and these are listed in Table 2-3. There is no dependence of Mp/ M0 and K on crosslink density. However, the diffusivity of drug decr eases considerably with increasing crosslink density. The decrease in the diffusivity is likely the result of a decrease in the bulk diffusivity due to a decrease in the pore size of the gels. 2.3.1.5 Release of DX into PBS All the results reported above for DX correspond to transport in gels hydrated in DI water. The transport in tear fluid or PBS which is a re asonable mimic of the tear fluid may be different from that in DI water partly because of di fferences in drug binding to the gel in the two

PAGE 35

35 mediums, and also due to differe nt degree of swelling of gels, a nd also due to small differences in bulk viscosity. To investigate these issues it was decided to load DX in PHEMA gels by direct entrapment, and measure th e release rates in PBS and in DI water. The DX release profiles (both 1st and 2nd) in water and PBS were shown in Figure 2-7. As seen in this figure, the two profiles are similar and the values of K and D which are noted in the figure caption, are comparable as well. These results are exp ected because DX is a nonionic drug and so its partitioning is not expected to depend strongly on the salt concentration. 2.3.2 Dexamethasone 21-acet ate (DXA) Loaded Gel Since DXA has a very low solubili ty in DI water and in PBS, it was loaded into the gel either by direct entrapment or by soaking in drug -ethanol solution. So the partition coefficients and diffusivities were only determined through the drug release experiments. 2.3.2.1 Release of DXA loaded by direct entrapment 2.3.2.1.1 Partition coefficient a nd dynamics of DXA release The release experiments were conducted with four different initial drug loadings ( M0 /Mg = 1.37, 2.74, 4.78, and 6.82 mg/g). The equilibriu m concentrations in the water phase and the corresponding values of the a pparent partition coefficient Kapp are listed in Table 2-4. Using equation (12), the intrinsic partition coefficient K and the mass of drug permanently entrapped in the gel Mp were calculated and these are also give n in Table 2-4. The intrinsic partition coefficient K of DXA in PHEMA gel is also relativel y independent of concentration, and its value is 79.82 4.44 which is much higher than that of DX (34.30 2.82). For DXA, only about 33.6% of initial drug is available for drug rel ease since most of drug (66.4% of initial drug loading) is permanently entrapped. The D for the 1st and the 2nd releases are 1.29 10-11 3.67 10-13 m2/sec and 1.08 10-11 1.13 10-13 m2/sec, respectively, which are comparable to that for

PAGE 36

36 DX. As in Figure 2-8, the overall time of DXP re lease is about 32 hours whic h is longer than that of DX (about 16 hours) even though th ey have comparable diffusiv ities since the time scale of diffusion is K H2/ D and K value of DXA is about 2 times of that of DX. Since both DX and DXA have comparable sizes, the bulk diffusivity of these two molecules is expected to be similar. Furthermore, the effective diffusivity D values are also comparable even though the K value differ by a factor of larger than 2 suggesting that the dominant transport mechanism is bulk diffusion, or alternatively the surface diffusivity is larger for DX by a factor of 2 and so the increased partition coefficient for DXA is balanced by a decrease in surface diffusivity leading to similar effective diffusivities for DX and DXA. 2.3.2.2 Release of DXA loaded by DXA-ethanol presoaking As shown above, direct entrapment of DXA re sults in about 67% irre versible entrapment. To avoid losing a large fraction of the loaded drug to permanent entrapment, drug DXA was loaded in the gels by soaking in drug-ethanol solution. Release experiments from gels loaded with DXA by soaking are described below. 2.3.2.2.1 Effect of gel thickness on DXA release DXA release profiles from thick gel ( H = 0.13 mm) and thin gel ( H = 0.06 mm) are shown in Figure 2-9. In this figure, Cw / Cw,f is plotted as a function of time/H2 to scale out the effects of changes in volume ratio of gel and water and gel thicknesses. The scaled re lease profiles overlap and the fitted values of D are 1.12 10-11 m2/sec and 1.16 10-11 m2/sec for the thick and the thin gels, respectively, which both are in reasonable agreement with the values obtained from the direct entrapment studies. Based on the fact that transport time scales ar e quadratic in thickness, it can be concluded that DXA transport in the PHEMA gels is also diffusion limited.

PAGE 37

37 2.3.2.2.2 Release of DXA into PBS To mimic the transport of DXA in the tear fl uid, release studies into PBS were done and compared with that in water (Fi gure 2-10). The results show that DXA transport is slightly faster in gels soaked in PBS compared to those soaked in DI water. The fitted K and D values of DXA for PBS are 85.49 and 1.20 10-11 m2/sec, respectively which are about 10% larger than those for DI water (82.79 and 1.12 10-11 m2/sec). Therefore, it can be concluded that the ionic strength and pH have a negligible effect on tran sport of DXA, which is perhaps due to the fact that DXA is a nonionic drug. 2.3.3 Dexamethasone 21-disodium phosphate (DXP) Loaded Gel 2.3.3.1 Partition coefficient and dyna mics of DXP loading and release The partition coefficients of DXP between the gel and the aqueous phase were determined by following the same procedures as described earlier for DX. DXP can get ionized with two pKa ’s of 1.89 and 6.4 [37] and so its behavior in DI water is expected to be significantly different from that in PBS. Accordingly, all the loading and release experiments for DXP were conducted in PBS in which a majority of the dr ug is expected to be charged because the pH ranges from 7.0-7.5. Drug loading studies were conducted by soaking gels in DXP-PBS solution at 0.115, 0.086, and 0.038 mg/mL. After equilibri um was attained, the PBS was replaced and the release experiments were conducted. The K values obtained from DXP loading and release are listed in Table 2-5. The K values from the loading experiments are relatively similar and the K values from the release experiments are sim ilar, but these are larger than the loading K values by an order of magnitude. It is expected that if experiments are conducted in which the equilibrium concentrations are in between 0.005 and 0.036 mg/m L, the partition coefficients will transition smoothly from about 30 to about 3. This behavi or suggests that there may be two types of

PAGE 38

38 adsorption sites on the gel, and the first type bind s the drug strongly but saturates at rather small concentrations. Since the K values in the release experime nts are relatively concentration independent, the model proposed in section 3.1.1.2 can still be used to determine the diffusivities for DXP loading into PHEMA gel. In the loading experiments the concentration inside the gel is close to zero at short times and so in principle the dependence of K on concentration needs to be taken into account to determin e the diffusivity. However since K becomes concentration independent at relatively small c oncentrations we neglect this i ssue and use constant values of K from the loading experiments to fit the dynamic loading data. Some simulations were conducted by fitting the partition coefficient-concentration data to a sum of two Langmuir isotherms and then using this form in the model. However thes e simulations yielded diffusivities close to those obtained by using a constant value of K The experimental profiles for loading and release and those predicted by model for DXP loading are pl otted in Figure 2-11 and in Figure 2-12. The fitted value of D from DXP loading is 1.14 10-12 8.98 10-14 m2/sec, which is an order of magnitude lower than that of DX or DXA loading. Furthermore the fitted value of D from the release data is 2.12 10-11 m2/sec, which is much larger than diffusivity obtained from the loading experiments. The differences between th e diffusivities in loading and release can be explained by noting that the fitted values re present the average diffusivities, and include contributions from both bulk diffusion and surface diffusion, i.e., ) ( f K D fD Ds f The values of Ds and Df are likely to be independent of drug concentrations but the partition coefficient could depend on concentration. So concentration dependence of the effective diffusivity likely arises from the surface diffusion. For the case of DXP, the partition coefficients at release concentrations are an order of magnitude larger than those in the loading concentrations and D is also an order of magnitude larg er, suggesting that the surface diffusion

PAGE 39

39 may be the dominant mechanism for transport of dexamethasone phosphate. It is noted though that this is only a plausible argument and we cannot conclusively prove this hypothesis with the indirect measurements reported in this paper. 2.3.3.2 Effect of gel thickness on DXP loading The effect of gel thickness on DXP transport was investigated by performing DXP loading experiments at initial concentr ation 0.086 mg/mL with hydrated ge ls of two different thicknesses (half thickness H = 0.13 and 0.06 mm). The scaled Cw / Cw,f vs. time/ H2 profiles are plotted Figure 2-13. The two loading profiles are relatively similar, and the fitted values of K (3.14 for thick and 2.89 for thin) and D (1.21 10-12 m2/sec for thick and 1.35 10-12 m2/sec for thin) are also in reasonable agreement suggesting that the process of DXP tran sport is also diffusion limited. 2.3.3.3 Effect of ionic strength of outer solution on DXP loading and release The molecular weight of DXP (516.4) is only marginally larger than those of DX (392.5) and DXA (434.5) but one significant differen ce between DXP and the other two forms of dexamethasone is the fact that DXP exists in charged form in PBS. To investigate whether electrostatic effects contribute to the differences in diffusiviti es between the three forms of dexamethasone, the transport of DXP was investig ated in NaCl-PBS solutions with different ionic strengths ( I = 173, 594, 1022 and 1202 nm). In itial concentration of DXP ( Cw,i) was fixed at 0.087 mg/mL in each of the solutions. The DXP loading and release profiles for these experiments are shown in Figure 2-14. To clearly observe the effect of ionic strength on loading and release timescale, the DXP concentrati on in aqueous phase was normalized by final equilibrium concentration and plotted in Figure 2-14 (a) and (b), for the loading and release, respectively. It is clearly evident that an increase in the ionic streng th leads to an increase in time needed to achieve equilibrium during both the loading and the release phases. The fitted values

PAGE 40

40 of K and D are listed for these experiment s in Table 2-6. The values of K are relatively independent of the ionic st rength but the values of D for both the loading and the release decrease with increasing ionic st rength. This tendency can be seen more clearly in Figure 2-15 which plots the diffusivities of DXP for loading a nd release as a function of the ionic strength. The changes in D with ionic strength suggest that electrostatics affect the transport of charged drugs like DXP even in PBS. The effect of elect rostatics is expected to reduce with increasing ionic strength due to screening effects and so the D values at the very high ionic strength are expected to be the true diffusivity values. The D values at the highest ionic strength are similar from both loading and release but these are an order of magnitude lower compared to the diffusivities of the uncharged forms of dexamethasone described earlier. In an effort to understand the mechanisms of transport of charged drugs, it was decided to explore transport of timolol maleate, which also exists in a charged form at physiological pH and its molecular weight is comparable to that of dexamethasone phosphate. Transport of timolol maleate in PHEMA and other t ypes of hydrogels has been in vestigated by a number of researchers. However, to our knowledge the eff ect of ionic strength on transport of timolol maleate in PHEMA hydrogels has not been invest igated. We conducted loading experiments for timolol maleate at a fixed loading concentrati on of 0.0857 mg/mL and at three different salt concentrations with protocols identical to those for dexameth asone phosphate. Subsequently, release experiments were perfor med also with protocols described earlier for dexamethasone phosphate. The loading and release data was fit in the same manner as described earlier. The results of loading and release expe riments along with the fitted curves and the best fit values of diffusivities are shown in Figure 2-16 and Table 2-7. Timolol maleate diffusivities obtained by fitting the loading data are different from those obtained by fitting the release data. Furthermore,

PAGE 41

41 the diffusivities decrease with an increase in ioni c strength and at the highest ionic strength the diffusivities from the loading and the release are similar (Figure 2-17). Each of these trends is similar to those for dexamethasone phosphate. Howeve r, the values of diffusivities obtained for timolol are much larger than those for dexameth asone phosphate. At the highest ionic strength where electrostatics are expected to be screened the partition coefficients for timolol maleate during the release is about 60% higher than that during loading, and the diffusivities are about twice, which is in accordance with the earlier obs ervations for other drugs. However the values of diffusivities for timolol mal eate during loading are about 5 times that of DXP even though the partition coefficients are comparable. We speculate that the significant reducti on in diffusivity of DXP may be due to aggregation of the charged drug in to micelles. Since DXP is a rela tively linear molecule with a charged hydrophilic group on one side and the hydrophobic group on the other, it is reasonable to expect DXP to be surface active and also form micelles. In fact corticosteroid 21-phosphate esters including DXP have been reported to fo rm micelles in DI water and methylprednisolone 21phosphate whose structure is very simila r to that of DXP forms micelles above a concentration of 0.017 M [38]. To verify whether DXP forms micelles at concentrations corresponding to drug loading and release experime nts, the surface tension isotherm of DXP in PBS was measured. The surface tension vs log (concentration) data plotted in Figure 2-18 demonstrates that the CMC of DXP in PBS is above 0.1 M, which is higher than the concentrations explored in the lo ading-release studies. Thus aggregation of DXP into micelles is an unlikely reason for the small diffusiviti es. The mesh size of the PHEMA hydrogel was obtained by following the reported method by Canal and Peppas [39]. The mean value of the equilibrium volume degree of swelling of th e PHEMA hydrogel is 1.60, and accordingly the

PAGE 42

42 corresponding mesh size is a bout 1.7 nm. The size of DXP is about 1.2 nm and this is comparable to the mesh size of PHEMA hydroge l. Aggregation of DXP micelles in the bulk phase in the gel is again unlik ely the case for this reason. 2.3.4 Mathematical Model for Drug Transport to the Cornea We utilized the model for drug delivery by soaked contact lenses reported previously [6] and detailed description is in the Appendix. Be low we solve the coupled mass transfer problem for drug delivery from a contact lens in the eye. 2.3.4.1 Concentration profiles in the POLTF Figure 2-19 shows DX concentration in the post le ns tear film vs. time plots at three axial locations for case 1, i.e., no flux to the pre lens tear film. The inset in the figure shows the magnified view of the plots near t = 0. The concentration starts at lens center zero and then very quickly increases to a value of about 0.85. Si nce the concentrations are dedimensionalized by Ci/ K the maximum possible value of Cf in the POLTF is 1. During the period in which the concentration is increasing, the drug flux from the gel is larger than the sum of the drug loss from the sides ( x = L ) to the outer tear lake and the dr ug uptake by the cornea. The maximum value of the concentration is reached in a very short period of time because the volume of the POLTF is small as reflected in the large value of P2 At initial times, the drug flux into the cornea and the drug loss from the sides are much less than the drug flux from the lens causing drug concentration in the post lens tear film to increase. However, as the drug concentration in the POLTF builds up, the flux of the drug from the lens decreases and the drug loss from the sides due to dispersion and the drug flux into cornea increases. C onsequently, very quickly, the drug concentrations in the POLTF begin to decrease.

PAGE 43

43 2.3.4.2 Fraction of drug that enters cornea The amount of drug that diffuses into the cornea is simply equal to 002L f cdxdt C k. The mass of drug that diffuses into the tear lake from the edges of the POLTF is given by 02 dt h dx dC D Df L x f f. Finally, the amount of the drug that is lost to the PLTF is given by 00 02L y g gdxdt dy dC D. On dedimensionalizing each of these masses by the initial mass of drug in the gel (= 2 CiLhg), we get the following equations: 0 1 0 00 02 d d C P2 P3 dxdt C k M Ff L f c c (2-25) 00 1 02 d d dC D P2 P1 dt h dx dC D D M F f f L x f f s (2-26) 0 1 0 0 00 0 02 d d d dC dxdt dy dC D M F g L y g g p (2-27) where Fc is the fraction of the total drug that enters cornea, Fs is the fraction of the total drug wasted from the side to the outer tear lake, and FP is the fraction of the total drug wasted to the PLTF. After determining the concentration profiles, the various fractions can be determined by computing the integrals nu merically. The values of Fc, Fs, and Fp are listed in Table 2-8 for the three drugs for both case 1 (no flux to PLTF) a nd case 2 (zero concentration in PLTF). The values of Fc represent the bioavailability and these are much higher than 1-3%, which are the typical values for delivery by eye drops. The values of Fc are higher for case 1 because of neglect of drug loss to the pre lens tear film. The physio logical boundary condition in the pre lens tear

PAGE 44

44 film is expected to be in between case 1 and case 2, and so the average of these two cases may be a good approximation for the bioavailability. Amongs t the three drugs inve stigated here, the bioavailability is highest for DXA, primar ily due to the highest cornea permeability. 2.4 Conclusions The bioavailability of ophthalmic drugs delivered via contact lenses is significantly higher in comparison to that for topical application as eye drops. The bioava ilability of ophthalmic drugs delivered by contact lenses can be estima ted by solving the mass transport problem in the eye in the presence contact lenses. In order to solve the transport model, one requires the parameters that describe the drug transport in the contact lens. In this paper we have investigated the transport of three different forms of de xamethasone in PHEMA gels, which are a common contact lens material. The transport of the drugs is investigat ed by soaking PHEMA gels in aqueous drug solutions and monitoring the dynamic drug concentr ations. After reaching equilibrium, the gels are soaked in fresh solutions for the release experiments. Since DXA has very limited aqueous solubility, it is loaded into the gel via soaking in ethanol-drug solution. Furthermore drug release studies are also conducted for situations in which the drug was added directly to the polymerizing mixture. The equilibr ium concentrations in both the loading and the release studies are utilized to determine the partition coefficients Furthermore, the dynamic data is fitted to the diffusion equation to determine the mean diffus ivity, which includes contributions from both bulk and surface diffusion. The partition coefficients of DX and DXA are i ndependent of concentration, and are about 40 and 80, respectively. The partition coefficien ts are relatively similar from both the loading and the release studies. The partition coefficients estimated from the direct entrapment studies seem to be significantly different but the differences can be attribut ed to the fact that addition of

PAGE 45

45 drug to the polymerizing mixture re sults in some irreversible drug entrapment. The irreversible entrapment could be both physical and/or chemi cal. By utilizing the first, second and third release results, it was determined that a bout 17% of DX and 65% DXA gets irreversibly entrapped, and after taking into account the irreve rsible entrapment, the partition coefficients are in reasonable agreement with the results from soaking and release studies. The transport of all the three drugs is diffusion limited. The diffusivity of DX is 1.05 10-11 m2/sec from both the loading and the release studies, and the values ar e comparable in release studies with directly entrapped drug. The diffusivity for DXA is about 1.29 10-11 m2/sec, and this value too is similar for all uptake and release studies. There are slight differences in diffusivities estimated from the first and the second rele ases after direct entrapment, but these differences are within experimental errors. The partition coefficient of DXP is concentra tion dependent; it decreases from about 30 to 3 as the concentration in crease from 0.003 to 0.107 mg/mL. The diffusivity of DXP is larger for the release phase than for the loading phase by an order of magnitude, which can be explained by noting that the fitted values represent the average diffusivities, and include contributions from both bulk diffusion and surface diffusion, i.e., ) ( f K D fD Ds f To investigate whether electrostatic effects impact DXP transport, the transport of DXP was investigated in NaCl-PBS solutions with different ionic strengths. These ex periments show that the partition coefficient of DXP is relatively independent of the ionic strength but the mean diffusivity values decrease with an increase in ionic strength. At all ionic strengths, the diffusivity in loadin g is smaller by about a factor of 10, which can be attributed to the contribution from surface diffusion. The fact that diffusivities change on increasing ionic strength s uggests that electrostati c effects are important in transport of charged drugs even in PBS. If there is screening effect of salt on charged drug,

PAGE 46

46 DXP, we can assume that D of DXP will be close to true diffusivity at very high ionic strength of outer solution which would be similar to that of the other noncharged derivatives, but obtained D at very high ionic strength was still as low as 10-12 m2/sec, an order of magnitude lower than that of the others. The partition coefficient of timolol maleate, which is also charged in PBS is relatively independent of concentration and depends w eakly on the ionic strength. Timolol maleate diffusivities obtained by fitting the loading data are different from those obtained by fitting the release data. At the highest ionic strength where el ectrostatics are expected to be screened, the partition coefficients for timolol maleate during the release is a bout 60% higher than that during loading, and the diffusivities are about twice, which could perhaps be attributed to surface diffusion. Furthermore, the diffusi vities decrease with an increase in ionic strength and at the highest ionic strength the diffusiv ities from the loading and the release are similar. These trends are similar to those for dexamethasone 21-disodium phosphate but values of diffusivities for timolol maleate during loading are about 5 tim es that of DXP even though the partition coefficients are comparable. We speculated that the significantly smaller di ffusivity of DXP is due to aggregation of DXP into micelles. However surface tension m easurements showed that the CMC of DXP in PBS is larger than concentrations explored in loading-release studies a nd so aggregation into micelles is unlikely to cause the reduction in diffusivities. It is how ever possible that DXP aggregates at lower concentr ations inside the constraint pores in the PHEMA hydrogel. The predicted values of Fc, which represent the bioavailab ility, are much higher than the typical values for delivery by eye drops. The values of Fc are higher for case 1 because of neglect of drug loss to the pre lens tear film. The physio logical boundary condition in the pre lens tear

PAGE 47

47 film is expected to be in between case 1 and case 2, and so the average of these two cases may be a good approximation for the bioavailability. Amongs t the three drugs inve stigated here, the bioavailability is highest for DXA, primarily due to the highest cornea permeability. Thus DXA delivery via soaked daily dispos able PHEMA contact lenses seem s like a much more efficient method of delivering dexamethasone to eyes in comparison to delivery through eye drops. However clinical tests are needed to firmly es tablish the safety and efficacy of drug-loaded contact lenses for ophthalmic drug delivery because issues such as continuous exposure of ocular tissue to the drug could possi bly evoke toxic response.

