Lens Adaptation to Glutathione Deficiency

LENS ADAPTATION TO GLUTATHIONE DEFICIENCY:

IMPLICATIONS FOR CATARACT

JEREMY A. WHITSON

Submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Department of Pathology

CASE WESTERN RESERVE UNIVERSITY

May, 2017 1

Table of Contents

List of Tables 6 List of Figures 7 Acknowledgements 11 List of Abbreviations 12 Abstract 17 1. Introduction 19 1.1. The Crystalline Lens - Biological Glass 20 1.1.1. Anatomy of the Lens 20 1.1.2. Functions of the Lens 21 1.1.3. Lens Cell Biology 22 1.1.4. Transport and Homeostasis in the Lens 23 1.1.5. Conclusions 25 1.2. Age-Related Cataract - A Disease of Protein Aggregation 26 1.2.1. Pathology of Age-Related Cataract 26 1.2.2. Risk Factors for and Impact of Age-Related Cataract 27 1.2.3. Clinical Treatment of Cataract 28 1.2.4. Posterior Capsular Opacification 29 1.2.5. Conclusions 30 1.3. Glutathione - The Guardian of Lens Redox Homeostasis 30 1.3.1. Structure and Biosynthesis of Glutathione 30 1.3.2. Regulation of Cellular Glutathione Content 32 1.3.3. The Essential Role of Glutathione in the Lens 32 1.3.4. Lenticular Glutathione Declines with Age 34 1.3.5. A Loss of Glutathione Activity Leads to Cataract 34 1.3.6. The LEGSKO Mouse Model 35 1.3.7. A Salvage Pathway for Glutathione Synthesis 36 1.3.8. Conclusions 37 1.4. Glutathione Transport – Known Mechanisms 38 1.4.1. Glutathione Transport in the Lens 38 1.4.2. Active Transport Systems 39 1.4.3. Passive Transport 41 1.4.4. Conclusions 42 1.5. Thesis Statement 44 2. Evidence of Dual Mechanisms of Glutathione Uptake 45 in the Rodent Lens: A Novel Role for Vitreous Humor in Lens Glutathione Homeostasis 2.1. Abstract 46 2.1.1. Purpose 46

2.1.2. Methods 46 2.1.3. Results 46 2.1.4. Conclusions 47 2.2. Introduction 47 2.2.1. Antioxidant Functions of Glutathione 47 2.2.2. Lenticular Glutathione Homeostasis 48 2.2.3. The LEGSKO Mouse Model of Cataract 48 2.2.4. Glutathione Transport 49 2.2.5. Purpose of Study 50 2.3. Materials and Methods 50 2.3.1. Chemicals 50 2.3.2. Animal Work 51 2.3.3. LC-MS/MS Analysis 51 2.3.4. Cultured Lens Uptake Experiments 52 2.3.5. Lens Wash Test 53 2.3.6. Cultured Lens GSH Efflux Assay 53 2.3.7. GSH Uptake Imaging 54 2.3.8. In Vivo Uptake Experiments 55 2.3.9. Vascular Eye Perfusion 56 2.3.10. Lens Surface Permeability Assay 56 2.3.11. Tissue Procurement and Dissection 57 2.3.12. Statistical Analysis 58 2.4. Results 58 2.4.1. Lenticular LEGSKO GSH Is Supplied by Circulating 58 GSH 2.4.2. Salvage Pathway Is Not the Source of LEGSKO Mouse 60 Lens GSH 2.4.3. Cultured Lens GSH Uptake Occurs by Passive Diffusion 62 2.4.4. Lens GSH Uptake From Aqueous Humor Is 72 Carrier-Mediated 2.4.5. Mouse Vitreous Contains High Levels of GSH 75 2.4.6. Vitreous but Not Aqueous Contributes Highly to Lens 77 GSH In Vivo 2.4.7. High Vitreous GSH Content Is Unique to Small Animals 81 2.5. Discussion 84 2.5.1. Evidence of Vitreous GSH Transport From Other Studies 84 2.5.2. Uptake From Vitreous but Not Aqueous Is Sufficient to 85 Maintain Steady State GSH Concentration in LEGSKO Lenses 2.5.3. Vitreous GSH Flows Into Lenses but Not Vice Versa 86 2.5.4. Vitreous GSH Varies Greatly Among Species 87

2.5.5. Role of Vitreous GSH in Health and Disease 88 2.5.6. Conclusions 89 2.6. Acknowledgements 90 3. Transcriptome of the GSH-Depleted Lens Reveals 91 Changes in Detoxification and EMT Signaling Genes, Transport Systems, and Lipid Homeostasis 3.1. Abstract 92 3.1.1. Purpose 92 3.1.2. Methods 92 3.1.3. Results 92 3.1.4. Conclusions 93 3.2. Introduction 93 3.2.1. The Role of Glutathione In Age-Related Cataract 93 3.2.2. Analysis of the LEGSKO Mouse 94 3.2.3. Purpose of Study 95 3.3. Materials and Methods 95 3.3.1. Animal Work 95 3.3.2. Sample Preparation 96 3.3.3. RNA-Seq 96 3.3.4. Cluster Generation and HiSeq Sequencing 97 3.3.5. Bioinformatic Analysis 98 3.3.6. Validation of RNA-Seq Data by Real Time PCR (qPCR) 98 3.3.7. Statistical Analysis 100 3.4. Results 100 3.4.1. Overview of Significant Gene Expression Changes 100 Resulting from GSH Depletion in the Lens 3.4.2. RNA-Seq Data Shows Strong Differential Expression of 103 Lens Epithelia and Fiber Cell Marker Genes and Excellent Agreement with qPCR Results 3.4.3. GSH Depletion Results in a Partial Downregulation of β- 106 and γ-Crystallins 3.4.4. Trends Identified in Upstream Regulators and 108 Molecular/Cellular Functions of Genes with Modulated Expression 3.4.5. Expression of Several Detoxification Genes Increases as a 111 Function of Decreasing Lenticular GSH Content 3.4.6. Several Transport Systems Are Modulated as a Function of 113 Decreasing Lenticular GSH Content 3.4.7. An Array of Lipid Metabolism Genes Show Modulated 116 Expression as a Function of Decreasing Lenticular GSH

Content 3.4.8. Lens GSH Depletion Induces Gene Expression Changes 118 Related to Epithelial-Mesenchymal Transition (EMT) Pathways. 3.5. Discussion 130 3.5.1. Interpretation of Data 130 3.5.2. The GSH-Deficient Lens Transcriptome Shows an Unusual 131 Oxidative Stress Response and Indicates Protective Detoxification Genes 3.5.3. The GSH-Deficient Lens Shows Numerous Changes in 134 Transport Systems 3.5.4. GSH-Deficient Lenses Have Altered Lipid Metabolism 136 3.5.5. GSH-Deficient Lenses Show Activation of EMT signaling 138 and a Loss of Differentiation 3.5.6. Analogies and Differences from Other Models of 139 Oxidative Stress and GSH-Depletion 3.5.7. Limitations of this Study 142 3.5.8. Conclusions 142 4. Analysis of Downstream Metabolic Effects of GSH-Dependent 143 Transcriptomic Changes in the Lens 4.1. Introduction 144 4.2. Materials and Methods 144 4.2.1. Animal Work 144 4.2.2. Lens Total Iron Assay 144 4.2.3. Urea Assay 145 4.2.4. BCA Normalization 145 4.2.5. Statistical Analysis 146 4.3. Results 146 4.3.1. Lens Total Iron Content 146 4.3.2. Urea Within and Surrounding the Lens 147 4.4. Discussion 148 4.4.1. Can Lens Fiber Cells Undergo Ferroptosis? 148 4.4.2. Urea Homeostasis in the Lens 149 4.4.3. Conclusions 150 5. Discussion 151 5.1. Summary of Findings 152 5.1.1. The Source and Mechanism of Residual GSH Levels in 152 the LEGSKO Lens 5.1.2. The Major Consequences of Glutathione Deficiency in 153 the Lens

5.1.3. Conclusions 153 5.2. Future Directions and Open Questions 153 5.2.1. Glutathione as a Potential Therapeutic Agent 153 5.2.2. The Source of Vitreous Glutathione 155 5.2.3. Identity of the Lens Epithelia Glutathione Transporter 156 5.2.4. The Glutathione-Depleted Lens Proteome 157 5.2.5. The Potential Protective Role of GSTK1 in the Lens 157 5.2.6. Conclusions 159 5.3. Concluding Remarks 159 6. Appendix 160 References 191

List of Tables Table 1.1 Risk factors for age-related cataract development. 27

Table 2.1. LC-MS/MS MRM Settings. 52

Table 3.1. Top Upstream Regulators and Molecular and Cellular 109

Functions of Gene Expression Changes in GSH-Deficient Lenses.

Table 3.2. Comparison of Overlapping Gene Expression Changes in 128

GSH-Deficient and EMT Lens (Medvedovic et al., 2006) Transcriptomes.

Table 6.1. All significant (P < 0.05; FDR < 0.1) gene expression changes 178 in LEGSKO and BSO-treated LEGSKO lenses compared to WT.

List of Figures Figure 1.1. Anatomy of the lens and its local environment. 21

Figure 1.2. Chemical structure of glutathione. 31

Figure 1.3. Diagram of glutathione transport, function, and metabolism. 43

Figure 2.1. Purity of synthesized GS-B tested by LC-MS/MS analysis. 54

Figure 2.2. Analysis of the source of LEGSKO lens GSH. 59

Figure 2.3. Analysis of the mechanism of LEGSKO lens GSH uptake. 61

Figure 2.4. Comparison of GSH uptake in cultured lenses. 62

Figure 2.5. Time course of uptake in cultured lenses. 63

Figure 2.6. Initial rate curves of uptake in cultured lenses. 65

Figure 2.7. Temperature dependence of transport in cultured lenses. 66

Figure 2.8. Test for competitive inhibition of uptake. 67

Figure 2.9. Wash test to assess cellular localization of 68

13 15 GSH-(glycine- C2, N)

Figure 2.10. Measurement of cultured lens GSH efflux. 69

13 15 Figure 2.11. Comparison of GS-B and GSH-(glycine- C2, N) 71 uptake rate.

Figure 2.12. Visualization of GS-B uptake in cultured lenses. 72

Figure 2.13. Comparison of initial uptake rates from anterior chamber 73 in WT and LEGSKO. lenses.

Figure 2.14. Test for competitive inhibition of uptake from the anterior 74 chamber.

Figure 2.15. Permeability of lens surfaces. 75

Figure 2.16. Comparison of concentrations of aqueous, vitreous, and lens 76

GSH in LEGSKO and WT eyes.

Figure 2.17. Comparison of in vivo GSH transport and homeostasis in 78 mouse eyes.

Figure 2.18. Measurement of blood to lens GSH transport. 79

Figure 2.19. GSH content of LEGSKO mouse eyes treated with BSO. 80

Figure 2.20. Comparative ocular distribution of glutathione in aqueous, 82 vitreous, and lens in eyes of various species.

Figure 2.21. Regional distribution of GSH within human and porcine 83 vitreous.

Figure 3.1. Overview of significant gene expression changes. 101

Figure 3.2. Number of significantly up- and down-regulated genes 102

(relative to WT control) in major biological pathways.

Figure 3.3. Verification of RNA-Seq Results. 103

Figure 3.4. Confirmation of RNA-Seq Results by RT-qPCR. 105

Figure 3.5. Crystallin Gene Expression by Lens GSH Content. 107

Figure 3.6. Expression Changes in Detoxification Genes by Lens GSH 112

Content.

Figure 3.7. Top 10 Expression Changes in Small Molecule Transport 114

Genes by Lens GSH Content.

Figure 3.8. Top 10 Expression Changes in Lipid Metabolism Genes. 116

Figure 3.9. Top Expression Changes in Canonical EMT Pathway Genes 119 in Lens Epithelia by GSH Content.

Figure 3.10. Diagram of expression changes in canonical pathways of 121

EMT in lens epithelia.

Figure 3.11. Top 10 lens epithelia expression changes in non-crystallin 122 vision genes.

Figure 3.12. Lens epithelia expression changes in cell cycle genes. 124

Figure 3.13. Lens epithelia expression changes in extracellular matrix 125 organization genes.

Figure 3.14. Correlation of LEGSKO and BSO-treated LEGSKO lens 127 epithelia with EMT mouse lens transcriptional profiles (taken with permission from Medvedovic et al., 2006).

Figure 4.1. Total iron content of WT and LEGSKO lens epithelia and 146

Fiber cells.

Figure 4.2. Urea content of WT and LEGSKO lenses and surrounding 147 fluids.

Figure 6.1. Analysis of RNA-Seq Results. 161

Figure 6.2. Verification of qPCR primers. 162

Figure 6.3. Expression changes in transcription genes. 169

Figure 6.4. Expression changes in metabolism genes. 170

Figure 6.5. Expression changes in developmental biology genes. 171

Figure 6.6. Expression changes in immune response genes. 172

Figure 6.7. Expression changes in protein metabolism genes. 173

Figure 6.8. Expression changes in small molecule transport genes. 174

Figure 6.9. Expression changes in extracellular matrix organization genes. 175

Figure 6.10. Expression changes in cell cycle genes. 176

Figure 6.11. Expression changes in cellular stress response genes. 177

Acknowledgements

Firstly, I must express my gratitude to my mentors, Vincent M. Monnier,

M.D., and Xingjun Fan, Ph.D., to all members of the Monnier and Fan labs, and to the members of my thesis committee for their feedback, support, guidance, and friendship. I was fortunate to work with a number of collaborators throughout this project. Richard Armstrong, Ph.D., and Michael C. Goodman of Vanderbilt

University kindly provided glutathione analogues and advice on their use. Xiang

Zhang, Ph.D., Mario Medvedovic, Ph.D., and Jenny Chen at the University of

Cincinnati aided us in obtaining and analyzing the lens glutathione-responsive transcriptome. This work would not have been possible without the help of many great researchers.

I am thankful to have received generous travel awards from both the

Association for Research in Vision and Ophthalmology and the Barshop Institute for Longevity and Aging Studies in order to travel to their annual conferences and present my research and to have received support from the CWRU Visual Sciences

Training Program (T32 EY007157) and the Cole Eye Institute Visual Sciences

Training Grant (T32 EY024236) in addition to my mentors’ funding sources

(EY07099 to VMM, EY024553 to XF).

I must also thank my family for their continued support, especially my mother who was my first science teacher and fostered my love of science and research from an early age. Lastly, I am overwhelmingly indebted to my wife and constant companion, Anastasia Whitson, who provided me with love, support, advice, and distraction throughout the entirety of my Ph.D. training.

List of Abbreviations

13C Carbon-13 stable isotope 15N Nitrogen-15 stable isotope 35S Sulfur-35 radioisotope 3H Tritium (hydrogen-3) radioisotope AAPH 2,2'-Azobis(2-amidinopropane) dihydrochloride ABCA1 ATP-binding cassette transporter subfamily A member 1 ABCA13 ATP-binding cassette transporter subfamily A member 13 ABCG1 ATP-binding cassette transporter subfamily G member 1 ABCG2 ATP-binding cassette transporter subfamily G member 2 ADP Adenosine diphosphate AEBP1 AE binding protein 1 AKR1B10 Aldo-keto reductase family 1 member B10 ALDH1A1 Aldehyde dehydrogenase 1 family member A1 ALDH1A3 Aldehyde dehydrogenase 1 family member A3 ALDH1A7 Aldehyde dehydrogenase 1 family member A7 ALDH1L1 Aldehyde dehydrogenase 1 family member L1 ALDH3A1 Aldehyde dehydrogenase 3 family member A1 APOE Apolipoprotein E As3mt Arsenite methyltransferase Asf Alternative splicing factor ATP Adenosine triphosphate bp Base pairs BSO Buthionine sulfoximine CaCl2 Calcium chloride CD36 Cluster of differentiation 36 cDNA Complementary DNA CES1G Liver carboxylesterase CFTR Cystic fibrosis transmembrane conductance regulator CINP Cyclin dependent kinase 2 interacting protein CO2 Carbon dioxide COL1A1 Collagen, type I, alpha 1 COL1A2 Collagen, type I, alpha 2 CPNE6 Copine-6 CPNE7 Copine-7 CRX Cone-rod homeobox CRYAA Crystallin αA CRYAB Crystallin αB CRYBA1 Crystallin βA1 CRYBA4 Crystallin βA4 CRYBB1 Crystallin βB1 CRYGA Crystallin γA CRYGB Crystallin γB CRYGC Crystallin γC CRYGD Crystallin γD

CRYGE Crystallin γE CRYGF Crystallin γF CRYGN Crystallin γN CRYGS Crystallin γS CSN3 Kappa-casein CST Cathepsin xCT Cystine/glutamate transporter Cx30 Connexin 30 Cx31 Connexin 31 Cx43 Connexin 43 Cx46 Connexin 46 Cx50 Connexin 50 CYP2J Cytochrome P450, family 2, subfamily k, polypeptide 9 Cys Cysteine Cys-Gly Cysteinyl glycine DHA Dehydroascorbic acid DIC Dicarboxylate ion carrier DMC1 DNA meiotic recombinase 1 DMGDH Dimethylglycine dehydrogenase DNA Deoxyribonucleic acid DNC Decorin dUTP Deoxyuridine triphosphate EAAT3 Excitatory amino acid transporter 3 ECM Extracellular matrix EGF Epidermal growth factor EMT Epithelial-mesenchymal transition EP300 E1A binding protein p300 FABP5 Fatty acid binding protein 5 FDR False discovery rate FGFR4 Fibroblast growth factor receptor 4 FHL2 Four and a half LIM domains protein 2 FOXE3 Forkhead box E3 FTH1 Ferritin heavy chain FTL Ferritin light chain G0S2 G0/G1 switch 2 GABA γ-Aminobutyric acid GABRR1 γ-Aminobutyric acid receptor subunit rho-1 GABRR2 γ-Aminobutyric acid receptor subunit rho-2 GAL Galanin GCL γ-Glutamylcysteine ligase GCLC γ-Glutamylcysteine ligase catalytic subunit GCLM γ-Glutamylcysteine ligase modifier subunit γ-GC γ-Glutamylcysteine γ-EAG γ-Glutamyl-alanine-glycine γ-ESG γ-Glutamyl-serine-glycine GGT γ-Glutamyltransferase

GJB3 Gap junction protein beta 3 GJB6 Gap junction protein beta 6 Glu Glutamate GLUT1 Glucose transporter 1 GLUT3 Glucose transporter 3 Gly Glycine Gp130 Glycoprotein 130 GPX Glutathione peroxidase GR Glutathione reductase GRX Glutaredoxin GRX(SH2) Reduced glutaredoxin GRX(S-S) Oxidized glutaredoxin GS Glutathione synthase GS-B Glutathione-bimane conjugate GSH Reduced glutathione GSH-EE Glutathione ethyl ester GSSG Oxidized glutathione GST Glutathione S-transferase GSTK1 Glutathione S-transferase kappa 1 GSTM1 Glutathione S-transferase mu 1 GSTP1 Glutathione S-transferase pi 1 GSTT1 Glutathione S-transferase theta 1 H2O Water H2O2 Hydrogen peroxide HCl Hydrogen chloride HIBCH β-hydroxyisobutyryl-coenzyme A hydrolase HPLC High performance liquid chromatography HPRT Hypoxanthine phosphoribosyltransferase IER3 Immediate early response 3 Il-1β Interleukin-1β IPA Ingenuity Pathway Analysis K+ Potassium ion KCl Potassium chloride KCND1 Potassium voltage-gated channel subfamily D member 1 KCND3 Potassium voltage-gated channel subfamily D member 3 KCNIP Kv channel-interacting protein Km Michaelis constant KOH Potassium hydroxide LC-MS Liquid chromatography-mass spectrometry LDL Low-density lipoprotein LDLR Low-density lipoprotein receptor LEGSKO Lens epithelia glutathione synthesis knockout LIM2 Lens intrinsic membrane protein 2 Mg2+ Magnesium ion MgCl Magnesium chloride MRM Multiple reaction monitoring

15 mRNA Messenger ribonucleic acid MRP Multidrug resistant protein MT1 Metallothionein 1 MT2 Metallothionein 2 MT-CYB Mitochondrially encoded cytochrome B MT-ND4 Mitochondrially encoded NADH:ubiquinone oxidoreductase core subunit 4 MT-ND5 Mitochondrially encoded NADH:ubiquinone oxidoreductase core subunit 5 MT-RN2 mitochondrial 16S ribosomal RNA Na+ Sodium ion NAC N-acetylcysteine NaCl Sodium chloride NaDC3 Sodium-dependent dicarboxylate transporter NADP+ Oxidized nicotinamide adenine dinucleotide phosphate NADPH Reduced nicotinamide adenine dinucleotide phosphate NaH2PO4 Sodium phosphate monobasic NaHCO3 Sodium bicarbonate NaOH Sodium hydroxide NPAS2 Neuronal PAS domain-containing protein 2 NR4A1 Nuclear receptor subfamily 4 group A member 1 Nrf2 Nuclear factor (erythroid-derived 2)-like 2 O2 Oxygen OAT-3 Organic anion transporter 3 OATP Organic anion transporting polypeptide OGC 2-oxoglutarate carrier OTX2 Orthodenticle homeobox 2 PAH Phenylalanine hydroxylase PBS Phosphate buffered saline Pc Permeability coefficient PCO Posterior capsular opacification PCR polymerase chain reaction PCSK9 Proprotein convertase subtilisin/kexin type 9 pH Potential of hydrogen Pi Phosphate ion PI3K Phosphoinositide 3-kinase PLIN4 Perilipin 4 PLIN5 Perilipin 5 Poly(A) Polyadenylated PPARG Peroxisome proliferator-activated receptor gamma Protein-SH Protein with reduced cysteine residue Protein-SSG Protein with glutathionylated cysteine residue QC Quality control QPCT Glutaminyl-peptide cyclotransferase RARA Retinoic acid receptor alpha RBP3 Retinol binding protein 3

RcGshT Sodium-independent rat canalicular GSH transporter RCN1 Reticulocalbin 1 RER1 Retention in endoplasmic reticulum sorting receptor 1 RHO Rhodopsin RHOH Ras homolog family member H RNA Ribonucleic acid RNA-Seq RNA sequencing RPKM Reads per kilobase of transcript per million mapped reads RPS10 40S ribosomal protein S10 RsGshT Rat sinusoidal GSH transporter RT-qPCR Reverse transcription quantitative PCR SCNN1B Sodium channel epithelial 1 beta SD Standard deviation SEM Standard error of the mean SERPINE1 Serpin family E member 1 SLC14A1 Solute carrier family 14 member 1 SLC40A1 Solute carrier family 40 (iron-regulated transporter) member 1 SMARCA4 SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily A, member 4 SNCA α-Synuclein SOS Son of sevenless SPOCK2 SPARC/osteoectin, cwcv and kazal like domains proteoglycan 2 SPP1 Secreted phosphoprotein 1 SVCT2 Sodium-dependent vitamin C transporter 2 SYK Spleen tyrosine kinase TGFBI Transforming growth factor, beta-induced TGF-β1 Transforming growth factor beta 1 TNC Tenascin C TNF Tumor necrosis factor TRF Transferrin Tris Tris(hydroxymethyl)aminomethane TXNIP Thioredoxin-interacting protein UT-B Urea transporter B UV Ultraviolet UVR Ultraviolet radiation WNT10A Wnt family member 10A WT Wild-type

Lens Adaptation to Glutathione Deficiency:

Implications for Cataract

Abstract

JEREMY A. WHITSON

The antioxidant glutathione (GSH) protects lens proteins from post- translational modifications that result in their aggregation and cataract formation.

With age, the human lens becomes increasingly depleted of GSH, which contributes to the development of age-related cataract. In order to gain a comprehensive understanding of the role of GSH in the pathogenesis of age-related cataract, I set out to study the consequences of and adaptations to GSH-deficiency in the lens using the Lens Glutathione Synthesis Knockout (LEGSKO) mouse model of cataract. The questions addressed in this thesis are: 1) How does the LEGSKO lens maintain >1 mM GSH despite a complete lack of GSH synthesis? and 2) What gene expression and signaling changes are associated with lens GSH deficiency? The first of these questions was addressed by measuring lens uptake of isotopically- labeled GSH using an LC-MS/MS system. I determined that mouse lenses obtain exogenous GSH in two ways: from the aqueous humor via an active transport mechanism and from the vitreous humor via passive diffusion. It was found that mouse eyes have a high concentration of GSH in their vitreous humor and a low

18 concentration of GSH in their aqueous humor and, because of this, nearly all the

GSH in the LEGSKO lens is derived from equilibration with the vitreous pool. It was also found that the eyes of humans and other large animals lack this high vitreous GSH concentration. The second question of this thesis was addressed by comparing the transcriptomic profiles of wild-type control lenses, chronically

GSH-deficient LEGSKO lenses, and acutely/severely GSH-deficient buthionine sulfoximine-treated (BSO;GSH synthesis inhibitor) LEGSKO lenses using RNA-

Seq technology. These data show that the most robust responses to GSH-deficiency in the lens are upregulation of detoxifying genes, including metallothioneins, aldehyde dehydrogenases, and carboxylesterase, activation of epithelial- mesenchymal transition (EMT) signaling, and alterations to lipid homeostasis and transport systems. These findings suggest that GSH plays a role in EMT-mediated posterior secondary cataract and implicate new potential targets for cataract therapeutics. This body of work greatly expands knowledge of the benefits and regulation of GSH in the lens, consequences of its loss, and other genes which promote lens clarity.

1. Introduction

1.1 The Crystalline Lens – Biological Glass 1.1.1 Anatomy of the Lens

The mature human lens is a highly transparent elliptical organ located directly behind the iris in the anterior eye and anchored to the ciliary body via zonule fibers (Augusteyn et al., 2010). The internal lens structure, which is wrapped in a unique basement membrane known as the lens capsule (Danysh and Duncan,

2009), features a monolayer of cuboidal epithelium at the anterior surface, differentiating fiber cells along the equator, and a large mass of mature fiber cells filling most of the lens (McAvoy et al., 1999). Fiber cells assume a bow-like shape, such that cells originating from opposite positions along the equator of the lens meet in the middle of the lens at their apical and basal surfaces, which forms sutures that run through the lens (McAvoy et al., 1999; Augusteyn et al., 2010). This results in concentric layers of stacked fiber cells, giving the internal lens an onion-like appearance. The lens lacks vascularization and is bathed in aqueous humor at its anterior surface and vitreous humor at its posterior surface. Figure 1 illustrates these basic features of the lens.

Figure 1.1. Anatomy of the lens and its local environment.

1.1.2 Functions of the Lens

The primary function of the lens is to focus light on the retina in order to facilitate vision. In order to successfully carry out this essential task, the lens must maintain three primary features: clarity, a high refractive index, and the ability to accommodate, that is to change shape in order to focus on near and far images

(Hejtmancik and Shiels, 2015).

Clarity is partly achieved from the unique features of lens fiber cells, as described in the next section, and due to the disappearance of the tunica vasculosa lentis, an extensive capillary network along the lens posterior and lateral surfaces, shortly after birth (Barishak, 1992). This makes the mature lens a uniquely avascular organ and allows light to pass through the back of the lens without being scattered.

The lens has a refractive index of around 1.4, whereas air has a refractive index of 1.0 and glass has a refractive index of just above 1.5 (Kasthurirangan et al., 2008). This high refractive index means that the path of incoming light is curved and focused on the fovea. This refractive index is maintained by a high concentration of crystallin proteins in the lens (Andley, 2007).

Accommodation is facilitated by the activity of the ciliary muscle, which is attached to the lens through zonule fibers (Chien et al., 2006). The stretching and relaxing of these fibers induces the lens to take on a more elliptical or spherical shape, respectively, altering the path of incoming light to best fit the distance of the image so that it remains focused on the fovea.

1.1.3 Lens Cell Biology

Growth of the lens occurs when the epithelia at the equatorial regions of the lens undergo mitosis and the resulting daughter cells elongate to become fiber cells

(McAvoy et al., 1999; Augusteyn et al., 2010). This process continues such that cells are laid down on top of one another, with fiber cells in the lens nucleus being the oldest in the lens while cells near the surface of the lens are the youngest

(McAvoy et al., 1999; Augusteyn et al., 2010). Growth of the lens continues throughout life, although at a greatly reduced rate after early development

(McAvoy et al., 1999).

Beyond their elongated shape, fiber cells are also structurally distinct from lens epithelia in that, during the process of differentiation, they highly express the crystallin family of proteins and then degrade their organelles, including their nuclei (McAvoy et al., 1999; Augusteyn et al., 2010). Crystallins are incredibly

23 abundant in the lens, constituting approximately 90% of its water soluble protein, and form a stable latticework within the lens that prevents light scattering and increases the refractive index of the lens (Andley, 2007). In addition to the structural role provided by all crystallins, α-crystallins also have important chaperone activity that protects other crystallins and lens proteins from misfolding

(Andley, 2007). These adaptations make the mature fiber cells of the lens very unique within the body since they are very metabolically and mitotically inactive, with little to no protein turnover. Because of this, mature lens fiber cells and their associated proteins are among the longest lived cells in the entire body, with the nuclear lens fiber cells having been present since embryonic stages of development

(McAvoy et al., 1999). Adherent to the lens epithelia and surrounding the entire lens is the capsule, a specialized basement membrane consisting primarily of type

IV collagen, laminin, nidogen, and perlecan, that is unique among basement membranes due to its high level of transparency (Danysh and Duncan, 2009).

1.1.4 Transport and Homeostasis in the Lens

To fulfill the metabolic needs of the lens without vascularization, the lens relies on nutrition from its surrounding ocular fluids (Kiel, 2010). The anterior of the lens is bathed by the nutrient rich aqueous humor while the posterior of the lens is in contact with the vitreous humor, which is traditionally believed to serve a largely structural role (Kiel, 2010; Kiel et al., 2011; Purves et al., 2001). These humors are produced by the ciliary body, which is vascularized and functions as a blood-aqueous barrier, and are generally referred to more simply as the aqueous and vitreous (Kiel et al., 2011). Uptake of many nutrients from the aqueous by the

24 anterior lens has been characterized. Known transport systems that exist within the lens include the glucose transporters GLUT1 and GLUT3, the ascorbic acid transporter SVCT2, the cystine transporter xCT, and various other amino acid transporters (Fan et al., 2006;Fan et al., 2011;Umapathy et al., 2013;Merriman-

Smith et al., 1999).

There are two models proposed to explain how nutrients taken up by peripheral lens cells are then distributed throughout the lens (Tamara et al., 2016).

According to the lens microcirculation model, it is proposed that the lens maintains an internal fluid circulation that carries water and solutes throughout the lens so that nutrients and waste products can be carried to and away from the lens nucleus, respectively (Donaldson et al., 2001). This circulation is purportedly generated by a Na+ current that results from the activity of Na+/K+-ATPases in the basal lens epithelia pumping Na+ out of the lens, creating a large electrochemical gradient

(Donaldson et al., 2001). Na+ in the extracellular spaces between fiber cells leaks into the cells, setting up an ion current that flows into the lens across the extracellular space and out of the lens by travelling through gap junction-coupled cells to the Na+/K+-ATPases of the lens epithelia (Donaldson et al., 2001). In this model, the Na+ current drives the movement of water and solutes throughout the intercellular and extracellular spaces between fiber cells and ensures exchange of nutrients and wastes between the lens periphery and the mature fiber cells in the center of the lens.

+ While there is clear evidence of such Na currents in the lens (Gao et al.,

2011), it is unclear whether they truly play a major role in lens metabolite

25 distribution. One recent study showed that, of 34 metabolites analyzed using a shotgun LC-MS approach, none matched the distribution expected for the lens microcirculation model (Tamara et al., 2016). As an alternative to the microcirculation model, many believe that the distribution of solutes in the lens entirely based on passive diffusion through gap junctions that extensively couple lens cells (Beyer and Berthound, 2014). Lens epithelia are coupled to fiber cells through Cx43 channels while fiber cells are highly interconnected through Cx46 and Cx50 connexins. This coupling of cells allows free flow of solutes throughout the lens and is essential to lens health in both models. It is possible that different solutes adhere more closely to one model or the other and that both modes of transport are necessary.

1.1.5 Conclusions

The lens is a unique organ of the eye which features a number of specialized functions for maintenance of transparency and focusing of light onto the retina.

These special features result in long lived cells with a unique need for robust transport and homeostatic processes.

