Caspase-1-Dependent Inflammatory Signaling in Retinal Müller Cells During the Development of Diabetic Retinopathy

Home , Adenosinergic

CASPASE-1-DEPENDENT INFLAMMATORY SIGNALING IN RETINAL MüLLER CELLS DURING THE DEVELOPMENT OF DIABETIC RETINOPATHY

KATHERINE EILEEN TRUEBLOOD

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Dissertation Advisor: Dr. George R. Dubyak

Department of Physiology and Biophysics

CASE WESTERN RESERVE UNIVERSITY

January 2012

CASE WESTERN RESERVE UNIVERSITY

SCHOOL OF GRADUATE STUDIES

We hereby approve the thesis/dissertation of

Katherine Eileen Trueblood ______

Doctor of Philosophy candidate for the ______degree *.

Dr. Corey Smith (signed)______(chair of the committee)

Dr. George R. Dubyak ______

Dr. Thomas McCormick ______

Dr. Patrick Wintrode ______

Dr. Margaret Chandler ______

Dr. Thomas Nosek ______

10/21/11 (date) ______

*We also certify that written approval has been obtained for any proprietary material contained therein.

DEDICATION

I dedicate this work to my brother, Jonathan Vern Trueblood, as he embarks on his own graduate school career this year. There will be many days where you will want to give up and throw in the towel; but I promise there will be many more to come where you will be glad you didn’t. Rom. 4:18-21 Always “hope against hope!”