PAGE 48

48 Figure 2-1. Molecular structures of model drugs A) DX B) DXA C) DXP Figure 2-2. The model geometry of the gel. Water y H Gel A C B

PAGE 49

49 0.0E+00 1.0E-02 2.0E-02 3.0E-02 4.0E-02 5.0E-02 6.0E-02 7.0E-02 8.0E-02 9.0E-02 05101520253035 Time (hr)Cw (mg/ml) Model prediction Cw,i=0.077 mg/ml (error=1.41%) Cw,i=0.077 mg/ml (error=1.31%) Cw,i=0.064 mg/ml (error=1.85%) Cw,i=0.058 mg/ml (error=1.66%) Cw,i=0.058 mg/ml (error=1.44%) Cw,i=0.051 mg/ml (error=0.79%) Cw,i=0.051 mg/ml (error=1.65%) Cw,i=0.038 mg/ml (error=1.62%) Cw,i=0.026 mg/ml (error=1.56%) Cw,i=0.026 mg/ml (error=1.97%) Figure 2-3. Comparison of the model predicti on and experimental data for DX loading into PHEMA gel soaked in drug solution. Initia l drug concentration in the aqueous phase and rms errors are indicated. K = 39.00 3.14 and D = 1.04 10-11 5.85 10-13 m2/sec (n=10).

PAGE 50

50 0.0E+00 5.0E-03 1.0E-02 1.5E-02 2.0E-02 2.5E-02 05101520253035 Time (hr)Cw (mg/ml) Model prediction Mo/Mg=2.14mg/g (error=1.60%) Mo/Mg=1.39mg/g (error=2.67%) Mo/Mg=0.96mg/g (error=3.01%) Mo/Mg=0.71mg/g (error=3.22%) Mo/Mg=0.68mg/g (error=3.29%) Mo/Mg=0.34mg/g (error=2.73%) Figure 2-4. Comparison of the m odel prediction and experimental data for DX release into fresh water from a PHEMA gels which have been soaked in drug solution. K = 37.00 3.41 and D = 1.06 10-11 1.48 10-12 m2/sec (n=6). Initial drug amount loaded in the gel and rms errors are indicated.

PAGE 51

51 0.0E+00 1.0E-02 2.0E-02 3.0E-02 4.0E-02 5.0E-02 6.0E-02 7.0E-02 8.0E-02 05101520253035 Time (hr)Cw (mg/ml) Model prediction Mo/Mg=6.82mg/g (error=2.18%) Mo/Mg=6.82mg/g (error=1.93%) Mo/Mg=4.78mg/g (error=1.86%) Mo/Mg=4.78mg/g (error=2.03%) Mo/Mg=2.74mg/g (error=3.59%) Mo/Mg=2.74mg/g (error=2.40%) Mo/Mg=1.37mg/g (error=2.55%) Mo/Mg=1.37mg/g (error=3.88%) Figure 2-5. Comparison of the model prediction and experimental data for DX release into the fresh water from PHEMA gels synthesized in a direct entrapment method. The solid legends and solid lines represent 1st release and the hollow legends and dashed lines 2nd release. K =34.30 2.82, D (1st release)=1.00 10-11 6.32 10-13 m2/sec, D (2nd release)=1.23 10-11 7.05 10-13 m2/sec, and Mp/M0=17.4 1.0 % (n=4). Initial drug amount loaded in the gel and rms errors are indicated.

PAGE 52

52 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0.0E+003.0E+126.0E+129.0E+121.2E+13 t / H2 (sec/m2)Cw/Cw,f Model prediction Thick gel (H=0.13mm), Loading Thin gel (H=0.06mm), Loading Thick gel (H=0.13mm), Release Thin gel (H=0.06mm), Release Figure 2-6. Effect of gel thickness on DX load ing and release of a PHEMA gel soaked in drug solution. Normalized DX loading and releas e profiles of thin gel and thick gel by final drug concentration in aqueous ph ase are plotted as a function of t/ H2. Cw,i = 0.058 mg/mL. For thick gel, K (loading) = 43.02 and D (loading) = 1.02 10-11 m2/sec (error=1.93%), K (release) = 59.86 and D (release)= 1.22 10-11 m2/sec (error=1.86%), for thin gel K (loading) = 44.44 and D (loading)= 9.51 10-12 m2/sec (error=2.03%), K (release) = 53.85 and D (release)= 9.51 10-12 m2/sec (error=4.93%). Half thicknesses of hydrated gels ( H ) are indicated.

PAGE 53

53 0.0E+00 2.0E-03 4.0E-03 6.0E-03 8.0E-03 1.0E-02 1.2E-02 1.4E-02 05101520253035 Time (hr)Cw (mg/ml) Model prediction 1st release in water 1st release in PBS 2nd release in water 2nd release in PBS Figure 2-7. Comparison of DX release profiles in water and PBS from a PHEMA gel synthesized in a direct entrapment method. M0/ Mg = 1.37mg/g. For release in water, K = 33.35, D (1st release) =1.02 10-11 m2/sec (error=2.55%), and D (2nd release) =1.24 10-11 m2/sec (error=3.88%), for release in PBS, K = 32.39, D (1st release) =9.21 10-12 m2/sec (error=4.47%). and D (2nd release) =1.17 10-11 m2/sec (error=3.66%).

PAGE 54

54 0.0E+00 2.0E-03 4.0E-03 6.0E-03 8.0E-03 1.0E-02 1.2E-02 1.4E-02 1.6E-02 1.8E-02 2.0E-02 01020304050 Time (hr)Cw (mg/ml) Model prediction Mo/Mg=6.82mg/g (error=2.88%) Mo/Mg=6.82mg/g (error=2.86%) Mo/Mg=4.78mg/g (error=1.91%) Mo/Mg=4.78mg/g (error=3.56%) Mo/Mg=2.74mg/g (error=2.63%) Mo/Mg=2.74mg/g (error=3.14%) Mo/Mg=1.37mg/g (error=2.62%) Mo/Mg=1.37mg/g (error=4.56%) Figure 2-8. Comparison of the model prediction and experimental data for DXA release into the fresh water from PHEMA gels synthesized in a direct entrapment method. The solid legends and solid lines represent 1st release and the hollow legends and dashed lines 2nd release. K = 79.82 4.44, D (1st release)=1.29 10-11 3.67 10-13 m2/sec, D (2nd release)=1.08 10-11 1.13 10-13 m2/sec, and Mp/ M0 = 66.4 1.2 % (n=4). Initial drug amount loaded in the gel and rms errors are indicated.

PAGE 55

55 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0E+005.0E+121.0E+131.5E+132.0E+13 t / H2 (sec/m2)Cw/Cw,f Model prediction Thick gel (H=0.13mm) Thin gel (H=0.06mm) Figure 2-9. Effect of gel thickness on DXA rel ease of a PHEMA gel which has been soaked in DXA-ethanol solution. Normali zed DXA release profiles of thin gel and thick gel by final drug concentration in aqueous phase are plotted as a function of t / H2. M0/ Mg = 1.76 mg/g. For thick gel, K = 85.49 and D = 1.12 10-11 m2/sec (error=2.09%), for thin gel K = 85.79 and D = 1.16 10-11 m2/sec (error=2.58%). Half thicknesses of hydrated gels ( H ) are indicated.

PAGE 56

56 0.0E+00 1.0E-03 2.0E-03 3.0E-03 4.0E-03 5.0E-03 6.0E-03 7.0E-03 8.0E-03 9.0E-03 1.0E-02 020406080 Time (hr)Cw (mg/ml) Model prediction Release in water Release in PBS Figure 2-10. Comparison of DXA release profile s in water and PBS from a PHEMA which has been soaked in DXA-ethanol solution. M0/ Mg = 1.76mg/g. For release in water, K = 85.49 and D = 1.12 10-11 m2/sec (error=2.09%), for release in PBS, K = 82.79 and D = 1.28 10-11 m2/sec (error=2.58%).

PAGE 57

57 0.0E+00 2.0E-02 4.0E-02 6.0E-02 8.0E-02 1.0E-01 1.2E-01 1.4E-01 01020304050 Time (hr)Cw (mg/ml) Model prediction Cw,i=0.115mg/ml (error=0.39%) Cw,i=0.086mg/ml (error=0.80%) Cw,i=0.038mg/ml (error=0.39%) Figure 2-11. Comparison of the model predicti on and experimental data for DXP loading by PHEMA gel soaked in DXP solution. K = 3.30 0.15 and D = 1.14 10-12 8.98 10-14 m2/sec (n=3). Initial drug concentration in the aqueous phase and rms errors are indicated.

PAGE 58

58 0.0E+00 1.0E-03 2.0E-03 3.0E-03 4.0E-03 5.0E-03 6.0E-03 05101520253035 Time (hr)Cw (mg/ml) Model prediction Mo/Mg=0.495mg/g (error=1.71%) Mo/Mg=0.343mg/g (error=1.91%) Figure 2-12. Comparison of the model prediction and experimental data for DXP release into the PBS from a PHEMA gels which have been soaked in drug solution. Mean K = 28.98 and mean D =2.12 10-11 m2/sec. Initial drug amount loaded in the gel and rms errors are indicated.

PAGE 59

59 5.0E-01 6.0E-01 7.0E-01 8.0E-01 9.0E-01 1.0E+00 1.1E+00 0.0E+002.0E+124.0E+126.0E+128.0E+121.0E+13 t/H2 (sec/m2)Cw (mg/ml) Theoretical Thick gel (H=0.13mm) Thin gel (H=0.06mm) Figure 2-13. Effect of gel thickness on DXP loading to a PHEMA soaked in DXP solution. Normalized DXP loading profiles for thin gel and thick gel by final drug concentration in aqueous phase are plotted as a function of t / H2. Cw,i = 0.086 mg/mL, For thick gel, K = 3.14 and D = 1.21 10-12 m2/sec (error=0.80%), for thin gel K = 2.89 and D = 1.35 10-12 m2/sec (error=0.23%). Half th icknesses of hydrated gels ( H ) are indicated.

PAGE 60

60 0.60 0.70 0.80 0.90 1.00 1.10 1.20 020406080100120140 Time (hr)Cw/Cw,f Model prediction I = 173 mM I = 594 mM I = 1022 mM I = 1202 mM 0.0 0.2 0.4 0.6 0.8 1.0 1.2 020406080100 Time (hr)Cw/Cw,f Model prediction I = 173 mM I = 594 mM I = 1022 mM I = 1202 mM Figure 2-14. Effect of ionic strength on DXP loading and releas e in a soaking method. A) Normalized DXP loading profiles by final drug concentration in outer solution with different ionic strength. B) Normaliz ed DXP release profiles by final drug concentration in outer solution. Cw,i = 0.086 mg/mL for loading. Ionic strength is indicated. A B

PAGE 61

61 Figure 2-15. Plot of diffusiv ities of DXP for loading and re lease in a soaking method as a function of ionic strength. Cw,i = 0.086 mg/mL. 0.0E+00 5.0E-12 1.0E-11 1.5E-11 2.0E-11 2.5E-11 050010001500 Ionic strength (mM)D (m2/sec) loading release 0.0E+00 5.0E-13 1.0E-12 1.5E-12 050010001500 loading

PAGE 62

62 0.80 0.85 0.90 0.95 1.00 1.05 1.10 1.15 0102030405060 Time (hr)Cw/Cw,f Model prediction I = 173 mM I = 594 mM I = 1022 mM I = 1202 mM 0.0 0.2 0.4 0.6 0.8 1.0 1.2 010203040 Time (hr)Cw/Cw,f Model prediction I = 173 mM I = 594 mM I = 1022 mM I = 1202 mM Figure 2-16. Effect of ionic st rength on timolol maleate loading a nd release in a soaking method. A) Normalized timolol maleate loading profiles by final drug concentration in outer solution. B) Normalized timolol maleate release profiles by fina l drug concentration in outer solution. Cw,i = 0.086 mg/mL for loading. Ioni c strength is indicated. A B

PAGE 63

63 0.0E+00 5.0E-12 1.0E-11 1.5E-11 2.0E-11 2.5E-11 050010001500 Ionic strength (mM)D (m2/sec) loading release Figure 2-17. Plot of diffusivit ies of timolol maleate for loadi ng and release in a soaking method as a function of ionic strength. Cw,i = 0.086 mg/mL. 50.0 55.0 60.0 65.0 70.0 75.0 1E-041E-031E-021E-011E+00 LOG (Molar concentration)Surface tension (mN/m) Figure 2-18. Plot of surface tension of DXP as a function of concentration for determination of CMC.

PAGE 64

64 Figure 2-19. Concentration transi ents of DX in the POLTF at different axial locations for case 1, i.e., no flux to the PLTF. The inset shows a ma gnified view near t = 0. The values P1, P2, and P3 are 0.138, 321.8, and 173.2, respectively. 0 5 10 15 20 25 30 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 Dimensionless TimeC/(Ci/K)x/L=0 0.5 0.95 0 0.2 0.4 0.6 0.8 1 1 x 10-3 0 0.2 0.4 0.6 0.8 1 Dimensionless TimeC/(Ci/K)0.5 0.95 x/L=0

PAGE 65

65 Table 2-1. Concentration a nd partition coefficients ( K ) of DX in a soaking method. No. Cw,f [mg/mL] Cw,i (loading) [mg/mL] M0/ Mg (release) [mg/g] K 1 2 3 4 5 6 7 8 9 10 1 4 10 1 4 10 Loading 0.014 0.015 0.020 0.027 0.028 0.033 0.033 0.035 0.043 0.045 1st release 0.006 0.013 0.018 2nd release 0.003 0.006 0.008 0.026 0.026 0.038 0.051 0.051 0.058 0.058 0.064 0.077 0.077 0.69 1.37 2.11 0.33 0.66 0.90 37.54 35.61 41.65 37.50 40.72 43.38 43.00 39.82 35.83 34.90 38.40 38.91 36.37 36.81 40.69 30.81

PAGE 66

66 Table 2-2. Apparent pa rtition coefficients ( Kapp), intrinsic partition coefficients ( K ), and permanently entrapped drug amount ( Mp) for DX release from the PHEMA gel synthesized in a dir ect entrapment method. No. Cw,f [mg/mL] M0/ Mg [mg/g] Mp/ M0 [%] Kapp K 1 2 3 4 1 2 3 4 1st release 0.011 0.020 0.038 0.053 average 2nd release 0.005 0.009 0.016 0.022 average 1.37 2.74 4.78 6.82 18.0 17.7 15.9 17.8 17.4 50.15 56.52 47.93 49.32 50.98 75.82 79.46 69.11 75.57 74.99 33.35 38.46 33.17 32.22 34.30 Table 2-3. Effect of crosslink density on partitio n coefficient and diffusivities of DX in the gel synthesized in a direct drug entrapping method. crosslink density [mol %] M0/ Mg [mg/g] Mp/ M0 [%] K 1st release K 2nd release K D1st release [m2/sec] D2nd release [m2/sec] 0.24 0.48 1.19 2.38 4.78 4.77 4.71 4.63 16.1 16.4 16.7 18.3 48.64 49.56 50.88 49.39 75.82 79.46 69.11 75.57 33.48 33.65 33.94 32.71 1.12 10-11 1.01 10-11 8.92 10-12 5.87 10-12 1.22 10-11 1.13 10-11 9.08 10-12 6.53 10-12

PAGE 67

67 Table 2-4. Apparent pa rtition coefficients ( Kapp), intrinsic partition coefficients ( K ), and permanently entrapped drug amount ( Mp) for DXA release from the PHEMA gel synthesized in a direct entrapment method. No. Cw,f [mg/mL] M0/ Mg [mg/g] Mp/ M0 [%] Kapp K 1 2 3 4 1 2 3 4 1st release 0.0029 0.0053 0.0095 0.0136 average 2nd release 0.0018 0.0033 0.0061 0.0090 average 1.37 2.74 4.78 6.82 65.81 68.09 66.39 65.42 66.43 310.36 339.17 329.52 330.19 327.31 449.32 498.77 469.17 453.99 467.81 76.29 76.55 80.70 85.76 79.82 Table 2-5. Concentration a nd partition coefficients ( K ) of DXP in a soaking method. No. Cw,f [mg/mL] Cw,i (loading) [mg/mL] M0/ Mg (release) [mg/g] K 1 2 3 2 3 Loading 0.036 0.081 0.107 average 1st release 0.0030 0.0052 average 0.038 0.086 0.115 0.343 0.495 3.33 3.14 3.44 3.30 29.83 28.13 28.98

PAGE 68

68 Table 2-6. Effect of ionic strength ( I ) on DXP loading and release. No. I [mM] Cw,f [mg/mL] Cw,i (loading) [mg/mL] M0/ Mg (release) [mg/g] K D [ 10-12 m2/sec] 1 2 3 4 1 2 3 4 Loading 173 594 1022 1202 Release 173 594 1022 1202 0.0816 0.0797 0.0785 0.0776 0.0031 0.0041 0.0047 0.0047 0.0857 0.0857 0.0857 0.0857 0.343 0.490 0.522 0.635 2.87 4.33 5.30 6.08 19.52 26.49 31.56 42.76 1.44 0.82 0.47 0.39 19.76 12.10 4.53 3.27 Table 2-7. Effect of ionic strength on timolol maleate loading and release. No. I [mM] Cw,f [mg/mL] Cw,i (loading) [mg/mL] M0/ Mg (release) [mg/g] K D [ 10-12 m2/sec] 1 2 3 4 1 2 3 4 Loading 173 594 1022 1202 Release 173 594 1022 1202 0.0779 0.0795 0.0788 0.0788 0.0069 0.0056 0.0058 0.0060 0.0857 0.0857 0.0857 0.0857 0.594 0.472 0.529 0.525 5.66 4.41 4.99 4.95 7.72 6.38 11.31 8.79 5.66 4.79 2.79 2.57 19.62 13.12 12.75 6.90

PAGE 69

69 Table 2-8. Various fractions ( Fc, Fs, and Fp) in the eye for three derivatives of dexamethasone. DX DXA DXP Case 1 (no flux to PLTF) Fc Fs Case 2 (zero concentration in PLTF) Fc Fs Fp 0.8291 0.1737 0.1573 0.8171 0.0486 0.9196 0.0799 0.2912 0.7202 0.0298 0.7893 0.2122 0.0711 0.9068 0.0375

PAGE 70

70 CHAPTER 3 EXTENDED DELIVERY OF OPHTHALMIC DRUGS BY SILICONE HYDROGEL CONTACT LENSES 3.1 Introduction The aim of current study was to develop new s ilicone hydrogel materials that can be used for extended delivery of ophthalmic drugs. This first of a kind study de monstrates that the composition of silicone hydrogels can be tuned to obtain extended drug release, while retaining all other important properties relevant to cont act lens applications including oxygen and ion permeability, mechanical and optical properties, and degree of hydr ation. Additionally we focus on understanding some fundamental i ssues related to transport m echanisms and the relationship between composition and physical a nd transport properties of the si licone hydrogels. It is noted that while our main focus is development of ex tended wear silicone hydro gel contact lenses for drug delivery, we also believe that the materials presented here could be used for drug delivery applications such as puncta plugs, ophtha coil s, retinal implants, transdermal patches, wound healing patches, etc. 3.2 Materials and Methods 3.2.1 Materials N,N-dimethylacrylamide (DMA, 99%), 1-vi nyl-2-pyrrolidone (NVP 99+%), ethylene glycol dimethacrylate (EGDMA), and dexamethas one (DX, 98%) were purchased from SigmaAldrich Chemicals (Milwaukee, WI). Dexamethasone 21-acetate (DXA, 99%), timolol maleate, ( 98%), ethanol ( 99.5%), and Dulbecco’s phosphate buffered saline (PBS) were purchased from Sigma-Aldrich Chemicals (St. Loui s, MO). The macromer bis-alpha,omega(methacryloxypropyl) polydimethylsiloxane (M w 7152, Mn 5460 as measured by gel permeation chromatography) was supplied by Clariant. 3methacryloxypropyltris(tri methylsiloxy)silane (TRIS) was gifted by Silar laboratoies (Scoti a, NY) and 2,4,6-trimethylbenzoyl-diphenyl-