1.2 Age-Related Cataract – A Disease of Protein Aggregation

1.2.1 Pathology of Age-Related Cataract

Because mature lens fiber cells lack the ability to degrade damaged proteins and synthesize new protein, lenticular proteins must remain very stable and unmodified throughout the long life of these cells (Petrash, 2013; Sharma and

Santhoshkumar, 2009). Damaged lens proteins, including crystallins, may become insoluble and aggregate, altering the refractive index of the lens and resulting in the light scattering and opacification that characterizes cataract (Petrash, 2013). The accumulation of growing insoluble protein aggregates over time defines age-related cataract, which is distinct from congenital forms of cataract that can arise from various defects in important lens genes, such as mutations in connexin or crystallin genes (Shiels and Hejtmancik, 2013).

Study of which protein modifications are involved in cataract development has been performed by analyzing protein found in the water-insoluble fractions of normal and cataractous lenses. Primarily, these fractions consist of crystallins, and modifications include methionine oxidation, deamidation, backbone cleavage, glycation, and extensive disulfide-linkage (Sharma and Santhoshkumar, 2009;

Hanson et al., 2000; Nagaraj et al., 2012). While sulfhydryl groups of lens proteins are mostly in a reduced state in young lenses, extensive oxidation of methionine and cysteine residues is found in aged lenses, with the majority of these residues being oxidized in cataractous lenses (Kodama and Takemoto, 1998; Garner and

Spector, 1980). Glycation of lens proteins, which occurs non-enzymatically through the Maillard reaction, results in the yellowing/browning of the aging lens and is another hallmark of age-related cataract (Fan et al., 2006). Due to the accumulation of these modifications, protein content of the water-soluble fraction

27 decreases in older lenses while the protein content of the water-insoluble fraction inversely increases, and this change is more dramatic in cataractous lenses, leading to growing light scattering and opacity (Sharma and Santhoshkumar, 2009).

1.2.2 Risk Factors for and Impact of Age-Related Cataract

Cataract has a widespread impact on aging populations and currently accounts for over half of the cases of blindness globally (Pascolini and Mariotti,

2010). While age remains the primary determinant of cataract formation, a number of other risk factors have been identified in various studies and are summarized in

Table 1.1.

Table 1.1 Risk factors for age-related cataract development.

Factor Increased Risk Citations Diabetes 74-534% Floud et al., 2016 Park et al., 2016 Mamatha et al., 2015 Hypertension 33-55% Park et al., 2016 Mamatha et al., 2016 Obesity 6-30% Floud et al, 2016

Smokeless Tobacco 204-362% Mamatha et al., 2015

Smoking Tobacco 13-58% Floud et al, 2016 Lindblad et al, 2014 Park et al., 2016 UVR>1000 J/m2 37-145% Yu et al., 2016

Vitrectomy 59% Effenterre et al., 1992

A major common denominator among these risk factors is oxidative stress.

Excessive ultraviolet radiation induces lipid peroxidation and aldehyde formation, major stressors of the lens, which is regularly exposed to UV light (Yu et al., 2016;

Choudhary et al., 2005). Vitrectomy is thought to result in oxidative stress of the lens due to the vitreous’ oxygen buffering capability (Milazzo, 2014). Removal of the vitreous leads to an elevation in lenticular partial pressure of oxygen, which in turn increases the production of reactive oxygen species. Tobacco use, of both smoking and smokeless varieties, induces widespread oxidative stress in the body and it is likely that this is the mechanism by which tobacco affects cataractogenesis

(Fletcher, 2010).

Another common trait among these risk factors is carbonyl stress, as found in diabetes and obesity (Miyata et al., 2003). Hypertension, which is associated with and can be induced by obesity and carbonyl stress (Chen et al., 2013), is also a major risk factor for cataract (Park et al., 2016). High circulating levels of glucose lead to an increase in glucose in the lens, resulting in extensive glycation of lens proteins that produces lens brunescence and protein aggregation (Hashim and

Zarina, 2011).

1.2.3 Clinical Treatment of Cataract

To date, no effective preventative treatment for human cataract exists and the only solution to treating cataract is surgical replacement of the lens, in part or in whole, with an artificial lens (Javitt et al., 1996; Song et al., 2014). The most common modern form of this surgery involves removing a small portion of the anterior lens capsule, disrupting and removing lens cells by ultrasonic phacoemulsification, and placing an intraocular lens implant into the remaining lens capsule (Pershing and Kumar, 2011).

Although this surgery is effective in restoring vision, it is not without costs and complications. There are many barriers inhibiting access to this surgery in low- income countries, where the majority of individuals with visual impairment live

(Pascolini and Mariotti, 2010). As a result, cataract goes untreated for more than one-hundred million individuals, a number that has been steadily increasing

(Foster, 2000). Because of this, an effective treatment that delays or reduces the risk for cataract formation could massively reduce the global burden of blindness due to cataract. Due to the strong age-relatedness of cataract, just a ten-year delay in cataract formation would be sufficient to reduce the prevalence of cataract by

50%, as estimated by the National Eye Institute (Panel, 1983).

1.2.4 Posterior Capsular Opacification

Even for patients who are able to receive a replacement lens, the most common complication of the surgery is posterior capsular opacification (PCO), a form of cataract caused by residual lens epithelia cells after surgery (Song et al.,

2014). These residual cells, which have disrupted organization and are under heavy stress, proliferate to the posterior lens capsule and undergo epithelial-mesenchymal transition (EMT) (Lovicu et al., 2016). This transition results in the formation of disorganized fibrotic lesions that block the passage of light and wrinkle the lens capsule. The formation of PCO, also known as secondary cataract, causes severe vision impairment and must be cleared by laser ablation surgery to restore vision

(Hayashi et al., 2003). PCO remains a major issue standing in the way of effective cataract treatment. 1.2.5 Conclusions

Age-related cataract is a visual impairment caused by aggregation of protein in the lens. This pathology results from post-translational modification of long lived lens protein; exacerbated by conditions including diabetes, vitrectomy, and others that cause oxidative or carbonyl stress. Cataract is treatable by replacement of the lens with an artificial implant, but this surgery is not readily available in many regions. In addition, while able to restore vision, it is associated with complications linked to the surgery itself as well as secondary cataracts. As a result, cataract remains the number one cause of blindness globally and has a heavy impact on aging populations.

1.3 Glutathione – The Guardian of Lenticular Redox Homeostasis 1.3.1 Structure and Biosynthesis of Glutathione

Glutathione (GSH) is a small tripeptide consisting of glutamate, cysteine, and glycine that exists at high millimolar concentration within the lens (Giblin,

2000). The synthesis of glutathione occurs in the cytosol of cells throughout the body in a two-step process (Lu, 2013). The first, and rate-limiting, step is the formation of a γ-peptide bond between glutamate and cysteine, linking the side chain carboxylic group of glutamate to the primary amino group of the cysteine, by the enzyme γ-glutamylcysteine ligase (GCL) (Lu, 2013). GCL has two subunits: a catalytic subunit (GCLC) and a modifier subunit (GCLM). Although GCLC retains catalytic activity in the absence of GCLM, it functions at a much higher rate in the presence of the modifier subunit, with an affinity for its substrates that is several times greater (Huang et al., 1993). Synthesized γ-glutamylcysteine (γ-GC) is

31 rapidly converted to glutathione by addition of glycine through a typical amide bond to the cysteine residue of the peptide by the enzyme glutathione synthase (GS)

(Lu, 2013). Both steps in this synthesis are energy dependent and require hydrolysis of one ATP molecule to ADP. The detailed chemical structure of GSH is shown in

Figure 1.2.

Figure 1.2. Chemical structure of glutathione. Glutathione consists of glutamate conjugated to cysteine through a γ-peptide bond and cysteine conjugated to glycine through an amide bond.

1.3.2 Regulation of Cellular Glutathione Content

Cysteine is the rate-limiting component in glutathione synthesis and its abundance is a major factor in the rate of GSH production (Sekhar et al., 2011).

Additionally, GCLC has a feedback mechanism wherein its activity is non- allosterically inhibited by the binding of GSH to a specific non-active site on the enzyme (Richman and Meister, 1975). Thus, as cytoplasmic GSH increases, the rate of GSH production slows.

The production of the GCLC, GCLM, and GS enzymes is governed by the transcription factor Nrf2, which binds to the antioxidant response elements in the promoters of these and other antioxidant genes when it is activated (Lu, 2009;

Jaiswal, 2004). Nrf2 is normally bound by Keap1 in the cytoplasm, where it is regularly degraded. Oxidative stress modifies a cysteine residue in Keap1, leading to the release of Nrf2, which can then travel to the nucleus and bind to DNA response elements (Jaiswal, 2004).

Through these pathways, cellular GSH content, in the lens and elsewhere, is maintained at a high steady state level and is responsive to oxidative stress.

1.3.3 The Essential Role of Glutathione in the Lens

The high levels of GSH in the lens are vital to preventing cataract due to the antioxidant functions of GSH. One major mechanism by which GSH carries out its antioxidant function is through a redox cycle in which reduced GSH is oxidized by glutathione peroxidase enzymes (GPX) in order to detoxify peroxides, preventing damage to cytosolic proteins (Giblin, 2000; Lu, 2013). Oxidized glutathione

(GSSG) consists of two GSH molecules linked by a disulfide bond between their cysteine residues. GSSG is restored to two GSH molecules by the action of glutathione reductase (GR), which requires NADPH as a cofactor (Giblin, 2000;

Lu, 2013).

Glutaredoxin (GRX) is an antioxidant enzyme that breaks disulfide bonds between glutathione and cysteine residues of proteins and detoxifies dehydroascorbic acid (DHA), which is the oxidized form of Vitamin C and a major component of lens carbonyl stress and glycation (Lu, 2013;Cheng et al., 2006).

These reactions generate an internal disulfide bond between cysteine residues in

GRX that inactivate the enzyme (Lillig et al., 2008). This internal disulfide bond is reduced by the oxidation of GSH to GSSG, restoring GRX activity.

Glutathione S-transferases (GSTs) catalyze nucleophilic attacks by GSH, a reaction known as glutathione adduction, to various xenobiotic substrates (Hayes et al., 2005). Glutathione adduction allows for detoxification and subsequent metabolism or export of many reactive compounds, including aldehydes and other lipid peroxidation products.

Additionally, GSH by itself has been shown to account for a large portion of hydrogen peroxide detoxification in the mouse lens and to be a highly efficient free radical scavenger and metal chelator in vitro (Spector et al., 1997; Galano and

Alvarez-Idaboy, 2011). These diverse functions allow the high concentrations of

GSH in the lens to protect proteins and maintain a reduced, stable, environment with minimal post-translational modification of crystallins.

1.3.4 Lenticular Glutathione Declines with Age

In humans, GSH content of the lens has been found to steadily decrease with age, and this change is even more dramatic in cataractous lenses (Harding,

1970). GCL and GS activity decreases with age, possibly due to accumulation of post-translational modifications to these enzymes, which is thought to account for the age-dependent loss of GSH (Rathbun et al., 1993). Additionally, a study on the diffusion of radiolabeled GSH in cultured lenses revealed that a barrier to GSH diffusion into the lens nucleus develops with age (Sweeney and Truscott, 1998).

The result is critically low GSH content in the lens nucleus, making protein in this region highly susceptible to oxidative damage and aggregation.

1.3.5 A Loss of Glutathione Activity Leads to Cataract

There are a number of lines of experimental evidence that directly link lens

GSH content with cataract formation. Lens culture experiments indicate that a GSH concentration of approximately 1 mM is a critical point for the lens nucleus, below which proteins rapidly form aggregates (Giblin, 2000).

Knockouts of GPX, GRX, and GCLC enzymes in mouse models have been demonstrated to result directly in cataract formation (Reddy et al., 2001; Wu et al.,

2014; Fan et al., 2012), indicating the essential role of GSH-utilizing enzymes in protecting the lens.

A null genotype for the GST gene GSTM1 has been indicated as a possible risk factor for age-related cataract formation, based on the genotyping of age- related cataract patients and healthy controls (Liao et al., 2015). Individuals with the GSTM1 null genotype develop cataract at double the frequency of others and this risk is further increased in individuals with certain GSTT1 and GSTP1 alleles

(Liao et al., 2015).

The findings points to the fact that GSH is an essential component of lenticular redox homeostasis and a loss in its activity predisposes the lens to cataract formation.

1.3.6 The LEGSKO Mouse Model

The Lens Glutathione Synthesis KnockOut (LEGSKO) mouse was generated to provide a robust animal model for the study of lens GSH deficiency in relation to cataract (Fan et al., 2012). These mice have Gclc, the gene encoding

GCLC, conditionally knocked out in their lenses. This conditional knockout is accomplished by utilizing a loxP/Cre system with the Cre recombinase gene having an αA-crystallin promotor containing an inserted Pax6 consensus sequence. This promoter leads to expression of the Cre recombinase, and thus deletion of the floxed

Gclc, only within lens cells (Fan et al., 2012).

These mice show age-related cataract formation that mimics the human phenotype, with extensive disulfide bonding, methionine oxidation, and aggregation of crystallins and other lens proteins (Fan et al., 2012;Fan et al., 2015).

GCLC knockout was confirmed by measuring GCLC mRNA and protein content in lenses and using a derivatization HPLC method to test for synthesis of GSH from its component amino acids. Despite a complete lack of GCLC expression and activity in the lenses of homozygous knockout mice, GSH is still present at concentrations above 1 mM in LEGSKO lenses (Fan et al., 2012). This observation contradicts the characterized mechanisms for a cell to accumulate GSH and indicates that the lens may be taking up GSH from an exogenous source via some transport system.

1.3.7 A Salvage Pathway for Glutathione Synthesis

Because LEGSKO lenses lack GCLC, but not GS, it is plausible that

LEGSKO lenses could synthesize GSH if they receive and take up γ-GC from an exogenous source. γ-GC is not normally present in extracellular fluids but lens, ciliary body, and corneal epithelium express γ-glutamyltransferase (GGT) on their surface, which cleaves the γ-glutamyl moiety from GSH and links it to other amino acids (Reddy and Unakar, 1973). If extracellular cysteine is present, this GSH degradation process can lead to the formation of γ-GC.

This mechanism of γ-GC production and uptake following GSH degradation has been proposed by others and investigated in a number of studies.

Griffith (1981), injected mice intraperitoneally with 2-vinylpyridine-derivatized cysteine and unlabeled GSH and measured 2-vinylpyridine-derivatized γ-GC in urine, demonstrating that γ-GC can be generated from GSH through this mechanism. It was also demonstrated that γ-GC can be used to recover intracellular levels of GSH in kidney cells of mice treated with buthionine sulfoximine (BSO), a specific inhibitor of GCLC, while equimolar amounts of cysteine and glutamic acid had no effect (Anderson and Meister, 1983). Furthermore, by using acivicin, an inhibitor of GGT, it was found that GGT must be active in order for extracellular

GSH to increase the concentration of intracellular GSH in lymphoid cells (Jensen and Meister, 1983).

Based on these findings, it was proposed that generation of γ-GC from GSH by GGT, followed by import of γ-GC and re-synthesis of GSH by GS, constitutes a salvage pathway of GSH synthesis/transport in many tissues. However, the

37 questions of the identity of the importer of γ-GC and whether this salvage pathway occurs in the lens have not been previously addressed. GGT knockout mice were found to develop cataracts and had reduced lenticular GSH concentration but this was found to be secondary to systemic cysteine deficiency, due to a loss of GSH recovery in the kidneys, and lenses recovered when N-acetylcysteine was administered, indicating that this pathway may not play an essential role in lenses

(Chevez-Barrios et al., 2000). The question of the potential existence of a GSH salvage pathway based on GC uptake has been investigated as part of this thesis work in the LEGSKO mouse, as described in Chapter 2.

1.3.8 Conclusions

GSH is an essential and abundant tripeptide in the lens that helps to prevent cataract formation by acting as a substrate for anti-oxidative and anti-glycative enzymes and by acting non-enzymatically as a free radical scavenger and metal chelator. With age, lenses become increasingly deficient in GSH and, thus, more susceptible to damaging post-translational modifications. In order to better study the role of GSH in cataractogenesis, the LEGSKO mouse model of cataract was generated. Despite a total lack of lenticular GSH synthesis, these mouse lenses maintain >1 mM GSH, indicating a transport or salvage mechanism for maintaining lens GSH content.

1.4 Glutathione Transport – Known Mechanisms

1.4.1 Glutathione Transport in the Lens

Transport of intact GSH into the lens has been reported by multiple groups

(Stewart-DeHaan et al., 1999; Zlokovic et al., 1994). These studies measured the concentration of radiolabeled GSH in the lens and aqueous humor of rats and guinea pigs after its injection into the systemic circulation. Uptake of intact GSH was distinguished from synthesis and the salvage pathway by utilizing acivicin to block

GGT activity (Zlokovic et al., 1994) or using HPLC to ensure that only intact GSH was measured (Stewart-DeHaan et al., 1999). It was found that rat lenses could obtain 12.3% of their total GSH from the injected GSH after only 4 hours, indicating that this transport occurs rapidly and is important for normal lenticular maintenance of high GSH concentration (Stewart-DeHaan et al., 1999).

The identity of the transport system responsible for this uptake has remained elusive. Two GSH transporters were reported to have been found in rats, known as rat sinusoidal GSH transporter (RsGshT) and sodium-independent rat canalicular

GSH transporter (RcGshT) (Kannan et al., 1995). These transporters were discovered by isolating poly(A)+ RNA from the lens, generating a cDNA library, and screening cDNA by injecting it into Xenopus oocytes and performing GSH uptake assays (Kannan et al., 1995). Oocytes positive for GSH uptake showed the expression of these transporters (Kannan et al., 1995). Unfortunately, it was later discovered that these cDNA sequences were cloning artifacts from the Escherichia coli K-12 genome and the characterization of these groups as mammalian GSH transporters has since been disregarded (Li et al., 1997).

Since this initial setback, no mammalian plasma membrane importers of

GSH have been fully characterized. The organic anion transporter OAT-3 was

39 shown to uptake GSH and p-aminohippurate (PAH) in exchange for 2-oxoglutarate when expressed recombinately from bacteria and reconstituted in proteoliposomes

(Lash, 2007). However, the ability of this transporter to facilitate GSH uptake in a cellular system has not yet been demonstrated. OAT-3 is expressed in lens epithelia at the mRNA and protein level but only at the apical surface of the epithelia, which is inconsistent with a model of lens GSH uptake from the aqueous humor (Li et al.,

2010). NaDC-3 has also been proposed as a possible GSH transporter based on its expression in the lens and kidney, but no experimental evidence exists to show that it is capable of transporting GSH (Li et al., 2010).

1.4.2 Active Transport Systems

Previously, it was reported that the mitochondrial dicarboxylate carrier

(DIC) and 2-oxoglutarate carrier (OGC) are responsible for transport of GSH into mitochondria, where it is present at a high concentration and undergoes reactions but is not synthesized (Lash, 2006). However, a recent study that attempted to characterize this transport by measuring GSH uptake in membrane vesicles containing these transporters found that neither transporter could facilitate GSH transport, even though they could transport their other known substrates (Booty et al., 2015). Additionally, transport of other substrates in this system could not be inhibited by excess GSH (Booty et al., 2015), indicating that GSH does not compete for binding to the transporters. These findings raise serious questions about the identity of the GSH transporters in the mitochondria and warrant a reevaluation of the proposed mitochondrial GSH import system.

Although the identity of any GSH importer in mammalian cells has yet to be confirmed, a number of transporters have been shown to be involved in the export of GSH, GSSG, and GSH conjugates. Several members of the multidrug resistance associated protein (MRP) family, including the cystic fibrosis transmembrane conductance regulator (CFTR), and the solute carrier organic anion-transporting polypeptide (OATP) family in mammals have been shown to export GSH from cells (Bachhawat et al., 2013). None of these transporters are specific or have a high affinity for GSH and may primarily serve the function of removing glutathione adducts and GSSG from cells or secreting GSH into the blood and ocular fluids. Although there is no evidence that any of these transporters can import GSH into cells, it is possible that a mammalian GSH importer exists that shares homology or mechanism with these exporters.

The only confirmed GSH importers in living cells are Hgt1p, an oligopeptide transporter in S. cerevisiae, and its homologues in plants, bacteria, and other yeast (Bachhawat et al., 2013; Bourbouloux et al., 2000). Unfortunately, no homologues of this transporter exist in higher eukaryotes beyond plants

(Bachhawat et al., 2013). Hgt1p was discovered by knocking out genes of the oligopeptide transporter family in S. cerevisiae and determining the effect on the ability of the yeast to uptake GSH (Bourbouloux et al., 2000). This transporter has a high-affinity (Km ~54 uM) for GSH and is not appreciably inhibited by other di- or tri-peptides. This transport is driven by electrochemical potential, with Hgt1p functioning as a proton/GSH symporter (Bourbouloux et al., 2000). Although this transporter does not exist in mammals, it does indicate that a similar system could

41 account for lens GSH transport. A separate family of peptide transporters exists in mammals that function by a similar proton/peptide symport mechanism and could potentially have the ability to transport GSH (Smith et al., 2013).

1.4.3 Passive Transport

Within the lens, it has been demonstrated that the GSH content of nuclear, but not cortical, fiber cells are dependent on the expression of Cx46 gap junction channels (Slavi et al., 2014). Furthermore, by inhibiting Na+/K+-ATPases, it was determined that GSH movement from the lens cortex to the nucleus is governed only by its concentration gradient and is not dependent on the lens microcirculation

(Slavi et al., 2014). Thus, whether obtained by uptake or synthesis, GSH within the lens seems to diffuse down its concentration gradient towards the lens nucleus from the lens periphery. Transport through connexins cannot account for GSH uptake from the aqueous humor since the GSH concentration within the lens is many fold higher than in the aqueous and thus requires energy to be taken up. Also lens epithelium lacks hemichannels on its basal (aqueous facing) surface (Bhat, 2001;

Stewart-DeHaan et al., 1999), inconsistent with such a mechanism of uptake.

1.4.4 Conclusions

A wealth of evidence suggests that active transport of GSH occurs in mammalian tissues, including the lens, but the identity of such transport systems has remained elusive and controversial. Regardless of the uptake mechanism, GSH within the lens appears to be distributed by passive diffusion through gap junction channels that link lens fiber cells and not via lens microcirculation. An integrated view of the characterized mechanisms of GSH production, homeostasis, and transport, as well as possible routes by which LEGSKO lenses obtain GSH, is illustrated in Figure 1.3.

Figure 1.3. Diagram of glutathione transport, functions, and metabolism. Glutathione (GSH) is synthesized in the cytosol by γ-glutamylcysteine ligase (GCL), which consists of a catalytic (GCLC) and a modifier (GCLM) subunit. Glutathione works with glutaredoxin (GRX), glutathione peroxidase (GPX), and glutathione S-transferase (GST) enzymes to detoxify harmful compounds such as dehydroascorbic acid (DHA), hydrogen peroxide (H2O2), and various xenobiotics, respectively. Oxidized glutathione (GSSG) is reduced by glutathione reductase (GR). GSH, GSSG, and glutathione adducts can be exported by the multidrug resistant protein transporters (MRP1-5) and GSH can also be exported by the cystic fibrosis transmembrane conductance regulator (CFTR), organic anion transporting polypeptides (OATP) 1 and 2, and ATP-binding cassette transporter subfamily G member 2 (ABCG2). GSH is broken down extracellularly by γ- glutamyltransferase (GGT) and dipeptidase. Glutamate and cysteine can be taken up by excitatory amino acid transporter 3 (EAAT3) and used for GSH synthesis. Buthionine sulfoximine (BSO) treatment and the LEGSKO knockout disrupt GSH synthesis by targeting GCLC. Direct transport of GSH or transport of γ- glutamylcysteine are possible mechanisms by which the LEGSKO lens maintains residual GSH. 1.5 Thesis Statement

The aim of this work was to obtain a clear and extensive understanding of how the aging lens adapts to a deficiency in GSH and how this impacts the development of age-related cataract. To that end, there are two primary questions that were investigated in this research:

1) How does the LEGSKO lens maintain >1 mM GSH despite a complete lack

of GCLC activity?

2) What gene expression and signaling changes are associated with lens GSH

deficiency?

These questions are investigated and discussed in the following chapters.

2. Evidence of Dual Mechanisms of Glutathione Uptake in the Rodent Lens: A Novel Role for Vitreous Humor in Lens Glutathione Homeostasis

Adapted from the article : “Whitson JA, Sell DR, Goodman MC, Monnier VM, Fan X. Evidence of Dual Mechanisms of Glutathione Uptake in the Rodent Lens: A Novel Role for Vitreous Humor in Lens Glutathione Homeostasis. Investigative Ophthalmology & Visual Science. 2016;57(8):3914-3925.”

2.1 Abstract

2.1.1 Purpose

Lens glutathione synthesis knockout (LEGSKO) mouse lenses lack de novo glutathione (GSH) synthesis but still maintain >1 mM GSH. We sought to determine the source of this residual GSH and the mechanism by which it accumulates in the lens.

2.1.2 Methods

Levels of GSH, glutathione disulfide (GSSG), and GSH-related compounds were measured in vitro and in vivo using isotope standards and liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis.

2.1.3 Results

Wild-type (WT) lenses could accumulate GSH from γ-glutamylcysteine and glycine or from intact GSH, but LEGSKO lenses could only accumulate GSH from intact GSH, indicating that LEGSKO lens GSH content is not due to synthesis by a salvage pathway. Uptake of GSH in cultured lenses occurred at the same rate for

LEGSKO and WT lenses, could not be inhibited, and occurred primarily through cortical fiber cells. In contrast, uptake of GSH from aqueous humor could be competitively inhibited and showed an enhanced Km in LEGSKO lenses. Mouse vitreous had >1 mM GSH, whereas aqueous had <20 μM GSH. Testing physiologically relevant GSH concentrations for uptake in vivo, we found that both

LEGSKO and WT lenses could obtain GSH from the vitreous but not from the

47 aqueous. Vitreous rapidly accumulated GSH from the circulation, and depletion of circulating GSH reduced vitreous but not aqueous GSH.

2.1.4 Conclusions

The findings described above provide, for the first time, evidence for the existence of dual mechanisms of GSH uptake into the lens, one mechanism being a passive, high-flux transport through the vitreous exposed side of the lens versus an active, carrier-mediated uptake mechanism at the anterior of the lens.

2.2 Introduction

2.2.1 Antioxidant Functions of Glutathione

In order to protect itself from oxidation, the lens, like most other cellular systems, maintains very high concentrations (∼3–5 mM) of the antioxidant glutathione (GSH) (Giblin, 2000). Glutathione has a number of unique functions, the best studied of which is its role as a cofactor for glutathione peroxidase (GPX) enzymes (Reddy et al., 2001). Glutathione peroxidase detoxifies H2O2 in a reaction that generates a disulfide linkage between two GSH molecules, resulting in production of glutathione disulfide (GSSG). Glutathione disulfide can be restored to its reduced form by an NADPH-dependent reaction catalyzed by glutathione reductase (GR). Glutathione can also protect proteins by acting directly as a free- radical scavenger and metal chelator, conjugating to toxic xenobiotic compounds by glutathione S-transferase (GST) enzymes, and regenerating glutaredoxin (GRX)

48 enzymes, which reduce compounds such as dehydroascorbic acid (DHA) (Kalinina et al., 2014).

2.2.2. Lenticular Glutathione Homeostasis

The lens GSH pool is thought to be produced by the lens epithelium and immature differentiating fiber cells in a two-step process (Umapathy et al., 2013).

The first and rate-limiting step is the generation of γ-glutamylcysteine (γ-GC) from cytosolic amino acids through an ATP-dependent reaction catalyzed by γ- glutamylcysteine ligase (GCL). Synthesis is then rapidly completed by the addition of glycine to the peptide through another ATP-dependent reaction catalyzed by glutathione synthase (GS). Glutathione synthesized in metabolically active regions of the lens is then distributed to the remainder of the lens by diffusion through the gap junctions that extensively couple lens fiber cells (Slavi et al., 2014). Deficiency in GSH has been linked to age-related nuclear cataract as lens GSH concentration, particularly in the lens nucleus, decreases with age (Harding, 1970). This is hypothesized to result from a reduced activity in GSH synthesis enzymes (Rathbun et al., 1993), as well as a barrier to GSH diffusion into the nucleus that develops with age (Sweeney and Truscott, 1988). It has been demonstrated that ∼1 mM GSH is a critical concentration for the lens nucleus, below which crystallins rapidly form disulfide bonds and aggregate (Giblin, 2000).

2.2.3. The LEGSKO Mouse Model of Cataract

The lens glutathione synthesis knockout (LEGSKO) mouse was generated to provide a model for the study of glutathione deficiency in the lens and its effect on cataract progression (Fan et al., 2012). Using the lens-specific Cre recombinase

MLR10, these mice lack expression of Gclc, the catalytic subunit of GCL, within their lenses. As expected, these mice develop a phenotype that mimics age-related nuclear cataract. Because homozygous knockout lenses were not found to have any residual Gclc mRNA, protein, or enzymatic activity, it was expected that these lenses would completely lack GSH. Surprisingly, although LEGSKO lens GSH content is significantly decreased compared to that in wild-type (WT) lenses, it is still maintained above 1 mM. This observation implies that lenses are able to obtain high concentrations of GSH from an exogenous source, that is, the aqueous or vitreous humors.

2.2.4. Glutathione Transport

The concept that the lens may obtain GSH through transport in addition to synthesis is not a novel one and has been reported in a number of studies (Steward-

DeHaan et al., 1999;Zlokovic et al., 1994;Li et al., 2010). Such experiments have led to speculation about the nature of a possible GSH transporter. Although two importers of GSH were previously reported (Kannan et al., 1995), they were later determined to be artifacts from the Escherichia coli genome and have since been discredited (Li et al., 1997). At the same time, others have suggested that specific

GSH transporters do not exist at all and instead that γ-GC can be generated at the external surface of cells by the breakdown of GSH through the enzyme γ- glutamyltransferase (GGT) (Griffith et al., 1981;Anderson and Meister, 1983). In this model, it is the γ-GC precursor, rather than intact GSH, that is taken up by cells.

If this pathway were active in the lens, it could account for the GSH content of

LEGSKO lenses, because they still have active GS and could generate GSH from cytosolic γ-GC if they can obtain it from their environment.

2.2.5. Purpose of Study

We set out to clarify the source and mechanism of GSH uptake in the lens by using an unbiased approach. To this end, we developed liquid chromatography- tandem mass spectrometry (LC-MS/MS) assays that could accurately and reliably measure GSH, GSSG, and isotopically-labeled compounds simultaneously. Unlike radiologic methods, this mass spectrometric analysis ensures that only intact compounds are measured, providing robust measurements of endogenous GSH levels in addition to labelled GSH derived from uptake and/or synthesis. Because it lacks de novo GSH synthesis in its lens, the LEGSKO mouse provides a unique model for investigating these mechanisms in detail.

2.3 Materials and Methods

2.3.1 Chemicals

Non-isotopic standards and other reagents of the highest available purity were obtained from Sigma-Aldrich Corp. (St. Louis, MO, USA). Isotopic compounds were obtained from Cambridge Isotope Laboratories, Inc. (Tewksbury,

MA, USA). γ-Glutamyl-serine-glycine (γ-ESG) and γ-glutamyl-alanine-glycine (γ-

EAG) peptides were synthesized at the Dr. Richard Armstrong Laboratory

(Vanderbilt University, Nashville, TN, USA) following a published protocol (Chen et al., 1985).

2.3.2 Animal Work

All animals were used in accordance with the guidelines of the Association for Research in Vision and Ophthalmology for the Use of Animals in

Ophthalmology and Vision Research, and experimental protocols for this study were approved by the Institutional Animal Care and Use Committee (IACUC) of

Case Western Reserve University.

2.3.3. LC-MS/MS Analysis

A mass spectrometer (MicroMass Quattro Ultima; Waters Corp., Milford,

MA, USA) equipped with an electrospray ionization source coupled to a separation module (Alliance 2695; Waters Corp.) with a reversed-phase C18 column

(Discovery HS model; Supelco Analytics, Bellefonte, PA, USA) was used for LC-

MS/MS analysis. Multiple reaction monitoring (MRM) was performed using electrospray ionization in positive ion mode with a cone voltage of 60 V. Precursor and product ions were determined by infusing standards on MS at a concentration of 100 μM. Precursor and product ions used for MRM are shown in Table 2.1. Also as shown in Table 2.1, two different mass transitions were used for quantitation of

13 15 both GSH and GSH-(glycine- C2, N) in order to improve accuracy and reduce matrix effects.