iii TABLE OF CONTENTS

DEDICATION…………………………………………………………………………iii

LIST OF TABLES…………………………………………………………………….ix

LIST OF FIGURES…………………………………………………………………...x

ACKNOWLEDGEMENTS………………………………………………………….xiv

LIST OF ABBREVIATIONS……………………………………………………….xvi

ABSTRACT...... 1

CHAPTER 1: INTRODUCTION...... 3

SECTION 1.1: DIABETES: A PUBLIC HEALTH CRISIS ...... 4

SECTION 1.2: BASIC RETINAL PHYSIOLOGY ...... 6

SECTION 1.3: DIABETIC RETINOPATHY: THE LEADING CAUSE OF

ACQUIRED BLINDNESS ...... 7

SECTION 1.4: RETINAL MÜLLER CELLS AND DIABETIC RETINOPATHY ...... 12

SECTION 1.5: CASPASE-1/INTERLEUKIN-1-BETA SIGNALING CASCADE...... 13

SECTION 1.5.1: CASPASE-1 ACTIVATION...... 13

SECTION 1.5.2: INTERLEUKIN-1-BETA ...... 20

SECTION 1.6: PURINERGIC SIGNALING ...... 22

SECTION 1.6.1: ATP/ADENOSINE METABOLISM...... 22

SECTION 1.6.2: P1 ADENOSINE RECEPTOR FAMILY ...... 23

SECTION 1.6.3: P2 RECEPTOR FAMILY ...... 25

SECTION 1.6.3: PURINERGIC SIGNALING WITHIN THE RETINA ...... 26

SECTION 1.7: SUMMARY...... 30

CHAPTER 2: MATERIALS AND METHODS...... 71

SECTION 2.1: MATERIALS...... 72

SECTION 2.2: METHODS...... 74

CHAPTER 3: PURINERGIC REGULATION OF HIGH GLUCOSE-

INDUCED CASPASE-1 ACTIVATION IN RAT RETINAL MÜLLER

CELLS (RMC-1) ...... 87

SECTION 3.1: INTRODUCTION...... 88

SECTION 3:2: RESULTS...... 90

Subsection 3.2.1: Treatment of rMC-1 cells with extracellular apyrase or

adenosine deaminase suppresses high glucose-induced caspase-1

activation…………………………………………………………………………...90

Subsection 3.2.2: Treatment of rMC-1 cells with exogenous ATP,

exogenous adenosine analog, or adenosine uptake inhibitors mimics

high glucose-induced caspase-1 activation ...... 90

Subsection 3.2.3: Treatment of rMC-1 cells with suramin or adenosine

receptor antagonists suppresses high glucose-induced caspase-1

activation………………………………………………………………………...... 93

Subsection 3.2.4: No significant role for P2X7 receptors in purinergic

regulation of high glucose-induced caspase-1 activation ...... 93

Subsection 3.2.5: Activation of caspase-1 by high glucose, NECA, or

ATP is mimicked by forskolin and correlated with increased gene

expression of caspase-1 and TXNIP………………………………………. .... 93

v Subsection 3.2.6: rMC-1 cells express Gq-coupled, Ca2+-mobilizing

P2Y2/P2Y4 and P2Y1 receptors and high glucose treatment increases

the efficacy of Ca2+ mobilization...... 97

SECTION 3.3: DISCUSSION ...... 99

SECTION 3.4: FIGURES ...... 107

CHAPTER 4: DIABETES-INDUCED CASPASE-1 ACTIVITY IS

REGULATED BY AN AUTOINFLAMMATORY FEED-BACK

MECHANISM ...... 129

SECTION 4.1: INTRODUCTION...... 130

SECTION 4:2: RESULTS...... 132

Subsection 4.2.1: Inhibition of diabetes-induced vascular damage

in caspase-1 knockout mice ...... 132

Subsection 4.2.2: Hyperglycemia-induced caspase-1 activity

pattern in the retinas of diabetic mice...... 132

Subsection 4.2.3: High glucose-induced caspase-1 activity and

IL-1β production patterns in rMC-1 ...... 134

Subsection 4.2.4: High glucose-induced caspase-1 activity and

IL-1β production patterns in hMC ...... 135

Subsection 4.2.5: Exogenous interleukin-1-β-induced caspase-1

activity in retinal Müller cells...... 135

Subsection 4.2.6: The effect of inhibition of early phase

caspase-1/IL-1β signaling on Müller cell viability ...... 136

vi Subsection 4.2.7: Glucose washout effect on late phase

caspase-1 activity...... 137

Subsection 4.2.8: Hyperglycemia-induced induction of

inflammasome component mRNA...... 137

SECTION 3.3: DISCUSSION ...... 140

SECTION 3.4: FIGURES ...... 145

CHAPTER 5: DISCUSSION AND FUTURE DIRECTIONS...... 166

SECTION 5.1: SUMMARY...... 167

SECTION 5.2: IS HIGH GLUCOSE-INDUCED CASPASE-1 ACTIVATION

MEDIATED BY INFLAMMASOME SIGNALING COMPLEXES IN RETINAL

MÜLLER CELLS? ...... 168

Subsection 5.2.1: Does P2X7R participate in the development of

diabetic retinopathy in vivo? ...... 170

Subsection 5.2.2: What are the sources of DAMPS that regulate

high glucose-induced caspase-1 activation?...... 171

Subsection 5.2.3: What mediates hyperglycemic induction of

NFκB activation?...... 172

SECTION 5.3: WHAT REGULATES AUTOCRINE ATP AND ADENOSINE

RELEASE AND/OR METABOLISM WITHIN THE CONTEXT OF

HYPERGLYCEMIA?...... 173

Subsection 5.3.1: ATP/Adenosine metabolism in the retina ...... 174

Subsection 5.3.2: Candidate conduits for autocrine ATP release ...... 175

vii SECTION 5.4: WHAT ROLE DO INCREASED LEVELS OF EITHER ACTIVE

CAMP OR PKA PLAY IN MODULATING HIGH GLUCOSE-INDUCED CASPASE-1

ACTIVATION WITHIN THE RETINA?...... 179

Subsection 5.4.1: Explore the mechanism of negative feedback by

ADP and UTP in regulating caspase-1 activation in this system...... 180

Subsection 5.4.2: Which specific adenosine receptor participates in

caspase-1 activation in retinal Muller cells?...... 181

SECTION 5.5: EXACTLY HOW DOES ELEVATED TXNIP EXPRESSION

PARTICIPATE IN MEDIATING CASPASE-1 ACTIVATION WITHIN THE RETINA? ...... 182

SECTION 5.6: CONCLUSION ...... 184

REFERENCES ...... 186

viii LIST OF TABLES

TABLE 1.1: SELECTIVE FUNCTIONS OF MÜLLER CELLS ...... 47

TABLE 1.2: CHARACTERISTICS OF REACTIVE MÜLLER CELLS IN VARIOUS

DISEASES ...... 49

TABLE 1.3: RELATIVE NUCLEOTIDE SENSITIVITIES OF P2 RECEPTORS ...... 61

TABLE 4.1: SUMMARY OF ANIMAL DATA ...... 147

ix LIST OF FIGURES

CHAPTER 1: INTRODUCTION

FIGURE 1.1: RETINAL STRUCTURE ...... 33

FIGURE 1.2: TRYPSIN DIGEST FROM DIABETIC HUMAN DEMONSTRATING

ACELLULAR CAPILLARY FORMATION...... 35

FIGURE 1.3: BIOCHEMICAL EVENTS DURING DIABETIC RETINOPATHY

(FROM RODENT STUDIES) ...... 37

FIGURE 1.4: MORPHOLOGICAL CHANGES DURING THE STAGES OF DIABETIC

RETINOPATHY ...... 39

FIGURE 1.5: RETINALGRAMS FROM NORMAL AND DIABETIC PATIENTS IN

VARYING STAGES OF DIABETIC RETINOPATHY ...... 41

FIGURE 1.6: NORMAL AND OBSTRUCTED VISION...... 43

FIGURE 1.7: RETINAL MÜLLER CELL MORPHOLOGY ...... 45

FIGURE 1.8: THE CASPASE CASCADE INVOLVED IN APOPTOSIS...... 51

FIGURE 1.9: ACTIVATION OF INFLAMMASOMES ...... 53

FIGURE 1.10: HIGH GLUCOSE CONDITIONS MODULATE TXNIP EXPRESSION ...... 55

FIGURE 1.11: MITOCHONDRIA: INTEGRATORS OF METABOLIC STRESS AND

ACTIVATORS OF THE NLRP3 INFLAMMASOME ...... 57

FIGURE 1.12: STRUCTURE AND FUNCTION OF P2X7R CHANNELS AND

INTRACELLULAR SIGNALING CASCADE ...... 59

FIGURE 1.13: SCHEMATIC ILLUSTRATION OF THE KEY COMPONENTS OF

PURINERGIC SIGNALING IN THE SUBRETINAL MICROENVIRONMENT...... 64

x FIGURE 1.14: SCHEME OF THE AUTOCRINE GLUTAMATERGIC-PURINERGIC

SIGNALING CASCADE INVOLVED IN THE VEGF-INDUCED INHIBITION OF

MÜLLER CELL SWELLING ...... 66

FIGURE 1.15: PROPOSED MODEL FOR HIGH GLUCOSE-INDUCED CASPASE-1

ACTIVATION IN RETINAL MÜLLER CELLS ...... 69

CHAPTER 3: PURINERGIC REGULATION OF HIGH GLUCOSE-INDUCED CASPASE-1

ACTIVATION IN RAT RETINAL MÜLLER CELLS (RMC-1)