PAGE 71

71 phophineoxide (Darocur TPO) was kindly provided by Ciba Specialty Chemicals (Tarrytown, NY). All the other chemicals were of reagent grad e. All the chemicals were used without further purification. 3.2.2 Preparation of Silicone Hydrogels The silicone hydrogels can be prepared by polymerizing a high ion permeability hydrophilic monomer along with a high oxygen perm eability silicone monomer. Typically, a macromer has to be added to ensure solubili zation of these two components. The silicone hydrogels are non-homogeneous structures, ofte n displaying discerna ble phase separation between a silicone rich phase and a hydrophilic monomer derived phase. We chose DMA as the hydrophilic monomer, TRIS as the h ydrophobic monomer and Bis-alpha,omega(methacryloxypropyl) polydimethylsiloxane as the macromer. Additionally, NVP was added to increase water content and EGDMA was added for controlled crosslinking. Silicone hydrogels of several different compositions were prepared by free radical bulk pol ymerization of the monomers using photoinitiation. To prepare the polymerizing mixt ure, 3 mL of a mixture of TRIS, Macromer, and DMA (amounts of the th ree components are listed in Table 3-1 was combined with 0.18 mL of NVP and 15 l of EGDMA. The mixture was purged with bubbling nitrogen for 15 min. The compositions for GEL5 and GEL6 phase separate and so ethanol was added after purging to aid in so lubilization. The ethanol was pur ged with nitrogen separately before addition. To each monomer mixture, 8 g of photoinitiator Darocur TPO was added with stirring for 5 min and the resulting mixture was immediately injected into a mold which composed of two 5 mm thick glass plates separa ted by a plastic spacer of a desired thickness. The glass plates were coated with thin film of FEP (a polymer of tetrafluoroethylene and hexafloropropylene) to minimize adhesion of the gels to the mold. The spacer thickness was

PAGE 72

72 chosen to be 0.1, 0.2, or 0.4 mm. The mold was then placed on Ultraviolet transilluminiator UVB-10 (UltraLum, Inc.) and the gel was cure d by irradiating by UVB light (305 nm) for 40 min. The molded gel was cut in either square piec es (about 1.5 1.5 cm) with scissors or circular pieces (about 1.65 cm dia.) with a cork borer and dried in air overnight before further use. 3.2.3 Drug Loading The drug was loaded into the gels by soaki ng the gel either in a drug-PBS solution. However, before soaking in the drug-PBS solution, the gel was soaked in ethanol for 3 hours to remove the unreacted monomer and then dried befo re the soaking step. Next the gel was soaked in drug-PBS solution. During soaking of gels the dynamic drug concen tration in PBS was monitored by measuring the absorbance spectra of the solution over the wavelength range of either 220-270 nm for DX and DXA or 261-321n m for timolol (base form) with a UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). The loading step was conducted till equilibrium was reached. The total amount of drug loaded into the gel was determined by finding the total amount of drug-loss from the aqueous solution. UV-VIS absorbance was converted to the corresponding concentration of drug by fo llowing the absorbance spectra deconvolution method reported previously [40]. The drugs DX and DXA are hydrophobic with limite d solubility in water. Furthermore, the silicone hydrogels have a limited swelling in wate r. These combined effects limit the uptake of hydrophobic drugs by silicone hydrogel s through soaking in drug-PB S solutions. The drugs have a much higher solubility in ethanol and the silicone hydrogels that were used in this study typically absorb more than 1.8 mL of ethanol pe r gram of dry gel while they absorb a maximum of 0.4 mL of water per gram dry gel, and so it was decided to use ethano l for loading drugs into the gels. The drugs were loaded by soaking the ge l pieces in 2 or 2.5mL of drug-ethanol solution for 3 hours. During the soaking in drug-ethano l solution, the dynamic concentration in the

PAGE 73

73 ethanol phase was not monitored. At the end of the loading stage the gel was taken out and excess drug-ethanol solution was blotted with wipe s from the surface of the piece. The gels were dried in air overnight, and subsequen tly used for release experiments. 3.2.4 Drug Release Experiments The drug release experiments were carried out by soaking a drug loaded gel (square or circular piece) in PBS. The volume ratio of the PBS to the gel was maintained about 70. During the release experiments, the dynamic drug concentr ation in the PBS was analyzed in the same way described in drug loading experiments. In so me cases, gels were transferred to fresh PBS during release and the process was repeated. These two releases are referred as the 1st and the 2nd release. 3.2.5 Packaging Tests Commercial contact lenses are typically packaged in small pl astic containers that contain about 1-1.5 mL of packaging solution such as PB S. To estimate the effects of packaging on drug release behavior of the silicone hydrogels, packaging tests were conducted. A 0.1 mm thick gel was cut into circular pieces of 1.75 cm diamet er and drug was loaded into gel by soaking in drug-ethanol solution. The drug-loaded gel was dr ied and then stored in 1 mL of PBS as a packaging solution in a sealed vial for 1 to 2 months. At the end of storage in the packaging solution, the gel piece was taken out and soaked in 2 mL of fresh PBS for the subsequent drug release studies. The drug concentration in the packaging soluti on and the dynamic drug concentration in the PBS were analyzed with the procedure described above 3.2.6 Gel Characterization 3.2.6.1 Dynamic mechanical analysis The mechanical properties of gels were an alyzed in tensile mode by using a dynamic mechanical analyzer (DMA Q800, TA instrume nts). A 0.4 mm thick rect angular hydrated gel

PAGE 74

74 was mounted on the tension clamp while submerged in water at room temp erature. A periodic tensile force was applied in th e longitudinal direction with va ried frequency and the response (storage modulus and loss modulus) of gel was de termined. A static preload force of 0.01 N was applied and a 115% of force track was used to keep the sample taut on the tension clamps. Strain sweep tests were conducted at room temperature at 1 Hz to determ ine linear viscoelastic range, and 20 micron strain (0.13% strain) which was conf irmed within linear range by the strain sweep test was chosen for subsequent strain controlled frequency sweep experiments. 3.2.6.2 Ion permeability measurements Ion permeability of gels was measured by m odified ionoflux measur ement technique that described previously by Nicolson et al. [41]. Specifically, a 0.1 mm thick silicone hydrogel was attached to the bottom of donor tube of 14.5 mm inner diameter, which was subsequently filled with 14 mL of 0.1 M NaCl solution. The tube was then placed in a receiving chamber containing 60 mL of DI water. The receiving chamber was well-stirred and placed on the heating plate which maintained the temperature of DI wate r in receiving chamber at about 35 C. The conductivity and temperature of the fluid in the receiving chamber was monitored by a conductivity meter with temperatur e sensor (Con 110 series, OAKTON). It is noted that the gel was fully hydrated in DI water for over 24 hours before getting mounted on the donor tube. 3.2.6.3 Surface contact angle measurements Surface wettability was measured by surface c ontact angle measurement with Drop Shape Analyzer (DSA100, KRSS). A sess ile drop of DI water was depos ited on the fully hydrated gel and contact angle ( ) the drop makes with the gel surface was measured. 3.2.6.4 Transmittance measurements The transparency of the gels was measured by light transmittance tests using UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). A fully hydrated gel of 0.1 mm in

PAGE 75

75 thickness was mounted on one side of outer surf ace of a quartz cuvette. The cuvette was placed in the spectrophotometer and the transmitta nce values were measured at 600 nm. 3.3 Results and Discussion 3.3.1 Timolol Loaded Gel 3.3.1.2 Loading and release of timolol by th e GELS 1 and 4 soaked in PBS solution Figure 3-1 shows the dynamics of timolol loading from drug-PBS solution for two different compositions of gel (G EL1 and GEL 4) and three differe nt initial concentrations of timolol (0.035, 0.070, 0.105 mg/mL). The shape of the profiles looks relatively independent of concentration; there is a brief period of rapid uptake lasting a day, followed by a slower uptake lasting a few weeks. The uptake experiments were stopped at 24 days at which time the concentrations seems to be approaching equi librium. GEL4 takes up about 50% more timolol than GEL1 for the same initial concentration of timolol-PBS solution. The equilibrium uptakes were utilized to determine the partition coefficien ts of drug in the gel. As described previously [40], the partition coefficients for the loading experiment can be calculated by w,f g w,f w,i w w,f g,fC V ) C (C V C C K (3-1) where Vw and Vg are the volumes of the aqueous phase a nd the fully hydrated gel, respectively, and Cg,f Cw,i and Cw,f are the equilibrium concentrations of th e drug in the gel, and the initial and equilibrium concentrations in the aqueous phase respectively, in the loading experiment. The calculated values of partition coefficients ( K ) are listed in Table 3-2. The values of K for both gels are relatively independent of concentrations, perhaps due to the small concentration range explored. The K values about 2 and 3 for GELS 1 and 4, respectively. Timolol has a pKa value of 9.2 [42], which is much highe r than the pH of the PBS which is about 7.4. Consequently, a

PAGE 76

76 majority of timolol is expected to be in the pr otonated form, which is expected to have a very low solubility in the silicone hydrogels, which leads to low K values. After 24 days of drug uptake, the samples of GELS 1 and 4 soaked in the highest concentration drug solution (0.105 mg/mL) were tr ansferred to fresh PBS for release experiment, and the release profiles are plotte d in Figure 3-1. The GEL1 rapi dly releases a bout 25% of the loaded drug in the initial 3 days and then releases the remaining drug slowly for about 80 days. The GEL4 exhibits a similar behavior with about 50% drug loss in the first 3 days, and then a gradual release for about 90 days. The long duration of the drug release from both GELS 1 and 4 is very encouraging but the mass of drug loaded and released is relatively small, and furthermore it took almost a month to load the drug into the gels. It is shown below th at all of these issues can be addressed by loading the drug by soaking the gels in drug-ethanol solutions. 3.3.1.2 Release of timolol from GELS 1-6 after l oading by soaking in drug-ethanol solution Timolol was loaded into the GELS 1-6 by soaking in drug-ethanol solution at a concentration of 0.64 mg/mL for a period of 3 hours. Subsequently timolol release experiments were conducted with 0.1 mm thick gels for each of the six different compositions (GEL1 to GEL6), and the release profiles ar e shown in Figure 3-2. The solid lines in the figure are model fits to the data, and are described later. It is noted that the repeat expe riments are ongoing and so the error bars have been calculated only for tim e less than 90 days. Since the data prior to 90 days is average of measurements from three gels prepared separately an d that after is from a single gel, there is an artificial jump in the profil es. The results clearly show that all the silicone hydrogels release timolol for a long period of time lasting a few months. The amount of drug released is substantially larger than that fo r loading from PBS solution proving the utility of ethanol as a loading medium. The total amount of drug released per weight of gel depends on the

PAGE 77

77 gel composition, with the amount being least for G EL 3, which is the gel with the highest TRIS fraction. Replacing TRIS with the macromer w ithout changing DMA fraction increases the drug loading and the release amounts as evident by comparison of release profiles from GELS 2 and 4. The duration of release is relatively similar for GELS 1-4 and is substantially shorter for GELS 5 and 6. The release profiles for GELS 5 and 6 are almost identical which is expected in view of the identical compositions. It is interesting that the duration of release is relatively independent of the DMA fraction (GELS 1-4) but on increasing it beyond a critical valu e, the release duration changes drastically (GELS 5, 6). This suggests that the morphology of GELS 1-4 is relatively similar and is different from that of GELS 5 a nd 6. The change in morphology likely arose from an increase in the hydrophilic DMA fraction, but the presence of ethanol in the polymerizing mixture may also have played some role. To further understand the contribution of the hydrophilic component (DMA) and the silicone components on drug transport, it was deci ded to prepare DMA, Tris and macromer gels and measure drug transport of timolol through each. These gels were prepared by using a monomer mixture of 3 mL of DMA or TRIS or macromer, al ong with 0.18 mL of NVP and 15 l of EGDMA. Timolol was loaded into these gels by soaking in drug-ethanol solutions at a concentration of 0.64 mg/mL. Ti molol loading and release experi ments were not conducted for pure TRIS gel due to handling difficulty. It shoul d be noted that these gels were 0.2 mm in thickness which is 2 times thicker than the gels in Figure 3-2. Figure 33 shows timolol release by the DMA and macromer gels as a function of time. The macromer gel (GEL8) releases 2.2 mg timolol/g gel over 5 months while DMA gel (G EL7) releases about 10 mg timolol/g gel very rapidly in period of less than 2 hours. This shows that the time scale for transport of timolol in macromer phase is significantly longer than th at in DMA, which is expected due to the

PAGE 78

78 significant differences in hydration of these mate rials. The structure of a silicone hydrogel is complex with possibly multiple phases and based on th e results in Figure 3-3, it is clear that the DMA phase offers negligible resistance to timolo l transport, while the macromer phase offers substantial resistance. Accordingl y, an increase in the DMA conten t leads to a reduction in the total duration of release. The data in Figure 3-3 also shows th at total amount of timolol released from DMA gel is about 5 fold th at from macromer gel, which e xplains the increase in the drug loading as the DMA fraction increases in GELS 1-6. After loading through ethanol, timolol exists as the base in both the silicone and the DMA phases of the silicone hydrogels. On hydration, the base in the DM A phase is protonated and is then quickly transported th rough the DMA phase. However, the DMA phase is likely not continuous due to its small mass fraction, and so the drug needs to convert to the base form, dissolve in the silicone phases, and then slowly diffuse. The slow diffusion through the silicone phase is likely the rate limiting step for transpor t. This speculative mechanism agrees with earlier studies such as by Kajihara, et al. [43]. 3.3.1.3 Rate limiting mechanism for drug transport Since silicone hydrogels have complex morphol ogy, it is plausible that processes other than diffusion, such as kinetics of binding-unbind ing to the phase boundaries and to the polymer could control the transport rates. Alternativ ely, the gel could cont ain fast diffusing DMA channels and the transport mechanism could be slow rate limiting diffu sion across microscopic silicone domains followed by rapid diffusion th rough the channels. The dominant mechanism can be inferred by measuring transport from gels of different thicknesses. If a process is diffusion controlled with homogeneous diffusivity, an incr ease in gel thickness by a factor of 2, while keeping fluid to gel ratio fixed, will increase the ti me scale of release by a factor of 4. A change in time scale by a factor of less than 4 signifies that the adsorption-deso rption process occurs on

PAGE 79

79 a time scale comparable to the diffusion or the transport may be occurring through DMA channels. To confirm that diffusion is indeed the rate limiting step, we pr epared gels of 0.2 mm and 0.4 mm thickness, loaded drug into the gels by ethanol soaking, and then conducted drug release in PBS while maintaining the same gelfluid volume ratio as for the 0.1 mm thick gels. We compare the results from gels of the thre e thicknesses by plotting the amount released vs. scaled time, which is defined as scaled time = 2a time (3-2) where a is a ratio of thickness (in mm) of gel to 0.1 mm, i.e., a is 1 for 0.1 mm thick gel, 2 for 0.2 mm, and 4 for 0.4mm. The timolol release prof iles for GEL1 to GEL4 verses scaled time are shown in Figure 3-4. It is noted that the experiments ar e still continuing for the thickest gels and so the comparison is shown only scaled time of about 50 days. Within the margins of experimental error, the release profiles overlap for all thicknesses for G ELS 1-3 proving that the drug transport is controlled by diffusion for thes e gels. The curves do not overlap for GEL4 but the differences are small and more importantly th e scaled release times match for all gels again suggesting that the transport is diffusion controlled and that the small differences between the curves can be attributed to experimental e rrors. This conclusion along with the arguments presented above, firmly establish that the rate limiting step for drug transport is diffusion through the silicone phases of th e gels, and that the microstructure does not contain DMA channels through which the molecules can diffuse very rapidly. For some of release experime nts discussed above, PBS solution was replaced with fresh PBS when concentration of timolol in the PB S solution reached the measurement limit of UVVIS spectrophotometer which is about 0.13 mg/mL. Interestingly, ther e were no significant change in release rate of tim olol between the end of the 1st release and the beginning of the 2nd

PAGE 80

80 release, which make all the curves smooth for w hole period of release time. The likely reason is because of the very high solubility of timolol in PBS, even a concentration of 0.13 mg/mL corresponds to a perfect sink condition, and so reducing the c oncentration to zero (perfect sink) does not alter the release profiles. The diffusion limited drug release from the gels into PBS can be modeled by assuming that the transport can be described by the diffusion equation, i.e., 2 2y C D t C (3-3) The boundary conditions for the drug release experiment are wKC h y t C y t y C ) ( 0 ) 0 ( (3-4) where h is the half-thickness of the gel, the first boundary condition assumes symmetry at the center of the gel and the s econd boundary condition assumes equilibrium between the drug concentration in the gel and th at in the PBS phase. A mass bala nce on the PBS in the beaker yields h y gel w wy C DA dt dC V 2 (3-5) where Vw is the PBS volume, Agel is the cross-sectional area of the gel, and C is the drug concentration in the release medium. Finally the initial conditions for the drug release experiments are 0 ) 0 ( ) 0 ( t C C t y Cw i (3-6)

PAGE 81

81 Since the fluid volume is much larger than the gel volume the concentration Cw is negligible, and in this perfect sink condition the a bove set of equations can be solved analytically to yield ) 1 ( 1 2 1 ) ( 8 ) 2 ) 1 2 ( cos( ) 1 2 ( 4 ) 1 (2 2 2 2 2 24 ) 1 2 ( 0 2 2 0 4 ) 1 2 ( Dt h n n f gel i w n Dt h n i ne n V V C C e y h n n C C (3-7) Thus the mass released (M) per mass of gel (Mg) is given by ) 1 ( 1 2 1 ) ( 8 ) (2 2 24 ) 1 2 ( 0 2 2 Dt h n n g gel i ge n M V C M t M (3-8) The data in Figure 3-2 was fitted to the above model to determine the diffusivity and the model prediction is compared with experimental data in Figure 3-2. The diffusivities were also calculated by fitting the short time data to Higuchi equation (Figure 3-6). The values of diffusivities obtained by both of the approaches ar e listed in Table 3-3. The fitted values from both approaches are similar which proves that the re lease is truly Fickian. Also the diffusivities are comparable for GELS 1-4 and are much higher for GELS 5 and 6. 3.3.1.4 Impact of concentration in drug-ethan ol solution on drug loading and release It is important to load suffici ent amount of drug in the silic one hydrogels so that these can release drugs at therapeutic doses for an ex tended period of time. The drugs have a high solubility in ethanol and so the drug loading in the gel could eas ily be controlled by manipulating the drug concentration in the drug-ethanol solution. To valid ate this hypothesis, drug was loaded into GELS 1-3 by soaking in drug-ethanol soluti on with 6.4 mg/mL drug concentration. The drug release profiles from these gels are shown in Fi gure 3-5. The thickness of the gels used in these experiments was 0.4 mm and the release profil es are plotted verses scaled time which

PAGE 82

82 corresponds to 0.1 mm thick gels because a typica l contact lens is about 0.1 mm thick. By comparing Figure 3-2 and Fi gure 3-5, we see that increase in the amount of timolol in the drugethanol solution by 10 times resulted in more than 10 times higher in an amount released for any given time. However, interestingly the durati on of drug release has reduced considerably compared to the data shown in Figure 3-2, or in other words, the effective diffusivity has increased. The increase in diffusivity could be caus ed due to alterations in the gel microstructure after soaking in ethanol solution with very high drug loading. It is also noted that the drug is present in the silicone hydroge ls in either free form or th e adsorbed form. The effective diffusivity is the average of th e surface and the bulk diffusivitie s weighted by the fraction of the bound and the free drug, respectively [40]. The fraction of the bound and the free drug depends on the total drug loading and typi cally the ratio of bound to free drug decreases with increasing total loading. It is noted that some drug may also be present as precipitate but that results in further slowdown of drug release with incr eased drug loading and so it is not likely. Furthermore, formation of precipitates may re duce transparency, which was not observed. The bulk diffusivity of the drug is expected to be larger than that of the surfac e diffusivity and so the effective diffusivity is expected to increase with an increase in drug loading, just as observed. This issue has to be taken into account while designing the optimal init ial loading to get the desired flux and the desire d duration of release. 3.3.2 DX or DXA Loaded Gel 3.3.2.1 Loading and release of DX by the gel soaked in PBS solution In the same manner as described in 3.1.1, DX loading experiments were conducted with the 0.1 mm thick gels for the two different composition gels (GEL1 and GEL4) with three different initial concentrations (0.0026, 0.052, 0.0.078 mg/mL). The loading prof iles are shown in Figure 3-7 and the partition coefficients of DX in the gels are listed Table 3-2. The total

PAGE 83

83 equilibrium loading time is about 110 to 130 days for all gels and all init ial concentrations. The total amount of DX uptake is very similar for both of the GEL1 and GEL4 w ith equivalent initial loading. The profiles in Figure 3-7 show a complex behavior with three clearly evident time scales. There is a period of rapi d uptake which lasts about a week, followed by a period of slower uptake that lasts about 60 days, and finally a period of very slow transport near the equilibration. Thus, this behavior is more complex than th at for timolol, and is a signature of complex microstructure of the hydrogels. As seen in Tabl e 3-2, the partition coefficients of DX is about 45-fold of those of timolol, which is expected because of the small solubility of DX in PBS in comparison to that of timolol. 3.3.2.2 Release of DX(or DXA) loaded in the gel by soaking in drug-ethanol solution DX was loaded in the gel by soaking in drug-et hanol solution and subse quently released in PBS. These experiments were done with the gels of two different thicknesses (0.1 and 0.2 mm) for four different composition gels (GEL1, GEL4 GEL5, and GEL6) The results in Figure 3-8 show an extended release lasting more than 200 days for GELS 1 and 4, and lasting about 20 days for GELS 5 and 6. The release durations are longer than that fo r timolol but the total amount of release is much smaller for DX since the partition coefficient of DX in the silicone hydrogel is about 45-fold that of timolol (Table 3-2). The rel ease profiles for GELS 1 and 4 almost overlap while there are some differences in the profiles for GELS 5 and 6. It is noted that GELS 1, 5 and 6 were tran sferred to fresh PBS for 2nd release at times indi cated in the figure. Contrary to the timolol release profiles, PBS replacement significantly alters the release profiles, perhaps because at the instance of the PBS change, the 1st release no longer corresponded to perfect sink condition. The change in the slope of the release profiles at the instant of PBS change is more clearly evident for the data shown in Figure 3-9A for GEL 1.