Table 2.1. LC-MS/MS MRM Settings.

Formic acid, 0.1%, was used as the ion-pairing agent, and 90% acetonitrile was used as the mobile phase. Compounds were separated using a 12-minute gradient with mobile phase increasing linearly from 2% to 40%. Concentration of compounds in samples was quantitated using external calibration of standards. In

13 15 measurements of endogenous glutathione content, GSH-(glycine- C2, N) was used as an internal standard.

All samples were diluted and homogenized in 0.1% formic acid solution for

LC-MS/MS analysis. Samples were centrifuged at 8000 × g for 10 minutes to precipitate protein and other large debris, and the supernatant was filtered over

0.45-µm cellulose acetate Spin-X columns (Corning, Inc., Corning, NY, USA).

2.3.4 Cultured Lens Uptake Experiments

Eyes were removed from mice immediately after they were euthanized by

CO2 asphyxiation. Lenses were dissected from eyes by carefully cutting away sclera and removing the lenses with forceps. Lenses were washed three times in sterile phosphate-buffered saline (PBS) to remove vitreous humor and other tissues

53 and then placed in sterile uptake buffer (140 mM NaCl, 25 mM d-(+)-glucose, 10 mM HEPES, 4.2 mM NaHCO3, 5 mM KCl, 1.3 mM CaCl2, 0.5 mM MgCl2, 0.36 mM Na2HPO4, 0.44 mM NaH2PO4, pH 7.4). Lenses were treated with 1 mM of the

GCLC inhibitor buthionine sulfoximine (BSO) and 500 μM of the γ-GT inhibitor acivicin for 1 hour prior to uptake experiments in order to prevent breakdown and synthesis of GSH during uptake experiments. Uptake buffer pH was adjusted using

NaOH to maintain a pH of 7.4 after addition of substrates. Unless otherwise noted, assays were performed at 37°C in 5% CO2. Uptake assays were stopped by washing lenses three times in ice-cold sterile PBS followed by immediate homogenization of the lenses in ice-cold 0.1% formic acid solution.

2.3.5 Lens Wash Test

13 15 Following 1 hour of incubation with 5 mM GSH-(glycine- C2, N), lenses were placed in 100 μL of ice-cold sterile PBS for 60 seconds with gentle agitation.

Lenses were then moved into fresh buffer, and this procedure was repeated a total of 6 times. Each wash and the post washed lenses were analyzed using LC-MS/MS after homogenization and dilution in 0.1% formic acid solution.

2.3.6 Cultured Lens GSH Efflux Assay

Lenses, obtained as stated for cultured lens uptake experiments, were placed in individual wells of a 96-well plate containing 200 μL of uptake buffer with 500

μM of acivicin and other inhibitors. Lenses were incubated at 37°C in 5% CO2 for

1 hour, after which medium was replaced and lenses were incubated for an additional hour. After incubation, a sample of medium was taken for analysis and diluted 1:3 in ice-cold 0.1% formic acid solution.

2.3.7 GSH Uptake Imaging

Glutathione was reacted with monobromobimane to generate a glutathione- bimane conjugate (GS-B) by adding GSH to a final concentration of 10 mM to a solution of 40 mM monobromobimane, 200 mM N-ethylmorpholine, 20 mM KOH, pH 8.0, and reacting at room temperature in the dark for 30 minutes. GS-B was isolated from the reaction by phase separation after the addition of methylene chloride. The extraction procedure was repeated four times, after which the aqueous phase maintained a bright yellow coloration, indicating the presence of the bimane conjugate, and the organic phase was completely clear. Purity of the GS-B product was assessed by LC-MS/MS (Figure 2.1).

Figure 2.1. Purity of synthesized GS-B tested by LC-MS/MS analysis. The GS-

B product was found to be 99.75% pure, with trace amounts of GSH and monobromobimane present.

Lenses were cultured as described previously and incubated with 1 mM GS-

B for 0, 5, 15, or 30 minutes, washed three times in cold PBS, and fixed for 30

55 minutes in 4% paraformaldehyde. Whole lenses were imaged using confocal microscopy (LSM 510 META model; Carl Zeiss AG, Oberkochen, Germany) with an excitation of 384 nm and emission of 470 nm for the bimane adduct.

2.3.8 In Vivo Uptake Experiments

Mice were anesthetized using intraperitoneal injection of 3 mg/20 g of body weight ketamine and 0.3 mg/20 g of body weight xylazine. Before injection of substrates into the eye, 1% atropine sulfate drops were applied topically to provide better visualization of the needle and internal eye structures.

Aqueous humor injections were performed by puncturing the cornea with a

27-gauge needle, drawing out endogenous aqueous with an ophthalmic sponge, and

13 15 injecting approximately 5 μL of uptake buffer containing GSH-(glycine- C2, N) using a 10-μL syringe equipped with a 33-gauge needle (Nanofil; World Precision

Instruments, Sarasota, FL, USA). A small air bubble was injected with the solution in order to prevent the solution from leaking out of the puncture site.

Vitreous humor injections were performed by puncturing mouse sclera with a 10-μL syringe (Nanofil) equipped with a 33-gauge needle at a 45° angle in order to avoid puncturing the lens, and injecting 1 μL of uptake buffer containing GSH-

(glycine-13C2,15N) directly into the vitreous body. Injection localization was confirmed by observation of small amounts of vitreous humor leaking from the injection site and by visual inspection of the lens following dissection.

2.3.9 Vascular Eye Perfusion

Mice were anesthetized by intraperitoneal injection of 3 mg/20 g of body weight ketamine and 0.3 mg/20 g of body weight xylazine, and eye perfusion was carried out based upon established surgical techniques to perfuse the eye and brain

(Cattelotte et al., 2008). Briefly, an incision was made in the chest to expose the throat and surrounding tissues. The right common carotid artery was ligated distal to the catheter insertion site. The artery was cut half way across its diameter, and a

32-gauge carotid artery catheter (World Precision Instruments, Sarasota, FL, USA) was inserted. Sutures were tied around the catheterized carotid artery to firmly maintain the catheter placement. The right internal and external jugular vein were cut fully across to allow for drainage of perfusion fluid. The catheter was connected to a syringe pump (model 11 plus; Harvard Apparatus, Holliston, MA, USA), and perfusion was carried out at a rate of 1 mL/min. Perfusion fluid consisted of a bicarbonate-buffered physiological saline solution (128 mM NaCl, 24 mM

NaHCO3, 4.2 mM KCl, 2.4 mM NaH2PO4, 1.5 mM CaCl2, 0.9 mM MgCl2, and 9

13 15 mM d-glucose, pH 7.4) containing 200 μM GSH-(glycine- C2, N). Mouse body temperature was maintained by using an electric heating pad. Perfusion fluid was heated to and maintained at 37°C for the duration of the perfusion.

2.3.10 Lens Surface Permeability Assay

Lucifer yellow uptake buffer, 100 μM, was administered to cultured lenses or injected into the anterior chamber of WT mouse eyes. After 1 hour of incubation, lenses were washed three times in PBS and homogenized in PBS. Lens samples were analyzed for fluorescence intensity alongside a lucifer yellow standard curve, with excitation at 485 nm and emission at 528 nm. Permeability coefficients (Pc)

were determined by the equation [Pc = (V/A × Ci) × (Cf/T)], where V is the lens

2 volume (in mL), A is the surface area of the cell layer assayed (in cm ), Ci is the concentration of lucifer yellow administered (100 μM), Cf is the concentration of lucifer yellow in the lens after incubation, and T is the incubation time (in seconds).

2.3.11 Tissue Procurement and Dissection

Frog, fish, and rat eyes were obtained as fresh discarded tissue from various laboratories at Case Western Reserve University. Aqueous humor was collected using a 10-μL syringe (Nanofil) equipped with a 33-gauge needle by fully puncturing the cornea just above the pupil at a 45° angle and drawing out the fluid.

The lens-vitreous-retina mass was removed from the eye by carefully cutting away the cornea and sclera. Vitreous, retina, and lens were separated using prewetted spin columns, following an established method (Skeie et al., 2011).

Fresh bovine and porcine eyes were obtained from local slaughterhouses.

Human donor eyes were obtained from the National Disease Research Interchange

(Philadelphia, PA, USA). Only healthy eyes without opacities were accepted.

Human eyes were from individuals 20 to 60 years of age who had not undergone radiation or chemical therapy.

Aqueous humor was collected from large mammal eyes by using a 1-mL syringe equipped with a 25-gauge needle by injection into the anterior chamber and drawing out the fluid. Vitreous humor and lenses were isolated by cutting along the sclera to access the posterior eye and removing the tissues with forceps.

2.3.12 Statistical Analysis

Bar graphs are expressed as means ± standard deviations (SD). Line graphs are expressed as means ± standard error of the means (SEM) for greater clarity.

Statistical significance of differences in mean values was assessed by Student's t- test. Only P values < 0.05 were considered statistically significant.

2.4 Results

2.4.1 Lenticular LEGSKO GSH Is Supplied by Circulating GSH

Using LC-MS/MS methodology, we first confirmed that lens GSH levels were suppressed by >60% (P < 0.01) in LEGSKO mouse lenses and that these levels could be further depleted by treating mice with BSO, an irreversible inhibitor of

GCLC (Figure 2.2). LEGSKO lenses were depleted of GSH by ~70% (P < 0.005) by using BSO treatment, although WT lenses lost only ∼15% of their GSH (P <

0.01). It has been reported that BSO has very poor penetration through the blood– brain and blood–eye barriers in adult mice (Steinherz et al., 1990) and will only have a large effect in circulation and organs not protected by such barriers. These experiments suggest that the residual GSH levels in the LEGSKO lens are linked to a GSH pool that is rapidly replenished from systemic circulation.

Figure 2.2. Analysis of the source of LEGSKO lens GSH. Treated mice were supplied exclusively with water containing 10 mM BSO for 2 months before analysis by LC-MS/MS. Wild-type mouse lenses were ∼15% depleted of GSH by

BSO treatment (P < 0.01). LEGSKO mouse lens GSH content was depleted ∼70% compared to WT before BSO treatment (P < 0.01) and ∼90% depleted from WT after BSO treatment (P < 5E-10). Values are means ± SD; n = 4.

2.4.2 Salvage Pathway Is Not the Source of LEGSKO Mouse Lens GSH

Next, we determined whether a GSH synthesis salvage pathway relying on

13 15 exogenous γ-GC was functioning in LEGSKO lenses. GSH-(glycine- C2, N) or

13 15 equal amounts of γ-GC and glycine-( C2, N) were injected into the anterior chambers of WT and LEGSKO mouse eyes, and the accumulation of GSH-

13 15 (glycine- C2, N) in lenses after injection was determined by LC-MS/MS.

13 15 Although WT lenses were found to accumulate GSH-(glycine- C2, N) equally for

13 15 both injections, LEGSKO lenses only accumulated GSH-(glycine- C2, N) when it was given in its intact form (P < 0.05) (Figure 2.3). This indicates that WT lenses

13 15 generated GSH-(glycine- C2, N) by de novo synthesis and confirms the fact that the high residual GSH content of LEGSKO lenses is due to uptake of intact GSH.

Because LEGSKO lenses were unable to synthesize GSH from exogenous γ-GC, we conclude that a salvage pathway relying on GGT activity cannot be the source of LEGSKO lens GSH.

Figure 2.3. Analysis of the mechanism of LEGSKO lens GSH uptake. To test

13 15 for GSH salvage pathway activity, we injected 250 μM of GSH-(glycine- C2, N)

13 15 or glycine-( C2, N) and γ-GC into the anterior chambers of WT and LEGSKO mouse eyes. After 30 minutes, lenses were removed and analyzed by LC-MS/MS.

13 15 Wild-type lenses accumulated equal amounts of GSH-(glycine- C2, N) from both

13 15 treatments. Conversely, LEGSKO mice only accumulated GSH-(glycine- C2, N) when it was administered in its intact form (P < 0.05). Values are means ± SD; n =

2.4.3 Cultured Lens GSH Uptake Occurs by Passive Diffusion

To obtain detailed kinetic measurements of lens GSH uptake, whole lenses were taken from mice and cultured ex vivo. Lenses were pretreated with BSO, an inhibitor of GCLC, and acivicin, an inhibitor of GGT (Smith et al., 1995), to prevent any potential breakdown or synthesis of GSH. These lenses were incubated with isotopically labeled compounds for analysis of uptake by LC-MS/MS. Cultured

13 15 LEGSKO and WT lenses showed no differences in GSH-(glycine- C2, N) uptake rates (Figure 2.4). Because this indicates that LEGSKO lenses take up GSH by the same mechanism as WT lenses, WT lenses were used for subsequent experiments.

Figure 2.4. Comparison of GSH uptake in cultured lenses. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any

13 15 breakdown or turnover of GSH-(glycine- C2, N). Wild-type and LEGSKO lenses

13 15 were incubated with 500 μM GSH-(glycine- C2, N) at 37°C in 5% CO2 for 1 hour.

No significant differences were observed between uptake rates of LEGSKO and those of WT lenses. Values are means ± SD; n = 4.

13 15 For comparative purpose, the uptake of C3-alanine and N2-arginine was

13 15 tested in addition to GSH-(glycine- C2, N), because both alanine (Kern et al.,

1977) and arginine (Fan et al., 2011) have been shown to be actively taken up by

63 mammalian lenses. A time course of uptake demonstrated that all three compounds were taken up by the lens with an initial rate period occurring in the first few minutes (Figure 2.5).

Figure 2.5. Time course of uptake in cultured lenses. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any breakdown or

13 15 turnover of GSH-(glycine- C2, N). Wild-type mouse lenses were incubated with

13 15 2 mM of each substrate and taken at various time points. GSH-(glycine- C2, N)

15 13 had the lowest initial uptake rate, followed by N2-arginine and C3-alanine (P <

0.005). Values are means ± SEM; n = 4.

Subsequent experiments were carried out using initial rate conditions to ensure first-order kinetics. Transport was found to be dose dependent in all cases

13 15 (Figure 2.6). C3-alanine and N2-arginine uptake rates were saturable and displayed Michaelis-Menten kinetics with estimated Km values of ∼1.9 mM and

∼3.3 mM, respectively, and Vmax of 22 μM/min and 18 μM/min, respectively. The rates found in this study match relatively well with previous studies which showed uptakes rates of ~0.27 µM/min for alanine when it was administered at a concentration of 100 µM (Kern et al., 1977) and ~14 µM/min for arginine when administered at a concentration of 10 mM (Fan et al., 2011). GSH-(glycine-

13 15 C2, N) uptake was very sluggish and did not appear to be saturable or fit well to the Michaelis-Menten equation, appearing instead to increase essentially linearly with concentration.

Figure 2.6. Initial rate curves of uptake in cultured lenses. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any

13 15 breakdown or turnover of GSH-(glycine- C2, N). Lenses were incubated in

65 various concentrations of substrates and taken for analysis within the initial rate period of uptake. Curves are best fits of the data to the Michaelis-Menten equation.

13 15 13 GSH-(glycine- C2, N) showed a significantly lower rate of uptake than C3-

15 alanine and N2-arginine at all concentrations (P < 0.01). Values are means ±

SEM; n = 4.

To determine whether GSH uptake was mediated by active transport, uptake

13 15 was tested at both 37°C and 4°C. Whereas C3-alanine and N2-arginine uptake rates were decreased by 75% (P < 0.05) and 81% (P < 0.01), respectively, lowering

13 15 temperature had no discernable effect on GSH-(glycine- C2, N) uptake (Figure

13 15 2.7A). Furthermore, C3-alanine and N2-arginine uptake rates were significantly

13 15 higher than the rate of GSH-(glycine- C2, N) uptake at 37°C (P < 0.05) but approximately equal at 4°C (Figure 2.7B), suggesting that GSH uptake by the lens is likely a nonspecific and passive process.

Figure 2.7. Temperature dependence of transport in cultured lenses. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any

13 15 breakdown or turnover of GSH-(glycine- C2, N). Lenses were incubated with 2 mM of each substrate at 37°C or 4°C. (A) Lowering temperature had no effect on

13 15 13 GSH-(glycine- C2, N) uptake but did significantly inhibit uptake of C3-alanine,

15 13 and N2-arginine (P < 0.05 and P < 0.001, respectively). (B) C3-alanine,

15 and N2-arginine were taken up at significantly higher rates than GSH-(glycine-

13 15 C2, N) at 37°C (P < 0.05), but all compounds were taken up similar rates at

4°C. Values are means ± SD; n = 4.

This was further supported by performing uptake experiments with a 5-fold excess of unlabeled GSH or the GSH analogs γ-EAG or γ-ESG, while maintaining

13 15 a constant concentration of GSH-(glycine- C2, N) (Figure 2.8). None of these

13 15 compounds had any significant effect on GSH-(glycine- C2, N) uptake, indicating a lack of competitive inhibition and confirming that lens GSH uptake is not carrier- mediated under conditions of the intact lens immersed in culture medium.

Figure 2.8. Test for competitive inhibition of uptake. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any breakdown or

13 15 turnover of GSH-(glycine- C2, N). Lenses were incubated with 1 mM of GSH-

13 15 (glycine- C2, N) at 37°C in 5% CO2 and 5 mM of GSH, γ-EAG, or γ-ESG. There

13 15 were no significant differences between the uptake rates of GSH-(glycine- C2, N) between these groups. Values are means ± SD; n = 4.

13 15 To clarify whether the measured GSH-(glycine- C2, N) is localized intracellularly or more loosely associated with the lens, lenses were washed

13 15 extensively in PBS after uptake assays, and the amount of GSH-(glycine- C2, N) in each wash was determined by LC-MS/MS (Figure 2.9). The amounts of GSH-

13 15 (glycine- C2, N) present in the washes were negligible relative to the amount

13 15 present in the lens, indicating that the GSH-(glycine- C2, N) measured in uptake assays was likely located intracellularly and not in the extracellular fluid surrounding lens fiber cells.

13 15 Figure 2.9. Wash test to assess cellular localization of GSH-(glycine- C2, N).

Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to

13 15 prevent any breakdown or turnover of GSH-(glycine- C2, N). To assess how

13 15 tightly bound to lenses GSH-(glycine- C2, N) was, lenses were incubated with 5

13 15 mM GSH-(glycine- C2, N) for 30 minutes and then washed 6 times in 100 μL of

13 15 PBS. Each wash fraction was saved and analyzed for GSH-(glycine- C2, N) content by LC-MS/MS. Values are means ± SD; n = 4.

In order for GSH uptake to be a passive process, endogenous GSH should efflux from the lens at the same rate at comparable concentrations. This was tested by culturing lenses in GSH-free buffer and measuring the amount of GSH released into the medium by using LC-MS/MS (Figure 2.10).

Figure 2.10. Measurement of cultured lens GSH efflux. Lenses were preincubated with 1 mM BSO and 500 μM acivicin for 1 hour to prevent any breakdown or turnover of GSH. The efflux of GSH from lenses was tested by measuring the GSH content of buffer after 1 hour of incubation. This efflux rate was significantly lower than the uptake rate at the GSH concentration found within lenses (P < 0.05), was unaffected by ouabain or Na+-free medium, and increased to

13 15 match the rate of uptake when GSH-(glycine- C2, N) was present in the medium at higher concentrations than within the lens (10 mM) but not when it was present at a lower concentration (2.5 mM). The increased efflux induced by excess GSH-

13 15 (glycine- C2, N) was inhibited by the connexin inhibitors 18β-glycyrrhetinic acid and octanol at concentrations of 10 μM and 100 μM, respectively. Values are means

± SD; n = 4.

Under these conditions, GSH was released from lenses but at a significantly

(P < 0.05) slower rate than anticipated based on the rate of uptake at GSH concentrations similar to those found in the lens. One possibility is that this was due to the activity of the lens microcirculation, which drives solute flow into the

70 center of the lens (Mathias et al., 1997). However, ouabain (an inhibitor of Na+-

K+-ATPase pumps) and Na+-free buffer had no effect on the GSH efflux rate, indicating that the microcirculation does not play a significant role in this system.

13 15 However, when GSH-(glycine- C2, N) was present in buffer at a concentration higher than the lenticular GSH concentrations, efflux rate increased significantly (P

< 0.05) and matched the expected rate. Furthermore, the connexin inhibitors 18β- glycyrrhetinic acid and octanol reduced the efflux rate back to that of the control (P

< 0.05), indicating that a facilitated but passive GSH exchange in lenses may be at least partially mediated by connexin hemichannels that are gated and influenced by relative GSH concentrations.

Localization of GSH uptake was assessed by incubating lenses with a fluorescent GS-B conjugate. Although this conjugate may have slightly different properties than GSH, it was determined that GS-B was taken up by lenses at the

13 15 same rate as GSH-(glycine- C2, N) (Figure 2.11) and is thus likely taken up by the same mechanism.

13 15 Figure 2.11. Comparison of GS-B and GSH-(glycine- C2, N) uptake rate. WT lenses were incubated with 2 mM of each compound in uptake buffer for 4 minutes and analyzing the intralenticular accumulation of the compound by LC-MS/MS.

There were no significant differences in uptake rates. Values are means ± SD; n =

A time course of GS-B uptake revealed that uptake was visible at both the anterior and poster lens within 5 minutes of incubation, but uptake along the lens posterior and equator appeared to occur much more rapidly than at the anterior, accounting for the majority of uptake seen at 15 and 30 minutes (Figure 2.12). GS-

B entering from the posterior lens appeared to progressively accumulate in cortical fiber cells and diffuse toward the anterior lens surface. Penetrance into the lens core was low, and GS-B only appeared significantly in the region after 30 minutes of incubation. This is not consistent with the lens microcirculation model of solute

72 delivery and further confirms that this system is unlikely to be responsible for GSH uptake in cultured lenses.

Figure 2.12. Visualization of GS-B uptake in cultured lenses. Representative images are shown of GS-B uptake by cultured lenses after a 0, 5, 15, or 30 minutes incubation with 1 mM GS-B.

2.4.4 Lens GSH Uptake From Aqueous Humor Is Carrier- Mediated

Glutathione uptake at the anterior lens was characterized by replacing endogenous aqueous humor with uptake buffer containing various concentrations

13 15 of GSH-(glycine- C2, N) (Figure 2.13). Anterior lens GSH uptake in both WT and LEGSKO appeared to be saturable and fit well to the Michaelis-Menten equation, with an apparent Vmax of ∼0.21 μM/min. LEGSKO lens uptake showed

an enhanced affinity for GSH with an apparent Km of ∼50 μM compared to Km of

∼250 μM in WT, indicating a faster uptake rate under physiological conditions.

Figure 2.13. Comparison of initial uptake rates from anterior chamber in WT and LEGSKO lenses. Uptake buffer containing various concentrations of GSH-

13 15 (glycine- C2, N) was heated to 37°C and injected into the anterior chambers of

WT and LEGSKO mouse eyes after aqueous humor was removed. Lenses were taken for analysis after 30 minutes of incubation. Curves are best fits to the

Michaelis-Menten equation. Both of the groups had approximately the same

Vmax of 0.21 μM/min, but LEGSKO mice had a lower Km value of ∼50 μM compared to ∼250 μM for WT lenses. Values are means ± SEM; n = 4.

Glutathione uptake at the anterior lens could be significantly (P < 0.005) inhibited by including an excess of unlabeled GSH or the closely analogous peptide

γ-EAG or γ-ESG in the uptake buffer, indicating a carrier-mediated mechanism

(Figure 2.14). Because the serine and alanine analogs of GSH could competitively inhibit the uptake to the same degree as GSH, it appears that this uptake is not highly selective for GSH and could be a more general peptide transport mechanism.

Figure 2.14. Test for competitive inhibition of uptake from the anterior

13 15 chamber. Uptake buffer containing 125 μM GSH-(glycine- C2, N) and 500 μM

GSH, γ-EAG, γ-EAG or none of these was injected into the anterior chambers of

WT mouse eyes after removal of the aqueous humor. Lenses were taken for analysis after 30 minutes of incubation. All three compounds significantly inhibited lens

13 15 uptake of GSH-(glycine- C2, N) (∼75%; P < 0.005). There were no significant differences in degree of inhibition among the compounds. Values are means ±

SD; n = 4.

The permeability of lens cell barriers was tested at the anterior lens and the whole cultured lens, using lucifer yellow (Figure 2.15). Whole, cultured lenses had a >20-fold higher Pc (P < 0.05) than the anterior lens surface. This indicates that the anterior lens surface is much less permeable than the posterior lens, which may account for the differences between GSH uptake mechanisms in cultured lenses and that in the anterior lens.

Figure 2.15. Permeability of lens surfaces. 100 μM lucifer yellow in uptake buffer was administered to cultured lenses or injected into the anterior chamber of WT mouse eyes. Cultured lenses showed a significant increase in Pc (P < 0.05) compared to that in lenses treated with lucifer yellow only at the anterior surface.

Values are means ± SD; n = 3.

2.4.5 Mouse Vitreous Contains High Levels of GSH

Glutathione content in aqueous, vitreous, and lenses from WT and

LEGSKO mice was analyzed by LC-MS/MS (Figure 2.16). These data revealed surprisingly high (>1 mM) vitreous GSH content and very low (<20 μM) aqueous

GSH content in both LEGSKO and WT lenses. LEGSKO vitreous and lens GSH concentrations were robustly the same, strongly implying an equilibration between the tissues.

Figure 2.16. Comparison of concentrations of aqueous, vitreous, and lens GSH in LEGSKO and WT eyes. Aqueous, vitreous, and lenses were dissected from WT and LEGSKO mice and analyzed for total glutathione content. Wild-type and

LEGSKO aqueous and vitreous had similar glutathione content, but LEGSKO lenses showed a significant loss of glutathione (P < 0.01). LEGSKO lens glutathione concentration matches LEGSKO vitreous glutathione concentration.

Values are means ± SD; n = 4.

2.4.6 Vitreous but Not Aqueous Contributes Highly to Lens GSH In Vivo

The ability of vitreous and aqueous GSH pools to contribute to the lens GSH pool in vivo was tested at physiologically relevant concentrations by injecting WT

13 15 and LEGSKO mice with GSH-(glycine- C2, N) to final concentrations of approximately 20 μM and 1 mM in aqueous and vitreous, respectively. Figure

2.17A shows that the lens can take up, in vivo, a significant (P < 0.05) amount of

GSH from the available pool in the vitreous but not in the aqueous and that

LEGSKO and WT uptake rates are the same. It should be noted that Figure 2.17A is presented as a linear time course to more clearly display the results but the linearity of this uptake remains unknown without further measurements. As in ex vivo lens culture experiments, excess unlabeled GSH could not inhibit the uptake

13 15 of GSH-(glycine- C2, N) from the vitreous (Figure 2.17B).

Figure 2.17. Comparison of in vivo GSH transport and homeostasis in mouse eyes. (A) Wild-type and LEGSKO mice were anesthetized, and GSH-(glycine-

13 15 C2, N) in physiological (pH 7.4) solution was injected to a final concentration of

1 mM in the vitreous body or 20 μM in the anterior chamber of eyes in order to mimic physiological conditions. After 1 hour, mice were euthanized, and lenses

13 15 were analyzed along with control lenses for GSH-(glycine- C2, N) content.

Anterior chamber injections did not lead to an increase in lens GSH-(glycine-

13 15 C2, N), but intravitreal injections did lead to a significant accumulation of the compound in lenses (P < 0.05). LEGSKO and WT lenses showed no differences in

13 15 rates of GSH-(glycine- C2, N) accumulation. Values are means ± SEM; n = 4. (B)

13 15 Mice were injected with GSH-(glycine- C2, N) intravitreally as in A, but one group was also injected with unlabeled GSH at a 10-fold excess. Solutions were

13 15 formulated such that the final concentration of GSH-(glycine- C2, N) injected remained the same and was not diluted. In both the WT and the LEGSKO mice, this excess unlabeled GSH had no effect on the lens uptake of GSH-(glycine-

13 15 C2, N). Values are means ± SD; n = 4.

To determine the relative rates at which circulating GSH moved into these compartments, mouse eyes were perfused by catheterization of the common carotid

13 15 artery, which was pumped with fluid containing GSH-(glycine- C2, N) (Figure

13 15 2.18). GSH-(glycine- C2, N) accumulated within the vitreous more rapidly than the lens or aqueous (P < 0.05). Furthermore, the vitreous and the lens both

13 15 accumulated a significant (P < 0.05) amount of GSH-(glycine- C2, N) before any was detectable in the aqueous.

Figure 2.18. Measurement of blood to lens GSH transport. Wild-type mouse carotid arteries were catheterized and pumped with bicarbonate-buffered

13 15 physiological solution containing 200 μM GSH-(glycine- C2, N) at a rate of 1 mL/min. Mice were anesthetized and kept warm throughout. Ocular tissue was

13 15 13 15 isolated and analyzed for GSH-(glycine- C2, N) content. GSH-(glycine- C2, N) was detectable in lens and vitreous samples before any accumulated in the aqueous

(P < 0.05) and accumulated in vitreous humor to a significantly higher degree than in lens or aqueous humor (P < 0.005). Values are means ± SEM; n = 4.

In order to further examine the relationship between systemic, vitreous, and lenticular GSH, circulating GSH was depleted in LEGSKO mice by replacing drinking water with water containing various concentrations of BSO. This treatment had little to no effect on aqueous humor GSH but significantly lowered lens and vitreous GSH contents (P < 0.05) (Figure 2.19). At all concentrations of

BSO given, the concentrations of GSH in the lens and vitreous of these mice were approximately equal, supporting the hypothesis that the vitreous provides GSH to the lens through a passive process.

Figure 2.19. GSH content of LEGSKO mouse eyes treated with BSO. Mice were given BSO-containing water of various concentrations for 2 months. Total glutathione contents of lens, vitreous, and aqueous were determined. A concentration of 10 mM BSO was sufficient to significantly lower (P < 0.05) lens and vitreous glutathione contents but had no effect on aqueous glutathione levels.

For each BSO treatment, there were no significant differences between the lens and vitreous glutathione concentrations.

2.4.7 High Vitreous GSH Content Is Unique to Small Animals

Finally, we investigated whether the above observation of high vitreous

GSH in mice was applicable to other species by determining glutathione content of ocular tissue compartments from WT human, rat, cow, pig, Xenopus frogs, minnows, and mice (Figure 2.20).

All small animals had ∼1 mM vitreous GSH, whereas large mammals, including humans, pigs, and cows, had much lower levels (<100 μM). All species had low levels (<100 μM) of GSH in aqueous humor. Intriguingly, fish and frogs had significantly (P < 0.05) lower lens GSH than other species, which was at approximately the same level as their vitreous humor. This could indicate that vitreous humor is a major source of lens GSH in these species.

Figure 2.20. Comparative ocular distribution of glutathione in aqueous, vitreous, and lens in eyes of various species. (A) Fresh tissue was taken from members of each species. Bars with different letters show a statistically significant difference (P < 0.05). Rat, mouse, minnow, and frog vitreous glutathione contents were significantly (P < 0.05) higher than those of human, bovine, or porcine vitreous. Lens glutathione contents were similar for human, rat, mouse, and porcine eyes but significantly elevated (P < 0.05) in bovine eyes and significantly lower in minnow and frog eyes (P < 0.05). Values are means ± SD. Bovine, mouse, and porcine n = 4; rat and human n = 3.

Unlike the vitreous humor in rodents and other small animals, the vitreous humor of large mammals is large in volume and can be dissected into distinct regions, including the base (near the lens), cortex (near the retina), and core. These regions were isolated and analyzed for GSH content in human and porcine eyes

(Figure 2.21). Base vitreous was found to have significantly (P < 0.005) higher

GSH than that in the core or cortex in humans and significantly (P < 0.01) higher

GSH than that in the core but not the cortex in porcine eyes.

Figure 2.21. Regional distribution of GSH within human and porcine vitreous.

Vitreous was sampled at the base (near lens), core, and cortex (near retina) regions.

Human base vitreous had significantly (P < 0.005) higher GSH content than that in core and cortex, which had similarly low levels of GSH. Porcine base vitreous had significantly higher (P < 0.01) GSH content than the core but not the cortex. Values are means ± SD. Porcine n = 4; human n = 3.