FIGURE 3.1: TREATMENT OF RMC-1 CELLS WITH EXTRACELLULAR APYRASE

OR ADENOSINE DEAMINASE SUPPRESSES HIGH GLUCOSE-INDUCED

CASPASE-1 ACTIVATION ...... 107

FIGURE 3.2: TREATMENT OF RMC-1 CELLS WITH EXOGENOUS ATP,

EXOGENOUS ADENOSINE ANALOG, OR ADENOSINE UPTAKE INHIBITORS

MIMICS HIGH GLUCOSE-INDUCED CASPASE-1 ACTIVATION ...... 109

FIGURE 3.3: TREATMENT OF RMC-1 CELLS WITH EXOGENOUS ADP OR

EXOGENOUS UTP ATTENUATES HIGH GLUCOSE-INDUCED CASPASE-1

ACTIVATION ...... 112

FIGURE 3.4: TREATMENT OF RMC-1 CELLS WITH SURAMIN OR ADENOSINE

RECEPTOR ANTAGONISTS, BUT NOT P2X7R ANTAGONISTS, SUPPRESSES

HIGH GLUCOSE-INDUCED CASPASE-1 ACTIVATION ...... 114

FIGURE 3.5: RMC-1 CELLS DO NOT EXPRESS IMMUNOREACTIVE P2X7

RECEPTOR PROTEIN OR FUNCTIONAL P2X7 RECEPTORS ...... 116

xi FIGURE 3.6: ACTIVATION OF CASPASE-1 BY HIGH GLUCOSE, NECA, OR

ATP IS MIMICKED BY FORSKOLIN AND CORRELATED WITH INCREASED GENE

EXPRESSION OF CASPASE-1 AND TXNIP ...... 121

2+ FIGURE 3.7: RMC-1 CELLS EXPRESS GQ-COUPLED, CA MOBILIZING

P2Y2/P2Y4 AND P2Y1 RECEPTORS AND HIGH GLUCOSE TREATMENT

2+ INCREASES THE EFFICACY OF CA MOBILIZATION ...... 123

FIGURE 3.8: RMC-1 CELLS DO NOT EXPRESS FUNCTIONAL P2X4R ...... 125

FIGURE 3.9: PROPOSED MODEL FOR HIGH GLUCOSE-INDUCED CASPASE-1

ACTIVATION VIA PURINERGIC SIGNALING ...... 127

CHAPTER 4: DIABETES-INDUCED CASPASE-1 ACTIVITY IS REGULATED BY AN

AUTOINFLAMMATORY FEED-BACK MECHANISM

FIGURE 4.1: RETINAL CAPILLARY DEGENERATION IN WILD-TYPE AND

CASPASE-1 KNOCKOUT MICE...... 145

FIGURE 4.2: HYPERGLYCEMIA-INDUCED CASPASE-1 ACTIVITY PATTERN IN THE

RETINAS OF DIABETIC WILDTYPE AND IL-1R-1-/- MICE ...... 149

FIGURE 4.3: HIGH GLUCOSE-INDUCED CASPASE-1 ACTIVATION, ACTIVITY, AND

IL-1β RELEASE PATTERSN IN RMC-1 ...... 151

FIGURE 4.4: HIGH GLUCOSE-INDUCED CASPASE-1 ACTIVITY AND IL-1β

RELEASE PATTERNS IN HMC ...... 153

FIGURE 4.5: EXOGENOUS IL-1 β- INDUCED CASPASE-1 ACTIVATION IN

RMC-1 ...... 155

FIGURE 4.6: THE EFFECT OF INHIBITION OF EARLY PHASE CASPASE-1/IL-1β

ON RETINAL MÜLLER CELL VIABILITY ...... 158

xii FIGURE 4.7: GLUCOSE WASHOUT EFFECT ON LATE PHASE CASPASE-1

ACTIVITY ...... 160

FIGURE 4.8: HIGH GLUCOSE-INDUCED NLRP3 INFLAMMASOME COMPONENT

MRNA INDUCTION ...... 162

FIGURE 4.9: HYPERGLYCEMIA-INDUCED AUTO-INFLAMMATORY FEEDBACK

SIGNALING ...... 164

xiii ACKNOWLEDGEMENTS

Having pretty much grown up in the Department of Physiology and

Biophysics at Case Western Reserve since beginning my tenure first as an undergraduate student in the SURP program in 2004 and later as a full-fledged graduate student in 2006, it goes without saying that I am deeply indebted to many. Without the encouragement and guidance of Dr. David Essig, my Geneva

College Biology Advisor, mentor, and personal friend, I would never have taken that first step in trying on the “research shoe” that summer. Dr. Daryl Sas instilled a passion for learning, a willingness to engage science not only practically, but philosophically as well, and was always there to lend an empathetic and encouraging ear during the difficult thesis research years for which I am ever grateful. I know I would be a far less successful, (mostly) well-balanced and happy individual without their tireless investment in my life. Dr. George R.