PAGE 84

84 The effect of thickness of gel on drug release is shown in Figure 3-9. DX releases from two different gels (GEL1 and G EL4) of two different thicknesse s (0.1 mm and 0.2 mm) were conducted, and the release profiles are plotted as a function of s caled time. The release profiles overlap for both gels, which show DX transport in th e gels is also diffusion controlled. However, plots of drug released vs. square root of time (Figure 3-10) do not seem to be straight lines suggesting complicated diffusion m echanisms, or perhaps concentration dependent diffusivities. The release experiments were conducted fo r DXA loaded GEL1 of 0.4 mm thickness and the result is in Figure 3-11. Th e loading and release experiment s through PBS solution were not carried for DXA owing to very low solubility of DXA in aqueous solution. In Figure 3-11, the gel also shows long term release of DXA for a bout 150 days. However, total DXA release is 1.6 mg/g gel, which is lower than timolol loaded ge l and even lower than DX loaded gel for similar concentration of drug loading solu tion. This might be due to hi ghest hydrophobicity of DXA. At drug loading stage, solubility of DXA in ethano l is very high and partit ion coefficient of DXA between the gel and ethanol is small, which results in lower loading of DXA in gel. Moreover, at release stage, partition coefficient of DXA betw een the gel and PBS will be highest among the three drugs. These factors combine to cause th e smallest drug release from DXA loaded by soaking in ethanol am ongst the three drugs. 3.3.3 Packaging Effects A commercial contact lens is stored in packaging solution fo r long period of time before it is applied on an eye. To simulate the packag ing conditions, a 0.1 mm thick GEL1 was loaded with drugs by soaking in drug-ethanol soluti on of 6.4 mg/mL for timolol and 4.8 mg/mL for DXA. The gels were then stored in 1 mL of PBS in a sealed vi al up to 2 months, and then drug release in 2 mL of fresh PBS was conducted (F igure 3-12 and Figure 3-13 for timolol and DXA, respectively). For timolol, about 364 g and 404 g of drug was lost to packaging solution for

PAGE 85

85 1.3 months and 2 months packaging, respectivel y, and the remaining drug was released at an average rate of 0.09 mg/(g of ge l day ) for the first 20 days, and a lower rate for another 2 months. The weights of 20 commercial contact le nses of various types were measured to determine that a typical silicone contact lens weighs about 23 mg. Thus a release rate of 0.09 mg/(g of gel day) corresponds to release of about 2 g/day. The usual dose of timolol is one drop of 0.25% timolol maleate in the affected ey e(s) twice a day [44] and the daily dosage of timolol is 125 g each day if we assume a volume of 25 l for each drop [45]. The therapeutic requirement of timolol can be estimated about 1 g/day from the common knowledge that only 1% of the drug applied topically as eye drops goe s to the cornea. From the transport model of ophthalmic drug delivery by contact lenses [6], about 50% of the applied drug goes into cornea, which suggests that only about 2 g/day of release from a contact lens is required, which is comparable to the release rates from GEL 1 for the first 20 days after insertion. DXA was also released for an extended period of time lasting longer than 90 days at a rate of 0.3 g/day. For DXA packaging, about 8 g of drug was extracted in packagi ng solution for both 1.5 months and 2 months of packaging. The release rates of DXA after packaging are not adequate for therapeutic effects but the exte nded release behavior is still very promising because the drug loading and release can be increased by increasi ng concentration of dr ug in ethanol soaking solution. Furthermore, GELS 5 and 6 release dr ugs at a much higher ra te and so these would certainly be able to deliver th erapeutic dosages, although for a s horter period of about 10-15 days. 3.3.4 Mechanical Properties of Gels The frequency dependent storage moduli (E ) and the loss moduli (E ) of six different composition gels are shown in Figure 3-14 and Figure 3-15. The gels exhibited significant dependency on compositions. As content of TRIS increases, both E and E increase. These

PAGE 86

86 viscoelastic properties also de pend on the frequency and we can clearly see the increase on E and E as frequency increases. Moreover, the gel with higher content of TRIS shows a more significant dependency on the frequency. The modul us of a lens has important consequences on vision correction and safety. A hi gher modulus makes it difficult for the lens to deform and conform to the shape of the cornea, and so the base curves have to be designed to take this issue into account. Also, higher modulus could lead to larger mechanical interactions between the lens and the cornea that could lead to complications such as superior epithelial arcuate lesions (SEALS), which in turn could in itiate bacterial infection. The m odulus of silicone hydrogels is larger than that of hydrogels and the most suitable lens is one th at balances the advantages of silicone hydrogels while not significantly impacting modulus. The modulus of commercial silicone hydrogel lenses lies in the range of 1 1.5 MPa, which is close to the zero frequency intercept of the storage modulus for all the gels. 3.3.5 Ion Permeability, Equilibrium Water Content, and Surface Contact Angle of Gels According to Domscheke et al. [46], ion permeabili ty of contact lenses is a critical variable for lens motion on the eye. In this study, ion pe rmeabilities of six different gels (GEL1 to GEL6) were measured with using ionoflux measuremen t techniques described in the method section. The ionoflux diffusion coefficient, Dion, is determined by follows [41]: dx dc A n Dion / (3-9) where Dion = ionoflux diffusion coefficient (mm2/min), n' = rate of ion transport (mol/min), A = area of ion transport (mm2), dc = concentration difference (mol/mm3), dx = thickness of lens (mm).

PAGE 87

87 The transient conductivity was measured in the receiving chamber and then it was related to the NaCl concentration by using the calibratio n curve. The rate of ion transport was then determined by plotting the total amount of salt in the chamber (mol) vs. time (min) and then determining the slope. The values of Dion for all the gels are listed in Table 3-4. Nicolson et al. [41] reported that a Dion value greater than 6.0 10-6 mm2/min is required to ensure adequate motion of the lens after insertion in the eye. As shown in the table, all ge ls except for GEL3 have sufficient value of Dion i.e., contact lenses made of these gels would not stick to the eye on insertion. The value of Dion decreases as the DMA content decreases, which is reasonable because the ions are expected to move through the hydrated DMA phase. It is interesting that GEL4 shows higher Dion value than GEL2 which has same DMA content as GEL4 and even larger than GEL1 which has higher DMA content. Noting that GEL4 contains the largest content of macromer, we speculate that the macromer enables DMA to distribute more uniformly in the gel resulting in larger effective hydro philic region for the ion transport. The effect of compositions on water content (Q) and surface contact angle () of gel can be seen in Table 3-4. The equilibrium water content (Q) is defined as 100 [%] eq d eqW W W Q (3-10) where Weq is mass of hydrated ge l at equilibrium and Wd is mass of dry gel. In Table 3-4, Q decreases as DMA content decreases for GEL1 to GEL4 and the Q of GEL2 and GEL4 are relatively comparable (4.7% and 5.4%, respectively ) since they have same DMA content in bulk phase. Also GEL5 and GEL6 have significantly high Q compared to the other gels, which could be attributed to the higher DMA content. Interestingly, the water content of GEL 5 is larger than that of GEL 6 even though they both have the same monomer composition and the only difference between the two gels is the amount of ethanol that was adde d to the polymerizing

PAGE 88

88 mixture to help in solubilization of the monomers. This suggests th at the presence of ethanol in the polymerizing mixture may alter the mor phology of the gel. The water content of NIGHT&DAY™ lens (Ciba Vision) is listed on the package as 24%, which is much larger than that for GELS 1-4. Since water content is an im portant parameter for contact lenses, we conclude that amongst all the lenses explor ed here, GEL 5 is perhaps most appropriate for contact lens application. Surface angles listed in Table 3-4 do not show strong dependency on compositions of gels and varies 82 to 92 o. As expected these gels are relativ ely hydrophobic and so contact lenses made from these gels these will need surf ace treatment to make them more wettable. 3.3.6 Transparency of Gels All gels reported in this study have transparencies lager than 90%, and so are suitable for contact lens application (Table 3-4). The gels with the lower DMA content exhibit the highest transparency. GEL4 is a little hazy at the dry state (data no t shown here) but it has highest transmittance (99%) at hydrated gels. GEL5 and GEL6 which contai n about 29% DMA have comparatively lower transmittance of 92 and 94%, re spectively, but these are still visually clear and suitable for contact lens application [47]. 3.4 Conclusions Drug delivery via contact lenses is known to significantly in crease the bioavailability of ophthalmic drugs compared to that via eye drops. While hydrogel contact lenses lead to higher bioavailability, these cannot be used for exte nded drug delivery because these cannot be worn overnight and also these cannot release drugs for an extended period of time. Silicone hydrogel contact lenses can be worn for extended period s but commercial silicon e hydrogel lenses are not suitable for extended drug delivery [36]. This is not surprising because these lenses were developed to optimize other parameters includi ng oxygen and ionic permeability, water content

PAGE 89

89 and modulus. In this paper, we prepared silic one hydrogel lenses of various compositions and explored transport of ophthalmic drugs through thes e materials. Each lens releases drugs for an extended period varying from 10 days to a few months. The transport of timolol and DX in the silicone gels are diffusion limited, which were confirmed by comparing the drug release profiles from gels of three different thicknesses. The release profiles are co mplex particularly for dexamethasone which is evidence of complex mi crostructure of the gels. Also the release duration depends on the total drug loading which is likely due to the effect of total loading on partitioning of the drug between the free and polymer bound forms. For lenses that deliver drugs for months, dr ug loading by soaking in aqueous solutions of drugs is not appropriate, and etha nol or other organic liquids are more suitable mediums for drug loading. The method of using drugethanol solution for drug loading is particularly efficient for hydrophobic drugs since higher drug load ing in a gel can be achieved in comparison to that of using drug-PBS solution when solubility of drug in aqueous solution is limited. The effect of compositions of gels on m echanical properties, ion permeability, surface contact angle, equilibrium water content, and tran smittance of gels were explored. An increase in the TRIS content of the gels increases the storage and loss m odulus leading to stiffer gels. Hydrophilic DMA content enables ions and water to transport though gels, so increase in DMA content leads to higher ion perm eability and equilibrium water cont ent. Moreover, the amount of macromer also impacts ion permeability perhap s by altering the microstructure. The modulus, ion permeability and water content of GEL 5 are most suitable for contact lens applications. These gels also release drugs for about 15-20 days and so these materials may be best candidates for making contact lenses for extended ophtha lmic drug delivery. Th e silicone hydrogels

PAGE 90

90 developed in this study may also be suitable for other drug delivery applications such as puncta plugs, ophtha coils, retinal implants, transd ermal patches, wound healing patches, etc. While these systems are promising, it is noted th at some contact lens wearers prefer to take off lenses at night and soak them in liquids. This issue and other issues such as protein adsorption, tear turn over, mixing in the tear film, etc could potentiall y impact drug delivery dynamics in the eye. Furthermore, the continuous exposure of the corneal tissue to drugs could potentially be toxic. These issues are currently be ing explored and will be presented in a separate publication.

PAGE 91

91 0.00 0.05 0.10 0.15 0.20 0.25 0.30 020406080100 Time (days)Drug uptake or release (mg/g gel) Durg uptake, 0.105 mg/mL Durg uptake, 0.070 mg/mL Durg uptake, 0.035 mg/mL Drug release from gel 0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 020406080100 Time (days)Drug uptake or release (mg/g gel) Drug uptake, 0.105 mg/mL** Drug uptake, 0.070 mg/mL Drug uptake, 0.035 mg/mL Drug release from gel** Figure 3-1. Profiles of timolol uptake and release from 0.1 mm th ick A) GEL1 B) GEL4 in PBS solution. *, ** in legends indicate the gels that were used in the release experiments. The initial concentrations of timolol in the loading solution are indicated in the legend. A B

PAGE 92

92 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 050100150200 Time (days)Drug release (mg/g gel) Model prediction GEL1 GEL2 GEL3 GEL4 GEL5 GEL6 Figure 3-2. Profiles of timolol release from 0.1 mm thick GELS 1-6. Timolol was loaded in the gel by soaking in drug-ethanol solution of 0.64 mg/mL. The solid symbols indicate 1st release and the hollow symbols indicate 2nd release. Data is presented as mean S.D. with n = 3. The solid lines represent model predictions for release by Fickian diffusion into an infinite sink.

PAGE 93

93 0.0 2.0 4.0 6.0 8.0 10.0 12.0 01234567 Time (hours)Drug release (mg/g gel) GEL7 0.0 0.5 1.0 1.5 2.0 2.5 050100150200 Time (days)Drug release (mg/g gel) GEL8 Figure 3-3. Profiles of timolo l release from 0.2 mm thick A) GEL7 B) GEL8. Timolol was loaded in the gel by soaking in dr ug-ethanol solution of 0.64 mg/mL. A B

PAGE 94

94 0.0 0.5 1.0 1.5 2.0 2.5 3.0 050100150200 Scaled time (days)Drug release (mg/gel) 0.1 mm 0.2 mm 0.4 mm 0.00 0.50 1.00 1.50 2.00 2.50 050100150200 Scaled time (days)Drug release (mg/g gel) 0.1 mm 0.2 mm Figure 3-4. Effect of gel thic kness on release from A) GEL1 B) GEL2 C) GEL3 D) GEL4. Timolol was loaded in the gel by soaking in drug -ethanol solution of 0.64 mg/mL. The solid and the hollow symbols indicate the 1st release and the 2nd release, respectively. Data is presented as mean S.D. with n = 3. A B

PAGE 95

95 0.00 0.50 1.00 1.50 2.00 2.50 050100150200 Scaled time (days)Drug release (mg/g gel) 0.1 mm 0.2 mm 0.4 mm 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 050100150200 Scaled time (days)Drug release (mg/g gel) 0.1 mm 0.2 mm 0.4 mm Figure 3-4. Continued C D

PAGE 96

96 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 051015 Scaled time (days)Drug release (mg/g gel) GEL1 GEL2 GEL3 Figure 3-5. Profiles of timolol re lease from three different gels (0.4 mm thick) vs. scaled time. Timolol was loaded in the gel by soaking in drug-ethanol solution of 6.4 mg/mL. The solid and the hollow symbols indicate the 1st release and the 2nd release, respectively. Data is presented as mean S.D. with n = 3. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 01234567 Time0.5 (days0.5)Drug release (mg/g gel) GEL1 GEL2 GEL3 GEL4 GEL5 GEL6 Figure 3-6. Plot of mass of timolol released vers es square root of time. The solid lines are best fit straight lines to the data.

PAGE 97

97 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 050100150200 Time (days)Drug uptake (mg/g gel) 0.078 mg/mL 0.052 mg/mL 0.026 mg/mL 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 050100150200 Time (days)Drug uptake (mg/g gel) 0.078 mg/ml 0.052 mg/ml 0.026 mg/ml Figure 3-7. Profiles of DX uptake by 0.1 mm thic k A) GEL1 B) GEL4 in PBS solution. The initial concentrations of DX for uptake solution are indicated in the legend. A B

PAGE 98

98 0.0 2.0 4.0 6.0 8.0 10.0 12.0 050100150200250 Time (days)Drug release (mg/g gel) GEL1 GEL4 GEL5 GEL6 Figure 3-8. Profiles of DX rel ease from four different compos ition gels (0.1 mm thick). DX was loaded in the gel by soaking in drug-ethan ol solution of 4.99 mg/mL. The solid and the hollow symbols indicate the 1st release and the 2nd release, respectively.

PAGE 99

99 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 050100150200250 Scaled time (days)Drug release (mg/g gel) 0.1 mm 0.2 mm 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 050100150200250 Scaled time (days)Drug release (mg/g gel) 0.1 mm 0.2 mm Figure 3-9. Effect of gel thickness on DX releas e from (a) GEL1 (b) GEL4. DX was loaded in the gel by soaking in drug-ethanol solution of 4.99 mg/mL. The solid and the hollow symbols indicate the 1st release and the 2nd release, respectively. A B

PAGE 100

100 0.0 1.0 2.0 3.0 4.0 5.0 6.0 0.05.010.015.0 Time0.5 (days0.5)Drug release (mg/g gel) GEL1, 1st release GEL1, 2nd release GEL4 GEL5 GEL6 Figure 3-10. Plot of mass of DX released ve rses square root of time (0.1mm thick gels). 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 050100150200 Time (days)Drug release (mg/g gel) 1st release 2nd release Figure 3-11. Profile of DXA release from a 0. 4 mm thick GEL1. DXA was loaded in the gel by soaking in drug-ethanol solution of 4.8 mg/mL.

PAGE 101

101 0.0 0.5 1.0 1.5 2.0 2.5 3.0 020406080100120 Time (days)Drug release (mg/g gel) 1.3 month packaged 2 month packaged Figure 3-12. Profiles of timolol release from the GEL1 packaged for 1.3 and 2 month. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 020406080100120 Time (days)Drug release (mg/g gel) 1.5 month packaged 2 month packaged Figure 3-13. Profiles of DXA release from the GEL1 packaged for 1.5 and 2 month.

PAGE 102

102 0 10 20 30 40 50 60 020406080100 Frequency (Hz)Storage modulus (MPa) GEL1 (2.10 MPa) GEL2 (2.90 MPa) GEL3 (3.15 MPa) GEL4 (1.81 MPa) GEL5 (1.30 MPa) GEL6 (0.85 MPa) Figure 3-14. Dependence of storag e moduli of the gels on frequenc y. Storage modulus at 1Hz is indicated next to each GEL name in the legends. 0 10 20 30 40 50 60 020406080100 Frequency (Hz)Loss modulus (MPa) GEL1 (1.34 MPa) GEL2 (2.72 MPa) GEL3 (3.98 MPa) GEL4 (0.62 MPa) GEL5 (0.50 MPa) GEL6 (0.37 MPa) Figure 3-15. Dependence of loss moduli of th e gels on frequency. Lo ss modulus at 1Hz is indicated next to each GEL name in the legends.