2.5 Discussion

2.5.1 Evidence of Vitreous GSH Transport From Other Studies

This paper represents, to our knowledge, the first report of significant vitreous-to-lens GSH transport. To date, researchers studying lens uptake of GSH have focused their attention on the aqueous humor, which is conventionally considered the source of lens nutrition. However, several other studies have demonstrated that vitreous GSH is dynamic and that circulating GSH readily enters the vitreous. Measurement of GSH in rabbit ocular tissues by an HPLC method revealed that vitreous total glutathione nearly doubled 3 hours after treatment with

N-acetylcysteine (Nozal et al., 1997). Even without treatment, vitreous GSH concentration was found to be very high, just under 20% of that found in lenses.

Another study directly measured the production and movement of 35S-GSH in ocular tissues after injection of 35S-GSSG into the anterior chamber or vitreous body of rabbit eyes (Veltman et al., 2004). Although anterior chamber injections led to an accumulation of 35S-GSH in corneal cells, the lens did not appear to take up any significant amount of 35S-GSH. Conversely, the lens did uptake a significant amount of 35S-GSH from intravitreal injections, which increased steadily in the lens cortex and then slowly into the lens nucleus. In a study by Stewart-DeHaan et al.

(1997), rats were injected intraperitoneally with 3H- or 35S-GSH, and the uptake of these compounds into ocular tissues was measured by HPLC. From these measurements, it was found that the lens could obtain more than 12% of its total

GSH from circulating GSH in just 4 hours. To determine the source of GSH transported into the lens, the researchers compared the concentration of 3H-GSH and unlabeled GSH in the aqueous, vitreous, and lens after injection. Although the

3H-GSH-to-GSH ratio was much higher in the aqueous than in the vitreous, this was due only to the large difference in endogenous GSH content of these fluids, because the actual levels of 3H-GSH and 3H-GSSG found in the vitreous were

>200-fold greater than that in the aqueous. Additionally, the concentration of 3H-

GSH in the vitreous increased from 30 minutes to 4 hours, whereas aqueous 3H-

GSH did not. Thus, despite the authors' statement that aqueous humor must have been the source of lens 3H-GSH during that experiment, reevaluation of the results shows that the data appear to robustly support our mouse eye perfusion results.

2.5.2 Uptake From Vitreous but Not Aqueous Is Sufficient to Maintain Steady State GSH Concentration in LEGSKO Lenses

While it is clear that some uptake of GSH can and does occur at the anterior lens, it cannot account for the levels of GSH found in the whole LEGSKO lens.

Our measurements show that the rate of GSH transport into the LEGSKO lens from the aqueous humor at physiologically relevant concentrations is approximately 35 nM/min (Figure 2.13). Glutathione turnover in rodent lenses has been measured in the range of 0.014%−1hr to 0.018%−1hr (Reddy, 1990). In mice, this equates to a

GSH turnover rate of approximately 1 μM/min. Based on this, it is simply not possible for lenses to maintain a millimolar concentration of GSH by transporting

GSH from the aqueous, although it may be an important mechanism for protecting the epithelium from oxidative stress. Conversely, measurements of lens GSH uptake at concentrations found within the vitreous show a rate of ∼1 μM/min

(Figure 2.6). This shows that vitreous GSH content is sufficient to maintain a high steady state concentration in lenses if it is continuously supplied. Our perfusion

86 data showed circulating GSH entering the vitreous humor at an initial rate of ∼1 to

2 μM/min (Figure 2.18). Thus, vitreous GSH that is lost to the lens appears to be readily replaced by circulating GSH.

2.5.3 Vitreous GSH Flows into Lenses but Not Vice Versa

Interestingly, while our ex vivo and in vivo data are in agreement that lens

GSH uptake from the vitreous is not an energy-dependent process, transport appears to occur significantly in only one direction. LEGSKO and WT vitreous

GSH contents are robustly the same, indicating that the lens does not supply the vitreous with its high GSH content (Figure 2.16). If the lens and vitreous truly do exchange GSH through a passive process, why do WT lens and vitreous GSH pools not reach an equilibrium? One potential explanation is that lens microcirculation causes an inward flux of solutes along the posterior lens and that lenticular GSH cannot move against this current. However, we have determined that inhibiting the microcirculation has no effect on lens GSH release (Figure 2.10), and imaging of

GS-B indicates that microcirculation is not the delivery method for GSH (Figure

2.12). More likely, lens-vitreous GSH exchange is mediated by hemichannels, and the opening of these channels is gated. We have demonstrated that GSH release from lenses appears to be at least partially mediated by extralenticular GSH levels and that GSH release can be partially inhibited by connexin inhibitors (Figure 2.10).

This is consistent with the findings that gated hemichannels, which have a known permeability to GSH (Slavi et al., 2014), have been found on fiber cell membranes along the posterior lens (Shahidullah et al., 2012). However, the methods used here

87 were nonspecific, and further research will need to be performed to conclusively determine the exact mechanism.

Based on these data, it appears that mice have adapted a system wherein the vitreous can serve as a pool of GSH that will flow into the lens when lens GSH content is the same or lower than vitreous GSH content, but under normal conditions, lens GSH will not flow into the vitreous to a significant degree. This allows WT lenses to maintain a high lens-to-vitreous GSH gradient while also permitting LEGSKO lenses to equilibrate with the vitreous. However, vitreous

GSH is not a significant source of lens GSH in mouse lenses with active synthesis.

2.5.4 Vitreous GSH Varies Greatly Among Species

We found striking differences in the vitreous humor GSH content of small animals, including rodents and large mammals (Fig 2.20). Very similar results were reported by others when comparing rat, rabbit, and bovine vitreous GSH (Rose et al., 1997). The differences we report on vitreous GSH content of various species show no correlation with lens GSH content or temporal (i.e., diurnal/nocturnal) activity of the species. The most obvious explanation for the difference in vitreous

GSH content is that it results from morphologic differences in eyes based on species size. In large mammals, the vitreous humor typically makes up >80% of the eye volume, whereas in the eyes of smaller species, the lens makes up the major portion of the eye volume, and the vitreous humor volume is relatively small. In large- mammal eyes, GSH entering the vitreous from the retina or ciliary body, whether by diffusion or active transport, has a comparatively large volume to diffuse through relative to the surface area of cells secreting it. Thus, it is expected that

GSH could not easily reach a high concentration in the vitreous of these eyes.

Conversely, in mouse, rat, fish, and frog eyes, the volume of the vitreous is very small and GSH entering from the retina or the ciliary body, or both, could reach a much higher concentration. Rabbits form an interesting compromise in that they are a larger rodent with an eye volume of ∼60% vitreous humor. As would be expected, rabbit vitreous GSH content falls between that found in mice and rats and larger mammals, with ∼650 μM reported (Rose et al., 1997). Based on these findings, it is possible that vitreous GSH content is not highly regulated and may just be a result of the high (∼3 to 5 mM) GSH content in surrounding tissues, such as the retina and ciliary body. Further studies of GSH transport into and throughout the vitreous and an assessment of vitreous oxidative compounds in various species will be required to fully understand the role and regulation of vitreous GSH.

2.5.5 Role of Vitreous GSH in Health and Disease

Due to the striking differences in GSH contents of human and mouse vitreous, it is difficult to conclude the role of the vitreous GSH pool in human lens homeostasis. However, the elevated levels of GSH found in the vitreous base

(Figure 2.21) indicate that the vitreous GSH pool may be important for protecting lenses or providing constituent amino acids because these levels are nearly 50-fold higher than those found in the human aqueous. This may help explain why removal of the vitreous is tightly linked to the formation of cataracts in humans. Of patients who undergo vitrectomy surgery, 60% to 98% develop cataract within 2 years according to various reports (Cherfan et al., 1991;Holekam et al., 2005; Van

Effenterre et al., 1992). A loss of vitreous humor as a source of nutrients and

89 protective buffer against oxygen and oxidative species could greatly predispose lenses to cataract.

Additionally, it has been shown that human lens GSH synthesis is approximately 16-fold decreased over an 83-year timespan, even though lens GSH content only decreases ∼50% during this time (Rathbun et al., 1984). This high residual GSH cannot be accounted for by passive diffusion, as in LEGSKO, and implies that human lenses have a more robust lenticular GSH transport system than rodents. Such a difference would not be unprecedented because human lenses show a 10-fold greater propensity for the uptake of ascorbic acid, another important lenticular antioxidant, than rodents (Obrenovich et al., 2006). Given the much greater levels of GSH in the vitreous surrounding the lens compared to the aqueous, it would appear to be the more likely source of transported GSH. The relationship between the lens and vitreous in humans requires further investigation and could be of critical importance for maintaining human lenses in a healthy reduced state.

2.5.6 Conclusions

We have shown, for the first time that the vitreous humor can supply high levels of GSH to the mouse lens in the absence of its synthesis. Transport of GSH into the lens from this compartment is a passive process, in contrast to the anterior lens, where we show that an active carrier-mediated uptake of GSH occurs but with a significantly lower rate of transport that indicates it is not a major source of lens

GSH. Further study is needed to determine the mechanism by which rodent vitreous obtains such high concentrations of GSH. Our findings bring into question the concepts that lenses receive nutrients solely from the aqueous humor and that the

90 vitreous is not a highly dynamic fluid compartment, and indicate that the vitreous may be an essential component in the maintenance of healthy lenses.

2.6 Acknowledgments

Peptides γ-EAG and γ-ESG were kindly provided by the Dr. Richard

Armstrong Laboratory of Vanderbilt University, Nashville, TN, USA. Assistance in visualizing GSH uptake was provided by Catherine Doller and Scott Howell,

PhD, Department of Ophthalmology and Visual Sciences, Case Western Reserve

University, and Maryanne Pendergast, Department of Neuroscience, Case Western

Reserve University.

3. Transcriptome of the GSH-Depleted Lens Reveals Changes in Detoxification and Epithelial Mesenchymal Transition (EMT) Signaling Genes, Transport Systems, and Lipid Homeostasis

Adapted from submitted article in review: “Whitson JA, Zhang X, Medvedovic M, Chen J, Monnier VM, Fan X. Transcriptome of the GSH-Depleted Lens Reveals Changes in Detoxification and EMT Signaling Genes, Transport Systems, and Lipid Homeostasis. Investigative Ophthalmology & Visual Sciences.” 92

3.1 Abstract

3.1.1 Purpose

To understand the effects of chronic glutathione (GSH)-deficiency on genetic processes that regulate lens homeostasis and prevent cataractogenesis.

3.1.2 Methods

The transcriptome of lens epithelia and fiber cells was obtained from

C57BL/6 LEGSKO (lens GSH synthesis knockout) and buthionine sulfoximine

(BSO)-treated LEGSKO mice and compared to C57BL/6 wild-type mice using

RNA-Seq.

3.1.3 Results

RNA-Seq results were in excellent agreement with qPCR (correlation coefficients between 0.87-0.94 and P < 5E-6 for a subset of 36 mRNAs). Of the

24,415 transcripts mapped to the mouse genome, 441 genes showed significantly modulated expression. Pathway analysis indicated major changes in EMT signaling, visual cycle, small molecule biochemistry, and lipid metabolism. GSH- deficient lenses showed upregulation of genes relating to detoxification, including

Aldh1a1, Aldh3a1 (aldehyde dehydrogenases), Mt1, Mt2 (metallothioneins), Ces1g

(carboxylesterase), and Slc14a1 (urea transporter UT-B). These gene products act on many of the same toxins as GSH or glutathione-S-transferase and may protect

GSH-deficient lenses. Genes associated with canonical EMT pathways, including

Wnt10a, Egf, and Syk, showed upregulation in lens epithelia samples. Severely

GSH-deficient lens epithelia showed a broad downregulation of vision-related

93 genes (including Cryge, Crygf, and Rho). The BSO-treated LEGSKO lens epithelia transcriptome has significant correlation (r = 0.63,P < 0.005) to that of lens epithelia undergoing EMT.

3.1.4 Conclusions

These results show that GSH depletion of the lens leads to expression of detoxifying genes and activation of EMT signaling, in addition to changes in transport systems and lipid homeostasis. These data give new insight into the adaptation and consequences of GSH-deficiency in the lens and suggest that supplementation of GSH or a precursor after cataract surgery could potentially reduce the incidence of EMT-mediated posterior subcapsular opacification.

3.2. Introduction

3.2.1. The Role of Glutathione in Age-Related Cataract

Age-related cataract is a multifactorial disease associated with the accumulation of post-translational modifications on lens proteins that lead to the formation of insoluble aggregates which scatter light and reduce visual acuity

(Petrash, 2013). One defensive adaptation that helps prevent damage to lens proteins is maintenance of high concentrations of the versatile antioxidant glutathione (GSH), which exists at millimolar levels within the lens (Giblin, 2000).

It has been found that this lenticular GSH content decreases significantly with advanced age, particularly in the lens nucleus, which is thought to occur due to a decrease in its synthesis in the metabolically active lens cortex (Rathbun et al.,

1993) and a growing barrier around the lens nucleus that prevents its diffusion from the lens cortex (Sweeney and Truscott, 1998). This progressive loss of GSH may play an important role in age-related cataractogenesis by making lens proteins more susceptible to modifications that result in their insolubility and aggregation.

3.2.2. Analysis of the LEGSKO Mouse

In order to better study the role of glutathione in the lens and its relation to cataract, the Lens Glutathione Synthesis KnockOut (LEGSKO) mouse model was generated (Fan et al., 2012). These mice were developed using a lox/Cre system that specifically knocks out GCLC, the first step in GSH synthesis, in the lens. As a result, these mice show reduced lens GSH content and develop opacities that mimic human age-related cataract. However, severe opacities do not develop until

>9 months of age and, with subsequent generations, these mice show a further delayed onset of cataractogenesis.

Based on these results, we set out to determine how these mouse lenses adapt to GSH-deficiency and maintain the integrity of their lens protein. This question was partially addressed with the discovery that mouse eyes have >1 mM levels of GSH in their vitreous humor and that the lens and vitreous GSH pools equilibrate via passive diffusion when the lens becomes deficient in GSH (Whitson et al., 2016). This vitreous GSH pool is derived from circulating GSH and can be depleted by treating mice with the systemic GSH synthesis inhibitor buthionine sulfoximine (BSO), which also results in the depletion of LEGSKO lens GSH

(Whitson et al., 2016). The eyes of humans and other large animals lack this vitreous GSH reservoir, which may be one aspect of why species such as humans

95 and canines appear to be more susceptible to cataract formation than rodents

(Ohrloff and Hockwin, 1983). Such species may be especially reliant on the activity of other gene products to maintain a reduced and toxin-free lens environment with advanced age when the lenticular GSH pool diminishes.

3.2.3. Purpose of Study

In this study, we set out to determine the gene expression changes that occur in the lens in response to GSH-depletion in order to better understand how the lens adapts to the consequences of GSH-depletion with age.

3.3. Materials and Methods

3.3.1. Animal work

Both the LEGSKO and wild-type mice used in this study were of C57BL/6 genetic background and age-matched at 6 months. Groups were as follows: four male wild-type mice, four male LEGSKO mice, and four male LEGSKO mice exclusively receiving drinking water containing 10 mM buthionine sulfoximine

(BSO) for 1 month prior to sample collection. Mice were housed under diurnal lighting conditions and allowed free access to food and water. All animals were used in accordance with the guidelines of the Association for Research in Vision and Ophthalmology for the Use of Animals in Ophthalmology and Vision

Research, and experimental protocols for this study were approved by the

Institutional Animal Care and Use Committee (IACUC) of Case Western Reserve

University.

3.3.2 Sample Preparation

Eyes were removed from mice following CO2 asphyxiation and washed with RNAlater solution (Thermofisher Scientific, Waltham, MA). Lenses were carefully removed by cutting away cornea and sclera and were washed in ice cold nuclease-free water. Lens epithelia cells were isolated by removing the lens capsule with forceps and removing any visible fiber cells adherent to the capsule. Cortical lens fiber cells were isolated from the remaining fiber cell mass using forceps.

Samples of the same tissue from both eyes of each mouse were pooled together in

1 mL of RNAlater solution. Samples were kept at 4oC overnight and then frozen at

-80oC until RNA extraction.

3.3.3 RNA-Seq

RNA-Seq was performed in quadruplicate samples by the Genomics,

Epigenomics and Sequencing Core (GESC) in the University of Cincinnati as described below.

Sample tissue was homogenized in 0.4-0.8 mL Lysis/Binding Buffer from the mirVana miRNA Isolation Kit (Thermo Fisher, Grand Island, NY) using a

Precellys 24 homogenizer (Bertin Corp, Rockville, MD). Total RNA extraction was performed according to the mirVana protocol, and the RNA was eluted with 100

µL elution buffer. Quality of RNA was assessed using a 2100 Bioanalyzer (Agilent

Technologies, Santa Clara, CA).

NEBNext Poly(A) mRNA Magnetic Isolation Module (New England

BioLabs, Ipswich, MA) was used for initial poly(A) RNA purification with a total of 1 µg good quality total RNA as input. An Apollo 324 system (WaferGen,

Fremont, CA) was then used with PrepX PolyA Isolation Kit (WaferGen, Fremont,

CA) for automated poly(A) RNA isolation.

NEBNext Ultra Directional RNA Library Prep Kit (New England BioLabs,

Ipswich, MA) was used for library preparation, which uses dUTP in cDNA synthesis to maintain strand specificity. In short, the isolated poly(A) RNA was

Mg2+/heat fragmented to ~200 bp, reverse transcribed to 1st strand cDNA, followed by 2nd strand cDNA synthesis labelled with dUTP. The purified cDNA was end repaired and dA tailed, and then ligated to an adapter with a stem-loop structure.

The dUTP-labelled 2nd strand cDNA was removed by USER enzyme to maintain strand specificity. After indexing via PCR (~12 cycles) enrichment, the amplified library was cleaned up by AMPure XP beads for QC analysis.

To check the quality and yield of the purified library, 1 µL of the library was analyzed by Bioanalyzer (Agilent, Santa Clara, CA) using a DNA high sensitivity chip. To accurately quantify the library concentration for the clustering, the library was 1:104 diluted in dilution buffer (10 mM Tris-HCl, pH 8.0 with

0.05% Tween 20), and measured by qPCR using the Kapa Library Quantification

Kit (Kapabiosystem, Woburn, MA) with ABI's 9700HT real-time PCR system

(Thermo Fisher, Waltham, MA).

3.3.4 Cluster Generation and HiSeq Sequencing

To study differential gene expression, individually indexed and compatible libraries were proportionally pooled (~25 million reads per sample in general) for clustering in cBot System (Illumina, San Diego, CA). Libraries at the final concentration of 15 pM were clustered onto a single read (SR) flow cell using

Illumina TruSeq SR Cluster Kit v3, and sequenced for 50 bp using the TruSeq SBS

Kit on an Illumina HiSeq system.

3.3.5 Bioinformatic Analysis

To analyze differential gene expression, sequence reads were aligned to the mouse genome using standard Illumina sequence analysis pipeline, which was analyzed by The Laboratory for Statistical Genomics and Systems Biology at the

University of Cincinnati. The report consists of: 1) RNA-seq data, QC, and sample clustering analyzing results; 2) All gene expression level in RNA samples, normalized as Read Per Kilobase of transcript per Million mapped reads (RPKM);

3) Significantly differentially expressed genes between groups with P < 0.05 and false-discovery rate (FDR) < 0.1.

3.3.6 Validation of RNA-Seq Data by Real Time PCR

(qPCR)

An independent group of 6 month old male C57Bl/6 mice of the same genotype/treatment used in RNA-Seq were used for qPCR confirmation of RNA-

Seq results. Lens epithelia and cortical fiber cells were dissected in ice cold nuclease-free water and pooled together from both eyes of each mouse. Samples were immediately frozen in liquid nitrogen. RNA was extracted and purified using a standard Trizol protocol (Thermofisher, Waltham, MA). RNA purity and concentration was analyzed using a Nanodrop 2000c and only samples showing a

260/280 ratio of ≥1.8 and a 260/230 ratio of ≥2.0 were used (Thermofisher,

Waltham, MA). 1 µg of fiber cell RNA and at least 200 ng of epithelia RNA, was

99 treated with amplification grade DNase I (Thermofisher, Waltham, MA), for each sample, to remove genomic DNA. RNA was converted to cDNA using M-MuLV

Reverse Transcriptase and murine RNase inhibitor (New England BioLabs,

Ipswich, MA). An equal mixture of oligo(dT) and random oligo primers were used at a concentration of 20 µM for the synthesis.

Primers for qPCR were predesigned KicqStart primers ordered from Sigma-

Aldrich (St. Louis, MO), with the exception of primers for Mt1 and Aldh1a7, which were designed using Primer-BLAST software (National Center for Biotechnology

Information, Bethesda MD). Primers were verified by performing qPCR in triplicate on WT C57BL/6 mouse whole lens cDNA at three different concentrations and with a no template control in order to determine linearity of amplification and reproducibility. Specificity of primers was determined based on the presence of a single peak in melt curve analysis. All primer data are reported

(Figure 6.2).

Hprt and Rer1 were used in tandem as reference genes for relative quantification of expression (ΔΔCt method) using KicqStart SYBR Green Master

Mix with ROX (Sigma-Aldrich). 5, 15, or 30 ng of lens epithelia or fiber cell RNA was used for each reaction, based on the established linear range of amplification for each primer. Standard cycling was used with an initial 10-minute hold at 95oC followed by 40 cycles of 95oC for 15 seconds and 60oC for 1 minute.

3.3.7. Statistical Analysis

100

All values are expressed as means ± standard deviations (SD) with n = 4.

Only P-values < 0.05 were considered statistically significant. P-values for correlations were derived from the Pearson correlation coefficient (r) using the equation t = r √[(1−r2)/(n−2)] and determining the corresponding two-tailed P- value.

3.4 Results

3.4.1 Overview of Significant Gene Expression Changes Resulting from GSH Depletion in the Lens

Reads were mapped to 24,415 unique transcripts. All samples had a high

(>0.986) correlation coefficient to all other samples of the same cell type without any clear outliers (Figure 6.1). A total of 441 genes showed significantly (>2-fold,

P < 0.05, FDR < 0.1) modulated gene expression (Figure 3.1). 54 genes were significantly upregulated in LEGSKO epithelia while 38 were downregulated.

LEGSKO fiber cells had 84 genes upregulated and 60 downregulated genes. The results were more dramatic in LEGSKO lenses treated with BSO, with the epithelia having 78 upregulated genes and 37 downregulated genes and the fiber cells having

125 upregulated genes and 97 downregulated genes.

101

Figure 3.1. Overview of significant gene expression changes. Total number of significantly (min. 2-fold regulation change, P < 0.05, FDR < 0.1) up- and downregulated genes for each group compared to WT.

These gene changes were sorted into major pathways of metabolism (Figure

3.2A), transcription (normally referred to as gene expression but changed here to avoid confusion) (Figure 3.2B), cell cycle (Figure 3.2C), immunity (Figure 3.2D), developmental biology (Figure 3.2E), transport of small molecules (Figure 3.2F), extracellular matrix organization (Figure 3.2G), protein metabolism (Figure 3.2H), and cellular stress response (Figure 3.2I) by Reactome software (reactome.org).

Metabolism, transcription, and cell cycle were the most robustly altered pathways while, surprisingly, cellular stress response was the least affected pathway.

All significant gene expression changes with >2-fold change are listed in the Appendix in Table 6.1 and specific gene changes in major biological pathways are shown in Figures 6.3-6.11.

102

A B C

D E F

G H I

Figure 3.2. Number of significantly up- and down-regulated genes (relative to

WT control) in major biological pathways. (A) metabolism, (B) transcription,

(C) cell cycle, (D) immunity, (E) developmental biology, (F) transport of small molecules, (G) extracellular matrix organization, (H) protein metabolism, (I) cellular stress response.

103

3.4.2 RNA-Seq Data Shows Strong Differential Expression of Lens Epithelia and Fiber Cell Marker Genes and Excellent Agreement with qPCR Results

Based on the work of the Robinson laboratory (Hoang et al., 2014) pointing to the existence of region specific gene expression in the lens, we validated our protocol by determining the expression of established lens epithelia and fiber cell marker genes (Figure 3.3). For all marker genes, expression was much greater in the appropriate lens tissue, indicating successful isolation of lens epithelia and fiber cell compartments.

Figure 3.3. Verification of RNA-Seq Results. Differential expression of (A) lens epithelium and (B) fiber cell marker genes shown as relative fold changes based on means with standard deviations from WT RPKM-normalized RNA-Seq values.

104

RNA-Seq results were further validated by obtaining cDNA from the lens epithelia and fiber cells of independent but identical groups of mice and performing qPCR on a subset of transcripts (Figure 3.4). A total of 36 transcripts were tested, which were chosen from among the most robustly significantly up- and downregulated genes and unchanged genes with a high baseline expression in the lens.

At least 16 transcripts were analyzed for each group. In all cases, the qPCR results were highly consistent with RNA-Seq results, with Pearson correlation coefficients ranging from 0.87-0.94 and P < 5E-6. This dataset confirms the veracity and reproducibility of the transcriptome.

105

Figure 3.4. Confirmation of RNA-Seq Results by RT-qPCR. Each point represents one gene. (A) LEGSKO lens epithelia. (B) BSO-treated LEGSKO lens epithelia. (C) LEGSKO lens fiber cells. (D) BSO-treated LEGSKO lens fiber cells.

Values are log2 converted means of relative fold change. (E-H) Direct comparison of RNA-Seq and qPCR results for each gene tested. Values in bar graphs are means of relative fold changes +/- SD. n = 4.

106

3.4.3 GSH Depletion Results in a Partial Downregulation of β- and γ-Crystallins

Expression of all crystallin genes was analyzed, as crystallins are essential to the function of the lens and any expression changes may have a major effect on cataractogenesis. LEGSKO lens epithelia showed a slight trend of decrease in all crystallin transcripts and this difference was more exaggerated in BSO-treated epithelia (Figure 3.5).

Only downregulation of Cryge (Figure 3.5M) and Crygf (Figure 3.5O) reached significance (P < 0.05, FDR < 0.1) in LEGSKO epithelia while downregulation of Cryba1 (Figure 3.5C), Cryba4 (Figure 3.5E), Crybb1 (Figure

3.5F), Cryge (Figure 3.5M), and Crygf (Figure 3.5N) was significant in BSO- treated LEGSKO epithelia. However, only changes in Cryge and Crygf reached a

2-fold threshold. Lens fiber cells did not show the same general downregulation of all crystallin genes and only showed significant (P < 0.05, FDR < 0.1) decreases in

BSO-treated LEGSKO lens fiber cells, where Cryga (Figure 3.5I), Crygb (Figure

3.5J), Cryge (Figure 3.5M), Crygf (Figure 3.5N), and Crygn (Figure 3.5O) were downregulated. Of these, Cryge, Crygf, and Cryga had an >2-fold downregulation.

107

Figure 3.5. Crystallin Gene Expression by Lens GSH Content. (A) Cryaa, (B)

Cryab, (C) Cryba1, (D) Cryba2, (E) Cryba4, (F) Crybb1, (G) Crybb2, (H) Crybb3,

(I) Cryga, (J) Crygb, (K) Crygc, (L) Crygd, (M) Cryge, (N) Crygf, (O) Crygn, (P)

Crygs. Values are means +/- SD. 3.8 mM GSH = WT, 1.3 mM GSH = LEGSKO,

0.35 mM GSH = BSO-treated LEGSKO. Fold change and significance is relative to WT.

3.4.4 Trends Identified in Upstream Regulators and Molecular/Cellular Functions of Genes with Modulated Expression

108

The major upstream regulators of the lens’ transcriptomic response to GSH depletion were determined using Ingenuity Pathway Analysis (IPA) Software

(QIAgen, Hilden, Germany) (Table 3.1).

These data provide information on the pathways affected by the gene expression changes within these lenses. LEGSKO lens epithelia showed responses regulated upstream by transforming growth factor β1 (TGF-β1), tumor necrosis factor (TNF), low-density lipoprotein (LDL), AE binding protein 1 (AEBP1), and peroxisome proliferator-activated receptor gamma (PPARG). Based on IPA analysis readout, TGF-β1, TNF, AEBP1, and PPARG all play roles in proliferation, epithelial-mesenchymal transition (EMT), and epithelial cancers while LDL,

AEBP1, and PPARG are all indicated to regulate lipid homeostasis.

The top upstream regulators of BSO-treated LEGSKO lens epithelia were

AEBP1, tumor suppressor p53, cone-rod homeobox protein (CRX), rhodopsin

(RHO), and orthodenticle homeobox 2 (OTX2). p53 is a major tumor suppressor and a regulator of EMT (Termén et al., 2013). IPA indicates that CRX, RHO, and

OTX2 regulate the visual cycle and eye development and are associated with vision defects. Similar results were found in LEGSKO fiber cells, where the primary regulators were RHO, PPARG, TNF, son of sevenless homolog (Sos), and interleukin 1β (IL-1β). Sos and IL-1β were also indicated as EMT regulators.

109

) ) ) ) )

)

12 12 12 12 06

)

- - - - -

)

1.92E 1.92E 3.25E 3.25E 1.93E

3.05E

(

- - - - -

4.10E

6.97E

4.85E

(

2.44E

(

LEGSKO Fiber Cells Fiber LEGSKO

(

Repair

03 03 04 10 03 03

- - - - -

TNF

RARA

EP300

Protein Synthesis Protein

8.50E 4.16E 3.68E 6.20E 8.61E

SMARCA4

( ( ( ( (

Translational Modification Translational

Treated Treated

Post

BSO

Cellular Function and Maintenance and Function Cellular

Cellular Assembly and Organization and Assembly Cellular

DNA Replication, Recombination, and and Recombination, Replication, DNA

) ) ) ) )

)

06 06 06 06 05

)

- - - - -

)

5.92E 5.92E 5.92E 5.92E 1.84E

- - - - -

2.60E

(

Production

1.48E

1.17E

4.13E

(

9.10E

value range) value

(

03 03 03 03 03 03

(

- - - - -

value)

Sos

TNF

RHO

PPARG

Lipid Metabolism Lipid

Drug Metabolism Drug

Energy Energy

LEGSKO Fiber Cells Fiber LEGSKO

7.17E 9.15E 9.35E 8.05E 4.85E

( ( ( ( (

Small Molecule Biochemistry Molecule Small

Vitamin and Mineral Metabolism Mineral and Vitamin

) ) ) ) )

)

10 08 06 06 06

)

Upstream Regulators (p Regulators Upstream

)

- - - - -

)

1.18E 1.76E 2.61E 2.61E 3.07E

- - - - -

1.67E

(

Molecular and Cellular Functions (p Functions Cellular and Molecular

1.60E

1.23E

(

1.80E

(

03 03 03 03 03 03

- - - - -

CRX

RHO

OTX2

Cell Morphology Cell

AEBP1

8.47E 5.28E 8.47E 6.21E 7.66E

Cellular Compromise Cellular

( ( ( ( (

Treated LEGSKO Epithelia LEGSKO Treated

Translational Modification Translational

Cellular Development Cellular

BSO

Post

Cellular Growth and Proliferation and Growth Cellular

) ) ) ) )

)

07 07 07 06 06

- - - - -

)

3.72E 6.79E 6.79E 2.12E 2.12E

- - - - -

1.19E

7.36E

7.13E

(

1.51E

1.01E

(

03 03 03 03 03 03

β1(

- - - - -

LDL

TNF

LEGSKO Epithelia LEGSKO

AEBP1

PPARG

Lipid Metabolism Lipid

TGF

3.80E 3.80E 3.80E 3.80E 3.80E

( ( ( ( (

Cellular Development Cellular

Cell Death and Survival Survival and Death Cell

Small Molecule Biochemistry Molecule Small

Table Table 3.1. Top Upstream Regulators and Molecular and Cellular Functions of Gene Expression Changes in GSH Lenses. Deficient

Cellular Growth and Proliferation and Growth Cellular

110

The major regulators of the response in BSO-treated LEGSKO fibers were

TNF, IL-1β, retinoic acid receptor alpha (RARA), SWI/SNF Related, Matrix

Associated, Actin Dependent Regulator of Chromatin, Subfamily A, Member 4

(SMARCA4), and E1A binding protein p30 (EP300). IPA reports all of these molecules as regulators of cell growth/survival and epithelial cancers.

IPA software was also used to determine the most common molecular and cellular functions of the differentially regulated genes for each group (Table 3.1).

The most common functions found in the LEGSKO lens epithelia genetic response matched well with the upstream regulators that were noted and had a particular emphasis on cellular integrity. These include death and survival, cellular development, cellular growth and proliferation, lipid metabolism, and small molecule biochemistry.