Dubyak willingly took my rough around the edges self in when my first thesis advisor, Dr. Susanne Mohr, earned an incredible opportunity at Michigan State

University in 2009. It’s doubtful I would have persevered to the end of this process without his invaluable guidance, support, and example. I am still not entirely sure when he sleeps as I’ve never met anyone more up to date in their reading or more involved in all of his lab members’ projects and it is his example that daily encourages me to be a better scientist and person. I especially thank

Dr. Susanne Mohr for graciously giving me the caspase-1 project and allowing me to continue in my study despite our different locales. My thesis committee members- Drs. Corey Smith, Thomas Nosek, Patrick Wintrode, Margaret

xiv Chandler, and Thomas McCormick-have always been my advocates and a source of great support. I am grateful for their thoughtful comments throughout this process. Life at the bench would have been far less bearable without the people I’ve had the distinct pleasure of working with- E. Chepchumba K. Yego,

Jason Vincent, Domenick Prosdocimo, Andrew Blum, Andrea Boyd Tressler,

Christina Antonopolous, Daniel Chopyk, Caroline El-Sanadi, and Michael

Katsnelson: Thank you for keeping me sane and focused; for being insane when it was necessary for levity (which, given the nature of science, was quite often); for being not only my coworkers, but also my close friends and personal cheerleaders. Whatever you do, dont forget “Back Up Friday!” or to “Push it real good” on those extraordinarily good data days. Though Dr. Kristen Doud and I never technically worked together, I would have been lost without her personal and professional advice and support. Friends like her are hard to find, and I am grateful to call her both friend and colleague. Thank you also to the invaluable departmental Support Staff (Jean Davis, Morley Schwebel, Sylvia Hart and Ruth

Washington) for making life as a student a little more bearable by taking care of the nitty-gritty, behind the scenes work. And finally, thank you especially to my family and friends whose unwavering support saw me through to the very end.

It is the beauty of the created order which gives an answer to our

questionings about God~ Oxford Bishop Richard Harries

Soli Deo Gloria

xv LIST OF ABBREVIATIONS

ADA Adenosine deaminase

AGE Advanced glycation end product

AFC 7-amino-4-trifluoromethylcoumarin

Apaf-1 Apoptosis protease activating factor-1

AMP Adenosine monophosphate

ASC Apoptosis speck-like containing protein

Bcl-2 B-cell lymphoma 2

BSA Bovine Serum Albumin

CAD caspase activated DNAse

Cas-1 -/- Caspase-1 knockout mice

CARD Caspase recruitment domain

CHAPS 3-[(3-Cholamidopropyl)dimethylammonio]-1-

propanesulfonate

DAG Diacyclglycerol

DAMP Damage associated molecular pattern

DD Death domain

DHA Docosahexaenoic acid

DMEM Dulbecco’s Modified Eagles Medium

DMT1/DMT2 Diabetes Mellitus Type 1/Type 2

DR Diabetic retinopathy

DTT Dithiothreitol

EDTA Ethylenediaminetetraacetic Acid

xvi EtBr Ethidium bromide

FBS Fetal Bovine Serum

FSK Forskolin

GPCR G-protein coupled receptor hMC Isolated Human Müller Cells

HEK 293 Human Embyonic Kidney 293 cells

HEK 293-rP2X7R HEK 293 rat P2X7R expressing cells

HRP Horseradish Peroxidase

ICAD inhibitor of caspase activated DNAse

ICE Interleukin-converting enzyme

IL-1β Interleukin-1-beta

IL-6 Interleukin-6

IL-18 Interleukin-18

IL-1B NA Interleukin-1-beta neutralizing antibody

IL-1R1 Interleukin-1-beta receptor

IL-1R1 -/- Interleukin-1-beta receptor knockout

IL-1RA Interleukin-1-beta receptor antagonist

IKK IκB kinase

IPAF ICE-protease activating factors

IRAK IL-1R1 associated kinase

INL Inner Nuclear Layer kDa kilodalton

LRR Leucine-rich-repeat

xvii MSU Monosodium urate crystals mtATP Mitochondrial ATP mtDNA Mitochondrial DNA

NACHT Nucleotide binding and oligomerization domain

NECA 5’-N-ethlcarboxamido-adenosine

NFκB Nuclear factor kappa-light-chain-enhancer of activated B

cells

NIK NFκB inducing kinase

NLR Nucleotide binding domain

NLRP Nucleotide binding domain and LRR containing protein

NOD Nucleotide binding oligomerization domain

NO Nitric oxide

ONL Outer Nuclear Layer

P2Y1/2/6R P2Y1/2/6 Receptor

P2X7R P2X7 receptor

PAMP Pathogen associated molecular pattern

PBS Phosphate Buffered Saline

PKA Protein kinase A

PKC Protein kinase C

PLC Phospholipase C

PMSF Phenylmethanesulphonylfluoride

P/S Penicillin/Streptomycin

PUFA Polyunsaturated fatty acids

xviii RAGE Receptor for advanced glycation end products rMC-1 Transformed rat retinal Müller cells