PAGE 103

103 Table 3-1. Compositions of th e monomer mixtures (in mL) for various gels (0.18 mL of NVP and 15 L of EGDMA was added to each compos ition before preparing the gels). TRIS Macromer DMA Ethanol GEL1 GEL2 GEL3 GEL4 GEL5 GEL6 GEL7 GEL8 2.00 2.14 2.40 1.71 1.72 1.72 0.50 0.43 0.30 0.86 0.42 0.42 3.00 0.50 0.43 0.30 0.43 0.86 0.86 3.00 0.3 0.6 Table 3-2. Drug concentrati on and partition coefficients (K) of drug No. Cw,f [mg/mL] Cw,i (loading) [mg/mL] K Timolol GEL1 1 2 3 GEL4 4 5 6 0.101 0.068 0.033 0.100 0.066 0.033 0.105 0.070 0.035 0.105 0.070 0.035 2.06 1.99 2.76 3.11 3.65 3.86 DX GEL1 1 2 3 GEL4 4 5 6 0.035 0.024 0.0094 0.029 0.018 0.0084 0.079 0.052 0.026 0.079 0.052 0.026 99.57 100.54 128.86 141.08 147.10 163.60

PAGE 104

104 Table 3-3. Diffusivities of timolol in gels From straight line fit to short time data in Figure 3-6 From mathematical model Slope of line [mg/gday0.5] R2 D 1016 [m2/sec] Error D 1016 [m2/sec] GEL1 GEL2 GEL3 GEL4 GEL5 GEL6 0.3146 0.2818 0.2142 0.4214 1.3348 1.4306 0.9973 0.9969 0.9979 0.9909 0.9975 0.9933 3.79 3.32 3.00 5.18 27.8 28.0 0.0515 0.0197 0.0426 0.0867 0.0358 0.0566 4.16 3.46 3.25 5.61 27.9 32.6 Table 3-4. Physical properties of gels Ionoflux diffusion coefficients Dion 106 [mm2/min] Contact angle [o] Water content Q [%] Transmittance [%] GEL1 GEL2 GEL3 GEL4 GEL5 GEL6 8.3 6.4 3.2 14.0 70.6 79.9 84.1 1.9 87.9 3.1 91.6 0.3 85.7 0.6 83.6 0.3 85.7 4.0 8.7 4.7 0.9 5.4 18.1 13.6 98.0 98.3 98.3 99.0 92.0 94.1

PAGE 105

105 CHAPTER 4 EXTENDED DRUG DELIVERY FROM SILICONE-HYDROGEL CONTACT LENSES CONTAINING DIFFUSION BARRIERS 4.1 Introduction Vitamin E is powerful antioxidant and in fact it has been shown in some animal studies that the topical application of vitamin E inhibi ts a lot of eye diseas es including keratocyte apoptosis after surgery, ethanol-induced apoptosis in the corneal epithelium, etc. [48, 49]. Especially, there have been a number of in vivo studies suggesting vitamin E retard cataract development [5054]. Also there have been several attempts to develop ophthalmic solutions containing vitamin E [55, 56]. The aim of this study was to explore the possibi lity of controlling release of drug from the commercially available lenses by loading a bioc ompatible nutraceutical vitamin E. Vitamin E was loaded in five different commercially ava ilable extended wear contact lenses by soaking the lens in vitamin E-ethanol solutions. Two di fferent ophthalmic drugs, i.e., timolol and dexamethasone were incorporated into the vitamin E laden lens either by so aking the lens in drug solutions or by direct addition of the drug in th e vitamin E-ethanol solu tions at the vitamin E loading step. The drug releases by th e lens into PBS were measured. 4.2 Materials and Methods 4.2.1 Materials Five commercial silicone contact lenses (diopt er -6.50) are used and described in Table 49. Dexamethasone (DX, 98%), was purchased fr om Sigma-Aldrich Chemicals (Milwaukee, WI). Timolol maleate, (98%), sodium hydroxide pe llets (97+%), ethanol (99.5%), and Dulbecco’s phosphate buffered saline (PBS) were purchased from Sigma-Aldrich Chemicals (St. Louis, MO). Vitamin E (D-alpha tocopherol, Covitol F1370) was gifted by Cognis Corporation. 3methacryloxypropyltris(trimethylsiloxy)silane (TRI S) was kindly provided by Silar laboratoies

PAGE 106

106 (Scotia, NY). All the other chemicals were of reag ent grade. All the chemicals were used without further purification. 4.2.2 Drug Loading into Pure Lenses The commercial silicone contact lenses were rins ed with DI water and then dried in air for further use. Timolol used in this study is ba se form converted from timolol maleate salt as follows. 6 mL of 1.04 M sodium hydroxide was added to 0.12 g of timolol ma leate in a test tube. After gentle shaking for a while, the timolol maleat e forms the base form of timolol at the bottom of the tube at such high pH. To concentrate th e base form, 5 mL of the upper water phase was pipetted out and then all remaining water phase a nd timolol base was transf erred to a petri-dish. The excess water was gently blown out with air stream and then timolol was dried in the air overnight. Next, the drug (DX or tim olol) was loaded into the lenses by soaking the lens either in 2 mL of a drug-PBS solution or in the same volume of a drug-ethanol solution. During the soaking the lens in either solu tion, the dynamic concentration in the solution was not monitored. For soaking in drug-PBS solution, loading st ep was conducted for 24 hours or 7 days. For soaking in drug-ethanol solution, the lens was soaked for 3 hours. At the end of the loading stage the lens was taken out and excess drug-ethanol so lution was blotted with wipes from the surface of the lens. The lens was dried in air overnight, and subsequently used for release experiments. 4.2.3 Vitamin E Loading into Pure Lenses Vitamin E was loaded into lenses by soaking the lens in 3 mL of a vitamin E-ethanol solution for 24 hours. Vitamin E-ethanol solutions of various concentrations were prepared by simply mixing vitamin E and ethanol under moderate magnetic stirring for several minutes. After loading step, the lens was taken out and excess vitamin E-ethanol solution on the lens surface was blotted with wipes and th en dried in air overnight.

PAGE 107

107 4.2.4 Drug Loading into Vitamin E Loaded Lenses The drug was loaded in vitamin E loaded lenses either by directly adding drug in a vitamin E-ethanol solution before soaki ng the pure lens in the solutio n or by soaking the vitamin E loaded lens in a drug-PBS solution. For the case of adding drug in a vitamin E-ethanol solution, the drug was dissolved in 3 mL of a vitamin E-ethanol solution and then the pure lens was soaked in this drug/vitamin E-ethanol for 24 hour s. And next, the lens was taken out and excess solution was blotted with wipes and dried overnight for subsequent release experiments. For the case of soaking in drug-PBS solu tion, the vitamin E loaded lens was soaked in 2 mL of a drugPBS solution until equilibrium. While loading DX in to lenses, changes of drug concentration of soaking solution were monitored. The total amount of DX loaded into the gel was determined by finding the total amount of drug-loss from the aqueous solution by measuring the absorbance spectra of final solution afte r soaking over the wavelength at 241 nm for DX with a UV-VIS spectrophotometer (Thermospectronic Genesys 10 UV). The changes of timolol concentration at the loading stage were not measured since the absorbance of timolol in this concentration range was beyond measurement limit of the UV-VIS spectrometer. 4.2.5 Drug Release Experiments The drug release experiments were carried out by soaking a drug loaded lens in 2 mL of PBS. During the release experiments, the dynami c drug concentration in the PBS was analyzed in the same manner described in drug loading experiments.

PAGE 108

108 4.3 Results and Discussion 4.3.1 Dynamics of Drug By Pure Contact Lenses 4.3.1.1 DX loaded lenses DX releases from five different contact lens es for three different loading methods are shown in Figure 4-1. To see the effect of di fferent loading methods on DX release dynamics, mass of drug released was divided by total drug re leased and then plotted as a function of time. Since DX is hydrophobic drug and has limited sol ubility in PBS, DX-PBS solution of 0.08 mg/mL which is about maximum solubility of DX in PBS at room temperature was used for DX loading into lenses. The concentration of DX-etha nol was the same, i.e., 0.08 mg/mL as that of DX-PBS solution for comparison, though the solubili ty of DX in ethanol about 1 mg/mL. In the figure, ACUVUE ADVANCE™ lens releases DX rapidly showing the same DX release profiles for all three methods and release time fo r 90% of total drug released is about 4.5 hours. O2OPTIX lens release about 90% of total DX re leased for initial 8 hours and then release remaining drug up to 24 hours. This release time by this lens is independent of loading methods (Figure 4-1C). ACUVUE ADVANCE™ and AC UVUE OASYS™ also shows same release profiles for different loading methods. Howe ver, the DX release behaviors by NIGHT&DAY™ and PureVision™ lenses exhibit a little depende ncy on loading methods. For these lenses, there is not much significant difference in release am ount of DX from the lenses soaked in DX-PBS solution for two different soaking times, but they show slower DX release for soaked for 7 days than for 24 hours. This suggests that equilibrium time for DX loading for these two lenses are longer than 24 hours. Among five lenses NIGHT& DAY™ lens shows the longest release time (16 hours) followed by ACU VUE OASYS™ (10.5 hours), O2OPTIX™ (9.5 hours), and PureVision™ (8.5 hours), and then ACUVUE ADV ANCE™ has the shortest (4.5 hours) based on the lenses soaked in DX-PBS solution for 7 days. There is a good correlation between the

PAGE 109

109 water content of the lenses (listed in Table 4-9) and the duration of rele ase as shown in Figure 41F, with increasing water conten t resulting in shorter release dur ations. For total release amount of DX, PureVision™ and ACUVUE OASYS™ lens releases relatively smaller amount (about 28 g and 35 g, respectively) than the ot her three lenses (about 38-41 g). There is no correlation between amount of drugs released and the water content, which is expected because the hydrophobic drugs are expected to partition in the silicon rich phases, and so the partition coefficients in the gels will be influenced by th e silicone composition of the gels. All the lenses soaked in DX-ethanol solution release substantially low amount of DX (2 8 g) which is one magnitude order lower than that released by th e lenses soaked in timolol-PBS solution. The solubility of DX in ethanol is very high and the partition coefficient of timolol between lens and ethanol is very low at drug loading step, and cons equently DX exists a lot more in ethanol phase than in lens polymer which re sults in low loading of DX. 4.3.1.2 Timolol loaded lenses Figure 4-2 shows the dynamics of timolol releas e by each of five different contact lenses soaked in 0.8 mg/mL of timolol-P BS solution or timolol-ethanol solution. It is noted that this concentration of timolol is ten times that of DX for loading into lenses in the previous section. In timolol-PBS solution lenses were soaked for e ither 24 hours or 7 days and for timolol-ethanol solution 3 hours. All the lenses release 90% of to tal released timolol for time less than 1.5 hours. In addition, timolol release profiles for different loading methods overlap for each lens except for PureVision™ lens which shows a little faster re lease by the lens soaked in timolol-ethanol solution than that soaked in PBS medium. AC UVUE OASYS™ lens releases 90% of timolol relatively slow for initial 1.2 hours compared to the other lenses. ACUVUE ADVANCE™ lens exhibits rapid timolol release less than 0.5 hour and the other three le nses show comparable

PAGE 110

110 release duration. The release dura tions of timolol are not correlated to the water content of the lenses. For total amount of drug released is the highest by PureVision™ (about 57 g), the lowest by NIGHT&DAY™ (about 18 g), and those of the other lenses are similar ranging 2630 g based on PBS medium soaking method. The amounts of timolol uptake and release are also uncorrelated to the water co ntent likely due to differences in the hydrophilic components of the lenses, which leads to differences in drug binding to the hydrophilic component rich phases in the lenses. It is interesting that all the lenses so aked in ethanol solution for 3 hours release substantially high total amount of timolol 2.5 3 times more than those soaked in PBS solution. For example, ACUVUE OASYS™ lens soaked in PBS solution for 7 days releases 28 g of timolol, but that soaked in ethanol solution for 3 hours release about 95.7 g during the same release time. The increased uptake of timolol from ethanol soaking is likely due to the fact that timolol does not ionize in ethanol and so it prefer entially binds to the polymer. In PBS, the drug in almost entirely ionized, which leads to a very large solubility in water, and consequently to small binding to the gel. Therefore, 3 hour etha nol soaking method is efficient for higher timolol loading into all four lenses from soaking solution of the same c oncentration, while durations of timolol release are maintained. 4.3.2 Effect of Vitamin E Loading in Water Content of Lenses Vitamin E loadings into each lens for different initial concentration of vitamin E loading solutions are listed in Table 4-10 and Figure 4-3. As seen in the figure, vitamin E loading has linear dependency on initial con centration of vitamin E loadi ng solutions. In addition, among four lenses ACUVUE OASYS ™ takes up highest amount of vitamin E and NIGHT&DAY™ lowest from the vitamin E solution of the same concentration. The vitamin E loaded lenses were transparent for all loadings.

PAGE 111

111 Water contents (Q) of lenses are listed on each lens package and also measured in this study. Water content is determined as follows. Water content 100 ) ( eq ve l eqW W W W Q (4-1) where Weq, Wl, and Wve are mass of hydrated lens at equilib rium, mass of dry pure lens, and mass of vitamin E loaded in the lens, respectively. Both the listed and measured Q’s are shown in Table 4-9. Another way to express water cont ent of lens is equilibrium water content (EW) which is defined as mass of water absorbed by unit mass of pure lens and expressed by followings. Equilibrium water content 100 ) ( l ve l eqW W W W EW (4-2) EW ’s are also listed in Table 4-9. In the table, ACUVUE ADVACNE ™ lenses show very high EW (86.0 2.3) and NIGHT&DAY™ relatively low EW (31.1 5.5). Q and EW were measured also for vitamin E loaded lenses and sown in Table 4-10. The effect of vitamin E loading on Q and EW are clearly seen in Figure 4-4. In Figure 4-4A, water content of vitamin E loaded lenses tends to decrease relatively lin early as vitamin E loading increases. However, Weq of vitamin E loaded lenses increases as vitamin E loading which results decrease in Q values. To exclude the effect of total mass and to see the effect of vitamin E loading only on water amount absorbed in lens polymers, we plotted EW verses vitamin E loading in Figure 4-4B. This also shows decrease in EW for vitamin E loaded lenses compared to pure lenses. EW’s of ACUVUE OASYS™ and PureVision™ lenses linearly decreas e and the values of EW are 46% and 44% respectively for about 20% vitamin loading. Ho wever, for about 10 to 30 % of vitamin E loading, NIGHT&DAY™ and O2OPTIX lenses show relatively constant EW values which are about 24% and about 35%, respectively.

PAGE 112

112 4.3.3 Effect of Vitamin E Loadin g in Size of Contact Lenses Size changes of contact lenses after loadi ng the vitamin E are shown in Figure 4-5 and Table 4-10. The dry diameter increa se is difference of diameter of dry lenses between before and after loading vitamin E and the wet diameter increas es is that of hydrated lenses. As seen in the figure, change in wet diameter is smaller than th at in dry one. The lines in the figure are the best fit straight line to data for all lenses passing zero. Figure 4-5A shows dry diameter change of lenses and the slope of straight line is about 30 %/(g vitamin E/ g pure lens). For example, about 30 % vitamin E loaded lens shows increase about 10% in diameter in dry state, i.e., about 30 % volume change of lens leads to 10% diameter ch ange, which suggests that the expansion of lens by vitamin E loading is isotropic in geometry. In Figure 4-5B, wet diameter change is much less than dry diameter change, for example, lenses with about 30% vitamin E loaded lens expand about only 6.5 % in diameter since vitamin E doe s not take up PBS and only lens polymer does. For practical usage, wet diameter is critical and all the lenses show less than 8 % increase in wet diameter for about 40% of vitamin E, wh ich is likely to be tolerated by eyes. 4.3.4 Effect of Vitamin E Loading on Dynamic of Drug 4.3.4.1 DX-vitamin E loaded lenses Dynamics of DX uptake and release by vitami n E loaded lenses with three different vitamin E loadings are shown in Figure 4-6. The in sets in the figure show the magnified views of the plots for drug release at initial hours. Vitamin E was loaded in the lens first and the lens was dried and then DX was loaded in the vitamin E lo aded lens by soaking the lens in the DX-PBS solutions. The method of loading by direct addi tion of DX in vitamin E-ethanol was not used since DX loading through PBS me dium was much more efficient method for DX loading as shown in previous section. The uptake and rele ase experiments were done for four different lenses except for ACUVUE ADVANCE™ lens because it was too weak to handle after

PAGE 113

113 swelled in vitamin E solution and also AC UVUE ADVANCE™ lens showed most rapid release for both drugs (DX and timolol) as seen in the previous sections. In the figure, all the lenses exhibit increase in loading or release time as vitamin E loading increases. DX loading time is longest for ACUVUE OASYS™, followed by NIGHT&DAY™, O2OPTIX, and then shortest for PureVision™ for similar vitamin E loading. This order maintains for all the vitamin E loading range including no vitamin E loading. Total loading or release time increase for DX by vitamin E loaded lenses is more clearly seen in Figure 4-9A and B, respectively. Total loading (or release) time (T) is the duration in loading (or releas e) of 90% of total drug loaded (or release). T was divided by total loading (o r release) time by pure lens (T0) and then plotted versus vitamin E loading in this figure. For dexame thasone loading, time increases for NIGHT&DAY™ and O2OPTIX are very similar and about 2-fold load ing time increase for about 10% vitamin E loading, about 10-fold for about 30% loading. It is noted that percentage of vitamin E loading is percentage of vitamin E to pure lens in dry weight. The effect of about 10% loading for PureVision™ lens on loading duration is neglig ible and even about 40% loading shows only 6fold increase. These behaviors are similar for DX release time increase. However, increases in release duration are a little le ss than in loading duration, for example, NIGHT&DAY™ lenses with 27% vitamin E release shows 5-fold increase in release duration which is much less effect compared to 10-fold increase in loading durati on with the same vitamin E loading. As for increase in DX uptake duration is speculated to be likely because increase in partition coefficient (K) of DX in vitamin E loaded lenses compared to pure lenses since vitamin E is very hydrophobic in nature, hydrophobic DX can exist more in lenses. If we assume that DX transport is diffusion limited with a diffusivity D, time scale of diffusion through the lens of a half thickness h is Kh2/D, and increase in partition coefficient resu lts in increase in the transport time.

PAGE 114

114 In the table, the values of KVlens tend to increase as vitamin E loading increases for each lens. However, in the figure total amount of DX loaded and released are about same for each lens with various vitamin E loading within the error margins except for ACUVUE OASYS™ lens, which means partition coefficient remains about constant before and after vitamin E loading. Also the products of partition coefficient a nd lens volume were calcu lated since the volume lenses were hard to be determined due to their specific geometry and the values (KVlens) are listed in Table 4-11. However, the Vlens also increases (Table 4-10), which means increase in K value is not significantly high. So there is another caus e for slow down of drug uptake and release. This is discussed in the next section. It is noted that vitamin E was not detected in the release medium, perhaps due to very small solubility of vitamin E in PBS. 4.3.4.2 Timolol-vitamin E loaded lenses Figure 4-7 shows timolol release dynamics by tim olol-vitamin E loaded lenses for four different initial concentration of vitamin E loadi ng solutions. Vitamin E loadings are indicated in the figure and the insets show the magnified vi ews of the plots for dr ug release for initial 8 hours. The release experiments were done for four different lenses. Timolol and vitamin E were loaded into lenses as the same time by soaki ng the lens in 0.8 mg/mL of timolol/vitamin Eethanol solution for 24 hours. Fo r pure lenses (no vitamin E lo ading), timolol was loaded by soaking in timolol/ethanol solution of 0.8 mg/mL for 3 hours. It is clearly seen in the figure that the rate of timolol release by all the lenses decr eases as vitamin E loading increases while total drug release amount does not change significan tly. For comparable vitamin E loadings, ACUVUE OASYS™ shows longest dur ation of timolol release and PureVision™ shortest. For example ACUVUE OASYS™ lens releases about 100 g of timolol for the period of 7.5 hours for 12% vitamin E loading and 30 hours for 25% vitamin E loading, while PureVision™ lens

PAGE 115

115 releases a little higher amount of timolol (about 150 g) for only about 1 hour for all the vitamin E loadings. The effect of vitamin E loading in re lease duration increase compared to pure lens is greatest for NIGHT&DAY™ and there’s minimal ef fect for PureVision™ for whole range of vitamin E loadings as seen in Figure 4-9C. Specifically, NIGHT&DAY™ shows 10-fold release time for 16% vitamin E loading which is release time of about 6 hours, 62-fold for 27% which is 35 hours, and 390-fold for 74% which is 220 hou rs. Total amount of timolol released by NIGHT&DAY™ is lowest such as about 50 g which is likely because the diameter of lens is smallest leading smallest volume of lens among four lenses. O2OPTIX™ with 19% vitamin E loading releases about 90 g of timolol for about 7 hours and 34% vitamin E loading, 36 hours. So 20-30% vitamin loaded lenses such as ACUVUE OASY S™, NIGHT&DAY™, and O2OPTIX™ can deliver timol ol for about 1 day. Timolol was also loaded separately with vitamin E into lenses. First vitamin E was loaded into lenses by vitamin E-ethanol soaking for 24 hours and the lens was dried. And then timolol was loaded into the vitamin E loaded lens by timolol-PBS soaking for 7 days. Timolol release profiles of the NIGHT&DAY™ and O2OPTIX™ lenses are shown in Figure 4-8. It can be clearly seen that this method also increases timolol release duration and for higher vitamin E loading (74% for NIGHT&DAY™ and 99% for O2OPTIX™) also increases release amount of timolol from vitamin E loaded lenses compared to pure lenses. For NIGHT&DAY™ lenses, release rate from the lenses where timolol and v itamin E were separately loaded are almost the same as the one where timolol and vitamin E were loaded at the same time through ethanol solution for comparable vitamin E loading. As seen in the release profiles in this figure, it takes about 220 hours for the NIGHT&DAY™ lens with vi tamin E of 99% loading to release timolol, which means it likely takes much longer than 7 da ys for timolol to be loaded into the lens

PAGE 116

116 through the PBS medium though we load the timolol into the lenses fo r 7 days. It is noted that in Figure 4-7 the timolol and vitamin E were loaded into NIGHT&DAY™ and O2OPTIX™ lens for only 24 hours through ethanol medium and it shows almost the same profiles of timolol release as in Figure 4-8 for comparable vitamin E loadings. And total release amount of timolol is about 90 g for 19% vitamin E loaded O2OPTIX in Figure 4-8C while 24 g for the same vitamin E loading in Figure 4-7C. Therefore, lo ading timolol and vitamin E at the same time through ethanol medium is much more efficient way for preparation of timolol-vitamin E loaded lenses. Increase in timolol release duration can be also likely because increase in partition coefficient of timolol in vitamin E loaded lenses like the case of DX. However, it is noted that the effect of vitamin E loading or release duration increase of timolol is much larger than on that of DX for comparable vitamin E loading. Fo r example, NIGHT&DAY™ with 27% vitamin E loading has 62 times increase in timolol releas e time while it has only 5 times increase in DX even though actual release time is longer fo r DX (90 hours) than timolol (35 hours). O2OPTIX™ with 34% vitamin loading also shows larger incr ease as 36 times for timolol while just 5 times for DX. This is surprising since DX is expected to be more hydrophobic than timolol at pH 7.4 of PBS medium where most of timolol is charged and partition coefficient increase of DX due to vitamin E loading is also expected to be larger than that of timolol, which would lead to larger increase in release duration for DX. The reduction in release rates is thus likely due to presence of vitamin E particles that act as diffusion barrie rs. It is also possible that vitamin E does not form macroscopic aggregates but is present largely adsorbed on the polymer. The effective diffusivity of the drug is the average of the su rface and the bulk diffusivities weighted by the