BSO-treated LEGSKO lens epithelia showed similar functions, with the major ones being cell morphology, cellular compromise, cellular development, cellular growth and proliferation, and post-translational modification (relating to oxidation and tetramerization of proteins).

The response in LEGSKO lens fiber cells showed functions that were more metabolically focused, with categories of energy production, lipid metabolism, small molecule biochemistry, vitamin and mineral metabolism, and drug metabolism showing significant changes.

BSO-treated LEGSKO lens fiber cells showed different major functions of cellular assembly and organization, DNA replication, recombination, and repair, post-translational modification, protein synthesis, and cellular function and

111 maintenance. The difference between the treated and untreated LEGSKO fibers could imply a more robust stress response in the BSO treated animal.

3.4.5 Expression of Several Detoxification Genes Increases as a Function of Decreasing Lenticular GSH Content

Because a loss of GSH is expected to result in increased oxidative stress and a decreased clearance of toxic species, such as H2O2 and lipid peroxidation-derived aldehydes, genes related to detoxification were expected to show regulation changes and were investigated (Figure 3.6).

Both epithelia and fiber cells showed changes in aldehyde dehydrogenase genes, with Aldh1l1 (Figure 3.6E) and Aldh3a1 (Figure 3.6F) having an >2-fold upregulation in BSO-treated LEGSKO epithelia (P < 0.0005) and Aldh1a7 (Figure

3.6D) and Aldh1a1 (Figure 3.6B) showing a similar upregulation in both LEGSKO and BSO-treated LEGSKO fiber cells (P < 0.0005). Both fiber cell groups also showed ~3-fold downregulation of Aldh1a3 (Figure 3.6C)(P < 5E-8), the only significantly downregulated gene in this category for either tissue.

112

Figure 3.6. Expression Changes in Detoxification Genes by Lens GSH Content.

(A) Akr1b10, (B) Aldh1a1, (C) Aldh1a3, (D) Aldh1a7, (E) Aldh1l1, (F) Aldh3a1,

(G) As3mt, (H) Ces1g, (I) Gstk1, (J) Mt1, (K) Mt2, (L) Slc14a1. Values are means

+/- SD. 3.8 mM GSH = WT, 1.3 mM GSH = LEGSKO, 0.35 mM GSH = BSO- treated LEGSKO. Fold change and significance is relative to WT.

113

This category of genes also showed upregulation, in both epithelia and fiber cell groups, of the essential detoxification gene carboxylesterase (Ces1g)(Figure

3.6H)(P < 0.05), the urea transporter UT-B (Slc14a1)(Figure 3.6L)(P < 5E-12), and metal chelator/antioxidant metallothionein 1 (Mt1)(Figure 3.6J)(P < 5E-15). Mt1 had a particularly robust upregulation, with a~4-7-fold increase in epithelia and

>15-fold upregulation in fiber cells across both GSH-deficient groups. Epithelia showed a similar upregulation of metallothionein 2 (Mt2)(Figure 3.6K)(P < 5E-8), which is nearly identical in sequence to Mt1, but this change was not present in fiber cells. Additionally, untreated LEGSKO lens epithelia showed >2-fold upregulation of aldo-keto reductase family member B10 (Akr1b10)(Figure 3.6A)(P

< 0.0005) and both fiber cell groups showed an >2-fold upregulation of arsenite methyltransferase (As3mt)(Figure 3.6G)(P < 5E-5) and glutathione S-transferase kappa 1 (Gstk1)(Figure 3.6I)(P < 0.005) that was not present in the epithelial groups.

3.4.6 Several Transport Systems Are Modulated as a Function of Decreasing Lenticular GSH Content

The top 10 most robust changes in the category of small molecule transport are shown in Figure 3.7.

114

Figure 3.7. Top 10 Expression Changes in Small Molecule Transport Genes by

Lens GSH Content. (A) Abcg1, (B) Csn3, (C) Gjb3, (D) Gabrr1, (E) Gabrr2, (F)

Kcnd1, (G) Slc14a1, (H) Slc27a6, (I) Snca, (J) Trf. Values are means +/- SD. 3.8 mM GSH = WT, 1.3 mM GSH = LEGSKO, 0.35 mM GSH = BSO-treated

LEGSKO. Fold change and significance is relative to WT.

These changes include 5.2-fold upregulation of the connexin gene, Gjb3

(Figure 3.7C), which encodes Cx31, in BSO-treated LEGSKO epithelia (P <

0.005). Cx31 has not been previously characterized to be expressed within the lens.

Potassium channel gene Kcnd1 (Figure 3.7F), which could potentially play a role in lens microcirculation, showed an >2-fold downregulation in both epithelia and fiber cells (P < 0.005). The expression of this gene showed a strong positive correlation with lens GSH content in both tissue compartments.

115

Several changes were seen in iron transport genes, including a 6.0-8.5-fold upregulation of α-synuclein (Snca)(Figure 3.7I)(P < 5E-9), which has been recently characterized as a ferrireductase (Davies et al., 2011), and a 3-4-fold upregulation of transferrin (Trf)(Figure 3.7J)(P < 5E-13) expression in fiber cells. These changes track well with lens GSH content and imply enhanced uptake of iron by GSH- deficient lens fiber cells.

Expression changes were noted in multiple gamma-aminobutyric acid receptors, with Gabrr1 being ~6-fold upregulated (Figure 3.7D)(P < 0.005) and

Gabrr2 being 2-5-fold downregulated (Figure 3.7E)(P < 0.005) in LEGSKO and

BSO-treated LEGSKO fiber cells, respectively.

Kappa casein (Csn3), a transmembrane regulator without a well-defined role in the lens, was 14.4-fold upregulated in BSO-treated LEGSKO epithelia

(Figure 3.7B)(P < 0.0005), but not in the untreated mouse, and was not expressed at all in fiber cells.

Additionally, expression changes in lipid transport genes were found. Fatty acid transporter Slc27a6 was 5.5-fold upregulated in BSO-treated LEGSKO epithelia (Figure 3.7H)(P < 0.0005) and there was a 9-13-fold upregulation of the cholesterol and phospholipid ATP-binding cassette transporter Abcg1 in fiber cells

(Figure 3.7A)(P < 1E-8). These regulation changes imply enhanced uptake of lipids and potential membrane composition changes.

3.4.7 An Array of Lipid Metabolism Genes Show Modulated Expression as a Function of Decreasing Lenticular GSH Content

116

Many other changes occurred in lipid metabolism gene expression and the top 10 most robust of these changes are shown in Figure 3.8.

Figure 3.8. Top 10 Expression Changes in Lipid Metabolism Genes. (A)

Abca13, (B) Abcg1, (C) Alox15, (D) Cpne6, (E) Cyp2j9, (F) Dmgdh, (G) Fhl2, (H)

G0s2, (I) Gal, (J) Npas2. Values are means +/- SD. 3.8 mM GSH = WT, 1.3 mM

GSH = LEGSKO, 0.35 mM GSH = BSO-treated LEGSKO. Fold change and significance is relative to WT.

An additional ATP-binding cassette lipid transporter, Abca13, was 4.5-fold upregulated in BSO-treated LEGSKO fiber cells (Figure 3.8A)(P < 0.05).

LEGSKO lens epithelia showed an 8.7-fold downregulation of the membrane trafficking gene copine-6 (Cpne6)(Figure 3.8D)(P < 0.005). This transcript was not present in lens fiber cells and does not have a well-defined role in the lens.

117

Several changes were noted in genes related to the breakdown of lipids. All samples showed a 2.5-5.4-fold upregulation of the lipase inhibitor gene G0/G1 switch 2 (G0s2)(Figure 3.8H)(P < 5E-5). Cytochrome P450, family 2, subfamily j, polypeptide 9 (Cyp2j9), which is an oxidoreductase involved in cholesterol and steroid metabolism, was 8.0-10.7-fold upregulated in fiber cells (Figure

3.8E)(P<5E-7). Dimethylglycine dehydrogenase (Dmgdh), which catabolizes the phospholipid head group component choline, was 18.1-fold upregulated in BSO- treated LEGSKO fiber cells (Figure 3.8F)(P < 0.005).

Four and a half LIM domains 2 (Fhl2) was 5.7-fold downregulated in BSO- treated LEGSKO fiber cells (Figure 3.8G)(P < 5E-5). Fhl2 is thought to play a role in the development of extracellular matrix and its downregulation is associated with cellular transformation (Cao et al., 2015).

Gal encodes galanin, a neuroendocrine peptide that plays a role in consummatory behavior, and was 6.3-fold upregulated in BSO-treated LEGSKO epithelia (Figure 3.8I)(P < 0.0005). Although upregulation of galanin is associated with increased consumption of dietary fats (Barson et al., 2010), and thus lipid balance, it is unclear why a neuroendocrine peptide would be expressed in the lens and what effect it has. Similarly, the circadian metabolism-regulating gene Npas2, which encodes the transcription factor neuronal PAS domain-containing protein 2, was 5.8-fold downregulated in BSO-treated LEGSKO epithelia (Figure 3.8J)(P <

0.0005), but the effect on the lens is unclear.

3.4.8 Lens GSH Depletion Induces Gene Expression Changes Related to Epithelial-Mesenchymal Transition (EMT) Pathways.

118

Based on the identification of TGF-β and TNF as primary upstream regulators of the gene expression changes found in GSH-deficient epithelia (Table

3.1), activation of EMT signaling in lens epithelia was investigated. A number of changes were found in genes associated with canonical EMT pathways (Figure 3.9).

Constant between the LEGSKO and BSO-treated LEGSKO epithelia was robust upregulation of Wnt10a, with an >20-fold increase in expression in both sample groups compared to WT (Figure 3.9J)(P < 5E-5).

Additionally, LEGSKO lens epithelium showed a slightly over 2-fold upregulation of other EMT signaling initiators, including epidermal growth factor

(Egf)(Figure 3.9A)(P < 5E-5), fibroblast growth factor receptor 4 (Fgfr4)(Figure

3.9B)(P < 0.005), and IL-1B (Figure 3.9D)(P < 0.0005), as well as downstream regulator ras homolog family member h (Rhoh)(Figure 3.9F)(P < 1E-9) and a robust

5.9-fold upregulation of downstream regulator spleen associated tyrosine kinase

(Syk)(Figure 3.9H)(P < 0.005). LEGSKO lenses also showed a 3.8-fold downregulation of nuclear receptor subfamily 4 group member 1 (Nr4a1)(Figure

3.9E)(P < 5E-5), an inhibitor of TGF-β signaling.

119

Figure 3.9. Top Expression Changes in Canonical EMT Pathway Genes in

Lens Epithelia by GSH Content. (A) Egf, (B) Fgfr4, (C) Ier3, (D) Il-1B, (E)

Nr4a1, (F) Rhoh, (G) Spp1, (H) Syk, (I) Tgfbi, (J) Wnt10a. Values are means +/-

SD. 3.8 mM GSH = WT, 1.3 mM GSH = LEGSKO, 0.35 mM GSH = BSO-treated

LEGSKO. Fold change and significance is relative to WT.

BSO-treated LEGSKO epithelium did not recapitulate all of these findings; instead showing a 3.8-fold upregulation of transforming growth factor, beta-

120 induced (Tgfbi)(Figure 3.9I)(P < 5E-5) and a 2.3-fold upregulation of immediate early response 3 (Ier3)(Figure 3.9C)(P < 0.005), both of which are more tangentially related to EMT pathways. Secreted phosphoprotein 1 (Spp1), which inhibits AKT signaling, was 9.3-fold upregulated (Figure 3.9G)(P < 1E-5). This could mean that most of these proteins are only minor contributors to EMT in the lens, while WNT10A is the major driver, or that BSO-treated and untreated

LEGSKO epithelia are in different stages of the EMT process, with the untreated cells being in the initiating stages and the treated cells being at a more terminal/transformed stage.

Expression changes were overlaid on a diagram of canonical EMT pathways to allow for better visualization of the effect on these pathways (Figure

3.10). LEGSKO epithelia show expression changes effecting the initiating portions of many EMT pathways, as well as downstream changes focused on the PI3K complex. In contrast, BSO-treated LEGSKO epithelia show less clear results outside of WNT/Frizzled signaling.

121

Figure 3.10. Diagram of expression changes in canonical pathways of EMT in lens epithelia.

122

EMT signaling is expected to result in a loss of marker genes, so expression of genes related to vision and the eye were investigated in lens epithelia (Figure

3.11).

Figure 3.11. Top 10 lens epithelia expression changes in non-crystallin vision genes. (A) Cngb1, (B) Fscn2, (C) Gnat1, (D) Opn1sw, (E) Pde6g, (F) Pde6h, (G)

Prph2, (H) Rpb3, (I) Rho, (J) Rp1. Values are means +/- SD. 3.8 mM GSH = WT,

1.3 mM GSH = LEGSKO, 0.35 mM GSH = BSO-treated LEGSKO. Fold change and significance is relative to WT.

123

While expression of vision genes showed a trend of downregulation in

LEGSKO epithelia, these changes did not reach significance. However, these changes were more robust in BSO-treated LEGSKO epithelia with a broad downregulation of genes involved in visual transduction ranging from 3-20-fold changes (P < 0.005). All of these genes are strongly expressed in the eye and play a role in vision but do not have a known function within the lens

A number of genes associated with cell cycle progression showed significant regulation changes (Figure 3.12). Six genes encoding histone subunits showed 2-3-fold upregulation in epithelia samples when compared to WT (Figure

3.12B-G)(P < 0.005), indicating active cellular proliferation. The DNA repair gene

Dmc1 showed a 5.2-fold increase in BSO-treated LEGSKO epithelia (Figure

3.12A)(P < 0.005), while tubulin genes showed an >2-fold downregulation (Figure

3.12H and I)(P < 0.005).

124

Figure 3.12. Lens epithelia expression changes in cell cycle genes. (A) Dmc1,

(B) Hist1h4a, (C) Hist1h4b, (D) Hist1h4c, (E) Hist1h4d, (F) Hist1h4k, (G)

Hist1h4n, (H) Tuba8, (I) Tubb2b. Values are means +/- SD. 3.8 mM GSH = WT,

1.3 mM GSH = LEGSKO, 0.35 mM GSH = BSO-treated LEGSKO. Fold change and significance is relative to WT.

125

Extracellular matrix (ECM) organization gene changes were investigated since EMT is associated with changes in ECM (Figure 3.13).

Figure 3.13. Lens epithelia expression changes in extracellular matrix organization genes. (A) A2M, (B) Col1a1, (C) Col1a2, (D) Dcn, (E) Lamb3, (F)

Spock2, (G) Spp1, (H) Tnc. Values are means +/- SD. 3.8 mM GSH = WT, 1.3 mM

GSH = LEGSKO, 0.35 mM GSH = BSO-treated LEGSKO. Fold change and significance is relative to WT.

126

Type I collagen expression increases with EMT and Col1a1 and Col1a2 are commonly used markers of EMT signaling (Medici and Nawshad, 2010)

Interestingly, Col1a1 was 2.4-fold downregulated in BSO-treated LEGSKO epithelia (Figure 3.13B)(P < 0.005) while Col1a2 was 4.8- and 6-fold upregulated in untreated and BSO-treated LEGSKO epithelia, respectively (Figure 3.13C)(P <

5E-5). Laminin subunit beta 3 (Lamb3) a component of lens capsules (Danysh and

Duncan, 2009), was >2-fold upregulated in epithelia samples (Figure 3.13E)(P <

0.0001). Tnc encodes tenascin C, an ECM protein associated with wound healing

(Jensen et al., 2014) and was 9.6-fold upregulated in BSO-treated LEGSKO epithelia (Figure 3.13H)(P <5 E-5). Dcn encodes decorin, an extracellular protease that interacts with TGF-β (Jarvinen and Prince, 2015), and was 15.2-fold upregulated in untreated LEGSKO epithelia (Figure 3.13D)(P < 0.0005). Spock2 encodes an ECM protein with similarity to SPARC/osteonectin, an essential protein for lens clarity (Norose et al., 1998), and was 2.4-fold downregulated in BSO- treated LEGSKO epithelia (Fig. 3.13F)(P < 0.0005).

Finally, to determine whether the transcriptional profile of GSH-deficient lens epithelium is similar to the transcriptional profile of lens epithelium undergoing EMT, transcriptomic data from this study were compared to transcriptomic data from a study by Medvedovic et al. (2006) in which gene expression was measured in C57BL/6 mouse lens epithelia undergoing EMT after removal of lens fiber cells.

127

Figure 3.14. Correlation of LEGSKO and BSO-treated LEGSKO lens epithelia with EMT mouse lens transcriptional profiles (taken with permission from Medvedovic et al., 2006). (A) LEGSKO lens epithelia. (B) BSO-treated

LEGSKO lens epithelia. Values are log2 converted means of relative fold-changes.

Each point represents one gene.

Comparing genes that show significant (P < 0.05, FDR < 0.1) regulation changes in both studies, regardless of fold change, it was determined that LEGSKO lens epithelia trends towards a positive correlation with EMT lenses but this correlation does not reach significance (r = 0.353, P = 0.0906)(Figure 3.14A), whereas BSO-treated LEGSKO lens epithelia show a robust positive correlation with epithelia from EMT lenses (r = 0.633, P < 0.001)(Figure 3.14B). The specific overlapping gene changes that were compared are shown in Table 3.2, with changes occurring in the same direction bolded. These data indicate that, as lens GSH becomes increasingly deficient, lens epithelia increasingly develop EMT-like transcriptional profiles.

128

Table 3.2. Comparison of Overlapping Gene Expression Changes in GSH-

Deficient and EMT Lens (Medvedovic et al., 2006) Transcriptomes.

EMT Lens LEGSKO BSO-Treated Epithelia Gene Description Lens LEGSKO Lens (Medvedovic Epithelia Epithelia et al., 2006) 1110032A03Rik RIKEN cDNA 1110032A03 gene 1.34 -1.14 Abcb1a ATP-binding cassette, sub- 1.65 1.90 1.70 family B (MDR/TAP), member 1A Actg1 actin, gamma, cytoplasmic 1 -1.18 1.33 Agrn agrin -1.24 1.66 Alas1 aminolevulinic acid synthase 1 -1.25 1.73 Aldh3a1 aldehyde dehydrogenase 1.63 2.15 -3.05 family 3, subfamily A1 Apoe apolipoprotein E 2.03 2.13 1.63 Ccl2 chemokine (C-C motif) ligand 2 -1.59 2.25 Cldn10 claudin 10 -1.40 -2.00 Crybb1 crystallin, beta B1 -1.56 -4.78 Ctgf connective tissue growth 1.30 2.64 factor Dbnl drebrin-like -1.25 -1.37 Fabp5 fatty acid binding protein 5, -1.69 -3.48 epidermal Fscn2 fascin homolog 2, actin- -3.60 -1.75 bundling protein, retinal Gosr2 golgi SNAP receptor complex 1.26 -1.42 member 2 Gpnmb glycoprotein 5.85 1.80 (transmembrane) nmb Hebp1 heme binding protein 1 1.44 1.59 Hexb hexosaminidase B -1.21 1.93 Hist1h2bc histone 1, H2bc 1.40 1.56 Homer2 homer homolog 2 (Drosophila) 1.76 -1.78 Krt15 keratin 15 2.58 1.83 Lctl lactase-like -1.51 -1.38 -4.41 Ldlr low density lipoprotein -1.11 -1.31 -1.26 receptor Lim2 lens intrinsic membrane -1.53 -4.89 protein 2 Nfatc2 nuclear factor of activated T- -1.85 -3.39 cells, cytoplasmic 2 Oplah 5-oxoprolinase (ATP- 1.39 1.48 hydrolysing) Osmr oncostatin M receptor 1.42 3.14

Pir pirin 1.79 1.76 Rasgef1b GPI-GAMMA 4 -2.23 -1.38 Rbp3 retinol binding protein 3, -4.81 -1.57 interstitial Rho rhodopsin -3.75 -1.93 Sat1 spermidine/spermine N1- 1.25 1.34 1.68 acetyl transferase 1 Serpina3n serine (or cysteine) 9.62 3.96 129

Table 3.2. Comparison of Overlapping Gene Expression Changes in GSH-

Deficient and EMT Lens (Medvedovic et al., 2006) Transcriptomes

(Continued).

EMT Lens LEGSKO BSO-Treated Epithelia Gene Description Lens LEGSKO Lens (Medvedovic et Epithelia Epithelia al., 2006) Pir pirin 1.79 1.76 Rasgef1b GPI-GAMMA 4 -2.23 -1.38 Rbp3 retinol binding protein 3, -4.81 -1.57 interstitial Rho rhodopsin -3.75 -1.93 Sat1 spermidine/spermine N1- 1.25 1.34 1.68 acetyl transferase 1 Serpina3n serine (or cysteine) 9.62 3.96 proteinase inhibitor, clade A, member 3N Slc9a3r1 solute carrier family 9 1.69 1.37 (sodium/hydrogen exchanger), isoform 3 regulator 1 Slco2a1 solute carrier organic anion 1.93 1.94 1.79 transporter family, member 2a1 Stab1 stabilin 1 -1.32 2.68 Stat5b signal transducer and activator 1.21 -10.98 of transcription 5B Sulf1 sulfatase 1 1.33 2.13 Tcea3 transcription elongation -2.44 -1.16 factor A (SII), 3 Tln2 talin 2 1.28 1.40 Tnc tenascin C 9.59 5.24 Tsc22d1 TSC22 domain family, 1.35 1.52 member 1

130

3.5 Discussion

3.5.1 Interpretation of Data

While the impact of GSH depletion in the aging and cataractous human and mouse lens is widely understood in regards to oxidative protein changes (Fan et al.,

2015), its impact on gene expression is poorly understood. To date, no other transcriptomic studies on lens GSH depletion have been reported. The emerging question is how the observed transcriptomic changes should be interpreted, given that the LEGSKO mouse lens is a model of chronic adaption to GSH depletion from birth on, while BSO treatment induces a more acute GSH depletion by inhibiting systemic GSH formation, and thus its uptake into the lens (Whitson et al., 2016).

Moreover, and importantly, the lens epithelial layer is the major active site of gene expression in the lens, while nucleated cortical fiber cells have more limited transcription and translational capabilities (Hoang et al., 2014). Thus, in order to gain a clearer picture of the expression changes occurring, lenses were separated into distinct epithelial and cortical fiber regions before RNA extraction and transcriptomes of both regions were analyzed. All of these aspects are important as context and should be considered when interpreting transcriptomic data from the

GSH-deficient lens.

131

3.5.2 The GSH-Deficient Lens Transcriptome Shows an Unusual Oxidative Stress Response and Indicates Protective Detoxification Genes

While it was expected that depletion of GSH would result in enhanced expression of many genes regulated by the transcription factor Nrf2, the master regulator of the antioxidant response and GSH synthesis (Ma, 2013), this response was much more muted than expected. Nrf2, surprisingly, was not a significant regulator of the observed transcriptional changes (Table 3.1), and the only Nrf2 regulated genes (Ma, 2013) that showed a change in expression were Aldh3a1, Mt1, and thioredoxin-interacting protein (Txnip)(Table 6.1). Interestingly, Txnip, which acts as a pro-oxidant by inhibiting thioredoxin activity, was slightly upregulated, even though binding of Nrf2 to the Txnip promoter suppresses its expression (He and Ma, 2012). The one GSH-related enzyme that was upregulated, the glutathione-

S-transferase Gstk1, has not been shown to be regulated by Nrf2 (Wu et al., 2012) and does not have a well-characterized function or substrate. Additionally, there was a general lack of regulation changes in traditional antioxidant genes, such as peroxidases and superoxide dismutase.

These results may indicate that lenses are unable to properly respond to

GSH-deficiency and produce a typical oxidative stress response. This apparent lack of robust Nrf2 activation may help to explain why lenses develop GSH-deficiency in the first place, since a lack of GSH production should result in Nrf2 activation and, thus, expression of new GSH synthetic machinery to replace the non- functional enzymes in metabolically active regions of the lens. Without the ability

132 to mount this response, aged lenses become deficient in GSH while other aged tissues remain abundant in the antioxidant.

Rather than a traditional oxidative stress response, the primary defensive adaptation that the lens appears to have undergone is the upregulation of a multitude of detoxifying genes that may be able to carry out many of the same functions as

GSH. The most robust response appears to be upregulation of the metallothioneins, which are small cysteine-rich cytosolic proteins. Traditionally, metallothioneins are thought of as heavy metal chelators but, more recently, they have been shown to be effective free radical scavengers (Viarengo et al., 2000) due to their abundance of free sulfhydryl residues. Both of these functions overlap with those of GSH (Giblin,

2000) and indicate that metallothioneins may be a highly effective substitute for

GSH in regards to its non-enzymatic functions. In fact, studies have shown that a reduction in GSH results in upregulation of metallothionein (Ding et al., 2002) and vice versa (Hidalgo et al., 1990), indicating a linkage between the two cytosolic pools. Additionally, regulation changes in metallothionein have been noted in human cataractous lenses (Hawse et al., 2004;Ruotolo et al., 2003) and other mouse cataract models (Mansergh et al., 2004), giving it a clear role in protecting the lens.

One of the major roles of GSH in the lens is the detoxification of toxic aldehydes, such as 4-HNE, that result from lipid peroxidation (Micelli-Ferrari et al., 1996), a major consequence of UV irradiation (van Kuijk, 1991). GSH acts as a cofactor for glutathione-S-transferases (GST) to detoxify these groups by forming glutathione adducts (Singhal et al., 2015) and thus prevent damage to lenticular proteins. Conversely, aldehyde dehydrogenase detoxify aldehydes by converting

133 them to less toxic carboxylic acids and use NAD+ as a cofactor (Yoshida et al.,

1998). Based on these known pathways, it was expected that a loss of GSH in the lens would lead to an upregulation of aldehyde dehydrogenase genes.

GSH-deficient lenses did show upregulation of four aldehyde dehydrogenases, Aldh1a1, Aldh1a7, Aldh1l1, and Aldh3a1, as well as the aldo-keto reductase Akr1b10, which can also detoxify aldehydes. Interestingly, Aldh1a3 was slightly downregulated. Aldh1a1 and Aldh3a1 have previously been shown to be expressed in lenses at the protein level (Choudhary et al., 2005), with ALDH1A1 being one of the major non-crystallin proteins of the lens (Chen et al., 2012), and knockout of these groups leads to cataract formation (Choudhary et al.,

2005;Lassen et al., 2007). ALDH1A7 is nearly identical in sequence to ALDH1A1 and has expectantly similar function and substrate affinity (Black et al., 2009).

Aldh1l1 encodes 10-formyltetrahydrofolate dehydrogenase, so this gene does not play the same role in the detoxification of aldehydes as its other family members.

Aldh1a3 is involved in both eye development and retinal metabolism (Molotkov et al., 2006), so its downregulation is consistent with the trend of other vision genes found in the lens GSH-deficient transcriptome. Aldose reductases have been found to be important in the formation of diabetic cataract (Snow et al., 2015), but studies indicate that AKR1B10 does not contribute to this pathology (Huang et al., 2010).

Thus, the upregulation of Akr1b10 found here most likely relates to its ability to detoxify lipid peroxidation products and is a protective adaptation.

Carboxylesterase hydrolyzes toxic carboxylic esters, many of which can also be detoxified by GSTs, and has a known expression in the lens (Sanghani et

134 al., 2009), so its upregulation here indicates a putative protective role in the lens.

Interestingly, it has been found that carboxylesterase activity in the lens decreases steadily with age due to glycation of the enzyme (Yan and Harding, 2005), indicating a possible role in human cataract.

Little inquiry into urea transport in the lens has been made, but production of urea and other nitrogenous waste products has been demonstrated in the lens

(Jernigan, 1983). The robust upregulation of urea transporter UT-B, encoded by

Slc14a1, found in the GSH-deficient lens transcriptome could mean that detoxification of nitrogenous waste products is an important function of GSH in the lens and that, in the absence of GSH, export of these groups is essential. It has been demonstrated in vivo that lens crystallins are susceptible to carbamylation when subjected to cyanate, a urea derivative (Yan et al., 2010).

AS3MT detoxifies the oxidative group arsenite. Arsenite has been shown to accumulate significantly in the lens (Kleiman et al., 2016) and GSH has been reported as a substrate for AS3MT (Dheeman et al., 2014), so its upregulation appears to be protective and it could play a large, but as of yet uncharacterized, role in lenticular homeostasis.

3.5.3 The GSH-Deficient Lens Shows Numerous Changes in Transport Systems

Due to its avascular nature and the lack of metabolic machinery throughout much of the lens, transport processes are essential to the function of the lens

(Donaldson et al., 2001). The lens relies on both gap junctional coupling and a microcirculation induced by Na+/K+ currents in order to ensure that all lens cells

135 receive nutrients. Major changes were seen in several different groups of transporters and transport-related genes in the GSH-deficient lens transcriptome.

Lenses are believed to exclusively express Cx43, Cx46, and Cx50 at the protein level (Mathias et al., 2010), so it is unclear whether the upregulation of Gjb3

(Cx31), and the less robust downregulation of Gjb6 (Cx30) (Table 6.1), have any actual impact on the lens. Inappropriate expression of connexins could lead to dysfunction, and thus cataract, in the lens. However, it should be noted that even at their highest expression level, the RPKM values of these transcripts are at least 500- fold less than the characterized lens connexin genes.

In addition to upregulation of Kcnd1, other ion channel gene changes with a potential role in lenticular microcirculation, such as Kcnd3, Kcnj13, Kcnip1, and

Scnn1b (Table 6.1), were noted but the end result is unclear since there was a mix of up- and down-regulation in these genes and the net effect on ion flow cannot be determined.

One transport system with a change in a clear direction is iron transport, which showed robust upregulation of the iron uptake proteins Trf and Snca, as well as a slight upregulation of ferrireductase Steap3 and downregulation of the iron exporter Slc40a1 (Table 6.1). The reason for this change is unclear but it indicates that GSH-deficient lenses have enhanced uptake and retention of iron. Because free iron can induce the Fenton reaction (Winterbourn, 1995), increased lenticular iron could be a major danger to lens protein. One possible explanation for this seemingly paradoxical adaptation is that GSH-deficiency, including deficiency induced by

BSO treatment, has been shown to induce ferroptosis, a form of programmed cell

136 death independent from apoptosis (Yu et al., 2016). This pathway relies on iron accumulation and oxidative stress, rather than caspase activation, to kill cells.

Given the total lack of changes seen in apoptotic genes, despite the cells being under oxidative stress conditions, these data could mean that ferroptosis is potentially the primary mode of programmed cell death in the lens. If this pathway is truly being activated in the lens, it could have a major impact on lens health and clarity. Iron accumulation has been noted in cataractous human lenses (Dawczynski et al.,

2002), indicating a relationship between these results and human disease.

There were also several changes in γ-aminobutyric acid (GABA) receptors.

Expression of GABA receptors has previously been characterized in the lens and it is believed that these proteins play a role in fiber differentiation since they can be found in abundance at the tips of elongating fiber cells but not in mature differentiated fiber cells (Frederikse and Kasinathan, 2015). The large regulation changes in GABA receptors found in the GSH-deficient lens transcriptome could disrupt proper fiber cell differentiation and, thus lens clarity, and may relate to a potential dysfunction in differentiation discussed elsewhere in this paper.

3.5.4 GSH-Deficient Lenses Have Altered Lipid Metabolism

Lipid metabolism is essential to lens clarity for several reasons. Lens cell membranes have among the highest concentration of cholesterol of any tissue in the body (Subczynski et al., 2012). A major reason for this adaptation is that cholesterol-rich membranes are less permeable to O2, which is not needed in the mitochondrial-free regions of the lens and can potentially lead to protein oxidation.

This membrane composition produces an O2-depleted environment, with minimal

137

levels of O2 present in the lens nucleus (Subczynski et al., 2012). Additionally, it has recently been found that sterols have a direct role in preventing aggregation of lens proteins and that 25-hydroxycholesterol (Makley et al., 2015) and lanosterol

(Zhao et al., 2015) treatment can even reverse lens protein aggregation in vivo.

Further evidence for the importance of cholesterol metabolism in the lens is demonstrated by the fact that the use of cholesterol-lowering statin drugs significantly increases the risk of cataract development (Wise et al., 2014).

Abca1 and Apoe were both upregulated in fiber cells (Table 6.1) and are known to work together to export cholesterol from cells (Getz and Reardon, 2009).