RIPA Radio-immunoprecipitation

RIP2 Receptor-interacting-protein-2

ROS Reactive oxygen species

RPE Retinal pigmented epithelium

SDS Sodium dodecyl sulfate

STZ Streptozotocin

TLR Toll-like receptor

TNFα Tumor necrosis factor alpha

TRAF6 TNF-Receptor associated factor 6

TXNIP Thioredoxin-interacting protein

xix Caspase-1-Dependent Inflammatory Signaling in Retinal Müller Cells During the Development of Diabetic Retinopathy

Abstract

KATHERINE EILEEN TRUEBLOOD

Diabetic retinopathy, a common complication of diabetes, is the leading cause of blindness in adults in the United States. It is known that high blood glucose levels result in initiation of an inflammatory response in the eye, contributing to the development of disease. This inflammatory response is characterized by elevated levels of activated of caspase-1, an inflammatory mediator protein, and increased production of the inflammatory cytokine, interleukin-1-β (IL-1β). Retinal Müller cells are one source of inflammatory signaling molecules in the retina and possibly participate in the development and progression of diabetic retinopathy. However, how high blood glucose levels regulate caspase-1 activation is incompletely understood. In this dissertation, I characterized: (1) how extracellular purines act as autocrine/paracrine mediators of high glucose-induced signaling cascades that converge on activation of caspase-1 in a cultured rat Müller cell model (rMC-1); and (2) the role of an IL-1β driven autocrine feedback mechanism in mediating acellular capillary degeneration. My findings support a novel role for autocrine P1 and P2 purinergic receptors coupled to cAMP signaling cascades and transcriptional induction of caspase-1 in mediating high glucose-induced activation of caspase-1 and secretion of IL-1β in a cell culture model of retinal Müller cells. My studies also

1 indicate that sustained caspase-1 activation is crucial for the development of diabetic retinopathy as caspase-1 knock-out mice did not develop acellular capillaries, a hallmark of the disease. High glucose-induced a multiphasic pattern of caspase-1 activation in the retina in vivo and in rMC-1 in vitro. Blocking IL-1β mediated signaling using IL-1 receptor antagonist prevented the sustained, degenerative phase of caspase-1 activation suggestive of an IL-1β autocrine feedback mechanism in vitro. In vivo studies supported this model by demonstrating that knock-out of the interleukin-1 receptor prevented sustained activation of caspase-1 in the diabetic retina in addition to protecting against acellular capillary formation. Future experiments defining a specific role for the

NLRP3 inflammasome, a multi-protein complex that regulates caspase-1 activation in hematopoietic cells during microbial infection, may provide additional insight regarding the molecular mechanisms by which high glucose induces caspase-1 activation in these non-hematopoietic Müller cells that contribute to sterile inflammation in the retina.

CHAPTER 1

INTRODUCTION

Chapter 1

Introduction

1.1 Diabetes: A Public Health Crisis

Diabetes Mellitus is a growing epidemic with nearly 2 million new cases diagnosed in people aged 20 years or older during 2005-2008 and a total number of 25.8 million cases in the US alone (160). The Center for Disease Control

(CDC) recently reported that 35% of adults >20 years of age are prediabetic representing an additional 79 million people at significant risk for developing the disease. Economically, treatment of diabetes and its numerous complications costs on average $174 billion annually (160). There is an emerging and alarming trend between increases in sedentary behavior and obesity with the prevalence of diabetes. Diabetes remains the 7th leading cause of death in the US and few therapies exist to adequately address the effects of this chronic disease.

Diabetes Mellitus (DM) is a chronic condition characterized by elevated fasted blood glucose levels above 126 mg/dl. High blood glucose is the result of either significantly reduced or absent insulin production and/or secretion

(Diabetes Mellitus Type 1, DMT1), or due to reduced insulin sensitivity (Diabetes

Mellitus Type 2, DMT2). DMT1 is primarily characterized as an auto-immune disease where pancreatic β islet cells, the cells responsible for insulin production, are destroyed as a result of the immune response. It is still not entirely clear what

4 initiates this auto-inflammatory immune response, though genetic and environmental risk factors have been correlated with DMT1 onset. Of particular interest to the present study, recent reports have shown a specific role for the inflammatory cytokine, interleukin-1- beta (IL-1β) in mediating β islet cell death in both types of diabetes (120). DMT2 is not initially insulin dependent as pancreatic

β cells are still functioning in these patients, but due to a variety of external factors such as age, poor diet and low activity levels, chronic peripheral insulin insensitivity results in increased blood glucose levels. In response to prolonged insulin insensitivity, the pancreas initially increases insulin production but ultimately without significant lifestyle changes like exercise, weight loss, and diet changes (and sometimes despite these changes), β islet cells die and are no longer able to produce insulin.