PAGE 117

117 fraction of the bound and the free drug, respectiv ely [40]. If vitamin E is adsorbed on the polymer surfaces, it can hinder surface diffusion of drug. In an effort to understand the effect of hydrophobic property of vitamin E on timolol transport, it was decided to investigate timolol release behavior by the lens with TRIS loaded which is also very hydrophobic oil. In the same wa y, TRIS and timolol was loaded together into NIGHT&DAY™ lens through ethanol medium. The timolol release profiles by TRIS loaded NIGHT&DAY™ is shown in Figure 4-10. The ti molol release duration increased 2.5 times compared to pure lens. However the duration increas e effect is not as great as vitamin E loaded lens likely because TRIS has a high solubility in the lens and so it does not form a separate phase which is impermeable to the drug. Also about 17% of TRIS loaded in the gel was released into the release medium during the experiment. On th e contrary, vitamin E was not detected in the release medium during the experiments. 4.3.4.3 Scaling theory The increase in release times is likely due to the presence of vitamin E particles inside the gel that act as diffusion barriers leading to an increase in the le ngth of the path that molecules take to diffuse from inside the gel to the fluid reservoir. The path length of the tortuous path l should scale as )) ( 1 (* h, where depends on the microstructure of the vitamin E distribution in the gel, is the volume ratio of vitamin E in the dry gel, and ) (* is the fraction that is present as the vitamin E particles. The fraction *is assumed to be either existing as bound to the polymer gel or as pa rticles but in regions of the gel that do not contribute to drug transport. The time scale for release scales as D l2. The gel thickness increases due to vitamin E

PAGE 118

118 uptake, and assuming isotropic expansion and small vitamin E loading, ) 3 / 1 (0 h h, where h0 is a half thickness of pure lens. Th e time of release thus scales as 2 2 2 01 3 1 ~ D h (4-3) The term 23 1 does not make a significant contributio n to increase in release time as for as large as 1, this term is less than 2. On neglecting this term we get 2 2 2 2 02 1 ~ (4-4) It is noted that equati on (4) is only valid for* The parameters and can be obtained by fitting the data shown in Figure 4-9D to the above model. These parameters for timolol are listed in Table 4-12. The model proposed above is likely not valid for hydrophobic drugs that can partition into the vitamin E phase because for these drugs the transport occurs partially by diffusion around the vitamin E aggreg ates and partially by dissolution and diffusion through these aggregates. Accordingly, the incr ease in release time is much larger for hydrophilic drugs such as timolol compared to hydrophobic drugs such as dexamethasone. 4.3.4.4 Diffusivities of DX or timolo l in vitamin E loaded lenses Contact lenses have a specific geometry such as curvature with variable thicknesses from center to edge depending on power. However, a diam eter of a lens (about 14 mm) is much larger than its thickness (about 100 m) and so we can simplify the geometry of lens as thin flat film with different thickness on the axis perpendicula r to direction of overa ll diffusion. Under this assumption, the mass transfer problem for transpor t in the contact lens can be described by the following equations:

PAGE 119

119 2 2y C D t C (4-5) where C is the drug concentration in the gel, D is the diffusivity and y and t denote the transverse coordinate and time, respectivel y. The boundary conditions for th e drug release experiment are wKC x h y t C y t y C )) ( ( 0 ) 0 ( (4-6) where h is the half-thickness of the gel, which depends on the curved lateral coordinate x, Cw is the drug concentration in the release medium. The first bound ary condition assumes symmetry at the center of the gel and the second boundary condition assumes equilibrium between the drug concentration in the gel and th at in the PBS phase. A mass bala nce on the PBS in the beaker yields dx y C x P D dt dC Vh y S w w ) ( 20 (4-7) where Vw is the PBS volume, P(x) is the perimeter of the lens at the coordinate x, and S is a half of maximum arc length. Finally the initial conditions for th e drug release experiments are 0 ) 0 ( ) 0 ( t C C t y Cw i (4-8) The fluid volume is much larger than lens volum e and the solubility of timolol is very high in PBS of this pH 7.4, which satisfies perfect si nk condition. Even though th e solubility of DX is not as high as timolol in PBS, the release of DX also is at pe rfect sink condition for very short initial time. Under perfect sink conditions, the set of equations listed above can be solved analytically to give the following solution for the concentration profile in the lens: 0 ) ( 4 ) 1 2 (2 2 2) ) ( 2 ) 1 2 ( cos( ) 1 2 ( 4 ) 1 (n Dt x h n i ne y x h n n C C (4-9)

PAGE 120

120 In short time limit, the concentration profile can also be expressed as Dt y h id e C C4 022 (4-10) This result is only valid for times shorter than the Dt x h 4 / ) (. By using Eq.4-7 and Eq.4-10, we obtain the following equation: surface i S i w wA Dt C D dx x P Dt C D dt dC V4 2 ) ( 4 2 20 (4-11) where Asurface is the total surface area of the lens. The above equati on can be integrated to give, w surface i wV A C Dt C 2 (4-12) The fractional release i gel w wC V C V f can thus be expressed as 22 2h Dt V A Dt fgel surface (4-13) where h is the mean thickness of the gel defined as surface gelA V h The above equation is only valid for times shorter than Dt h4 /min, where hmin is the minimum gel thickness, which typically equals the center thickness for negative power contact lenses. Figure 4-11 plots % DX (or timolol) release by vitamin E loaded lenses as a function of square root of time. The lines in the figure are the best fit straight line to short time release data with R2 of 0.98. All the data for both DX and timolol show great fit with straight line, which potentially suggest that the tran sport of both DX and timolol in the vitamin E loaded lens is diffusion controlled. However, further investigat ion such as comparison of release profiles by different thickness lens should be conducted to confirm the diffusi on controlled transport, which

PAGE 121

121 are challenging experiments for commercial contact lenses. The straight lines can be expressed by 100 2 (%)2 h t D M Mt (4-14) where Mt is the mass of drug released at time t, M the mass of drug release as time approaches infinity and for perfect sink condition M = M0 (initial drug loading). Th erefore, from the slope of the straight line in the Figure 4-11, diffusivities of drug in the vitamin E loaded lens can be determined if the mean gel thickness is known. The slope for DX release and timolol release is plotted versus vitamin E loading in Figure 4-12A and B, respectively. As we expected, the slope decreases as vitamin E loading increases for all the lenses for both DX and timolol release, which means ( h D /) decreases. For example, DX release by O2OPTIX lens with 34% vitamin E loading shows slope of about 34% of that for pure lens. The h also increases as vitamin E loading increases and lens expands symmetrically for all expansion axes (isotropic expansion) as mentioned in the previous section and so increase of h might be 7.8% for hydrated O2OPTIX lens with 34% vitamin E loading from Figure 4-5B. Consequently, D is 37% of that of pure lens, i.e., D of DX in the 34% vitamin E loaded O2OPTIX lens is 14% of th at in pure lens. In the same manner, D of timolol in the 34% vitamin E loaded O2OPTIX lens is 4% of that in pure lens. The ratios of diffusivities of DX (or timo lol) in vitamin E loaded lens to pure lens (D/D0) for the other lenses are computed in the same way and listed in Table 4-13. In the table the values of D/D0 for DX and timolol for each lens are sim ilar for comparable vitamin E loading, but it is noted that 0D of timolol is larger than that of DX by an order of magnitude, i.e., D0 of timolol is larger by a factor of about 3, the eff ect of decrease on diffusivity is much larger for timolol.

PAGE 122

122 For two systems with identical geometry bu t different diffusivities, the profiles for fractional drug release versus th e product of diffusivity and tim e are identical. Thus if the introduction of vitamin E leads to a uniform decr ease in diffusivity everywhere in the gel, the fractional release curves with a nd without vitamin E are likely to overlap when plotted against the Dt. It is however noted that in troduction of vitamin E leads to an increase in gel thickness and so there is no exact sca ling to take this into account but it can be accounted for approximately by plotting th e release fraction against t h h D D2 2 0 0. These plots are shown in Figure 4-14 for timolol release from the ACUVUE OASYS™ and NIGHT&DAY™, which are the two lenses with largest effect of vitamin E upt ake on timolol diffusivity. The release curves do not overlap for both lens types. Th is suggests that the diffusivity in vitamin E laden gels of these two types is position dependent with smaller diffusivities near the center of the gel, possibly due to a higher concentration of vitamin E in the ge l center. A non uniform distribution of vitamin E is possible because during the process of ethanol evaporation, vitamin E may diffuse towards the center as ethanol evaporates. Further investigat ions are needed to unde rstand this issue. The dependence of diffusivity on positio n causes the ratios of release ti mes shown in Figure 4-9 to be much larger than D0/D shown in Figure 4-13. 4.4 Conclusions This study focuses on the concept of increasi ng the release duration of ophthalmic drugs from silicone-hydrogel contact lenses by creation of transport barriers. Vitamin E, which is considered as an important ocular nutraceuticals is loaded in the gels by soaking the gels in ethanol-vitamin E solution. On evaporation of et hanol, vitamin E is expect ed to phase-separate. Since hydrophilic drugs such as timolol (charged form) are almost insoluble in vitamin E, these

PAGE 123

123 have to diffuse around the vitamin E aggregates lead ing to an increase in effective path length, and thus an increase in the duration of the drug release. To explore the feasibility of this concept, vitamin E was loaded into five different commercially available silicone contact lenses, a nd transport of dexamethasone and timolol were measured. The drugs were loaded into the lenses by soaking in aqueous solutions for 1 day and 7 days, and also by soaking in drug-ethanol solu tions. Silicone hydrogel cont act lenses released DX and timolol for a period of time less than 16 and 2 hours, respectively, irrespective of the method of drug loading. However, soaking in et hanol-drug solution is useful for rapid loading and also the amount of drug loaded is highe r for ionizable drugs such as timolol. Vitamin E loaded lenses exhibit slower releas e rate of both drugs. Th e ratio of duration of release with and without vitamin E exhibits a qua dratic relationship with vitamin E loading, which is also predicted by a simple scaling theory. The increase in release time is much more for timolol compared to DX, perhaps because DX is partially soluble in vitamin E and so it can diffuse through vitamin E rich regions of the gel. The effect of vitamin E loading on the release rates is qualitatively similar fo r ACUVUE OASYS™, NIGHT&DAY™, and O2OPTIX™. The effect of vitamin E loading in dynamics of both DX and timolol is minimal for PureVision™. ACUVUE OASYS™, NIGHT&DAT™, and O2OTPIX™ with 10-13% vitamin E loading can deliver DX for about 24 hours, and NIGHT&DAY™ with 27% vita min E loading can deliver DX for about 4 days. The amount of drug releas ed could be controlled by manipulating the loading concentrations. For timolol delivery, the effect of vitamin E loading on release duration is highest by NIGHT&DAY™. However, actual timolol release duration is similar for ACUVUE OASYS™ and NIGHT&DAY™ for compar able vitamin E loading since the pure ACUVUE OASYS™ lens releases timolol longest among five lenses. The ACUVUE

PAGE 124

124 OASYS™ and NIGHT&DAY™ lenses can deliver timolol for about 35 hours with vitamin E loadings of about 25% and 27%, respectively, an d for about 6 hours with loadings of 12% and 16%, respectively. The change of lens diameter and equilibrium water content by vitamin E loading was also explored. The results show that there is diamet er increase for vitamin E loaded lenses and the lenses expand isotropically. All the lenses expand similarly and the lenses with about 30% vitamin E loading show about 6.5% of wet diam eter increase. The lenses with higher water content such as ACUVUE OASYS™ and PureVi sion™ experience larger decrease in EW, while EW of NIGHT&DAY™ is not significantly change for vitamin E loading up to 27%. The vitamin E loaded lenses stay transparent for all loadings reported here. During the release experiments, vitamin E was not released in dete ctable amounts into the release medium due to very low solubility of vitamin E in PBS. The results reported here conclusively show that vitamin E loading can substantially increase the release durat ion of drugs, particularly for hydrophi lic drugs. While it is reasonable to assume that the effect is caused by the presence of particles of vitamin E, it is also possible that vitamin E does not form macroscopic aggregates a nd is simply adsorbed on the polymer gel. The surface adsorption could impede surface diffusion of the drug along the polymer leading to a reduction in effective diffusion rates. Furthe r investigations are needed to obtain the microstructure of the vitamin E phases in the gels, and relate it to the drug transport. While vitamin E loading helps in extending dr ug release, it also likely reduces oxygen and ion permeability, which are important parameters for silicone-hydrogel lenses. Thus, clinical studies are needed to determine the impact of vitamin E addition on comfort and lens motion. Also, the release profiles from the vitamin E laden contact lenses are not zero-order and that may

PAGE 125

125 have significant clinician implications. Furthe rmore, addition of vitamin E to prefabricated lenses alters the base curve and the power and so it may be preferable to add vitamin E during the fabrication so that the geometri c parameters of the lens are kept at the desirable values. It is noted that in our studies, vitamin E does not diffuse out of the lenses due to very small solubility in PBS. It may be useful to design lenses which also contain some surfactants that could facilitate diffusion of vitamin E so that the le nses could provide both ophthalmic drugs and the nutraceuticals vitamin E.

PAGE 126

126 0 20 40 60 80 100 120 0102030405060 Time (hr)Drug release, M/Mf (%) DX-PBS for 24 hours (41.0 ug) DX-PBS for 7 days (41.0 ug) DX-ethanol for 3 hours (7.4 ug) 0 20 40 60 80 100 120 0102030405060 Time (hr)Drug release, M/Mf (%) DX-PBS for 24 hours (32.0 ug) DX-PBS for 7 days (34.7 ug) DX-ethanol for 3 hours (5.7 ug) Figure 4-1. Effect of DX loading met hod on profile of DX release by A) ACUVUE ADVANCE™ B) ACUVUE OASYS™ C) NIGHT&DAY™ D) O2OPTIX™ E) PureVision™ contact lenses. F) The plot of (DX release time)-1 versus water content of contact lenses. Drug release (M) divided by total amount released (Mf) are plotted as a function of time. DX was loaded by soak ing the lens in 0.08 mg/mL of indicated medium for indicated duration of time. Total amount of drug released for each lens is marked in parenthesis on the legends. A B

PAGE 127

127 0 20 40 60 80 100 120 0102030405060 Time (hr)Drug release, M/Mf (%) DX-PBS for 24 hours (37.4 ug) DX-PBS for 7 days (41.3 ug) DX-ethanol for 3 hours (2.6 ug) 0 20 40 60 80 100 120 0102030405060 Time (hr)Drug release, M/Mf (%) DX-PBS for 24 hours (37.6 ug) DX-PBS for 7 days (38.6 ug) DX-ethanol for 3 hours (3.7 ug) Figure 4-1. Continued C D

PAGE 128

128 0 20 40 60 80 100 120 0102030405060 Time (hr)Drug release, M/Mf (%) DX-PBS for 24 hours (29.2 ug) DX-PBS for 7 days (27.7 ug) DX-ethanol for 3 hours (3.2 ug) 0.00 0.05 0.10 0.15 0.20 0.25 01020304050 Water content (%)(DX release time)-1 (hr-1) ACUVUE ADVANCE™ ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-1. Continued F E

PAGE 129

129 0 20 40 60 80 100 120 02468 Time (hr)Drug release, M/Mf (%) Timolol-PBS for 24 hours (30.5 ug) Timolol-PBS for 7 days (29.1 ug) Timolol-ethanol for 3 hours (91.5 ug) 0 20 40 60 80 100 120 02468 Time (hr)Drug release, M/Mf (%) Timolol-PBS for 24 hours (26.0 ug) Timolol-PBS for 7 days (28.0 ug) Timolol-ethanol for 3 hours (95.7 ug) Figure 4-2. Effect of timolo l loading method on profile of timolol release by A) ACUVUE ADVANCE™ B) ACUVUE OASYS™ C) NIGHT&DAY™ D) O2OPTIX™ E) PureVision™ contact lenses. Drug release (M) divided by total amount released (Mf) are plotted as a function of time. Timolo l was loaded by soaking the lens in 0.8 mg/mL of indicated medium for indicated duration of time. Total amount of drug released for each lens is marked in parenthesis on the legends. A B

PAGE 130

130 0 20 40 60 80 100 120 02468 Time (hr)Drug release, M/Mf (%) Timolol-PBS for 24 hours (18.0 ug) Timolol-PBS for 7 days (17.7) Timolol-ethanol for 3 hours (46.4 ug) 0 20 40 60 80 100 120 02468 Time (hr)Drug release, M/Mf (%) Timolol-PBS for 24 hours (29.2 ug) Timolol-PBS for 7 days (27.6 ug) Timolol-ethanol for 3 hours (71.8 ug) Figure 4-2. Continued C D

PAGE 131

131 0 20 40 60 80 100 120 02468 Time (hr)Drug release, M/Mf (%) Timolol-PBS for 24 hours (57.4 ug) Timolol-PBS for 7 days (56.2 ug) Timolol-ethanol for 3hours (141.3 ug) Figure 4-2. Continued 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.000.050.100.150.200.250.300.35 Concentraition of soaking solution (g vitamin E/ml ethanol)Vitamin E loading (g vitamin E/g pure lens) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-3. Correlation of vitamin E loading and concentration of soaking solution for different lenses. The lines are the best fit st raight line to data. The slope and R2 of the line are 5.03, 0.9982 (ACUVUE OASYS™), 2.54, 0.9845 (NIGHT&DAY™), 3.24, 0.9910 (O2OPTIX™), 4.35, 0.9998 (PureVision™), respectively. E

PAGE 132

132 0 5 10 15 20 25 30 35 40 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)Q (%) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0 10 20 30 40 50 60 70 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)EW (%) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-4. Plot of A) water content (Q) B) EW of vitamin E loaded lenses versus vitamin E loading. B A

PAGE 133

133 0 5 10 15 20 25 30 35 0.00.20.40.60.81.0 Vitamin E loading (g vitaming E/g pure lens)% Diameter increase (dry) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0 5 10 15 20 25 30 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)% Diameter increase (wet) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-5. Percent increase in di ameter of A) dry lenses B) we t lenses before and after loading vitamin E. Lines are best fit straight lines passing zero to the data. B A

PAGE 134

134 0 10 20 30 40 50 60 70 80 90 100 0100200300400500 Time (hr)Drug uptake & release (g) 0 g vitamin E/g pure lens 0.13 g vitamin E/g pure lens 0.44 g vitamin E/g pure lens 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0 100.0 0100200300400500 Time (hr)Drug uptake & release (g) 0 g vitamin E/g pure lens 0.10 g vitamin E/g pure lens 0.27 g vitamin E/g pure lens Figure 4-6. Profiles of DX uptake and release by vitamin E loaded contact lenses A) ACUVUE OASYS™ B) NIGHT&DAY™ C) O2OPTIX™ D) PureVision™. Solid legends represent uptake and the hollow legends release. Vitamin E was loaded first by soaking pure contact lens in vita min E-ethanol solution and the lens was dried. And then DX was loaded by soaki ng the vitamin E loaded lens in DX-PBS solution (0.08 mg/mL). Some of data are presented as mean S.D. with n = 3. Vitamin E loadings are indicated. A B Drug release 0 10 20 30 40 020406080100 Drug release 0.0 10.0 20.0 30.0 40.0 020406080100

PAGE 135

135 0.0 20.0 40.0 60.0 80.0 100.0 120.0 0100200300400500 Time (hr)Drug uptake & release (g) 0 g vitamin E/g pure lens 0.12 g vitamin E/g pure lens 0.34 g vitamin E/g pure lens 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 050100150200250 Time (hr)Drug uptake & release ( g) 0 g vitamin E/g pure lens 0.13 g vitamin E/g pure lens 0.39 g vitamin E/g pure lens Figure 4-6. Continued 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 0244872 0 10 20 30 40 020406080100 C D Drug release Drug release