The related transporters Abca13 and Abcg1 were also upregulated in fiber cells and may be involved in this process. At the same time, other changes indicate changes to the uptake of lipids, as evidenced by the upregulation of lipoprotein lipase and downregulation of lipase inhibitors Plin4 and Plin5 (Table 6.1), indicating enhanced uptake of lipids. However, lipase inhibitor G0s2 (Fig. 3.8) and Pcsk9

(Table 6.1), which induces degradation of LDL receptor (Lagace, 2014), were upregulated and the fatty acid importer Cd36 was downregulated (Appendix), indicating depressed uptake of lipids. Thus, the end result of these regulation changes is somewhat unclear but any change to lens membrane composition or sterol synthesis could have a major effect on lens clarity. The lipid content of lenses will need to be investigated further to determine how these regulation changes affect cataractogenesis and how they relate to GSH-deficiency.

3.5.5 GSH-Deficient Lenses Show Activation of EMT signaling and a Loss of Differentiation

138

Posterior capsular opacification (PCO), also known as secondary cataract, is a major consequence of cataract surgery (Wormstone et al., 2009) PCO occurs when residual lens epithelia left on the lens capsule after surgery proliferate, migrating to the posterior lens capsule, and undergo epithelial-mesenchymal transition (EMT), forming fibrotic lesions that prevent focusing of light onto the retina (Wormstone et al., 2009) Thus, understanding factors that govern lens EMT is essential to reducing the incidence of PCO.

We were surprised to find a robust EMT-like response in the epithelia of

GSH-deficient lenses. This response is evidenced by the top upstream regulators of the transcriptomic changes being EMT-regulators including TGF-β1, TNF, and p53

(Table 3.1) and by the upregulation of genes associated with canonical EMT pathways, such as Egf, Fgfr4, Wnt10a, and Il-1B (Figure 3.9). Accompanying the upregulation of these EMT activators were changes indicative of EMT actually occurring in lens epithelia, including an upregulation of histones that demonstrates cellular proliferation (Figure 3.12) and changes to ECM organization (Figure 3.13).

BSO-treated lenses also showed a broad downregulation of genes associated with vision, which could be considered eye marker genes (Figure 3.11). This change seems to strongly indicate a loss of proper differentiation in GSH-deficient lens epithelia. This is supported by the decrease in β- and γ-crystallin expression (Fig.

3.5), which are restricted to lens tissue and thus highly specific markers.

Interestingly, one exception to this trend was the slight upregulation of lens epithelia marker Foxe3 (Table 6.1). This gene encodes a transcription factor that is essential to lens development but its activity induces proliferation of lens epithelia

139 and inhibits the proper differentiation of lens fiber cells (Landgren et al., 2008).

Thus, overexpression of Foxe3 may promote inappropriate lens epithelia cell growth and disrupt lenticular morphology, contributing to EMT-mediated pathology.

In comparing the GSH-deficient lens transcriptome to the transcriptional profile of mouse lenses actively undergoing EMT (Table 3.2), a major similarity was in the loss of vision genes, including Crybb1, Rho, Rbp3 and Lim2, as well as changes in genes relating to lipid homeostasis, such as Apoe, Ldlr, and Fabp5, and histone upregulation. These results show that the GSH-deficient mouse lens transcriptome recapitulates many of the same features of the EMT lens transcriptome.

3.5.6 Analogies and Differences from Other Models of Oxidative Stress and GSH-Depletion

One important question is the extent to which the above findings serve as a blue print for genetic response to oxidative stress in other systems. While a strong relationship exists between low GSH and oxidative stress, the reverse is not necessarily true. One gene expression study on the response of the human lens epithelial cell line SRA 01-04 to acute H2O2 exposure (Carper et al., 1999) found differential expression of glutamine cyclotransferase (Qpct), cytokine inducible nuclear protein (Cinp), glycoprotein 130 (Gp130), ribosomal protein S10 (Rps10), mitochondrial NADH dehydrogenase 4 (Mt-Nd4), mitochondrial NADH dehydrogenase 5 (Mt-Nd5), mitochondrial cytochrome b (Mt-Cyb), mitochondrial

16S rRNA (Mt-Rnr2), cathepsin (Cst), alternative splicing factor (Asf), and β- hydroxyisobutyryl-coenzyme A hydrolase (Hibch) in response to H2O2. These

140 specific gene changes do not have any overlap with the data presented in this paper and primarily relate to a downregulation in mitochondrial activity, likely as a means to reduce production of oxidative species. However, the authors do note that some of these changes, such as upregulation of Asf and Cinp are associated with epithelial cancers and regulated by TNF, indicating some similarity in signaling.

Another study determined differentially regulated genes in the mouse lens epithelial cell line αTN4 after chronic H2O2 exposure (Carper et al., 2001) This study found upregulation of catalase (Cat), glutathione peroxidase (Gpx), ferritin light chain (Ftl), ferritin heavy chain (Fth1), reticulocalbin (Rcn1), and Cryab and downregulation of Cryaa. This study reported a lack of change in mitochondrial genes and in Mt2. Once again, there was a lack of specific overlap between these results, which show a more traditional oxidative stress response, and what we found in GSH-depleted lenses. However, changes in iron homeostasis and binding are indicated by the upregulation of ferritin, similar to how GSH-depleted lenses show an upregulation in Trf and other iron homeostasis genes. Additionally, the downregulation of Cryaa is similar to the downregulation of crystallins and other eye genes found in the GSH-depleted lens and also indicates a possible movement of the cells away from their proper lineage. No studies on gene expression responses to oxidative stress in lens fiber cells exist for comparison to our data.

Looking more generally at global transcriptomic responses to oxidative stress outside of the lens, one study treated cultured bovine blastocysts with either the pro-oxidant 0.01 mM 2,2′-azobis (2-amidinopropane) dihydrochloride (AAPH) or BSO and measured the response using microarrays (Cagnone and Sirard, 2013).

141

The study found that of the 231 gene changes in AAPH-treated blastocysts and 481 gene changes in BSO-treated blastocysts, 69 of the gene changes overlapped. This demonstrates that GSH-depletion causes many changes outside of traditional oxidative stress responses and may not induce many expected oxidative stress responses, as we have shown in the current study. Interestingly, the transcriptomic profile of AAPH-treated blastocysts which showed TGF-β and TNF signaling and extracellular matrix organization as top altered pathways and changes in Col1a2,

Dcn, and Serpine1, was more similar to that of GSH-depleted lenses than that of

BSO-treated blastocysts, which instead showed a traditional Nrf2-mediated response.

There are a number of possible reasons for the differences between other studies and our results. One reason is that GSH-depletion alone may not truly induce oxidative stress in the lens, which exists in an oxygen-depleted environment with a low amount of metabolic activity, which is in direct contrast to cultured cell lines and blastocysts, which are rapidly dividing and exposed to much higher levels of oxygen. Another is that GSH, which can be conjugated to a multitude of proteins and other compounds, has many effects outside of antioxidant defense pathways.

The differences and similarities between this study and others indicates that culture systems and treating lenses with high levels of pro-oxidants may not be accurate methods to model the environment of the aging lens.

3.5.7 Limitations of this Study

The data presented above represents a comprehensive picture of the genetic response by the lens to GSH depletion. However, the degree to which the changes

142 noted here are carried through to protein expression remains unknown and, conversely, additional changes may be occurring at the protein level that are not represented in the transcriptomic data, including post-translational modifications.

In that regard, a follow-up proteomics study is in progress that will examine the extent to which the observed genetic changes indeed result in translational changes and any differences between the GSH depleted lens transcriptome and proteome.

3.5.7. Conclusions

The GSH-deficient mouse lens shows significant changes in lipid homeostasis, transport systems, and detoxifying genes, as well as a clear activation of EMT signaling. These data give new insights into the adaptation to and consequences of GSH-deficiency and shows that lenses do not mount a typical oxidative stress response to GSH depletion. These results also indicate that supplementation of GSH or a precursor after cataract surgery could potentially reduce the incidence of EMT-mediated posterior subcapsular opacification.

4. Analysis of Downstream Metabolic Effects of GSH-Dependent Transcriptomic Changes in the Lens

Original unpublished research

144

4.1. Introduction

To determine whether the noted transcriptional changes in iron and urea transport systems result in measurable metabolic changes in the lens, the iron and urea contents of wild-type and GSH-deficient lenses were analyzed using colorimetric assays.

4.2. Materials and Methods

4.2.1 Animal Work

Both the LEGSKO and wild-type mice used were of C57BL/6 genetic background and age-matched at 3 months. Mice were housed under diurnal lighting conditions and allowed free access to food and water. Mice were euthanized by

CO2 asphyxiation. Experimental protocols for this study were approved by the

Institutional Animal Care and Use Committee (IACUC) of Case Western Reserve

University.

4.2.2 Lens Total Iron Assay

The TECO Diagnostics (Anaheim, CA) Iron Assay Kit was used and the protocol was modified to allow for quantitation of iron from tissue rather than serum. Briefly, tissue was homogenized in purified water (10% or less weight to volume) and 10-15 µL of the homogenate were added to 200 µL of iron buffer

(acetate buffer containing 220 mM hydroxylamine hydrochloride, pH 4.5). These samples were vortexed and heated to 37oC for 10 minutes to promote the reduction and release of iron from proteins and chelation by hydroxylamine hydrochloride.

145

Samples were spun down at 8,000 x g for 10 minutes to precipitate protein and cellular debris and 200 µL of supernatant was added to wells of a 96-well plate. 10

µL of iron color reagent (16.6 mM ferrozine in iron buffer) was added to each well.

Plates were mixed for 30 seconds and then heated at 37oC for 10 minutes to develop color. Absorbance at 560 nM was read using a plate reader (Tecan Trading,

Switzerland). Iron in samples was quantitated by running a set of iron standard dilutions with samples and performing linear regression.

4.2.3 Urea Assay

Urea content of samples was determined using the standard protocol of the urea assay kit from Sigma-Aldrich Corp. (St. Louis, MO, USA). Absorbance at 570 nM was read using a plate reader (Tecan Trading, Switzerland). Urea in samples was quantitated by running a set of urea standard dilutions with samples and performing linear regression. Sample blanks without added enzyme were included to adjust for background signal. 4.2.4 BCA Normalization

Solid tissue samples for all assays were normalized based on mg protein as determined by the standard protocol of the Pierce BCA protein assay (Thermo

Fisher, Grand Island, NY). Absorbance at 560 nM was read using a plate reader

(Tecan Trading, Switzerland).

146

4.2.5 Statistical Analysis

Values are expressed as means ± standard deviations (SD). Statistical significance of differences in mean values was assessed by Student's t-test. Only P values < 0.05 were considered statistically significant.

4.3 Results

4.3.1 Lens Total Iron Content

The total iron (ferric and ferrous) content of lens compartments was assayed by a colorimetric method and compared between WT and LEGSKO mice (Figure

4.1)

Figure 4.1. Total iron content of WT and LEGSKO lens epithelia and fiber cells. Values are means ± SD. n = 4.

Strikingly, lens epithelia had much higher iron content than the rest of the lens, although there was no significant difference between WT and LEGSKO. This may be due to the low amount of protein in the epithelia compared to fiber cells, resulting in a large difference when normalizing by protein content. WT cortical

147 and nuclear fiber cells had only trace amounts of iron while LEGSKO fibers had significantly (P < 0.05) greater iron, with LEGSKO cortical fibers having >15-fold higher levels than WT and LEGSKO nuclear fibers having nearly 100-fold higher levels of iron.

4.3.2 Urea Within and Surrounding the Lens

The urea contents of the lens, aqueous, and vitreous were analyzed using a colorimetric assay and compared between WT and LEGSKO lenses (Figure 4.2).

P(a,b) < 0.01 P(a,c) < 0.005 P(b,c) < 5E-6 P(a,d) < 0.05 P(b,d) < 0.01 P(c,d) < 0.05

Figure 4.2. Urea content of WT and LEGSKO lenses and surrounding fluids.

Values are means ± SD. n = 4.

148

These results show that the lens has significantly (P < 0.05) lower urea content than its surrounding fluids, consistent with a model of active urea extrusion from the lens, and that the urea content of these fluids is not significantly different between WT and LEGSKO mice. Aqueous humor had significantly less urea than vitreous humor (P < 0.05) in both groups. There was a significant (P < 0.01) decrease in the urea content of the LEGSKO lens when compared to WT lenses, resulting in a nearly 50% loss of lens urea content.

4.4 Discussion

4.4.1 Can Lens Fiber Cells Undergo Ferroptosis?

The increase in the total iron content of GSH-deficient lens fiber cells

(Figure 4.1) confirms that the transcriptional changes to iron homeostatic genes found in the GSH-deficient lens (Figure 3.7) result in the accumulation of lenticular iron. This large increase in iron following depletion of GSH is highly indicative of ferroptosis. However, ferroptosis has not previously been described in the lens and it is unknown whether this large increase in iron will actually progress to ferroptotic cell death, particularly in organelle-free mature fiber cells. Unfortunately, as it is currently characterized, ferroptosis lacks clear markers that can be used to confirm its activation. One recent study demonstrated that autophagy is a central feature of the ferroptosis pathway (Gao et al., 2016). Based on these data, a study that combines analysis of autophagic activity, cellular iron content, redox status, and measurement of cell death would robustly confirm whether ferroptosis truly occurs in the lens and could provide important new insights into cataractogenesis.

149

4.4.2 Urea Homeostasis in the Lens

Initially, it was assumed that the robust upregulation of Slc14a1, encoding urea transporter UT-B, in the GSH-deficient lens (Figure 3.6) was a protective mechanism to clear urea produced in the lens, which can break down into toxic cyanate and carbamylate lens protein (Yan et al., 2010). However, analysis of the urea content of the lens and its surrounding fluids revealed that the lens actually has a significantly lower urea concentration than the aqueous and vitreous, which appear to have similar urea concentrations to mouse blood plasma (Rodrigues et al., 2014). Because UT-B is a facilitative transporter that simply promotes urea transport down a concentration gradient, UT-B cannot assist in the active clearance of urea from the lens. In fact, an increase in UT-B would be expected to increase the urea content of the lens, although the opposite was found to be true when comparing the urea content of LEGSKO and WT lenses (Figure 4.2). This result raises three major questions: 1) What is the function of UT-B in the lens if it does not enhance the uptake of urea? 2) How does the lens maintain significantly lower levels of urea than its surrounding fluids? And 3) How do LEGSKO lenses maintain lower levels of urea than WT?

The answer to the first question could be that UT-B plays a role in lens microcirculation, since it has been shown to have an additional function as a water channel (Yang and Verkman, 2002). This would be consistent with the changes seen in other genes with a putative role in lenticular microcirculation in response to

GSH deficiency (Figure 3.7). Of these putative genes, Slc14a1 was the most

150 robustly and consistently upregulated, indicating that it could play a major, but presently uncharacterized, role in coupling fiber cells.

In regards to the second and third questions, it is possible that the lens may have some urease-like activity, which could be a major route by which urea is cleared from the lens. However, given that ammonia is more toxic than urea, it is unclear why the lens would preferentially degrade urea and produce ammonia, especially within the LEGSKO lens. Another possibility is that an uncharacterized active transporter of urea exists in the lens and is responsible for the extrusion of urea.

Overall, it is evident that a study on the production, transport, and degradation of urea in the lens is needed in order to fully understand the role that this compound plays in lens physiology and if it has any relevancy to cataract.

4.4.3 Conclusions

Increased levels of iron were found in LEGSKO lens fiber cells compared to WT fiber cells, confirming that GSH-deficiency enhances uptake and retention of iron in the lens. Urea levels were found to be decreased in LEGSKO lenses and it was determined that the lens maintains significantly lower levels of urea than its surrounding fluids. These findings open new questions and demonstrate that further study on metabolites within the lens is necessary in order to fully understand the consequences of GSH-deficiency and the unique adaptations of the lens.

5. Discussion

152

5.1 Summary of Findings

5.1.1 The Source and Mechanism of Residual GSH Levels in the LEGSKO Lens

LEGSKO lenses are able to maintain >1 mM GSH levels even in the complete absence of its local synthesis (Figure 2.2). This residual GSH does not result from a salvage pathway or alternative synthesis (Figure 2.3) but instead derives from the GSH present in the vitreous humor of these animals, which exists at high levels (Figure 2.16) and appears able to equilibrate with the lens GSH pool via diffusion. This vitreous pool is derived from circulating GSH (Figure 2.19) and is rapidly replenished so that it is always maintained at >1 mM (Figure 2.18) The lens epithelia also feature an active transport mechanism to take up GSH from the aqueous humor (Figure 2.13) but, due to the low affinity of the system and the low concentration of GSH in the aqueous, this mechanism does not contribute significantly to LEGSKO lens GSH content.

Large animals appear to lack a highly concentrated vitreous GSH pool

(Figure 2.20), which may predispose their lenses to cataract formation. This difference in vitreous GSH content does not appear to be an evolutionary adaptation since distantly related small animals such as fish and frogs also maintain a high level of vitreous GSH. Rather, it appears that this difference is due primarily to eye morphology and the surface area/volume ratio of the vitreous.

153

5.1.2 The Major Consequences of Glutathione Deficiency in the Lens

GSH deficiency in the lens induced major changes in lipid homeostasis

(Figure 3.8) and transporter genes (Figure 3.7), and upregulation of detoxification genes (Figure 3.7), including metallothioneins, aldehyde dehydrogenases, and carboxylesterase. Additionally, the most robust changes were in pathways related to EMT, including upregulation of canonical EMT genes (Figure 3.9), a loss of vision genes and crystallins (Figures 3.5 and 3.11), and changes to cell cycle

(Figure 3.12) and ECM genes (Figure 3.13). These changes imply that GSH- deficient lenses are primed for EMT, which could be an important factor in PCO development.

5.1.3 Conclusions

The data presented here provide a robust and clear view of how the mouse lens adapts to GSH deficiency by identifying the source of the residual GSH levels in LEGSKO mouse lenses and the major transcriptional consequences of lenticular

GSH deficiency.

5.2 Future Directions and Open Questions

5.2.1 Glutathione as a Potential Therapeutic Agent

Glutathione supplementation through eye drops has previously been tested as a clinical therapeutic approach for age-related cataract patients but was found to have no beneficial effect on visual acuity or the density of opacities (Sharma et al.,

154

1989). This may be partially due to the fact that, as a highly polar molecule, GSH cannot easily pass through membranes (Garcia et al., 2015) and thus will not penetrate the cornea to gain access to the lens. Because of this and the blood- aqueous/blood-retinal barriers, it is predicted that supplementation with unmodified

GSH, whether locally or systemically, will not have any significant preventative effect on cataract formation.

A more promising alternative is glutathione ethyl ester (GSH-EE), a modified form of GSH that is more permeable to cell membranes and converted into GSH once taken up by cells (Levy et al., 1993). GSH-EE has been shown to have a slight inhibitory effect on cataract formation in diabetic rats (Zhang et al.,

2008) and in mouse lenses cultured in BSO (Calvin et al., 1997). GSH-EE has not yet been tested in a clinical setting.

Another alternative is N-acetylcysteine (NAC), which acts as a GSH precursor by providing cells with cysteine (Rushworth and Megson, 2014), and has been approved for clinical trials to treat a number of conditions (Hildebrant et al.,

2015;Deepmala et al., 2015). NAC also appears to have some minor antioxidant function of its own, independent of GSH production (Rushworth and Megson,

2014), which may be due to its free sulfhydryl group. However, given that the activity of glutathione synthesis enzymes decreases dramatically in the lens with age (Rathbun et al., 1993), it is unlikely that supplementation with an early precursor to GSH would have a substantial beneficial effect in aged human lenses.

Given that activation of EMT signaling is tightly linked to lens GSH content in the LEGSKO mouse model (Chapter 3), GSH supplementation is more attractive

155 as a potential therapeutic approach to reduce the risk of PCO after cataract surgery.

This treatment could involve injection of GSH directly into the anterior chamber of the eye during cataract surgery, the coating of the intraocular lens implant with a

GSH-containing gel, or ongoing supplementation with GSH-EE or NAC after surgery. Such treatments could potentially reduce the occurrence of PCO and the need for laser ablation surgery, greatly reducing the number of individuals with visual impairment. Such preventative treatments would need to first be tested in animal models and in a clinical setting to determine their effectiveness and safety, but could represent a simple way to reduce the incidence of PCO.

5.2.2 The Source of Vitreous Glutathione

One major unanswered question raised by this work is how the vitreous humor obtains its high levels of GSH. Because the vitreous humor contains only a sparse collection of cells, known as hyalocytes (Qiao et al., 2005), and because oral administration of BSO, which poorly penetrates blood-tissue barriers (Fekete et al.,

1990), is effective at lowering vitreous GSH concentration (Figure 2.19), it is unlikely that the vitreous GSH pool is synthesized within the vitreous. Rather, it must be transported from either the ciliary body, the retina, or both. Which tissue is the main contributor, whether this is an active or passive process, and the identity of the transport system are all lines of inquiry that remain unaddressed. Thus, this question could first be addressed by culturing each tissue and measuring their GSH efflux rates. The role of different transport systems could be assessed by siRNA or

CRISPR/Cas9-mediated knockdown and energetics could be determined by altering temperature and utilizing ATP synthase inhibitors.

156

If the transport system responsible for secreting GSH into the vitreous can be characterized, it represents still another therapeutic target that could be targeted for modulation to increase the GSH content in human vitreous.

5.2.3 Identity of the Lens Epithelium Glutathione Transporter

While I have shown that a low affinity active transport system for GSH exists in lens epithelia (Chapter 2), as have others (Stewart-DeHaan et al., 1999;

Zlokovic et al., 1994), the identity of this transport system remains unknown. While this transporter does not appear to be important for GSH homeostasis in the rodent lens, it may play a more essential role in human lenses, which lack the concentrated reservoir of GSH in vitreous humor found in rodents (Chapter 2). Thus identification and characterization of this transporter could produce another target for therapeutic agents to increase lenticular GSH levels in the aging human eye.

Because there are still no confirmed GSH importers in mammals, this will likely require a broad screening of potential transporters. This could be tested by determining the lens epithelia membrane proteome and then testing all the candidate importers in cell culture or proteoliposomes. Another method would be to generate a cDNA library from the lens and test these sequences for GSH transport activity in a model system such as yeast or oocytes. This cDNA method was used previously (Kannan et al., 1995), but led to false identification of a transporter (Li et al., 1997). Thus, any attempts to characterize the lens GSH transporter must be carried out carefully to ensure that any transport is not a result of cloning artifacts but instead represents the activity a true mammalian transporter.

157

5.2.4 The Glutathione-Depleted Lens Proteome

While the GSH-depleted lens transcriptome gives a good insight into the consequences of GSH deficiency in the lens, not all changes to mRNA regulation will result in changes at the protein level and, conversely, protein expression changes may occur that are not significantly regulated by their corresponding mRNA concentration. Thus, in order to gain a more comprehensive picture of the effect of GSH-deficiency in the lens, proteomic data will need to be obtained using a similar methodology to the transcriptomic data collection. The resulting dataset may help to clarify lingering questions raised by the transcriptome, such as the apparent inactivation of Nrf2 and the net effect on lipid homeostasis and transport systems. 5.2.5 The Potential Protective Role of GSTK1 in the Lens

Glutathione S-transferase kappa 1 (Gstk1) was identified as the only GSH- related gene upregulated in the GSH-depleted mouse lens (Table 6.1), making it an intriguing gene to consider for a potential protective role. This increase in production of GSTK1 could allow its reactions to proceed at the normal rate despite lower levels of its substrate, GSH, and thus could be an adaptive regulation change to maintain lenticular homeostasis.

GSTK1 is the least well understood GST enzyme, particularly within the lens, where it has yet to be studied in any capacity. Recent studies in other systems have demonstrated that GSTK1 has some peroxidase activity, including the ability to detoxify cumene hydroperoxide, tert-butyl hydroperoxide, and 15-S- hydroperoxy-5,8,11,13-eicosatetraenoic acid (Morel et al., 2004). This is likely due

158 to GSTK1 sharing a homologous thioredoxin-like domain with GPX (Li et al.,

2005). In fact, an analysis of the evolutionary relatedness of the GST superfamily reveals that GSTK1 did not evolve from the same lineage as other GST family members and is instead more closely related to GPX and mitochondrial glutathione transferases (Nebert and Vasiliou, 2004). Given the lack of peroxidases and GSH- related enzymes upregulated in GSH-deficient lenses, this could indicate GSTK1 is a major unidentified protective enzyme for maintaining lenticular redox.

However, unlike other GST enzymes, GSTK1 is localized to the endoplasmic reticulum, peroxisomes, and mitochondria (Morel et al., 2004;Liu et al., 2015). While these organelles are important sites of oxidative stress and damage, this indicates that GSTK1 likely only exists in the periphery of the lens and does not play a direct role in protecting the organelle-free lens nucleus.

Additionally, GSTK1 plays an important role in lipid homeostasis, as it binds to and regulates the concentration of adiponectin, which breaks down fatty acids (Liu et al., 2015). Thus, upregulation of Gstk1 may not be a protective response to oxidative stress but, instead, part of the same shift in lipid homeostasis seen throughout the GSH-deficient lens transcriptome (Figure 3.8).

Overall, it is still unclear whether GSTK1 plays a protective role in the lens and an in depth study of GSTK1 activity in the lens will need to be carried out to assess its putative beneficial role.

159

5.2.6 Conclusions

The data presented here provide a robust insight into lenticular redox homeostasis but also opens up many new lines of inquiry that should be investigated in relation to age-related cataract and its clinical treatment.

5.3 Concluding Remarks

This body of work, we believe, greatly expands our understanding of glutathione homeostasis in the lens and the consequences of its depletion. Chapter

2 details a major difference between humans and rodents in relation to glutathione homeostasis and its effect on the fitness of the aging lens. Chapter 3 identifies a set of genes responsive to low GSH conditions that may help provide continued protection to the lens, as well as clarifying the non-oxidative consequences of GSH depletion. This includes a robust activation of EMT signaling, indicating GSH as a strong candidate for the preventative treatment of PCO. Chapter 4 analyzes the downstream effects of some of these regulation changes and gives new insight into iron and urea homeostasis in the lens. Taken together, this work opens up many new avenues for cataract research from the bench to the clinic.

6. Appendix

Supplemental data from submitted article in review: “Whitson JA, Zhang X, Medvedovic M, Chen J, Monnier VM, Fan X. Transcriptome of the GSH-Depleted Lens Reveals Changes in Detoxification and EMT Signaling Genes, Transport Systems, and Lipid Homeostasis. Investigative Ophthalmology & Visual Sciences.”

161

Figure 6.1. Analysis of RNA-Seq Results. (A and B) Correlation coefficients between epithelial (A) and fiber cell (B) samples. All samples had a correlation of at least 0.986 to all other samples of the same cell type. (C and D) Heatmaps and clustering of samples for epithelia (C) and fiber cells (D). (E-H) Volcano plots showing significant gene expression changes in LEGSKO epithelia (E), LEGSKO fiber cells (F), LEGSKO epithelia treated with BSO (G), and LEGSKO fiber cells treated with BSO (H) compared to wild-type controls. Each point represents a single gene.

162

Figure 6.2. Verification of qPCR primers. Sequence, product size, melt curve, and linear regression of cDNA standard dilutions are shown for each primer set used in qPCR analysis.

163

Figure 6.2. Continued.

164

Figure 6.2. Continued.

165

Figure 6.2. Continued.

166

Figure 6.2. Continued.

167

Figure 6.2. Continued.

168

Figure 6.2. Continued

169

Figure 6.3. Expression changes in transcription genes. (A) Lens epithelia, (B) lens fiber cells. Values are means +/- SD. *, p<0.05, compared to WT; **, p<0.005, compared to WT; ***, p<0.0005, compared to WT; ****, p<5E-5, compared to WT †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, p<5E-5, compared to LEGSKO.

170

p<5E

SD.

5, compared to WT to compared 5,

ber cells. Values are means +/ Values means berare cells.

(A) Lens epithelia, (B) lensfi (B) epithelia, Lens (A)

4. Expression changes in changes Expression genes. metabolism 4.

gure gure

*, p<0.05, compared to WT; **, p<0.005, compared to WT; ***, p<0.0005, compared to WT; ****, p<5E ****, WT; to compared p<0.0005, ***, WT; to comparedp<0.005, **, WT; to compared p<0.05, *, ††††, LEGSKO; to compared p<0.0005, †††, LEGSKO; to compared p<0.005, ††, LEGSKO; to compared p<0.05, †, LEGSKO. to compared Fi

171

5, 5,

5, compared to to compared 5, WT

p<5E

(A) Lens epithelia, (B) lens fiber cells. Values are means are Values cells. means lens (B) epithelia, fiber Lens (A)

5. Expression changes in developmental biology genes. genes. in changes Expression biology developmental 5.

SD.

Figure Figure +/ ****, WT; to compared p<0.0005, ***, compared p<0.005, **, WT; to WT; to compared p<0.05, *, p<5E ††††, LEGSKO; to compared p<0.0005, †††, LEGSKO; to compared p<0.005, ††, LEGSKO; to compared p<0.05, †, LEGSKO. to compared

172

SD.

p<5E

5, compared to to compared 5, WT

ells. Values are means +/ means are Values ells.

(A) Lens epithelia, (B) lens fiber c fiber lens (B) epithelia, Lens (A)

6. Expression changes in immune response genes. genes. response immune in changes Expression 6.

Figure Figure ****, p<5E WT; to compared p<0.0005, ***, compared p<0.005, **, WT; to WT; to compared p<0.05, *, †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, LEGSKO. to compared

173

Figure 6.7. Expression changes in protein metabolism genes. (A) Lens epithelia, (B) lens fiber cells. Values are means +/- SD. *, p<0.05, compared to WT; **, p<0.005, compared to WT; ***, p<0.0005, compared to WT; ****, p<5E-5, compared to WT †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, p<5E-5, compared to LEGSKO.

174

5, compared to to compared 5, WT

cells. Values are means cells.means are Values

(A) Lens epithelia, (B) lens fiber Lens (A) epithelia, (B) lens

8. Expression changes in changes Expression 8. small transport molecule genes.

SD.

Figure Figure +/ ****, p<5E WT; to compared p<0.0005, ***, compared p<0.005, **, WT; to WT; to compared p<0.05, *, †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, p<5 LEGSKO. to compared

175

Figure 6.9. Expression changes in extracellular matrix organization genes. (A) Lens epithelia, (B) lens fiber cells. Values are means +/- SD. *, p<0.05, compared to WT; **, p<0.005, compared to WT; ***, p<0.0005, compared to WT; ****, p<5E-5, compared to WT †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, p<5E-5, compared to LEGSKO.

176

SD.

5, compared to to compared 5, WT

cells. Values are means +/ are cells.means Values

(A) Lens epithelia, (B) lens fiber (B) epithelia, Lens (A)

10. Expression changes in cell cycle genes. in changes Expression cyclecell genes. 10.

177

Figure 6.11. Expression changes in cellular stress response genes. (A) Lens epithelia, (B) lens fiber cells. Values are means +/- SD. *, p<0.05, compared to WT; **, p<0.005, compared to WT; ***, p<0.0005, compared to WT; ****, p<5E-5, compared to WT †, p<0.05, compared to LEGSKO; ††, p<0.005, compared to LEGSKO; †††, p<0.0005, compared to LEGSKO; ††††, p<5E-5, compared to LEGSKO.

Table 6.1. All significant (P < 0.05; FDR < 0.1) gene expression changes in LEGSKO and BSO-treated LEGSKO lenses compared to WT.