In the past, DMT1 was largely characterized as a juvenile disease while

DMT2 was characterized primarily as a disease of old age. Alarmingly, the rate of

DMT2 incidence in adolescents <20 years of age is rapidly rising putting a much younger demographic at risk for developing the many complications of diabetes much earlier than has been previously observed. Increased risks for cardiovascular disease, neuropathy, nephropathy, stroke, defective wound healing, and even Alzheimer’s disease, have all been linked to diabetes indicating this is in fact a “whole body” disease. Elevated blood glucose levels have been shown to increase the amount of circulating inflammatory cytokines such as IL-1β (127, 138, 184, 195). Thus, the deleterious effects of elevated glucose may be due in large part to its ability to induce chronic inflammation and

5 subsequent tissue damage. Few therapies for diabetes and its complications exist and though strict blood glucose control is obviously important, for many patients this is difficult to achieve. Unfortunately, for some patients, even rigorous control is not enough to prevent tissue damage (72, 140). While it is imperative that we continue to understand not only the pathogenesis of diabetes itself, it is equally important to determine exactly how elevated glucose levels induce chronic inflammatory events.

1.2 Basic Retinal Physiology

The retina is a thin, specialized layer of neuronal tissue located at the most posterior region of the eye responsible for translating light into complex visual information (Fig. 1.1 A & B). Organized into several thin layers (described as “cake layers” by Helga Kolb in 1995) by cell type, the retina receives visual input through the pupil at the front of the eye (105). Vascularization of the retina divides it into four well-organized quadrants in order to meet the high energy demands of visual processing (Fig 1.5). Three neuronal cell body layers (the outer [ONL], inner [INL] nuclear and ganglion cell layers) are separated by 2 synaptic layers (the outer [OPL] and inner [IPL] plexiform layers). The ONL, located closest to the retinal pigmented epithelium (RPE) at the posterior most portion of the retina, contains the cell bodies for the primary light sensing cells, called photoreceptors, which are commonly referred to as either rods or cones.

Rods are primarily responsible for sensing light while cones mainly sense colors.

6 Photoreceptors are responsible for the first level of visual sensing and even some initial processing. The RPE provides metabolic support to the rods and cones as well as protects the retina from excess light.

Counter-intuitively, rods and cones are hyperpolarized (or essentially deactivated) by light. Secretion of the excitatory neurotransmitter, glutamate, onto bipolar cells at their synapse in the INL is suspended after hyperpolarization resulting in decreased electrical activation of bipolar cells. Thus, photoreceptor neurotransmission is most active in the dark and is suppressed by light. Bipolar, horizontal and amacrine cells are located within the INL where transfer of information from the light/color sensing cells to neuronal cells takes place.

Turning bipolar cells on or off is a necessary step in translating visual stimuli into meaningful data (104, 105). This signal then leaves ganglion cells and travels along the optic nerve to the lateral geniculate nucleus where it is finally passed on to the visual cortex for final processing.

1.3 Diabetic Retinopathy: The Leading Cause of Acquired Blindness

Diabetes affects both the macro- and micro- vasculature resulting in significant cardiovascular and peripheral capillary disease. One of the most common complications of diabetes is the relatively rapid (within 10-15 years of diabetes) onset of diabetic retinopathy (DR), a condition characterized by degeneration of the retinal vasculature (Fig. 1.2) (66, 159). DR results in 12,000 to 24,000 new cases of blindness each year making diabetes the leading cause

7 of acquired blindness (72, 160). Chronic high blood glucose levels induce significant changes in the retina such as widespread inflammation and increased vascular permeability after as little as 2 months of diabetes (166, 196).

Within 6 weeks, high blood glucose levels induce significant biochemical changes within the retina (Fig. 1.3). One of the first changes is induction of significant mitochondrial stress resulting in increased super oxide and reactive oxygen species (ROS) production (106). The contribution of mitochondrial stress in the context of diabetes is only partially understood (109, 111, 139, 151, 202).

Enhanced ROS production is a likely inducer of the inflammasome (a multi- protein complex known to participate in caspase-1 activation, see section 1.5 for more information), caspase-1 activation and NFκB activation (necessary for transcription of pro-inflammatory proteins) (190). Impaired mitochondrial function

2+ 2+ and membrane integrity alter intracellular Ca (Ca i) homeostasis and is thought to be another mechanism of inflammation.

Dyslipidemia in the retina, as characterized by increased retinal docosahexaenoic acid (DHA), and very-long-chain-polyunsaturated fatty acid

(PUFA) production, contribute to elevated secretion of vascular endothelial growth factor (VEGF) and other inflammatory cytokines such interleukin-1-beta

(IL-1β), interleukin-6 (IL-6), and tumor necrosis factor alpha (TNFα) under diabetic conditions (18, 215). The vitreous of diabetic patients contain elevated levels of IL-1β indicating that this inflammatory response may be largely due to elevated levels of active caspase-1 (See section 1.5 for information on caspases) found in retinas of diabetic patients (41). Activation of protein kinase C (PKC), a

8 pleuripotent 2nd messenger, is the direct result of hyperglycemia-induced diacylglycerol (DAG) production (3, 84, 99, 136, 166, 196, 199, 216). The development of DR is also closely correlated with increased non-enzymatic advanced glycation end (AGE) product production and activation of AGE receptors (RAGE) due to increased blood glucose levels (9, 152, 216). Chronic activation of RAGE promotes sustained NFκB activation and subsequent pro- inflammatory protein expression, increased VEGF secretion, elevated vascular cell adhesion molecule (VCAM) expression and leukocyte adherence to vascular endothelia promoting vascular hyper-permeability. In addition, RAGE signaling represents a feed-forward mechanism, promoting further production of AGE products (10, 194). Increased expression of inter-cellular adhesion molecule

(ICAM) and CD18 also promote significant leukostasis and vascular damage during this time (88-90, 166, 190).