PAGE 136

136 0.0 20.0 40.0 60.0 80.0 100.0 120.0 050100150200250 Time (hr)Drug release (g) 0 g vitamin E/g pure lens 0.12 g vitamin E/g pure lens 0.25 g vitamin E/g pure lens 0.44 g vitamin E/g pure lens 0.0 10.0 20.0 30.0 40.0 50.0 60.0 050100150200250 Time (hr)Drug release (g) 0 g vitamin E/g pure lens 0.09 g vitamin E/g pure lens 0.16 g vitamin E/g pure lens 0.27 g vitamin E/g pure lens Figure 4-7. Profiles of timolol release by vitamin E loaded contact lenses. Timolol and vitamin E were loaded together by soaking A) ACUVUE OASYS™ B) NIGHT&DAY™ C) O2OPTIX™ D) PureVision™ contact lens in timolol/vitamin E-ethanol solution (0.8 mg timolol in 1 mL of vitamin E-etha nol solution of various concentrations) for 24 hours. Vitamin E loadings are indicated Timolol was loaded by soaking in timolol/ethanol solution (0.8 mg timolol in 1 mL of ethanol) for 3 hours. Some of data are presented as mean S.D. with n = 3. A B 0.0 10.0 20.0 30.0 40.0 50.0 60.0 02468 0.0 20.0 40.0 60.0 80.0 100.0 120.0 02468

PAGE 137

137 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 0100200300400500600 Time (hr)Drug release (g) 0 g vitamin E/g pure lens 0.10 g vitamin E/g pure lens 0.19 g vitamin E/g pure lens 0.34 g vitamin E/g pure lens 0.95 g vitamin E/g pure lens 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0 180.0 020406080100120140 Time (hr)Drug release (g) 0 g vitamin E/g pure lens 0.11 g vitamin E/g pure lens 0.22 g vitamin E/g pure lens 0.39 g vitamin E/g pure lens Figure 4-7. Continued D 0.0 30.0 60.0 90.0 120.0 150.0 180.0 02468 0.0 20.0 40.0 60.0 80.0 100.0 02468 C

PAGE 138

138 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 0100200300400500 Time (hr)Drug release (g) 0 g Vitamin E/g pure lens 0.16 g Vitamin E/g pure lens 0.74 g Vitamin E/g pure lens 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 0200400600800 Time (hr)Drug release (g) 0 g vitamin E/g pure lens 0.19 g Vitamin E/g pure lens 0.99 g Vitamin E/g pure lens Figure 4-8. Profiles of timolo l release by vitamin E loaded co ntact lenses A) NIGHT&DAY™ B) O2OPTIX™. Vitamin E was loaded first by soaking pure contact lens in vitamin E-ethanol solution and the lens was dried. And then timolol was loaded by soaking vitamin E loaded lens in timolol-PBS so lution (0.8 mg/mL) for 7 days. Vitamin E loadings are indicated. Some of data are presented as mean S.D. with n = 3. A B 0.0 2.0 4.0 6.0 8.0 10.0 12.0 14.0 16.0 18.0 20.0 02468 0.0 5.0 10.0 15.0 20.0 25.0 30.0 02468

PAGE 139

139 0 2 4 6 8 10 12 14 0.00.10.20.30.40.5 Vitamin E loading (g vitamin E/g pure lens)DX uptake time increase, /0 ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0 1 2 3 4 5 6 7 8 0.00.10.20.30.40.5 Vitamin E loading (g vitamin E/g pure lens)DX release time increase, /0 ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-9. Duration increase of A) DX loading B) DX release C) and D) timolol release by vitamin E loaded contact lenses. The lines in (D) are best fit 2nd order polynomial curve to data of each lens. Drug loading (or release) time is the duration in loading (or release) of 90 % of total drug loaded (o r released). The inset is a magnified view for lower vitamin E loading. A B

PAGE 140

140 0 50 100 150 200 250 300 350 400 450 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)Timolol release time increase, /0 ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0 50 100 150 200 250 300 350 400 450 0.00.10.20.30.40.50.6Timolol release time increase, /0 ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-9. Continued C 0 10 20 30 40 50 60 70 0.00.10.20.30.4 D

PAGE 141

141 0.0 10.0 20.0 30.0 40.0 50.0 60.0 0510152025 Time (hr)Drug release (g) 0.67 g TRIS/g pure lens Figure 4-10. Profile of timolo l release by TRIS loaded NIGHT&DAY™ lens. Timolol and TRIS were loaded together by soaking lens in timolol/TRIS-ethanol solution (0.8 mg timolol in 1 mL of 25 vol % TRIS-ethanol solution) for 24 hours. TRIS loadings are indicated.

PAGE 142

142 0 10 20 30 40 50 60 0.02.04.06.08.010.012.014.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.13 g vitamin E/g pure lens 0 5 10 15 20 25 30 35 40 45 0.05.010.015.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.11 g vitamin E/g pure lens 0.27 g vitamin E/g pure lens Figure 4-11. Plot of % drug release by vitamin E loaded lenses versus square root of time. The lines are the best fit straight for short time data of A) DX release B) timolol release by 1) ACUVUE OASYS™ 2) NIGHT&DAY™ 3) O2OPTIX™ 4) PureVision™. All R2’s are larger than 0.98. A-1 A-2

PAGE 143

143 0 5 10 15 20 25 30 35 40 0.05.010.015.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.13 g vitamin E/g pure lens 0.346 g vitamin E/g pure lens 0 5 10 15 20 25 30 0.02.04.06.08.010.012.014.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.13 g vitamin E/g pure lens 0.39 g vitamin E/g pure lens Figure 4-11. Continued. A-3 A-4

PAGE 144

144 0.0 20.0 40.0 60.0 80.0 100.0 120.0 0.05.010.015.020.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.12 g vitamin E/g pure lens 0.25 g vitamin E/g pure lens 0.44 g vitamin E/g pure lens 0 20 40 60 80 100 120 0.05.010.015.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.09 g vitamin E/g pue lens 0.16 g vitamin E/g pure lens 0.27 g vitamin E/g pure lens Figure 4-11. Continued. B-1 B-2

PAGE 145

145 0 20 40 60 80 100 120 0.05.010.015.020.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.09 g vitamin E/g pure lens 0.19 g vitamin E/g pure lens 0.34 g vitamin E/g pure lens 0 20 40 60 80 100 120 0.02.04.06.08.010.012.0 Time0.5 (hr0.5)% Drug release 0 g vitamin E/g pure lens 0.11 g vitamin E/g pure lens 0.22 g vitamin E/g pure lens 0.39 g vitamin E/g pure lens Figure 4-11. Continued. B-3 B-4

PAGE 146

146 0 2 4 6 8 10 12 14 16 18 0.00.10.20.30.40.5Vitamin E loading (g vitamin E/g pure lens)Slope (% hr-0.5) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0 20 40 60 80 100 120 140 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)Slope (% hr-0.5) ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-12. Slope of best fit st raight line to the plot of A) % DX release B) % timolol release versus square root of time for short time by vitamin E loaded lenses (Figure 4-11). All R2’s are larger than 0.98. A B

PAGE 147

147 20 40 60 80 100 120 0.00.20.40.60.81.0 Vitamin E loading (g vitamin E/g pure lens)D0/D ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ Figure 4-13. Plot of D0 /D versus vitamin E loading for timolol release.

PAGE 148

148 0 20 40 60 80 100 120 051015202530 (Dh0 2/D0h2)t (hr)% Drug release 0 g vitamin E/g pure lens 0.12 g vitamin E/g pure lens 0.25 g vitamin E/g pure lens 0.44 g vitamin E/g pure lens 0 20 40 60 80 100 120 0510152025 (Dh0 2/D0h2)t (hr)% Drug release 0 g vitamin E/g pure lens 0.09 g vitamin E/g pue lens 0.16 g vitamin E/g pure lens 0.27 g vitamin E/g pure lens Figure 4-14. Plot of % drug release by vitamin loaded lenses versus (Dh0 2/D0h2)t. A) ACUVUE OASYS™ B) NIGHT&DAY™ cont act lens. Vitamin E loadings are indicated. B A

PAGE 149

149 Table 4-9. List of silico ne containing commercial contact lens (di poter -6.50) used in this study. (n=6) Diameter [mm] Commercial name a (manufacturer) Material a Dry weight measured [mg] Water content, Q measured (listed a) [%] EW measured [%] Wet measured (listed a) Dry measured ACUVUE ADVANCE™ (Johnson&Johnson Vision Care, Inc., Jacksonville, FL) ACUVUE OASYS™ (Johnson&Johnson Vision Care, Inc., Jacksonville, FL) NIGHT&DAY™ (Ciba Vision Corp., Duluth, GA) O2OPTIX™ (Ciba Vision Corp., Duluth, GA) PureVision™ (Bausch&Lomb, Inc., Rochester, NY) Galyfilcon A Senofilcon A Lotrafilcon A Lotrafilcon B Balafilcon A 19.7 0.3 21.7 0.1 22.2 0.3 25.9 0.2 21.0 0.2 46.2 0.7 (47) 36.9 0.9 (38) 23.6 0.3 (24) 31.5 1.3 (33) 35.0 0.7 (36) 86.1 2.3 58.4 1.5 27.3 0.6 46.0 2.7 53.9 1.7 14.40 0.31 (14.0) 14.12 0.26 (14.0) 13.92 0.07 (13.8) 14.43 0.23 (14.2) 14.18 0.15 (14.0) 11.46 0.34 12.18 0.29 12.85 0.15 12.78 0.12 12.49 0.17 a Referred from product packages

PAGE 150

150 Table 4-10. Changes in diameter and wate r content of vitamin E loaded contact lenses before and after loading vitamin E. Diameter increase [%] Contact lenses Vitamin E loading [g vitamin E /g pure lens] Conc. of vitamin E-ethanol solution for soaking [g vitamin E /mL ethanol] dry wet Water content Q [%] EW [%] ACUVUE OASYS™ 0.00 0.12 0.25 0.44 0.000 0.025 0.050 0.090 7.0 8.3 14.0 2.2 2.4 5.8 36.9 28.1 25.4 18.8 58.4 48.6 43.3 33.6 NIGHT&DAY™ 0.00 0.09 0.16 0.27 0.74 0.000 0.025 0.050 0.090 0.300 4.1 6.8 11.1 24.3 1.0 3.2 5.0 18.3 23.6 17.4 15.7 16.5 9.7 31.1 22.5 21.6 24.3 18.5 O2OPTIX™ 0.00 0.10 0.19 0.34 0.95 0.000 0.025 0.050 0.090 0.300 3.1 6.6 15.6 25.0 2.2 5.5 7.8 18.2 31.5 24.9 23.6 21.7 7.5 46.0 34.3 36.2 35.8 16.3 PureVision™ 0.00 0.11 0.22 0.39 0.000 0.025 0.050 0.090 6.4 9.7 12.8 1.4 3.7 6.7 35.0 31.5 26.8 22.8 53.9 52.0 45.8 42.0

PAGE 151

151 Table 4-11. Products of partition co efficient of DX and lens volume (KVlens) for lenses soaked in DX-PBS solution. Contact lenses Vitamin E loading [g vitamin E /g pure lens] KVlens for loading [mL] KVlens for release [mL] ACUVUE ADVANCE™ 0.00 1.3 1.2 ACUVUE OASYS™ 0.00 0.13 0.44 1.1 2.1 1.6 NIGHT&DAY™ 0.00 0.11 0.27 2.6 2.6 3.0 3.0 3.4 4.1 O2OPTIX™ 0.00 0.13 0.34 3.1 3.2 4.2 3.1 4.0 4.8 PureVision™ 0.00 0.13 0.39 6.0 4.0 5.4 6.2 6.1 6.0 Table 4-12. Parameters of scaling theory for timolol Contact lenses ACUVUE OASYS™ NIGHT&DAY™ O2OPTIX™ PureVision™ 0.039 0.096 0.121 0.404 28.04 56.79 36.12 2.62

PAGE 152

152 Table 4-13. Ratio of diffusiv ities of drug in vitamin E lo aded lens to pure lens (D/D0). D / D0 Contact lenses Vitamin E loading [g vitamin E /g pure lens] DX Timolol ACUVUE OASYS™ 0.00 0.12 0.13 0.25 0.44 1.00 0.24 1.00 0.32 0.14 0.09 NIGHT&DAY™ 0.00 0.09 0.11 0.16 0.27 0.74 1.00 0.57 0.15 1.00 0.55 0.29 0.11 0.01 O2OPTIX™ 0.00 0.10 0.13 0.19 0.34 0.95 1.00 0.56 0.14 1.00 0.53 0.24 0.04 0.02 PureVision™ 0.00 0.11 0.13 0.22 0.39 1.00 0.54 0.24 1.00 0.67 0.53 0.14

PAGE 153

153 CHAPTER 5 CONCLUSIONS Drug delivery via contact lenses is known to significantly in crease the bioavailability of ophthalmic drugs compared to that via eye dr ops. The bioavailability of ophthalmic drugs delivered by contact lenses can be estimated by solving the mass transport problem in the eye in the presence contact lenses. In order to solve th e transport model, one requires the parameters that describe the drug transport in the contact lens. In chapter 1, we have investigated the transport of three different forms of dexa methasone in PHEMA gels, which are a common disposable contact lens material. The transport of the drugs is investigated by soaking PHEMA gels in aqueous drug solutions and monitoring the dynamic drug concentrations. After reaching equilibrium, the gels are soaked in fresh solu tions for the release e xperiments. Since DXA has very limited aqueous solubility, it is loaded in to the gel via soaking in ethanol-drug solution. Furthermore drug release studies are also conducted for situations in which the drug was added directly to the polymerizing mixt ure. The equilibrium concentrati ons in both the loading and the release studies are utilized to determine the partition coefficien ts. Furthermore, the dynamic data is fitted to the diffusion equation to determine the mean diffusivity, which includes contributions from both bulk and surface diffusion. The partiti on coefficients of DX and DXA are independent of concentration, and are about 40 and 80, respec tively. The partition coefficients are relatively similar from both the loading and the release stud ies. The partition coefficients estimated from the direct entrapment studies seem to be si gnificantly different but the differences can be attributed to the fact that addition of drug to the polymeri zing mixture results in some irreversible drug entrapment. The irreversible entrapment could be both physical and/or chemical. By utilizing the first, second and thir d release results, it was determined that about 17% of DX and 65% DXA gets irreversibly en trapped, and after taking into account the

PAGE 154

154 irreversible entrapment, the partition coefficients are in reasonable agr eement with the results from soaking and release studies. The transport of all the three drugs is diffusion limited. The diffusivity of DX is 1.05 10-11 m2/sec from both the loading and the release studies, and the values are comparable in release studies with di rectly entrapped drug. The diffusivity for DXA is about 1.29 10-11 m2/sec, and this value too is similar for all uptake and release studies. The partition coefficient of DXP is concentration de pendent; it decreases from about 30 to 3 as the concentration increase from 0.003 to 0.107 mg/mL. The diffusivity of DXP is larger for the release phase than for the loading phase by an order of magnitude, which can be explained by noting that the fitted values represent the average diffusivities, and include contributions from both bulk diffusion and surface diffusion, i.e., ) (f K D fD Ds f Finally we utilized the transport model to pred ict the bioavailability of the three forms of dexamethasone for drug delivery via contact lens es. Amongst the three drugs investigated here, the bioavailability is highest for DXA, primarily due to the highest cornea permeability. Thus DXA delivery via soaked daily disposable PHEM A contact lenses seems like a much more efficient method of delivering dexamethasone to eyes in comparison to delivery through eye drops. While hydrogel contact lenses lead to highe r bioavailability, these cannot be used for extended drug delivery because these cannot be wo rn overnight and also these cannot release drugs for an extended period of time. Silicone hydrogel contact lenses can be worn for extended periods but commercial silicone hyd rogel lenses are not suitable for extended drug delivery. In chapter 3, we developed new silicone hydrogel lenses of various compositions and explored transport of ophthalmic drugs through these materi als. Each lens releases drugs for an extended period varying from 10 days to a few months. The transport of timolol and DX in the silicone

PAGE 155

155 gels are diffusion limited, which were confirme d by comparing the drug release profiles from gels of three different thicknesses. The re lease profiles are complex particularly for dexamethasone which is evidence of complex micros tructure of the gels. Fo r lenses that deliver drugs for months, drug loading by soaking in aqueous solutions of drugs is not appropriate, and ethanol or other organic liquids are more suit able mediums for drug lo ading. The method of using drug-ethanol solution for drug loading is pa rticularly efficient for hydrophobic drugs since higher drug loading in a gel can be achieved in comparison to that of using drug-PBS solution when solubility of drug in aqueous solution is limited. The effect of compositions of gels on m echanical properties, ion permeability, surface contact angle, equilibrium water content, and tran smittance of gels were explored. An increase in the TRIS content of the gels in creases the storage and loss m odulus leading to stiffer gels. Hydrophilic DMA content enables ions and water to transport though gels, so increase in DMA content leads to higher ion perm eability and equilibrium water cont ent. Moreover, the amount of macromer also impacts ion permeability perhap s by altering the microstructure. The modulus, ion permeability and water content of GEL 5 are most suitable for contact lens applications. These gels also release drugs for about 15-20 days and so these materials may be best candidates for making contact lenses for ex tended ophthalmic drug delivery. Chapter 4 focuses on the concept of increasi ng the release duration of ophthalmic drugs from silicone-hydrogel contact lenses by creation of transport barriers. Vitamin E which is considered as an important ocular nutraceuticals is loaded in silicone contact lenses by soaking the lenses in ethanol-vitamin E solution. On ev aporation of ethanol, vitamin E is expected to phase-separate. Since hydrophilic dr ugs such as timolol (charged form) are almost insoluble in vitamin E, these have to diffuse around the vita min E aggregates leading to an increase in

PAGE 156

156 effective path length, and thus an increase in the duration of the drug re lease. To explore the feasibility of this concept, vitamin E was load ed into five different commercially available silicone contact lenses, and transport of dexa methasone and timolol were measured. Vitamin E loaded lenses exhibit slower release rate of bot h drugs. The ratio of dura tion of release with and without vitamin E exhibits a qu adratic relationship with vitami n E loading, which is also predicted by a simple scaling theory. The increa se in release time is much more for timolol compared to DX, perhaps because DX is partia lly soluble in vitamin E and so it can diffuse through vitamin E rich regions of the gel. The eff ect of vitamin E loading on the release rates is qualitatively similar for ACU VUE OASYS™, NIGHT&DAY™, and O2OPTIX™. The effect of vitamin E loading in dynamics of both DX and timolol is minimal for PureVision™. ACUVUE OASYS™, NIGHT&DAT™, and O2OTPIX™ with 10-13% vitamin E loading can deliver DX for about 24 hours, and NIGHT&DAY™ with 27% vita min E loading can deliver DX for about 4 days. The amount of drug releas ed could be controlled by manipulating the loading concentrations. For timolol delivery, the effect of vitamin E loading on release duration is highest by NIGHT&DAY™. However, actual timolol release duration is similar for ACUVUE OASYS™ and NIGHT&DAY™ for compar able vitamin E loading since the pure ACUVUE OASYS™ lens releases timolol longest among five lenses. The ACUVUE OASYS™ and NIGHT&DAY™ lenses can deliver timolol for about 48 hours with vitamin E loadings of about 25% and 27%, respectively, an d for about 10 hours with loadings of 12% and 16%, respectively. The change of lens diameter and equilibrium water content by vitamin E loading was also explored. The results show that there is diamet er increase for vitamin E loaded lenses and the lenses expand isotropically. All the lenses expand similarly and the lenses with about 30%

PAGE 157

157 vitamin E loading show about 6.5% of wet diam eter increase. The lenses with higher water content such as ACUVUE OASYS™ and Pure Vision™ experience larger decrease in EW, while EW of NIGHT&DAY™ is not significantly cha nge for vitamin E loading up to 27%. The vitamin E loaded lenses stay transparent for all loadings reported here. During the release experiments, vitamin E was not released in dete ctable amounts into the release medium due to very low solubility of vitamin E in PBS. The results conclusively show that vitamin E loading can substantially increase the release duration of drugs, particularly for hydrophilic drugs While it is reasonable to assume that the effect is caused by the presence of particles of vitamin E, it is also possible that vitamin E does not form macroscopic aggregates and is simp ly adsorbed on the polymer gel. The surface adsorption could impede surface diffusion of th e drug along the polymer leading to a reduction in effective diffusion rates. Further investigations are needed to obtain the microstructure of the vitamin E phases in the gels, and relate it to the drug transport.