178

LEGSKOvWT BSOvWT LEGSKOvWT BSOvWT Symbol Description Epithelia Epithelia Fibers Fibers 1300017J02Rik RIKEN cDNA 1300017J02 gene 2.51 1700001O22Rik RIKEN cDNA 1700001O22 gene 3.60 2.97 1700027H10Rik RIKEN cDNA 1700027H10 gene -9.93 1700071M16Rik RIKEN cDNA 1700071M16 gene 2.30 2010204K13Rik RIKEN cDNA 2010204K13 gene 2.23 2210408F21Rik RIKEN cDNA 2210408F21 gene 2.27 2410141K09Rik RIKEN cDNA 2410141K09 gene -4.54 2610100L16Rik RIKEN cDNA 2610100L16 gene -5.55 2610305D13Rik RIKEN cDNA 2610305D13 gene 2.11 2810032G03Rik RIKEN cDNA 2810032G03 gene 2.05 4732416N19Rik RIKEN cDNA 4732416N19 gene 4.90 4921511M17Rik RIKEN cDNA 4921511M17 gene -100.76 4930444P10Rik RIKEN cDNA 4930444P10 gene 2.43 4930453H23Rik RIKEN cDNA 4930453H23 gene -4.10 4930502E18Rik RIKEN cDNA 4930502E18 gene 2.56 3.75 4930558K02Rik RIKEN cDNA 4930558K02 gene -3.94 4930593A02Rik RIKEN cDNA 4930593A02 gene -3.64 4930598F16Rik RIKEN cDNA 4930598F16 gene -4.71 4932435O22Rik RIKEN cDNA 4932435O22 gene -2.76 4932438H23Rik RIKEN cDNA 4932438H23 gene 2.28 2.10 4932702P03Rik RIKEN cDNA 4932702P03 gene -2.41 4933406I18Rik RIKEN cDNA 4933406I18 gene 6.77 6430584L05Rik RIKEN cDNA 6430584L05 gene 4.89 9130409I23Rik RIKEN cDNA 9130409I23 gene 2.48 A230077H06Rik RIKEN cDNA A230077H06 gene 2.67 A2m alpha-2-macroglobulin 2.68 A730056A06Rik RIKEN cDNA A730056A06 gene 3.02 3.63 A930006K02Rik RIKEN cDNA A930006K02 gene 2.94 3.08 A930009A15Rik RIKEN cDNA A930009A15 gene 4.37 Abca1 ATP-binding cassette, sub- 2.80 family A (ABC1), member 1 Abca13 ATP-binding cassette, sub- 4.25 family A (ABC1), member 13 Abcg1 ATP-binding cassette, sub- 9.25 13.04 family G (WHITE), member 1 Abhd16b abhydrolase domain containing 11.44 13.00 16B Ace angiotensin I converting enzyme (peptidyl-dipeptidase 2.20 A) 1 Actn3 actinin alpha 3 2.12 2.29 Adra1b adrenergic receptor, alpha 1b 2.37 3.83 Afp alpha fetoprotein 4.57 5.08 Akna AT-hook transcription factor 3.34

179

Akr1b10 aldo-keto reductase family 1, 2.16 member B10 (aldose reductase) Aldh1a1 aldehyde dehydrogenase family 2.18 2.66 1, subfamily A1 Aldh1a3 aldehyde dehydrogenase family -2.81 -3.04 1, subfamily A3 Aldh1a7 aldehyde dehydrogenase family 2.09 2.18 1, subfamily A7 Aldh1l1 aldehyde dehydrogenase 1 2.57 family, member L1 Aldh3a1 aldehyde dehydrogenase family 2.15 3, subfamily A1 Alk anaplastic lymphoma kinase 2.74 Alox15 arachidonate 15-lipoxygenase -4.49 Alpk3 alpha-kinase 3 -5.23 Amtn amelotin 8.15 Amy2a5 amylase 2a5 -2.19 -2.82 Amy2b amylase 2b -2.09 -2.82 Amz1 archaelysin family -2.98 -2.28 metallopeptidase 1 Ankrd34c ankyrin repeat domain 34C -2.25 Apobec1 apolipoprotein B mRNA editing -2.98 enzyme, catalytic polypeptide 1 Apoe apolipoprotein E 2.03 2.13 As3mt arsenic (+3 oxidation state) 2.24 methyltransferase Asb2 ankyrin repeat and SOCS box- 2.51 containing 2 Asphd2 aspartate beta-hydroxylase -2.05 domain containing 2 Astn2 astrotactin 2 2.70 2.29 Atf3 activating transcription factor 3 -2.99 Axl AXL receptor tyrosine kinase -2.26 BC048679 cDNA sequence BC048679 2.02 BC052688 cDNA sequence BC052688 -2.09 Best3 bestrophin 3 2.31 Bgn biglycan -4.06 Bmx BMX non-receptor tyrosine 2.62 2.50 kinase Brinp1 bone morphogenic protein/retinoic acid inducible 3.01 3.05 neural specific 1 Bst1 bone marrow stromal cell 2.62 antigen 1 Btbd17 BTB (POZ) domain containing 2.33 17 C430002E04Rik RIKEN cDNA C430002E04 gene 8.74 Cabp2 calcium binding protein 2 -2.37 Cabp5 calcium binding protein 5 5.21 7.00 Cacng4 calcium channel, voltage- 2.00 dependent, gamma subunit 4 Cacng8 calcium channel, voltage- -2.99 dependent, gamma subunit 8 Camk1g calcium/calmodulin-dependent 4.43 protein kinase I gamma

180

Camk2n2 calcium/calmodulin-dependent 3.06 protein kinase II inhibitor 2 Capsl calcyphosine-like 3.04 Car2 carbonic anhydrase 2 2.16 Car4 carbonic anhydrase 4 2.29 2.76 Car9 carbonic anhydrase 9 2.09 2.77 Ccdc11 coiled-coil domain containing -3.34 11 Ccdc177 coiled-coil domain containing -4.31 177 Ccdc23 coiled-coil domain containing -2.64 23 Ccr2 chemokine (C-C motif) receptor 22.97 2 Cd200 CD200 antigen 2.81 2.95 Cd36 CD36 antigen -3.50 Cdc7 cell division cycle 7 (S. -2.00 cerevisiae) Cdk5r2 cyclin-dependent kinase 5, 12.29 regulatory subunit 2 (p39) Cebpb CCAAT/enhancer binding -4.32 -2.31 protein (C/EBP), beta Cend1 cell cycle exit and neuronal 2.11 differentiation 1 Cenpm centromere protein M 2.88 Ces1g carboxylesterase 1G 2.55 3.29 2.97 Cfap58 cilia and flagella associated 2.97 3.39 protein 58 Cfi complement component factor 27.29 26.01 i Chrm2 cholinergic receptor, muscarinic 4.13 5.98 2, cardiac Chrng cholinergic receptor, nicotinic, -2.86 gamma polypeptide Ckm creatine kinase, muscle 2.01 Clrn1 clarin 1 4.41 4.02 Clstn3 calsyntenin 3 2.19 2.37 Cngb1 cyclic nucleotide gated channel -4.08 beta 1 Cntn1 contactin 1 2.24 Cobll1 Cobl-like 1 -2.62 -2.90 Col16a1 collagen, type XVI, alpha 1 -2.45 Col1a1 collagen, type I, alpha 1 -2.36 Col1a2 collagen, type I, alpha 2 5.96 4.77 Coro6 coronin 6 -2.30 Cpne6 copine VI -8.67 Cpne7 copine VII 2.06 2.02 Crabp2 cellular retinoic acid binding -7.73 -5.21 protein II Creb3l4 cAMP responsive element 2.11 4.31 binding protein 3-like 4 Crip2 cysteine rich protein 2 -2.50 Crisp2 cysteine-rich secretory protein 2.57 2.26 2

181

Cryga crystallin, gamma A -2.63 Cryge crystallin, gamma E -2.08 -2.32 -2.94 Crygf crystallin, gamma F -2.61 -3.03 -5.50 Csn3 casein kappa 14.35 Cxcl5 chemokine (C-X-C motif) ligand -32.36 5 Cyp2j9 cytochrome P450, family 2, 10.71 8.00 subfamily j, polypeptide 9 Dcdc2a doublecortin domain containing 2.37 2a Dcn decorin 15.16 Dct dopachrome tautomerase 2.41 Dgat2 diacylglycerol O-acyltransferase -2.49 2 Dmc1 DMC1 dosage suppressor of mck1 homolog, meiosis-specific 5.20 homologous recombination Dmgdh dimethylglycine dehydrogenase 18.05 precursor Dpys dihydropyrimidinase -2.97 Dpysl3 dihydropyrimidinase-like 3 2.26 Drd4 dopamine receptor D4 2.15 Dusp4 dual specificity phosphatase 4 -2.65 -2.20 E230016D10 uncharacterized protein -2.30 E230016D10 Edaradd EDAR (ectodysplasin-A receptor)-associated death 3.84 3.78 domain Efcab6 EF-hand calcium binding -3.67 -3.37 domain 6 Egf epidermal growth factor 2.03 Elmod1 ELMO/CED-12 domain 2.09 containing 1 Emp3 epithelial membrane protein 3 -2.01 Endou endonuclease, polyU-specific 2.62 Enpep glutamyl aminopeptidase 2.17 Entpd3 ectonucleoside triphosphate 2.26 diphosphohydrolase 3 Epha3 Eph receptor A3 -3.09 Ephb2 Eph receptor B2 -2.09 Epn3 epsin 3 -2.13 Eppk1 epiplakin 1 -2.55 Eps8 epidermal growth factor 2.11 2.35 receptor pathway substrate 8 Erich2 glutamate rich 2 2.15 Ermap erythroblast membrane- -2.48 -4.01 associated protein Etnk2 ethanolamine kinase 2 2.55 Etv4 ets variant 4 2.27 Eva1b eva-1 homolog B (C. elegans) -2.98 Exoc3l exocyst complex component 3- 2.37 like

182

Fam166b family with sequence similarity 9.84 166, member B Fam181a family with sequence similarity 2.48 181, member A Fam184b family with sequence similarity 2.90 184, member B Fam83d family with sequence similarity -2.65 83, member D Fam84a family with sequence similarity 4.85 -2.72 84, member A Fbxw15 F-box and WD-40 domain -2.96 protein 15 Fbxw19 F-box and WD-40 domain -2.98 -2.26 protein 19 Fgfr4 fibroblast growth factor 2.13 receptor 4 Fhl2 four and a half LIM domains 2 -5.70 Fignl1 fidgetin-like 1 -5.56 Foxe3 forkhead box E3 2.03 2.30 Frat1 frequently rearranged in 2.30 2.25 advanced T cell lymphomas Frem1 Fras1 related extracellular 3.13 3.16 matrix protein 1 Fscn2 fascin homolog 2, actin- bundling protein, retinal -3.61 (Strongylocentrotus purpuratus) G0s2 G0/G1 switch gene 2 3.23 2.54 4.35 5.42 Gabra4 gamma-aminobutyric acid (GABA) A receptor, subunit -2.96 alpha 4 Gabrr1 gamma-aminobutyric acid (GABA) C receptor, subunit rho 5.43 6.50 1 Gabrr2 gamma-aminobutyric acid (GABA) C receptor, subunit rho -5.02 -2.39 2 Gal galanin 6.28 Galnt12 UDP-N-acetyl-alpha-D- galactosamine:polypeptide N- -2.36 acetylgalactosaminyltransferase 12 Gata5 GATA binding protein 5 -3.03 Gbp9 guanylate-binding protein 9 -2.16 Gfra1 glial cell line derived neurotrophic factor family 2.34 2.25 receptor alpha 1 Gjb3 gap junction protein, beta 3 5.20 Gjb6 gap junction protein, beta 6 -3.37 Glcci1 glucocorticoid induced 3.61 transcript 1 Glt28d2 glycosyltransferase 28 domain -2.67 containing 2 Gm10007 predicted gene 10007 2.93 Gm10057 predicted gene 10057 -75.57 Gm12250 predicted gene 12250 -4.81 Gm13040 predicted gene 13040 13.34 Gm13057 predicted gene 13057 13.34

183

Gm14499 predicted gene 14499 4.27 Gm14548 predicted gene 14548 18.00 Gm15663 predicted gene 15663 2.01 Gm19522 predicted gene, 19522 2.32 Gm20831 predicted gene, 20831 -3.05 Gm20858 Sycp3 like Y-linked pseudogene -6.58 Gm4541 NLR family, pyrin domain -12.33 -10.34 containing 4E pseudogene Gm4710 predicted gene 4710 13.21 8.93 Gm4792 predicted gene 4792 -22.85 Gm7120 predicted gene 7120 -2.34 -2.13 Gm833 predicted gene 833 -5.13 -7.80 Gm9839 predicted gene 9839 -6.81 Gml GPI anchored molecule like -3.06 protein Gmnc geminin coiled-coil domain -2.53 containing Gna14 guanine nucleotide binding 2.98 protein, alpha 14 Gnat1 guanine nucleotide binding -3.51 protein, alpha transducing 1 Gpnmb glycoprotein (transmembrane) 5.85 5.41 nmb Gpr115 G protein-coupled receptor 115 -4.45 Gpr152 G protein-coupled receptor 152 -4.13 Gpr158 G protein-coupled receptor 158 -2.42 Gpr85 G protein-coupled receptor 85 -2.35 Gprc6a G protein-coupled receptor, 2.68 family C, group 6, member A Grin2a glutamate receptor, ionotropic, 2.71 2.08 NMDA2A (epsilon 1) Grin2c glutamate receptor, ionotropic, 3.75 NMDA2C (epsilon 3) Gstk1 glutathione S-transferase kappa 2.52 1 Guca1b guanylate cyclase activator 1B -2.92 Hdac9 histone deacetylase 9 -2.02 Hist1h3a histone cluster 1, H3a 3.52 Hist1h3b histone cluster 1, H3b 3.42 Hist1h3d histone cluster 1, H3d 2.90 Hist1h3e histone cluster 1, H3e 3.41 Hist1h3f histone cluster 1, H3f 4.03 Hist1h3g histone cluster 1, H3g 3.31 Hist1h3h histone cluster 1, H3h 3.19 Hist1h3i histone cluster 1, H3i 3.09 Hist1h4a histone cluster 1, H4a 2.32 Hist1h4b histone cluster 1, H4b 2.30 Hist1h4c histone cluster 1, H4c 2.45 2.51 Hist1h4d histone cluster 1, H4d 2.97 Hist1h4k histone cluster 1, H4k 2.22 2.38

184

Hist1h4n histone cluster 1, H4n 2.52 Hist2h3b histone cluster 2, H3b 2.17 Hist2h3c1 histone cluster 2, H3c1 2.17 Hist2h3c2 histone cluster 2, H3c2 2.17 Hnmt histamine N-methyltransferase 2.06 Hsd11b1 hydroxysteroid 11-beta 2.31 dehydrogenase 1 Hspa12b heat shock protein 12B 3.82 3.61 Hspb8 heat shock protein 8 -3.41 Htra4 HtrA serine peptidase 4 -2.37 Ier3 immediate early response 3 2.25 Igf2bp2 insulin-like growth factor 2 -9.74 mRNA binding protein 2 Igfbp5 insulin-like growth factor -3.66 binding protein 5 Igfl3 IGF-like family member 3 -2.39 Iigp1 interferon inducible GTPase 1 -5.05 Ikbke inhibitor of kappaB kinase -2.13 -2.38 -2.95 epsilon Il1b interleukin 1 beta 2.05 Ildr1 immunoglobulin-like domain 2.35 2.26 containing receptor 1 Ip6k3 inositol hexaphosphate kinase 3 2.03 Kcnd1 potassium voltage-gated channel, Shal-related family, -2.00 -2.66 -3.46 -3.95 member 1 Kcnd3 potassium voltage-gated channel, Shal-related family, -2.32 -2.84 member 3 Kcne1l potassium voltage-gated channel, Isk-related family, 6.33 member 1-like, pseudogene Kcnip1 Kv channel-interacting protein 1 3.22 Kcnj13 potassium inwardly-rectifying channel, subfamily J, member 2.62 13 Kctd12b potassium channel tetramerisation domain -2.52 containing 12b Kif26a kinesin family member 26A 2.28 2.46 2.56 3.22 Klhl32 kelch-like 32 -2.17 Klhl33 kelch-like 33 6.18 Krt15 keratin 15 2.58 2.27 Krt76 keratin 76 -3.03 Krt78 keratin 78 -2.33 Lamb3 laminin, beta 3 3.07 2.23 4.22 6.25 Laptm5 lysosomal-associated protein -7.25 transmembrane 5 Lcp1 lymphocyte cytosolic protein 1 14.68 Lgi2 leucine-rich repeat LGI family, -2.32 member 2 Lin28b lin-28 homolog B (C. elegans) -6.27 -5.85

185

Lincred1 long intergenic non-protein coding RNA of erythroid -4.16 differentiation 1 LOC100505025 uncharacterized LOC100505025 3.67 Loxhd1 lipoxygenase homology -4.99 domains 1 Lpl lipoprotein lipase 2.18 Lpxn leupaxin 2.47 Lrmp lymphoid-restricted membrane 13.13 14.85 protein Lrp1b low density lipoprotein-related -5.48 protein 1B (deleted in tumors) Lrp5 low density lipoprotein -3.39 receptor-related protein 5 Lrrc73 leucine rich repeat containing 4.39 73 Lrrc9 leucine rich repeat containing 9 3.04 3.15 Ltf lactotransferrin -3.06 Lyz2 lysozyme 2 7.62 Manba mannosidase, beta A, lysosomal 2.00 2.37 Marcksl1 MARCKS-like 1 -2.81 Mir181a-2 microRNA 181a-2 7.71 Mir218-2 microRNA 218-2 12.37 Mir31 microRNA 31 -6.16 Mir3113 microRNA 3113 3.17 Mir762 microRNA 762 Mir99a microRNA 99a 6.30 Mkx mohawk homeobox -3.69 -2.32 -6.52 -2.63 Mpped2 metallophosphoesterase -2.29 domain containing 2 Mreg melanoregulin -3.95 Mt1 metallothionein 1 4.79 7.06 15.67 18.35 Mt2 metallothionein 2 3.91 6.49 Mtnr1a melatonin receptor 1A -2.80 Mybpc1 myosin binding protein C, slow- -3.48 type Myo18b myosin XVIIIb -7.44 Myo5c myosin VC 2.42 N28178 expressed sequence N28178 -3.77 Nap1l5 nucleosome assembly protein 3.12 3.12 1-like 5 Nlgn1 neuroligin 1 3.17 Npas2 neuronal PAS domain protein 2 -5.81 Nptx1 neuronal pentraxin 1 7.06 Nr1h3 nuclear receptor subfamily 1, 2.36 group H, member 3 Nr1h5 nuclear receptor subfamily 1, -2.04 group H, member 5 Nr4a1 nuclear receptor subfamily 4, -3.83 group A, member 1 Nrxn3 neurexin III 2.59

186

Nsg2 neuron specific gene family 2.11 member 2 Nt5e 5' nucleotidase, ecto 2.34 3.44 Ntn5 netrin 5 -3.73 Nxf7 nuclear RNA export factor 7 2.07 Nxph3 neurexophilin 3 -3.22 -2.43 Odf3l2 outer dense fiber of sperm tails -4.28 -8.74 3-like 2 Olfr1372-ps1 olfactory receptor 1372, -6.99 pseudogene 1 Olfr461 olfactory receptor 461 -2.34 Opn1sw opsin 1 (cone pigments), short- wave-sensitive (color blindness, -4.23 tritan) Osgin1 oxidative stress induced growth 2.28 inhibitor 1 P2rx3 purinergic receptor P2X, ligand- 2.07 gated ion channel, 3 Pak7 p21 protein (Cdc42/Rac)- 5.13 5.16 activated kinase 7 Parvb parvin, beta -2.06 Pcdh17 protocadherin 17 2.75 Pcsk2os2 proprotein convertase subtilisin/kexin type 2, opposite -2.53 strand 2 Pcsk9 proprotein convertase 2.22 3.14 2.31 subtilisin/kexin type 9 Pde10a phosphodiesterase 10A -2.23 Pde1c phosphodiesterase 1C 2.20 Pde3b phosphodiesterase 3B, cGMP- 2.02 inhibited Pde4b phosphodiesterase 4B, cAMP 4.40 specific Pde6a phosphodiesterase 6A, cGMP- -3.19 specific, rod, alpha Pde6b phosphodiesterase 6B, cGMP, -2.95 2.35 rod receptor, beta polypeptide Pde6g phosphodiesterase 6G, cGMP- -3.16 specific, rod, gamma Pde6h phosphodiesterase 6H, cGMP- -18.48 specific, cone, gamma Pdgfrl platelet-derived growth factor -2.04 receptor-like Pdhx pyruvate dehydrogenase 2.17 2.24 complex, component X Pik3cg phosphoinositide-3-kinase, 3.09 2.53 catalytic, gamma polypeptide Pipox pipecolic acid oxidase 2.40 Pira1 paired-Ig-like receptor A1 8.99 Pira11 paired-Ig-like receptor A11 11.70 Pira6 paired-Ig-like receptor A6 32.31 Pitpnm3 PITPNM family member 3 -2.51 Pkd2l1 polycystic kidney disease 2-like 2.32 1 Pla2r1 phospholipase A2 receptor 1 -2.04 Plch2 phospholipase C, eta 2 -3.02

187

Plcl2 phospholipase C-like 2 2.12 2.13 Plekhf1 pleckstrin homology domain containing, family F (with FYVE -2.05 domain) member 1 Plin4 perilipin 4 -2.21 Plin5 perilipin 5 -4.73 Plk3 polo-like kinase 3 3.44 Pm20d1 peptidase M20 domain 2.03 2.67 2.69 containing 1 Ppfia2 protein tyrosine phosphatase, receptor type, f polypeptide 3.71 (PTPRF), interacting protein (liprin), alpha 2 Ppm1e protein phosphatase 1E (PP2C domain containing) -2.40

Ppp2r2c protein phosphatase 2, 2.53 regulatory subunit B, gamma Prima1 proline rich membrane anchor 2.21 1 Prph2 peripherin 2 -3.82 Prrg3 proline rich Gla (G- carboxyglutamic acid) 3 -2.01 (transmembrane) Ptprm protein tyrosine phosphatase, -2.10 receptor type, M Ptprn2 protein tyrosine phosphatase, 2.21 receptor type, N polypeptide 2 Pvt1 plasmacytoma variant 2.39 2.53 translocation 1 Pygl liver glycogen phosphorylase 2.10 R3hdml R3H domain containing-like -2.27 Ramp2 receptor (calcitonin) activity -6.48 modifying protein 2 Rasgef1b RasGEF domain family, member -2.23 1B Rasl10b RAS-like, family 10, member B -3.24 Rbms3 RNA binding motif, single -2.94 -3.23 stranded interacting protein Rbp3 retinol binding protein 3, -4.81 -4.26 interstitial Rd3l retinal degeneration 3-like -5.03 Ret ret proto-oncogene -2.20 Rho rhodopsin -3.75 Rhoh ras homolog gene family, 2.17 member H Ribc2 RIB43A domain with coiled-coils -4.12 2 Rmst rhabdomyosarcoma 2 associated transcript (non- -2.27 coding RNA) Rnf208 ring finger protein 208 4.58 Rp1 retinitis pigmentosa 1 (human) -3.67 Rrh retinal pigment epithelium 2.61 derived rhodopsin homolog Rundc3a RUN domain containing 3A 2.02

188

Runx1t1 runt-related transcription factor 1; translocated to, 1 (cyclin D- -2.72 related) S1pr5 sphingosine-1-phosphate 2.57 receptor 5 Sall3 sal-like 3 (Drosophila) -2.37 -2.98 Samt1 spermatogenesis associated multipass transmembrane -100.76 protein 1 Sbpl spermine binding protein-like 2.53 2.62 Scnn1b sodium channel, nonvoltage- -3.36 gated 1 beta Scube1 signal peptide, CUB domain, 2.65 EGF-like 1 Sec14l5 SEC14-like 5 (S. cerevisiae) 2.31 2.00 Sell selectin, lymphocyte 12.34 Sema6a sema domain, transmembrane domain (TM), and cytoplasmic -2.45 domain, (semaphorin) 6A Sept1 septin 1 6.81 Serpina3f serine (or cysteine) peptidase 13.90 inhibitor, clade A, member 3F Serpina3g serine (or cysteine) peptidase 7.63 inhibitor, clade A, member 3G Serpina3n serine (or cysteine) peptidase 9.62 inhibitor, clade A, member 3N Serpinb1c serine (or cysteine) peptidase 2.91 inhibitor, clade B, member 1c Serpine1 serine (or cysteine) peptidase -2.10 inhibitor, clade E, member 1 Serpine3 serpin peptidase inhibitor, clade E (nexin, plasminogen 2.74 activator inhibitor type 1), member 3 Serping1 serine (or cysteine) peptidase -3.05 inhibitor, clade G, member 1 Slamf8 SLAM family member 8 -2.81 Slc14a1 solute carrier family 14 (urea 6.19 6.82 10.26 10.96 transporter), member 1 Slc17a1 solute carrier family 17 (sodium 3.06 2.82 phosphate), member 1 Slc17a7 solute carrier family 17 (sodium-dependent inorganic -3.95 phosphate cotransporter), member 7 Slc18a2 solute carrier family 18 (vesicular monoamine), -2.13 member 2 Slc1a2 solute carrier family 1 (glial high affinity glutamate transporter), 2.14 member 2 Slc22a4 solute carrier family 22 (organic 2.13 cation transporter), member 4 Slc24a4 solute carrier family 24 (sodium/potassium/calcium -2.49 exchanger), member 4 Slc27a6 solute carrier family 27 (fatty 5.53 acid transporter), member 6 Slc30a3 solute carrier family 30 (zinc -2.34 transporter), member 3

189

Slc40a1 solute carrier family 40 (iron- regulated transporter), member -3.44 1 Slc44a5 solute carrier family 44, 2.02 member 5 Slc8a3 solute carrier family 8 (sodium/calcium exchanger), -2.71 member 3 Sly Sycp3 like Y-linked -5.61 Snca synuclein, alpha 6.11 8.46 Snhg11 small nucleolar RNA host gene 2.63 11 Snord1b small nucleolar RNA, C/D box -12.78 1B Snord45b small nucleolar RNA, C/D box -4.91 45B Snx31 sorting nexin 31 2.16 Socs2 suppressor of cytokine signaling -5.50 2 Sost sclerostin 2.31 Spns2 spinster homolog 2 -2.07 Spock2 sparc/osteonectin, cwcv and kazal-like domains proteoglycan -2.43 2 Spp1 secreted phosphoprotein 1 9.26 Srgap1 SLIT-ROBO Rho GTPase -2.24 activating protein 1 Sstr4 somatostatin receptor 4 2.82 St8sia5 ST8 alpha-N-acetyl- neuraminide alpha-2,8- 2.78 sialyltransferase 5 Stat6 signal transducer and activator 2.21 2.20 of transcription 6 Steap3 STEAP family member 3 2.15 Syk spleen tyrosine kinase 5.87 12.61 9.66 Synb syncytin b -5.74 -3.81 Synpo2 synaptopodin 2 -2.42 Tacr3 tachykinin receptor 3 170.64 140.48 Tcea3 transcription elongation factor -2.44 A (SII), 3 Tert telomerase reverse 2.27 transcriptase Tgfbi transforming growth factor, 3.75 beta induced Thbd thrombomodulin 2.21 Tinag tubulointerstitial nephritis 2.24 antigen Tmeff1 transmembrane protein with EGF-like and two follistatin-like -3.12 domains 1 Tmem120b transmembrane protein 120B 2.61 Tmem163 transmembrane protein 163 2.53 2.82 Tmem190 transmembrane protein 190 -6.53 Tmem236 transmembrane protein 236 2.63 Tmem238 transmembrane protein 238 2.01

190

Tmem45a transmembrane protein 45a 2.26 2.33 Tmem95 transmembrane protein 95 -3.42 Tmprss5 transmembrane protease, 2.54 2.32 serine 5 (spinesin) Tmtc1 transmembrane and tetratricopeptide repeat 2.35 3.82 containing 1 Tnc tenascin C 9.59 Tnfaip6 tumor necrosis factor alpha -3.79 -3.88 induced protein 6 Trf transferrin 3.47 3.97 Tuba8 tubulin, alpha 8 -2.91 Tubb2a-ps2 tubulin, beta 2a, pseudogene 2 -2.36 -2.69 -2.18 -2.88 Tubb2b tubulin, beta 2B class IIB -2.06 Txnip thioredoxin interacting protein 2.08 Upk1b uroplakin 1B 2.40 Vash2 vasohibin 2 -2.28 Vax2os ventral anterior homeobox 2, -5.21 opposite strand Vmn2r1 vomeronasal 2, receptor 1 -3.48 Vstm2b V-set and transmembrane 8.31 17.17 domain containing 2B Wfdc3 WAP four-disulfide core domain -2.34 3 Wisp3 WNT1 inducible signaling -2.88 pathway protein 3 Wnt10a wingless-type MMTV integration site family, member 20.19 20.49 5.93 10A Zan zonadhesin -3.97

191

References

1.Ahmed N, Thornalley PJ, Dawczynski J, Franke S, Strobel J, Stein G, Haik GM. Methylglyoxal-derived hydroimidazolone advanced glycation end- products of human lens proteins. Invest Ophthalmol Vis Sci. 2013;44(12):5287-92. 2.Anderson ME, Meister A. Transport and direct utilization of gamma- glutamylcyst(e)ine for glutathione synthesis. Proc Natl Acad Sci USA. 1983;80(3): 707–11. 3.Andley UP. Crystallins in the eye: Function and pathology. Prog Retin Eye Res. 2007;26(1):78-98. 4.Augusteyn, RC. On the growth and internal structure of the human lens. Exp Eye Res. 2010;90(6):643–54. 5.Bachhawat AK, Thakur A, Kaur J, Zulkifli M. Glutathione transporters. Biochim Biophys Acta. 2013;1830(5):3154-64. 6.Barishak YR. Embryology of the eye and its adnexae. Dev Ophthalmol. 1992;24:1-142. 7.Barson JR, Morganstern I, Leibowitz SF. Galanin and consummatory behavior: special relationship with dietary fat, alcohol and circulating lipids. EXS. 2010;102:87-111. 8.Bassnett S, Shi Y, Vrensen GFJM. Biological glass: structural determinants of eye lens transparency. Phil. Trans. R. Soc. B. 2011;366: 1250-64. 9.Beyer EC, Berthoud VM. Connexin hemichannels in the lens. Frontiers Physiol. 2014;5:20. 10.Bhat SP. The ocular lens epithelium. Biosci Rep. 2001;21(4):537-63. 11.Black WJ, Stagos D, Marchitti SA, et al. Human Aldehyde Dehydrogenase Genes: Alternatively-Spliced Transcriptional Variants and Their Suggested Nomenclature. Pharmaco Genomics. 2009;19(11):893-902. 12.Booty LM, King MS, Thangaratnarajah C, Majd H, James AM, Kunji ER, Murphy MP. The mitochondrial dicarboxylate and 2-oxoglutarate carriers do not transport glutathione. FEBS Lett. 2015;589(5):621-8.

192

13.Bourbouloux A, Shahi P, Chakladar A, Delrot S, Bachhawat AK. Hgt1p, a high affinity glutathione transporter from the yeast Saccharomyces cerevisiae. J Biol Chem. 2000;275(18):13259-65. 14.Cagnone GL, Sirard MA. Transcriptomic signature to oxidative stress exposure at the time of embryonic genome activation in bovine blastocysts. Mol Reprod Dev. 2013;80(4):297-314. 15.Calvin HI, Zhu G, Wu J, Banerjee U, Fu JS. Progression of mouse buthionine sulfoximine cataracts in vitro is inhibited by thiols or ascorbate. Exp Eye Res. 1997;65(3):341-7. 16.Cao CY, Mok SW, Cheng VW, Tsui SK. The FHL2 regulation in the transcriptional circuitry of human cancers. Gene. 2015;572(1):1-7. 17.Carper D, John M, Chen Z, Subramanian S, Wang R, Ma W, Spector A. Gene expression analysis of an H(2)O(2)-resistant lens epithelial cell line. Free Radic Biol Med. 2001;31(1):90-7. 18.Carper DA, Sun JK, Iwata T, Zigler JS Jr, Ibaraki N, Lin LR, Reddy V. Oxidative stress induces differential gene expression in a human lens epithelial cell line. Invest Ophthalmol Vis Sci. 1999;40(2):400-6. 19.Cattelotte J, André P, Ouellet M, Bourasset F, Scherrmann JM, Cisternino S. In situ mouse carotid perfusion model: glucose and cholesterol transport in the eye and brain. J Cereb Blood Flow Metab. 2008;28(8):1449-59. 20.Chen WJ, Boehlert CC, Rider K, Armstrong RN. Synthesis and characterization of the oxygen and desthio analogues of glutathione as dead-end inhibitors of glutathione S-transferase. Biochem Biophys Res Commun. 1985;128(1):233-40. 21.Chen X, Mori T, Guo Q, Hu C, Ohsaki Y, Yoneki Y, Zhu W, Jiang Y, Endo S, Nakayama K, Ogawa S, Nakayama M, Miyata T, Ito S. Carbonyl stress induces hypertension and cardio-renal vascular injury in Dahl salt- sensitive rats. Hypertens Res. 2013;36(4):361-7. 22.Chen Y, Koppaka V, Thompson DC, Vasiliou V. ALDH1A1: From Lens and Corneal Crystallin to Stem Cell Marker. Exp Eye Res. 2012; 102C: 105–6.