A sustained inflammatory response due to chronic hyperglycemia ultimately results in the loss of retinal cells, which specifically contributes to the development of diabetic retinopathy (88, 90, 99, 132, 207). There appears to be a distinct connection between caspase-1 mediated inflammation and a mechanism of cell death called “Pyroptosis” (13, 60, 61, 98, 99, 122). Aberrant caspase-1 activation, subsequent cytokine production and pyroptotic cell death have been shown to play integral roles in the development and progression of other neurodegenerative diseases such as Alzheimer’s disease, Parkinson’s disease, and Huntington’s disease (211). Treatment with anti-inflammatory drugs like minocycline, aspirin and etanercept, an anti-TNFα drug, show significant

9 promise in their ability to prevent the progression of retinal cell death (41, 91,

136). In addition, interleukin-1-receptor-1 (IL-1R1) knockout mice (IL-1R1-/-), which lack the receptor mediating the IL-1β response, are protected from diabetes-induced vascular damage (207). However, exactly how glucose mediates caspase-1 dependent inflammation within the retina, and more specifically, in the specialized Müller glial cells (to be discussed more in depth later, section 1.4) was the focus of this project.

If diabetes is left untreated and DR progresses, ganglion cell (retinal neural cells) loss can occur as early as 2-3 months after diabetes onset followed by loss of Müller cell glia after only 4 months (57, 80, 81, 125). Whether permanent vision loss is due to the effects of hypoxia (due to diabetes-induced death of retinal vasculature endothelia) alone or due to the combined effects of hypoxia and concurrent inflammatory mediated damage to retinal neuronal cells is a matter of much debate (8).

It is not uncommon for patients to be unaware of their diabetic status until clinically observable retinal damage occurs. Prior to diagnoses, the first detectable morphological changes include development of pericyte ghosts, foci of vessel sudanophilia, capillary basement membrane thickening and acellular capillary formation (Fig 1.4) (66, 71, 125, 135, 154). The disease progresses from these background stages to the mild/moderate and severe non-proliferative stages during which time microaneurysms begin to develop (Fig 1.5). Patients begin to have visual obstructions during this time when fluid and blood leak from the damaged vessels causing retinal edema (Fig 1.6). Vitrectomy, removal and

10 replacement of the nutrient rich vitreous humor, is commonly performed to clear the patient’s vision though this is only a temporary fix and does not address the underlying cause of progressive retinal damage. The proliferative stage, which is characterized by mass neovascularization and hemorrhage, is the most detrimental to retinal health resulting in irreversible vision loss due to retinal neuronal cell death and retinal detachment (66). Studying mechanisms of high glucose-induced retinal damage is made increasingly difficult since many events are species- and cell-type specific and, unlike humans, rodent models of the disease do not progress past acellular capillary formation into the final stages of neovascularization.

Currently the only treatments for DR include good control of blood glucose levels with insulin, controversial anti-VEGF treatments, anti-inflammatory therapy, and laser photocoagulation of neovascularized capillaries (136, 216).

However, problems exist with these treatments as it has proven difficult to maintain blood glucose at consistent levels and laser ablation usually results in some loss of vision simply from the procedure itself. The retinal blood barrier that restricts many drugs from gaining access to retinal cells further complicates treatment and patient compliance with ophthalmic drops is less than ideal. It is clear then that understanding the mechanisms of high glucose-induced retinal cell death is necessary for continued development of potential therapies.

11 1.4 Retinal Müller Cells and Diabetic Retinopathy

Although DR is often characterized as a primarily vasculature disease, it is also possible that an insult to the surrounding tissue may lead to the development and/or progression of the disease. Aside from the diverse neuronal cells that make up the retinal layers, several vital support cells are also located within this tissue and are known to be glucose responsive. Of particular interest to our studies are retinal Müller cells which are the principal glia of the retina and provide both structural and metabolic support to the retinal vasculature and neurons. Müller cells arise from the neuroepithelium embryonically and are generated by the same progenitor responsible for retinal neurons.

Morphologically, these cells span the retinal layers and wrap around the local vasculature contributing to their specialized function (Fig. 1.7). Müller cells are responsible for maintaining the homeostatic environment of the retina by sequestering free glucose and producing lactate for retinal neurons; by regulating ion homeostasis and subsequently fluid homeostasis; and preventing glutamate toxicity by scavenging glutamate from the synaptic space (as summarized in

Table 1.1) (16).