PAGE 158

158 CHAPTER 6 FUTURE WORK 6.1 Drug Transport in PHEMA Copolymer The transport model of dexamethasone derivatives in PHEMA gels presented in chapter 2 reported that the effective diffusivity of the drug is the average of the surface and the bulk diffusivities weighted by the fraction of the bound a nd the free drug, respectively. In the chapter we showed that the contribution of bulk di ffusion and surface diffusion on the drug transport depend on drugs, such as DX, DXA, and DXP and the transport of DXP is likely surface diffusion dominant mechanism according to the resu lt that diffusivities of DXP release are an order of magnitude larger than those of loadi ng with an order of magn itude larger partition coefficients for loading than those for release. In the future we w ill also investigate the dependency of the contribution of bulk diffusi on and surface diffusion as well as partition coefficients of drug on the differe nt gel materials. These fundame ntal investigations can give insight to design hydrogel lenses for controlled drug delivery. 6.2 Silicone Contact Lenses for Ex tend Delivery of Ophthalmic Drug All the silicone hydrogel materi als shown in chapter 3 are ve ry promising for extended drug delivery of ophthalmic drug since all the lens es exhibits release for a very long period time varying several weeks to several months and retaining good mechan ical properties and acceptable ion permeabilities. However, water conten ts of these gels are relatively low compared to current commercial silicone contact lenses, which may cause discomfort for lens wearer. Higher water content can be achieved by increa sing content of hydrophilic component such as DMA in this study. Increase in DMA content with same other components, however, has limitations such as consequent decrease in transmittance and furthermore phase separation. We explored the different macrom er, acryloxy terminated dimethylsiloxane-ethylene oxide block

PAGE 159

159 copolymer which makes it possible to increase DMA content up to 40% with visually clear transparency. However, further manipulation in com positions of gels is needed since theses gels have very weak in ethanol solutions. In addi tion, higher content of hydrophilic components is expected to cause rapid release of drug, which is not suitable for extended drug delivery. We will adopt the method of drug laden particle load ing into PHEMA hydrogels which have been developed for controlled release of ophthalmic drugs in our group and the particles include drug laden microgels, microparticles, micells, microemuls ions, and etc. As mentioned in the chapter, discontinuous lens wear at night and storage in lens solutions as well as other issues such as protein adsorption, tear turn over, mixing in the tear film, etc. can potentially affect drug delivery dynamics in the eye and these issues will be investigated. The transport of drugs in sili cone hydrogel lenses is expect ed to depend on microstructure of the gels in this study and furthermore the sign ature of complex microstructure of lenses for DX transport has been invoked. We plan to obser ve actual microstructures of silicone hydrogels in hydrated state with using cryo-scanning electron microscopy (cryo-SEM) which has been studied preliminary and further investigation will be done. The deposition of silicone lenses with substances from the tear fluid such as proteins, lipids, etc. can result in clinical complications includ ing discomfort and inflammatory responses. The silicone lenses synthesized in this study s how relatively hydrophobic su rface and marginally lower water content, which may lead to lipid depo sition rather than protei n deposition on the lens surface. In vitro or in vivo studies on soiling of these silicone lenses are needed for contact lens application. 6.3 Drug Transport in Commerc ial Silicone Contact Lenses We explored dynamics of timolol and dexa methasone in commercial silicone contact lenses in chapter 4 and they exhibited depende ncy on drug and contact materials. We plan to

PAGE 160

160 explore dynamics of various ophthalmic drugs which have different chemical and physical properties such as hydrophobicity, charge existence in molecule, and molecular weight, etc. and investigate relationship between these propertie s and dynamics to understand transport of these drugs in silicone contact lenses. The drug delivery by the vitamin loaded lens es is very promising since vitamin E is nutraceutical for eye and it substantially increase s release duration of ophthalmic drugs from the lenses by acting as a diffusion barrier for the drug s. By exploring transport of various ophthalmic drug in the vitamin E loaded lenses, transpor t mechanism of drug could be more clearly confirmed. Actual microstructures such as exis tence of vitamin E on the polymer surface will be observed with the cryo-SEM and relate it to drug transport mechanism. Additionally, the vitamin E in the contact lenses likely reduces oxygen and ion permeability and alters mechanical properties. These issues need clinical studies to determine the impact of vitamin E loading on the lens motion in the eye. In the chapter, we obtained the ratio of diffusivities of each drug for vitamin E loaded contact lenses to those for pure lenses. We plan to measure thickness distributions of pure lens with specific geometry of lenses so that by using mathematical model fitting to experimental dynamic data we can determine the diffusivity and partition coefficients of each drug for all the pure or vitamin E loaded contact lenses, which are important parameters for ocular transport model of drug. Animal experiments to evaluate toxicity and in vivo drug release. We do not expect the contact lenses described above to cause any ocul ar toxicity or irritation because the proposed chemicals have been used in medical devices in other tissues including the eye without problems. However, we will test the lenses for ocular t oxicity and irritation. A dditionally, we will conduct

PAGE 161

161 studies to measure dynamic drug release rates from the contact lenses described above in animals.

PAGE 162

162 APPENDIX A MODEL DESCRIPTION FOR DRUG DELIVERY BY SOAKED CONTACT LENSES Figure A-1 shows the real and the model geometry of the lens and the tear film. The postlens tear film (POLTF) is pi ctured as a flat, two-dimensional film bounded by an undeformable cornea and an undeformable but moving contact lens The lens is treated as a two-dimensional body of length L and thickness hg, and is assumed to extend infinitely in the third direction. The post-lens tear film has a thickness hf which may depend on x as the front surface of the eye has a complicated geometry, but for simplicity is taken to be independent of x in this paper. The curvatures of the cornea and the lens have been neglected because the thicknesses of the tear film (about 10 m) and of the contact lens (about 100 m) are much smaller than the corneal radius of curvature of about 1.2 cm. The assumption of a two-dimensional geometry has been made to simplify the problem. The effect of gravity is negligible in the POLTF. Thus, for our purposes the pre-lens tear film-contact lens-POLTF-cornea sy stem is a flat, horizontally oriented channel. These assumptions have been utilized in the past to model mass transfer in the POLTF [6]. The drug concentrations in the gel matrix of the contact lens, and the tear film are Cg and Cf, respectively. The time t = 0 corresponds to insertion of the lens in the eye, and so the initial concentration in the tear film is zero, and the ini tial concentration in the lens is the concentration obtained in the loading phase, which is now defined as Ci. To determine the drug flux to the cornea, we need to simultaneously solve the m odified diffusion equation in the gel matrix and the convection diffusion equation in the pre and the post-lens tear film. By using asymptotic techniques, and multiple ti me scale analysis, the transport problem in the post lens tear film can be simplifie d to a dispersion equation of the form 0*h C k j x C D x t Cf c f f (A-1)

PAGE 163

163 where D* is the effective dispersion coefficient, kc is the permeability of the cornea for the drug and j is the flux of the drug entering the post-lens tear film from the contact lens, which is determined below by solving the transport problem in the lens. The details of this derivation of Eq.A-1 are available in Ref.[6]. Also the e xpression for the dispersion coefficient, which depends on the motion of the contact lens that is caus ed by the blinks, is also available in Ref.[6]. As described in the previous section, the transport problem in the gel is 2 2) (y C D t KC (A-2) Since the value of partition coefficient is cons tant for timolol release in PBS, the above equation can be expressed as 2 2y C D t Cg e g (A-3) where K D De, and Cg = KC is the drug concentration in the gel. The above equation is subjected to the following boundary conditions, 0 ) 0 ( ) ( ) ( y y C j y C D x KC h y Cg g e f g g (A-4a,b,c) The boundary condition Eq.A-4a assumes equili brium between the concentration in the contact lens and that in the tear fluid in th e POLTF and Eq.A-4b imposes flux continuity, and thus couples the mass transfer problems in th e POLTF and in the contact lens. The boundary condition Eq.A-4c assumes that there is no loss of drug from the lens to the pre-lens tear film (PLTF) that lies in between the le ns and the air. This assumpti on may be reasonable because the

PAGE 164

164 PLTF breaks very rapidly and the breakup of the PLTF prevents any further drug loss from the front surface. Additionally, the PLTF breakup causes partial dehydrat ion of the lens in the region close to the front surface, and cons equently the front surface of the c ontact lens is expected to be glassy, which may further reduce drug flux from th e front surface. This is clearly the scenario that will maximize the fraction of the drug trapped in the lens that will even tually be delivered to the cornea. This extreme case in which the drug flux to PLTF is neglected is referred as case 1. To determine the fraction of trapped drug that wi ll go to the cornea for the other extreme, we investigate the case in which we assume that the drug can diffuse into the PLTF and that rapid mixing and drainage from the PLTF keeps the dr ug concentration in PLTF about zero. This case is referred as case 2, and for this case, th e boundary condition Eq.A-A4c is replaced by the following equation 0 ) 0 ( y Cg (A-5) Finally the initial conditions for the tear film and the contact lens are: i g fC t C t C ) 0 ( 0 ) 0 ( (A-6) The dimensionless forms of the equations are: f f fC P3 j P2 C D P1 C ~ ~ ~ ~ (A-7) where 2 g eh t D D D D* ~, L x K C C Ci f/ ~, e f g c f g e g fD h h k P3 h h K P2 L D h D P12 2 2 and g i eh C D j j / ~ is the dimensionless flux from the contact lens into the POLTF. By using convolution theorem [6] the gel problem can be so lved to yield the following expressions for the dimensional flux for cases 1 and 2,

PAGE 165

165 0 0 4 1 2 2 2 0 4 1 22 2 2 24 1 2 ~ 2 2 ~n f ) ( ) n ( f n ) n ( C d e ) n ( C e j (A-8) 0 0 2 2 0 1 2~ ~ 2 ~ 4 ~2 2 2 2n f ) ( n f n f ) n -( C d e n C C e j (A-9) The values of the dimensional parameters required to calculate P1, P2 and P3 are noted in Table A-1. The parameters De and K are the values determined above by fitting the dynamic concentration data in loading a nd release experiments to a model, and the other parameters are obtained from literature. The values of P1, P2 and P3 for the transport of the three drugs are noted in Table A-1.

PAGE 166

166 Figure A-1. Geometry of lens utilized in the model. A) The real geometry for the PLTF-lensPOLTF system. B) The idealized geometry Table A-1. Model parameters for three derivatives of dexamethasone DX DXA DXP K De 1013 [m2/sec] Df 1010 [m2/sec] kc 108 [m/sec] [22] P1 P2 P3 32.18 2.92 4.03 5.06 0.138 321.8 173.2 82.79 1.55 3.64 21.1 0.235 827.9 1364.7 28.98 7.32 3.06 3.87 0.042 289.8 52.9 L [cm] hg [m] hf [m] 1 100 10 B Contact lens POLTF PLTF Cornea hg hf x y Cg Cf A Cornea Pre-lens tear film (PLTF) Drug loaded contact lens

PAGE 167

167 LIST OF REFERENCES [1] Ogura Y. Drug delivery to the posterior segments of the eye. Adv Drug Deliv Rev 2001;52:1-3. [2] Ahmed I, Patton T. Importance of the noncorne al absorption route in topical ophthalmic drug delivery. Invest Ophtha lmol Vis Sci 1985;26:584-587. [3] Ahmed I, Patton T. Disposition of timolol a nd inulin in the rabbit eye following corneal versus non-corneal absorption. Int J Pharm 1987:9-21. [4] McNamara N, Polse K, Brand R, Graham A, Chan J, McKenney C. Tear mixing under a soft contact lens: effects of lens di ameter. Am J Ophthalmol 1999;127:659-665. [5] Creech J, Chauhan A, Radke C. Dispersive mi xing in the posterior tear film under a soft contact lens. Ind Eng Chem Res 2001:3015-3026. [6] Li C, Chauhan A. Modeling ophthalmic drug delivery by soaked contact lenses. Ind Eng Chem Res 2006:3718-3734. [7] Hillman J. Management of acute glaucoma w ith pilocarpine-soaked hydrophilic lens. Br J Ophthalmol 1974;58:674-679. [8] Ruben M, Watkins R. Pilocarp ine dispensation for the soft hydrophilic contact lens. Br J Ophthalmol 1975;59:455-458. [9] Arthur B, Hay G, Wasan S, Willis W. Ultras tructural effects of topical timolol on the rabbit cornea. Outcome alone and in conjunction with a gas permeable contact lens. Arch Ophthalmol 1983;101:1607-1610. [10] Wilson M, Shields M. A comparison of the clinical variations of the iridocorneal endothelial syndrome. Arch Ophthalmol 1989;107:1465-1468. [11] Schultz CL, Nunez IM, Silor DL, Neil ML. Cont act lens containing a leachable absorbed material. US Patent No. 5723131, 1998. [12] Rosenwald, PL, Occular device. US Patent No. 4484922, 1984. [13] Schultz, CL, Mint, JM. Drug delivery system for antiglaucomatous medication. US Patent No. 6410045, 2002. [14] Ward J, Peppas N. Preparation of cont rolled release systems by free-radical UV polymerizations in the presence of a drug. J Control Release 2001;71:183-192. [15] Colombo P, Bettini R, Peppas N. Observati on of swelling process and diffusion front position during swelling in hydroxypropyl methyl cellulose (HPMC) matrices containing a soluble drug. J Contro l Release 1999;61:83-91.

PAGE 168

168 [16] Ende M, Peppas N. Transport of ionizable dr ugs and proteins in cr osslinked poly(acrylic acid) and poly(acrylic acid-co-2-hydroxyethyl methacrylate) hydrogels .2. Diffusion and release studies. J Control Release 1997;47-56. [17] Clark A, Yorio T. Ophthalmic drug di scovery. Nat Rev Drug Discov 2003;2:448-459. [18] Melby J. Drug spotlight program: systemic corticosteroid therapy: pharmacology and endocrinologic considerations. Ann Intern Med 1974;81:505-512. [19] Schwartz B. The response of ocular pressure to corticosteroids. Int Ophthalmol Clin 1966;6:929-989. [20] Urban RJ, Cotlier E. Corticosteroid-induced cataracts. Surv Ophthalmol 1986;31:102-110. [21] Einmahl S, Zignani M, Varesio E, Heller J, Veuthey J, Tabatabay C, et al. Concomitant and controlled release of dexamethasone a nd 5-fluorouracil from poly( ortho ester). Int J Pharm 1999;185:189-198. [22] Civiale C, Bucaria F, Piazza S, Peri O, Mi ano F, Enea V. Ocular permeability screening of dexamethasone esters through combined cellular and tissue systems. J Ocul Pharmacol Ther 2004;20:75-84. [23] Weijtens O, Schoemaker R, Lentjes E, Romijn F, Cohen A, van Meurs J. Dexamethasone concentration in the subretinal fluid afte r a subconjunctival injection, a peribulbar injection, or an oral dos e. Ophthalmology 2000;107:1932-1938. [24] Weijtens O, Schoemaker R, Romijn F, Cohen A, Lentjes E, van Meurs J. Intraocular penetration and systemic absorption after to pical application of dexamethasone disodium phosphate. Ophthalmology 2002;109:1887-1891. [25] Weijtens O, Schoemaker R, Cohen A, Romijn F, Lentjes E, van Rooij J, et al. Dexamethasone concentration in vitreous a nd serum after oral administration. Am J Ophthalmol 1998;125:673-679. [26] Weijtens O, Feron E, Schoemaker R, Cohen A, Lentjes E, Romijn F, et al. High concentration of dexamethasone in aqueous a nd vitreous after subconj unctival injection. Am J Ophthalmol 1999;128:192-197. [27] Weijtens O, van der Sluijs F, Schoemaker R, Lentjes E, Cohen A, Romijn F, et al. Peribulbar corticosteroid injection: vitreal and serum concentrations after dexamethasone disodium phosphate injection. Am J Ophthalmol 1997;123:358-363. [28] Gulsen D, Chauhan A. Ophthalmic drug de livery through contact lenses. Invest Ophth Vis Sci 2004;45:2342-2347. [29] Gulsen D, Chauhan A. Dispersion of microemulsion drops in HEMA hydrogel: a potential ophthalmic drug delivery vehicle. Int J Pharm 2005;292:95-117.

PAGE 169

169 [30] Venkatesh S, Sizemore SP, Byrne ME. Biomimetic hydrogels for enhanced loading and extended release of ocular therapeuti cs. Biomaterials 2007;28:717-724, 2007. [31] 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;26:1293-1298. [32] 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;25:1105-1113. [33] Hiratani H, Mizutani Y, Al varez-Lorenzo C. Controlling drug release from imprinted hydrogels by modifying the characteristics of the imprinted cavities. Macromol Biosci 2005;5:728-733. [34] Alverez-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;91:2182-2192. [35] Hiratani H, Alvarez-Lorenzo C. Timolol upt ake and release by imprinted soft contact lenses made of N,N-diethylacrylamide and methacrylic acid. J Control Release 2002;83:223-230. [36] Karlgard C, Wong NS, Jones L, Moresoli C. In vitro uptake and rel ease studies of ocular pharmaceutical agents by silicon-containing and p-HEMA hydrogel contact lens materials. Int J Pharm 2003;257:141-151. [37] Banga AJ. Electrically assisted transdermal and topical drug deliver y. Bristol: Taylor & Francis, 1998. p. 57. [38] Flynn G, Lamb D. Factors influencing solvolys is of corticosteroid-2 1-phosphate esters. J Pharm Sci 1970;59:1433-1438. [39] Canal, T., Peppas, N.A. Correlation between mesh size and equilibrium degree of swelling of polymeric networks J Biomed Mat Res 1989;23:1183-1193. [40] Kim J, Chauhan A. Dexamethasone transport and ocular delivery from poly(hyroxyethyl methacrylate). Int J Pharm 2008;353:205-222. [41] Nicolson P, Baron RC, Chabrecek P, Court J, Domscheke A, Griesser HJ, et al. Extended wear ophthalmic lens, US Patent No. 5760100, 1998. [42] Jouyban A, Khoubnasabjafari M, Chan HK, Altria KD, Clark BJ. Predicting electrophoretic mobility of beta-blockers in a water-met hanol based electrolyte system. Chromatographia 2003;57:191-195. [43] Kajihara M, Sugie T, Maeda H, Sano A, Fujioka K, Urabe Y, et al. Novel drug delivery device using silicone: Controlled release of in soluble drugs or two ki nds of water-soluble drugs. Chem Pharm Bull 2003;51:15-19.

PAGE 170

170 [44] Timoptic prescribing information, s upplied by MERCK: Whitehouse Station, NJ. (http://www.merck.com/product/usa/pi_c irculars/t/timoptic/timoptic_xe_pi.pdf) [45] Lederer C, Harold RC. Drop size of co mmercial glaucoma medications. Am J Ophthalmol 1986;101:691-694. [46] Domschke A, Lohmann D, Winterton L. On-eye mobility of soft oxygen permeable contact lenses. Proceedings of the ACS Spri ng National Meeting, San Francisco: PMSE, 1997. [47] Liu L, Sheardown H. Glucose permeable pol y (dimethyl siloxane) poly (N-isopropyl acrylamide) interpenetrating networks as ophthalmic biomaterials. Biomaterials 2005;26:233-244. [48] Y lmaz T, Aydemir O, zercan IH, stnda B. Effects of vitamin E, pentoxifylline and aprotinin on light-induced retinal in jury. Ophthalmologica 2007;221:159–166. [49] Bilgihan K, Adiguzel U, Sezer C, Akyol G, Hasanreisoglu B. Effects of topical vitamin E on keratocyte apoptosis afte r traditional photorefractive keratectomy. Ophthalmologica 2001;215:192–196. [50] Ohta Y. Possibility of clini cal application of vitamin E to cataract prevention. J Clin Biochem Nutr 2004;35:35-45. [51] Nagata M, Kojima M, Sasaki K. Effect of vitamin E eye drops on naphthalene-induced cataract in rats. J Ocul Ph armacol Ther 1999;15(4):345-50. [52] Kojiman M, Shui YB, Murano H, Sasaki K. I nhibition of steroid-i nduced cataract in rat eyes by administration of vitamin-E ophthalmic solution Ophthalmic Res 1999;28(Suppl.2):65-71. [53] Ohta Y, Yamasaki T, Niwa T, Majima Y, Is higuro I. Preventive eff ect of topical vitamin E-containing liposome instillation on the pr ogression of galactose cataract. Comparison between 5-weekand 12-week-old rats fed a 25% galactose diet Exp Eye Res 1999;68: 747-755. [54] Ohta Y, Yamasaki T, Niwa T, Majima Y. Preventive effect of vitamin E-containing liposome instillation on cataract progression in 12-month-old rats fe d a 25% galactose. J Ocul Pharmacol Ther 2000;16:323-335. [55] Hofmann, RE, Bottoni, DJ. Glutathione antio xidant eye drops. US Patent No. 5817630, 1998 [56] Braswell AG, Absher K, Duarte A. Liquid eye drop composition. US Patent No. 6194457 B1, 2001.

PAGE 171

171 BIOGRAPHICAL SKETCH Jinah Kim was born in 1972 in Daegu, Sout h Korea. She did her undergraduate study in Chemical Engineering from Pohang University of Science and Technology (POSTECH) in South Korea and graduated in February 1995. And two year later she rece ived her Master of Science degree at POSTECH. After receiving he r Master of Science degree, she worked for Daelim Engineering Co., Ltd. in Seoul, S outh Korea as an engineer, and for POSTECH Environmental Laboratory as a research staff. A nd then she worked as a research engineer for LG Household & Healthcare Cosmetic R&D Institute in Daejon, South Korea. She then joined the department of Chemical En gineering at University of Fl orida in August 2004 and pursued a Ph.D. degree under the guidance of Dr. Anuj Ch auhan. She received a Doctor of Philosophy degree in May 2008. She resides in USA with her husband and a daughter.