193

23.Cheng R, Feng Q, Ortwerth BJ. LC-MS display of the total modified amino acids in cataract lens proteins and in lens proteins glycated by ascorbic acid in vitro. Biochim Biophys Acta. 2006;1762(5):533-43. 24.Cherfan GM, Michels RG, de Bustros S, Enger C, Glaser BM. Nuclear sclerotic cataract after vitrectomy for idiopathic epiretinal membranes causing macular pucker. Am J Ophthalmol. 1991;111(4):434-8. 25.Chien CH, Huang T, Schachar RA. Analysis of human crystalline lens accommodation. J Biomech. 2006;39(4):672-80. 26.Choudhary S, Xiao T, Vergara LA, Srivastava S, Nees D, Piatigorsky J, Ansari NH. Role of aldehyde dehydrogenase isozymes in the defense of rat lens and human lens epithelial cells against oxidative stress. Invest Ophthalmol Vis Sci. 2005;46(1):259-67. 27.Danysh BP, Duncan MK. The Lens Capsule. Exp Eye Res. 2009;88(2):151-64. 28.Davies P, Moualla D, Brown DR. Alpha-synuclein is a cellular ferrireductase. PLoS One. 2011;6(1):e15814. 29.Dawczynski J, Blum M, Winnefeld K, Strobel J. Increased content of zinc and iron in human cataractous lenses. Biol Trace Elem Res. 2002;90(1- 3):15-23. 30.Deepmala, Slattery J, Kumar N, Delhey L, Berk M, Dean O, Spielholz C, Frye R. Clinical trials of N-acetylcysteine in psychiatry and neurology: A systematic review. Neurosci Biobehav Rev. 2015;55:294-321. 31.Dheeman DS, Packianathan C, Pillai JK, Rosen BP. Pathway of human AS3MT arsenic methylation. Chem Res Toxicol. 2014;27(11):1979-89. 32.Ding HQ, Zhou BJ, Liu L, Cheng S. Oxidative stress and metallothionein expression in the liver of rats with severe thermal injury. Burns. 2002;28(3):215-21. 33.Donaldson P, Kistler J, Mathias RT. Molecular solutions to mammalian lens transparency. News Physiol Sci. 2001;16:118-23. 34.Fan X, Liu X, Hao S, Wang B, Robinson ML, and Monnier VM. The LEGSKO Mouse: A Mouse Model of Age-Related Nuclear Cataract

194

Based on Genetic Suppression of Lens Glutathione Synthesis. PLoS One. 2012;7(11): e50832. 35.Fan X, Reneker LW, Obrenovich ME, Strauch C, Cheng R, Jarvis SM, Ortwerth BJ, Monnier VM. Vitamin C mediates chemical aging of lens crystallins by the Maillard reaction in a humanized mouse model. Proc Natl Acad Sci U S A. 2006;103(45):16912-7. 36.Fan X, Xiaoqin L, Potts B, Strauch CM, Nemet I, Monnier VM. Topical application of L-arginine blocks advanced glycation by ascorbic acid in the lens of hSVCT2 transgenic mice. Mol Vis. 2011;17:2221-7. 37.Fan X, Zhou S, Wang B, Hom G, Guo M, Li B, Yang J, Vaysburg D, Monnier VM. Evidence of Highly Conserved β-Crystallin Disulfidome that Can be Mimicked by In Vitro Oxidation in Age-related Human Cataract and Glutathione Depleted Mouse Lens. Mol Cell Proteomics. 2015;14(12):3211-23. 38.Fekete I, Griffith OW, Schlageter KE, Bigner DD, Friedman HS, Groothuis DR. Rate of buthionine sulfoximine entry into brain and xenotransplanted human gliomas. Cancer Res. 1990;50(4):1251-6. 39.Fletcher AE. Free radicals, antioxidants and eye diseases: evidence from epidemiological studies on cataract and age-related macular degeneration. Ophthalmic Res. 2010;44(3):191-8. 40.Floud S, Kuper H, Reeves GK, Beral V, Green J. Risk Factors for Cataracts Treated Surgically in Postmenopausal Women. Ophthalmology. 2016;123(8):1704-10. 41.Foster A. VISION 2020: THE CATARACT CHALLENGE. Community Eye Health. 2000;13(34):17-19. 42.Frederikse PH, Kasinathan C. Lens GABA receptors are a target of GABA-related agonists that mitigate experimental myopia. Med Hypotheses. 2015;84(6):589-92. 43.Gao J, Sun X, Moore LC, White TW, Brink PR, Mathias RT. Lens intracellular hydrostatic pressure is generated by the circulation of sodium

195

and modulated by gap junction coupling. J Gen Physiol. 2011;137(6):507- 20. 44.Gao M, Monian P, Pan Q, Zhang Wei, Xiang J, Jiang X. Ferroptosis is an autophagic cell death process. Cell Res. 2016;26:1021–32. 45.Garcia A, Eljack ND, Sani MA, Separovic F, Rasmussen HH, Kopec W, Khandelia H, Cornelius F, Clarke RJ. Membrane accessibility of glutathione. Biochim Biophys Acta. 2015;1848(10A):2430-6. 46.Garner MH, Spector A. Selective oxidation of cysteine and methionine in normal and senile cataractous lenses. PNAS. 1980;77(3):1274-7. 47.Getz GS, Reardon CA. Apoprotein E as a lipid transport and signaling protein in the blood, liver, and artery wall. J Lipid Res. 2009;50:S156-61. 48.Giblin FJ. Glutathione: a vital lens antioxidant. J Ocular Pharma Thera. 2000;16(2): 121-35. 49.Griffith OW, Bridges RJ, Meister A. Formation of gamma- glutamycyst(e)ine in vivo is catalyzed by gamma-glutamyl transpeptidase. Proc Natl Acad Sci USA. 1981;78(5):2777-81. 50.Hanson SRA, Hasan A, Smith DL, Smith JB. The Major in vivo Modifications of the Human Water-insoluble Lens Crystallins are Disulfide Bonds, Deamidation, Methionine Oxidation and Backbone Cleavage. Exp. Eye Res. 2000;71(2):195–207. 51.Harding JJ. Free and protein-bound glutathione in normal and cataractous human lenses. Biochem J. 1970;117(5): 957–60. 52.Hashim Z, Zarina S. Advanced glycation end products in diabetic and non- diabetic human subjects suffering from cataract. Age. 2011;33(3):377-84. 53.Hawse JR, Hejtmancik JF, Horwitz J, and Kantorow M. Identification and functional clustering of global gene expression differences between age- related cataract and clear human lenses and aged human lenses. Exp Eye Res. 2004;79(6): 935–40. 54.Hayashi K, Hayashi H, Nakao F, Hayashi F. Correlation between posterior capsule opacification and visual function before and after Neodymium: YAG laser posterior capsulotomy. Am J Ophthalmol. 2003;136(4):720-6.

196

55.Hayes JD, Flanagan JU, Jowsey IR. Glutathione transferases. Annu Rev Pharmacol Toxicol. 2005;45:51-88. 56.He X, Ma Q. Redox regulation by nuclear factor erythroid 2-related factor 2: gatekeeping for the basal and diabetes-induced expression of thioredoxin-interacting protein. Mol Pharmacol. 2012;82(5):887-97. 57.Hejtmancik JF, Shiels A. Overview of the Lens. Prog Mol Biol Transl Sci. 2015;134:119-27. 58.Hidalgo J, Garvey JS, Armario A. On the metallothionein, glutathione and cysteine relationship in rat liver. J Pharmacol Exp Ther. 1990;255(2):554- 64. 59.Hildebrandt W, Sauer R, Bonaterra G, Dugi KA, Edler L, Kinscherf R. Oral N-acetylcysteine reduces plasma homocysteine concentrations regardless of lipid or smoking status. Am J Clin Nutr. 2015;102(5):1014- 24. 60.Hoang TV, Kumar PK, Sutharzan S, Tsonis PA, Liang C, Robinson ML. Comparative transcriptome analysis of epithelial and fiber cells in newborn mouse lenses with RNA sequencing. Mol Vis. 2014;20:1491- 517. 61.Holekamp NM, Shui Y, Beebe DC. Vitrectomy surgery increases oxygen exposure to the lens: A possible mechanism for nuclear cataract formation. Am J Ophthalmol. 2005;139(2):302-10. 62.Huang CS, Chang LS, Anderson ME, Meister A. Catalytic and regulatory properties of the heavy subunit of rat kidney gamma-glutamylcysteine synthetase. J Biol Chem. 1993;268(26):19675-80. 63.Huang SP, Palla S, Ruzycki P, et al. Aldo-Keto Reductases in the Eye. J Ophthalmol 2010;2010:521204. 64.Ishizaki Y, Jacobson MD, Raff MC. A role for caspases in lens fiber differentiation. J Cell Biol. 1998;140(1):153-8. 65.Jaiswal AK. Nrf2 signaling in coordinated activation of antioxidant gene expression. Free Radic Biol Med. 2004;36(10):1199-207.

197

66.Järvinen TAH, Prince S. Decorin: A Growth Factor Antagonist for Tumor Growth Inhibition. BioMed Res Internat. 2015;2015:654765. 67.Javitt JC, Wang F, and West SK. Blindness Due to Cataract: Epidemiology and Prevention. Annual Review of Public Health. 1996;17:159-77. 68.Jensen GL, Meister A. Radioprotection of human lymphoid cells by exogenously supplied glutathione is mediated by gamma-glutamyl transpeptidase. Proc Natl Acad Sci USA. 1983;80(15):4714-7. 69.Jensen MA, Wilkinson JE, Krainer AR. Splicing factor SRSF6 promotes hyperplasia of sensitized skin. Nat Struct Mol Biol. 2014;21(2):189-97. 70.Jernigan HM Jr. Urea formation in rat, bovine, and human lens. Exp Eye Res. 1983;37(6):551-8. 71.Kalinina EV, Chernov NN, Novichkova MD. Role of glutathione, glutathione transferase, and glutaredoxin in regulation of redox-dependent processes. Biochemistry (Mosc). 2014;79(13):1562-83. 72.Kannan R, Stolz A, Ji Q, Prasad PD, Ganapathy V. Vitamin C transport in human lens epithelial cells: evidence for the presence of SVCT2. Exp Eye Res. 2001;73(2):159-65. 73.Kannan R, Yi JR, Zlokovic BV, Kaplowitz N. Molecular characterization of a reduced glutathione transporter in the lens. Invest Ophthalmol Vis Sci. 1995;36(9):1785-92. 74.Kasthurirangan S, Markwell EL, Atchison DA, Pope JM. In vivo study of changes in refractive index distribution in the human crystalline lens with age and accommodation. Invest Ophthalmol Vis Sci. 2008;49(6):2531-40. 75.Kern HL, Ho CK, Ostrove SA. Comparison of transport at the anterior and posterior surfaces of the calf lens. Exp Eye Res. 1977;24(6):559-70. 76.Kiel JW, Hollingsworth M, Rao R, Chen M, Reitsamer HA. Ciliary Blood Flow and Aqueous Humor Production. Prog Retin Eye Res. 2011;30(1): 1–17. 77.Kiel JW. 2010. The Ocular Circulation. San Rafael (CA): Morgan & Claypool Life Sciences; Chapter 2, Anatomy.

198

78.Kleiman NJ, Quinn AM, Fields KG, Slavkovich V, Graziano JH. Arsenite accumulation in the mouse eye. J Toxicol Environ Health A. 2016;79(8):339-41. 79.Kodama T, Takemoto L. Characterization of disulfide-linked crystallins associated with human cataractous lens membranes. Invest. Ophthalmol. Vis. Sci. 1988;29(1):1145-149. 80.Lagace TA. PCSK9 and LDLR degradation: regulatory mechanisms in circulation and in cells. Curr Opin Lipidol. 2014;25(5):387-93. 81.Landgren H, Blixt A, Carlsson P. Persistent FoxE3 expression blocks cytoskeletal remodeling and organelle degradation during lens fiber differentiation. Invest Ophthalmol Vis Sci. 2008;49(10):4269-77. 82.Lash LH, Putt DA, Xu F, Matherly LH. Role of Rat Organic Anion Transporter 3 (Oat3) in the Renal Basolateral Transport of Glutathione. Chemico-biological interactions. 2007;170(2):124-34. 83.Lash LH. Mitochondrial glutathione transport: physiological, pathological and toxicological implications. Chem Biol Interact. 2006;163(1-2):54-67. 84.Lassen N, Bateman JB, Estey T, Kuszak JR, Nees DW, Piatigorsky J, Duester G, Day BJ, Huang J, Hines LM, Vasiliou V. Multiple and additive functions of ALDH3A1 and ALDH1A1: cataract phenotype and ocular oxidative damage in Aldh3a1(-/-)/Aldh1a1(-/-) knock-out mice. J Biol Chem. 2007;282(35):25668-76. 85.Levy EJ, Anderson ME, Meister A. Transport of glutathione diethyl ester into human cells. Proc Natl Acad Sci U S A. 1993;90(19):9171-5. 86.Li B, Li L, Donaldson PJ, Lim JC. Dynamic regulation of GSH synthesis and uptake pathways in the rat lens epithelium. Exp Eye Res. 2010;90(2):300-7. 87.Li J, Xia Z, Ding J. Thioredoxin-like domain of human kappa class glutathione transferase reveals sequence homology and structure similarity to the theta class enzyme. Protein Sci. 2005;14(9):2361-9.

199

88.Li L, Lee TK, Ballatori N. Functional re-evaluation of the putative glutathione transporters, RcGshT and RsGshT. Yale J Biol Med. 1997;70(4): 301–10. 89.Liao RF, Ye MJ, Liu CY, Ye DQ. An Updated Meta-Analysis: Risk Conferred by Glutathione S-Transferases (GSTM1 and GSTT1) Polymorphisms to Age-Related Cataract. J Ophthalmol. 2015;2015:103950. 90.Lillig CH, Berndt C, Holmgren A. Glutaredoxin systems. Biochim Biophys Acta. 2008;1780(11):1304-17. 91.Lindblad BE, Håkansson N, Wolk A. Smoking cessation and the risk of cataract: a prospective cohort study of cataract extraction among men. JAMA Ophthalmol. 2014;132(3):253-7. 92.Liu M, Chen H, Wei L, Hu D, Dong K, Jia W, Dong LQ, Liu F. Endoplasmic reticulum (ER) localization is critical for DsbA-L protein to suppress ER stress and adiponectin down-regulation in adipocytes. J Biol Chem. 2015;290(16):10143-8. 93.Lovicu FJ, Shin EH, McAvoy JW. Fibrosis in the lens. Sprouty regulation of TGFβ-signaling prevents lens EMT leading to cataract. Exp Eye Res. 2016;142:92-101. 94.Lu SC. Glutathione synthesis. Biochim Biophys Acta. 2013;1830(5):3143- 53. 95.Lu SC. REGULATION OF GLUTATHIONE SYNTHESIS. Mol Asp Med. 2009;30(1-2):42-59. 96.Ma Q. Role of Nrf2 in Oxidative Stress and Toxicity. An Rev Pharmacol Toxicol. 2013;53:401-26. 97.Makley LN, McMenimen KA, DeVree BT, Goldman JW, McGlasson BN, Rajagopal P, Dunyak BM, McQuade TJ, Thompson AD, Sunahara R, Klevit RE, Andley UP, Gestwicki JE. Pharmacological chaperone for α- crystallin partially restores transparency in cataract models. Science. 2015;350(6261):674-7.

200

98.Mamatha BS, Nidhi B, Padmaprabhu CA, Pallavi P, Vallikannan B. Risk Factors for Nuclear and Cortical Cataracts: A Hospital Based Study. J Ophthalmic Vis Res. 2015;10(3):243-9. 99.Mansergh FC, Wride MA, Walker VE, Adams S, Hunter SM, Evans MJ. Gene expression changes during cataract progression in Sparc null mice: differential regulation of mouse globins in the lens. Mol Vis. 2004;10:490- 511. 100.Mathias RT, Rae JL, Baldo GJ. Physiological properties of the normal lens. Phys Rev. 1997;77(1):21-50. 101.Mathias RT, White TW, Gong X. Lens gap junctions in growth, differentiation, and homeostasis. Physiol Rev. 2010;90(1):179-206. 102.McAvoy JW, Chamberlain CG, de Iongh RU, Hales AM, Lovicu FJ.. Lens development. Eye (Lond). 1997;13 (Pt 3b):425-37. 103.McMahon M, Thomas N, Itoh K, Yamamoto M, Hayes JD. Redox- regulated turnover of Nrf2 is determined by at least two separate protein domains, the redox-sensitive Neh2 degron and the redox-insensitive Neh6 degron. J Biol Chem. 2004;279(30):31556-67. 104.Medici D, Nawshad A. Type I collagen promotes epithelial-mesenchymal transition through ILK-dependent activation of NF-kappaB and LEF-1. Matrix Biol. 2010;29(3):161-5. 105.Medvedovic M, Tomlinson CR, Call MK, Grogg M, Tsonis PA. Gene expression and discovery during lens regeneration in mouse: regulation of epithelial to mesenchymal transition and lens differentiation. Mol Vis. 2006;12:422-40. 106.Merriman-Smith R, Donaldson P, Kistler J. Differential expression of facilitative glucose transporters GLUT1 and GLUT3 in the lens. Invest Ophthalmol Vis Sci. 1999;40(13):3224-30. 107.Milazzo S. Pathogenesis of cataract after vitrectomy. J Fr Ophtalmol. 2014;37(3):243-4. 108.Miyata T, Ishikawa N, van Ypersele de Strihou C. Carbonyl stress and diabetic complications. Clin Chem Lab Med. 2003;41(9):1150-8.

201

109.Molotkov A, Molotkova N, Duester G. Retinoic acid guides eye morphogenetic movements via paracrine signaling but is unnecessary for retinal dorsoventral patterning. Development. 2006;133(10):1901-10. 110.Morel F, Rauch C, Petit E, Piton A, Theret N, Coles B, Guillouzo A. Gene and protein characterization of the human glutathione S-transferase kappa and evidence for a peroxisomal localization. J Biol Chem. 2004;279(16):16246-53. 111.Nagaraj RH, Linetsky M, Stitt AW. The pathogenic role of Maillard reaction in the aging eye. Amino Acids. 2012;42(4):1205-20. 112.Nebert DW, Vasiliou V. Analysis of the glutathione S-transferase (GST) gene family. Hum Genomics. 2004;1(6):460-4. 113.Norose K, Clark JI, Syed NA, Basu A, Heber-Katz E, Sage EH, Howe CC. SPARC deficiency leads to early-onset cataractogenesis. Invest Ophthalmol Vis Sci. 1998;39(13):2674-80. 114.Nozal MJ, Bernal JL, Toribio L, Marinero P, Moral O, Manzanas L, Rodriguez E. Determination of glutathione, cysteine and N-acetylcysteine in rabbit eye tissues using high-performance liquid chromatography and post-column derivatization with 5,5'-dithiobis(2-nitrobenzoic acid). J Chromatogr A. 1997;778(1-2):347-53. 115.Obrenovich ME, Fan X, Satake M, Jarvis SM, Reneker L, Reddan JR, Monnier VM. Relative suppression of the sodium-dependent Vitamin C transport in mouse versus human lens epithelial cells. Mol Cell Biochem. 2006;293(1-2):53-62. 116.Ochiai H.Transporters of glycine, cystine, glutamate and glutamine in canine lens epithelial cells. Acta Ophthalmologica 2013;91:0. 117.Ohrloff C, Hockwin O. Lens metabolism and aging: enzyme activities and enzyme alterations in lenses of different species during the process of aging. J Gerontol. 1983;38(3):271-7. 118.Park SJ, Lee JH, Kang SW, Hyon JY, Park KH. Cataract and Cataract Surgery: Nationwide Prevalence and Clinical Determinants. J Korean Med Sci. 2016;31(6):963-71.

202

119.Pascolini D, Mariotti SPM. 2010. Global estimates of visual impairment: 2010. British J. Ophthalmol. 120.Pershing S, Kumar A. Phacoemulsification versus extracapsular cataract extraction: where do we stand? Curr Opin Ophthalmol. 2011;22(1):37-42. 121.Petrash JM. Aging and age-related diseases of the ocular lens and vitreous body. Invest Ophthalmol Vis Sci. 2013;54: 54-59. 122.Qiao H, Hisatomi T, Sonoda KH, Kura S, Sassa Y, Kinoshita S, Nakamura T, Sakamoto T, Ishibashi T. The characterisation of hyalocytes: the origin, phenotype, and turnover. Br J Ophthalmol. 2005;89(4):513-7. 123.Rathbun WB, Schmidt AJ, Holleschau AM. Activity loss of glutathione synthesis enzymes associated with human subcapsular cataract. Invest Ophthalmol Vis Sci. 1993;34(6):2049-54. 124.Rathbun WB. Lenticular glutathione synthesis: rate-limiting factors in its regulation and decline. Curr Eye Res. 1984;3(1):101-8. 125.Reddy VN, Giblin FJ, Lin LR, Dang L, Unakar NJ, Musch DC, Boyle DL, Takemoto LJ, Ho YS, Knoernschild T, Juenemann A, Lütjen-Drecoll E. Glutathione peroxidase-1 deficiency leads to increased nuclear light scattering, membrane damage, and cataract formation in gene-knockout mice. Invest Ophthalmol Vis Sci. 2001;42(13):3247-55. 126.Reddy VN, Unakar NJ. Localization of gamma-glutamyl transpeptidase in rabbit lens, ciliary process and cornea. Exp Eye Res 1973;17(5):405–8. 127.Reddy VN. Glutathione and its function in the lens—An overview. Exp Eye Res. 1990;50(6):771–8. 128.Richman PG, Meister A. Regulation of gamma-glutamyl-cysteine synthetase by nonallosteric feedback inhibition by glutathione. J Biol Chem. 1975;250(4):1422-6. 129.Rodrigues WF, Miguel CB, Napimoga MH, Oliveira CJ, Lazo-Chica J. Establishing standards for studying renal function in mice through measurements of body size-adjusted creatinine and urea levels. Biomed Res Int. 2014;2014:872827.

203

130.Rose RC, Gogia R, Richer SP. Properties of electrochemically active components in mammalian vitreous humor. Exp Eye Res. 1997;64(5):807- 12. 131.Ruotolo R, Grassi F, Percudani R, Rivetti C, Martorana D, Maraini G, Ottonello S. Gene expression profiling in human age-related nuclear cataract. Mol Vis. 2003;9:538-48. 132.Rushworth GF, Megson IL. Existing and potential therapeutic uses for N- acetylcysteine: the need for conversion to intracellular glutathione for antioxidant benefits. Pharmacol Ther. 2014;141(2):150-9. 133.Sanghani SP, Sanghani PC, Schiel MA, Bosron WF. Human carboxylesterases: an update on CES1, CES2 and CES3. Protein Pept Lett. 2009;16(10):1207-14. 134.Sekhar RV, Patel SG, Guthikonda AP, Reid M, Balasubramanyam A, Taffet GE, Jahoor F. Deficient synthesis of glutathione underlies oxidative stress in aging and can be corrected by dietary cysteine and glycine supplementation. Am J Clin Nutr. 2011;94(3):847-53. 135.Shahidullah M, Mandal A, Delamere NA. TRPV4 in porcine lens epithelium regulates hemichannel-mediated ATP release and Na-K- ATPase activity. Ameri J Physiol - Cell Physiol. 2012;302(12):C1751-61. 136.Sharma KK, Santhoshkumar P. Lens Aging: Effects of Crystallins. Biochim Biophys Acta. 2009;1790(10):1095-108. 137.Sharma YR, Vajpayee RB, Bhatnagar R, Mohan M, Azad RV, Kumar M, Nath R. Topical glutathione therapy in senile cataracts. Cataract-III. Indian J Ophthalmol. 1989;37(3):121-6. 138.Shiels A, Hejtmancik J. Genetics of human cataract. Clinical Genetics. 2013;84(2):120-7. 139.Singhal SS, Singh SP, Singhal P, Horne D, Singhal J, Awasthi S. Antioxidant role of glutathione S-transferases: 4-Hydroxynonenal, a key molecule in stress-mediated signaling. Toxicol Appl Pharmacol. 2015;289(3):361-70.

204

140.Skeie JM, Tsang SH, Mahajan VB. Evisceration of mouse vitreous and retina for proteomic analyses. J Visualized Exp. 2011;(50). 141.Slavi N, Rubinos C, Li L, Sellitto C, White TW, Mathias R, Srinivas M. Connexin 46 (Cx46) Gap Junctions Provide a Pathway for the Delivery of Glutathione to the Lens Nucleus. JBC. 2014;289(47):32694–702. 142.Smith TK, Ikeda Y, Fujii Y, Taniguchi N, Meister A. Different sites of acivicin binding and inactivation of gamma-glutamyl transpeptidases. Proc Natl Acad Sci U S A. 1995;92(6):2360–4. 143.Snow A, Shieh B, Chang KC, Pal A, Lenhart P, Ammar D, Ruzycki P, Palla S, Reddy GB, Petrash JM. Aldose reductase expression as a risk factor for cataract. Chem Biol Interact. 2015;234:247-53. 144.Song E, Sun H, Xu Y, Ma Y, Zhu H, Pan C-W. Age-Related Cataract, Cataract Surgery and Subsequent Mortality: A Systematic Review and Meta-Analysis. PLoS ONE. 2014; 9(11):e112054. 145.Steinherz R, Mårtensson J, Wellner D, Meister A. Transport into brain of buthionine sulfoximine, an inhibitor of glutathione synthesis, is facilitated by esterification and administration of dimethylsulfoxide. Brain Res. 1990;518(1-2):115-9. 146.Stewart-DeHaan PJ, Dzialoszynski T, Trevithick JR. Modelling cortical cataractogenesis XXIV: uptake by the lens of glutathione injected into the rat. Mol Vis. 1999;5:37. 147.Subczynski WK, Raguz M, Widomska J, Mainali L, Konovalov A. Functions of cholesterol and the cholesterol bilayer domain specific to the fiber-cell plasma membrane of the eye lens. J Membr Biol. 2012;245(1):51-68. 148.Sweeney MH, Truscott RJ. An impediment to glutathione diffusion in older normal human lenses: a possible precondition for nuclear cataract. Exp Eye Res. 1998; 67:587-95. 149.T Micelli-Ferrari, G Vendemiale, I Grattagliano, F Boscia, L Arnese, E Altomare, and L Cardia. Role of lipid peroxidation in the pathogenesis of myopic and senile cataract. Br J Ophthalmol. 1996; 80(9):840–3.

205

150.Tamara SO, Yanshole LV, Yanshole VV, Fursova AZ, Stepakov DA, Novoselov VP, Tsentalovich YP. Spatial distribution of metabolites in the human lens. Exp Eye Res. 2016;143:68-74. 151.Termén S, Tan EJ, Heldin CH, Moustakas A. p53 regulates epithelial- mesenchymal transition induced by transforming growth factor β. J Cell Physiol. 2013;228(4):801-13. 152.Umapathy A, Donaldson P, Lim J. Antioxidant delivery pathways in the anterior eye. Biomed Res Int. 2013;2013:207250. 153.Van Effenterre G, Ameline B, Campinchi F, Quesnot S, Le Mer Y, Haut J. Is vitrectomy cataractogenic? Study of changes of the crystalline lens after surgery of retinal detachment. J Fr Ophtalmol. 1992;15(8-9):449-54. 154.van Kuijk FJ. Effects of ultraviolet light on the eye: role of protective glasses. Environ Health Perspect. 1991; 96: 177–84. 155.Veltman JC, Podval J, Mattern J, Hall KL, Lambert RJ, Edelhauser HF. The disposition and bioavailability of 35S-GSH from 35S-GSSG in BSS PLUS in rabbit ocular tissues. J Ocul Pharmacol Ther. 2004;20(3):256-68. 156.Viarengo A, Burlando B, Ceratto N, Panfoli I. Antioxidant role of metallothioneins: a comparative overview. Cell Mol Biol. 2000;46(2):407- 17. 157.Whitson JA, Sell DR, Goodman MC, Monnier VM, Fan X. Evidence of Dual Mechanisms of Glutathione Uptake in the Rodent Lens: A Novel Role for Vitreous Humor in Lens Glutathione Homeostasis. Invest Ophthalmol Vis Sci. 2016;57(8):3914-25. 158.Winterbourn CC. Toxicity of iron and hydrogen peroxide: the Fenton reaction. Toxicol Lett. 1995;82-83:969-74. 159.Wise SJ, Nathoo NA, Etminan M, Mikelberg FS, Mancini GB. Statin use and risk for cataract: a nested case-control study of 2 populations in Canada and the United States. Can J Cardiol. 2014;30(12):1613-9. 160.Wisnewski AV, Liu Q, Liu J, Redlich CA. Glutathione protects human airway proteins and epithelial cells from isocyanates. Clin Exp Allergy. 2005;35(3):352-7.

206

161.Wormstone IM, Wang L, Liu CS. Posterior capsule opacification. Exp Eye Res. 2009;88(2):257-69. 162.Wu H, Yu Y, David L, Ho YS, Lou MF. 2014. Glutaredoxin 2 (Grx2) gene deletion induces early onset of age-dependent cataracts in mice. J Biol Chem 289(52):36125-39. 163.Wu KC, Cui JY, Klaassen CD. Effect of Graded Nrf2 Activation on Phase-I and -II Drug Metabolizing Enzymes and Transporters in Mouse Liver. PLoS ONE. 2012;7(7):e39006. 164.Yan H, Harding JJ. Carnosine protects against the inactivation of esterase induced by glycation and a steroid. Biochim Biophys Acta. 2005;1741(1- 2):120-6. 165.Yan H, Zhang J, Harding JJ. Identification of the preferentially targeted proteins by carbamylation during whole lens incubation by using radio- labelled potassium cyanate and mass spectrometry. Int J Ophthalmol. 2010;3(2):104–11. 166.Yang B, Verkman AS. Analysis of double knockout mice lacking aquaporin-1 and urea transporter UT-B. Evidence for UT-B-facilitated water transport in erythrocytes. J Biol Chem. 2002;277(39):36782-6. 167.Yoshida A, Rzhetsky A, Hsu LC, Chang C. Human aldehyde dehydrogenase gene family. European J Biochem. 1998;251(3):549-57. 168.Yu H, Guo P, Xie X, Wang Y, Chen G. Ferroptosis, a new form of cell death, and its relationships with tumourous diseases. J Cell Mol Med. 2016. 169.Yu JM, Yang DQ, Wang H, Xu J, Gao Q, Hu LW, Wang F, Wang Y, Yan QC, Zhang JS, Liu Y. Prevalence and risk factors of lens opacities in rural populations living at two different altitudes in China. Int J Ophthalmol. 2016;9(4):610-6. 170.Zhang J, Yan H, Harding JJ, Liu ZX, Wang X, Ruan YS. Identification of the primary targets of carbamylation in bovine lens proteins by mass spectrometry. Curr Eye Res. 2008;33(11):963-76.

207

171.Zhang S, Chai FY, Yan H, Guo Y, Harding JJ. Effects of N- acetylcysteine and glutathione ethyl ester drops on streptozotocin-induced diabetic cataract in rats. Mol Vis. 2008;14:862-70. 172.Zhao L, Chen XJ, Zhu J, Xi YB, Yang X, Hu LD, Ouyang H, Patel SH, Jin X, Lin D, Wu F, Flagg K, Cai H, Li G, Cao G, Lin Y, Chen D, Wen C, Chung C, Wang Y, Qiu A, Yeh E, Wang W, Hu X, Grob S, Abagyan R, Su Z, Tjondro HC, Zhao XJ, Luo H, Hou R, Perry JJ, Gao W, Kozak I, Granet D, Li Y, Sun X, Wang J, Zhang L, Liu Y, Yan YB, Zhang K. Lanosterol reverses protein aggregation in cataracts. Nature. 2015;526(7574):595. 173.Zlokovic BV, Mackic JB, McComb JG, Kaplowitz N, Weiss MH, Kannan R. Blood-to-lens transport of reduced glutathione in an in situ perfused guinea-pig eye. Exp Eye Res. 1994;59(4):487-96.