Due to their unique structure and function, Müller cell viability and functionality have a marked impact on the health of the retinal vasculature and neurons. Data suggest that retinal Müller cells of diabetic patients are dysfunctional and are one source of active caspase-1 and IL-1β in the retina during DR (12, 52, 60, 95, 96, 119, 164, 192, 195, 196, 202). Therefore, we postulate that high glucose-induced Müller cell damage results in subsequent

12 damage to endothelia of retinal capillaries either directly by secretion of active IL-

1β or indirectly by the loss of Müller cells leaving the vasculature unprotected.

Though previous work has been focused on the downstream effects of high glucose induced caspase-1 activity in the development of DR, exactly how high glucose initiates this inflammatory cascade has not been previously elucidated.

We have observed evidence of Müller cell death and caspase-7 (caspase-

3-like) and 6 activation in a mouse model of the disease (unpublished data) while others have demonstrated positive annexin 4 staining, increased trypan blue staining in cell culture models exposed to high glucose conditions, and a decrease in the pro-survival molecule, AKT (57, 112, 207, 221, 223, 224). In addition, hyperglycemia-induced Müller cell mitochondrial dysfunction and increased ROS production has been well-described and implicated in the progression of DR (18, 111). Müller cell loss and dysfunction has also been directly associated with vascular changes including basement membrane thickening and bulging of vessels observed immediately prior to micro aneurysm formation (Table 1.2) (80, 81).

1.5 Caspase-1/Interleukin-1-beta signaling axis

1.5.1 Caspase-1 Activation

The name “caspase” was given to a group of cysteine proteases which participate in elegant signaling cascades that mediate cell death and pro- inflammatory pathways (30). Caspases are expressed as 30-50 kDa precursor

13 zymogens with very little protease activity that must be cleaved into smaller subunits which then self-multimerize to generate tetrameric complexes with high proteolytic activity. (Fig. 1.8) Specialized domains within the caspase structure contribute to its unique activation processes and ultimately, function (52).

Caspases bind not only to other caspases, but also to other proteins containing caspase-recruitment domains (CARD) and death domains (DD).

Two classes of caspases are well described as either initiator or executioner caspases. Members of the initiator caspases include caspases-2, 8 and 9 while caspases-3, 6, and 7 are considered executioner caspases.

Executioner caspases cleave target proteins like inhibitor of caspase-activated

DNAse (ICAD), which, once cleaved, releases its inhibition of caspase-activated

DNAse (CAD). CAD plays an integral role in DNA laddering during apoptosis

(35). Both extrinsic (receptor-mediated) and intrinsic (mitochondrial-mediated) apoptotic pathways involve complex intracellular cascades involving activation of both initiator and executioner caspases (Fig. 1.8). In one example of receptor- mediated apoptosis, activation of Fas-associated receptor by Fas ligand induces trimerization of the receptor subsequently recruiting FADD adaptor proteins which recruit and activate initiator caspase-8. Active caspase-8 directly cleaves target executioner caspases-3 or 7 that are responsible for initiating DNA laddering and other apoptosis-related events. In some but not all cells, active caspase-8 can also cleave the Bcl2-family protein Bid that then translocates to the mitochondrial membrane to intersect with, and activate the intrinsic or mitochondrial apoptotic pathway. However, the intrinsic apoptotic pathway is

14 most commonly triggered by many types of DNA damage, which trigger kinase cascades that result in increased transcriptional activity of p53. In turn, p53 drives increased expression of pro-apoptotic Bcl-2 family members, Bax and

Bak, which (as in the similar case of Bid) are responsible for inducing cytochrome c release from the mitochondria. Cytochrome c recruits the adapter protein Apaf-

1 and pro-caspase-9 into a multimeric protein structure termed the apoptosome.

Caspase-9 is then activated by the apoptosome and is responsible for activating caspase-3 and -7.

Although the inititator and executioner caspases function as mediators of apoptotic cell death, a third group of caspases, which includes caspase-1, human caspase-4, human caspase-5, and murine caspase-11, that predominantly function as regulators of inflammation and are only secondarily involved in non- apoptotic cell death processes. Caspase-1 was originally named ICE (IL-1β converting enzyme) because of its major physiological function of producing active inflammatory cytokines like IL-1β and IL-18 (30, 35, 200). Caspase-1 activation has been best described in monocyte/macrophage-lineage leukocytes that are the predominant sources of IL-1β in most inflamed tissues.

Accumulation of active caspase-1 in these cells is mediated by the assembly of multiprotein ‘inflammasome’ complexes, or platforms, that recruit the zymogen form of caspase-1 (procaspase-1) to signaling complexes based on adapter proteins belonging to the NLR (nucleotide-binding domain and leucine rich repeat containing) family (184).

15 NLR proteins are characterized by their tripartite architectural structure and the presence of a central nucleotide binding and oligomerization domain

(NACHT), a c-terminus leucine rich repeat (LRR), and an N-terminal interaction domain (59, 63, 121, 126, 130, 202). Members of the NLR family are distinguished by their particular N-terminal interaction domain- containing PYRIN

(nucleotide-binding domain, leucine rich repeat, NLRPs), caspase recruitment domain (CARD) (Nucleotide-binding oligomerization domain, NODs), or a baculovirus inhibitor of apoptosis protein repeat (BIR) domains (NLRB1) (162).

NODs have been shown to interact with receptor interacting protein 2 (RIP2) via their CARD domain and subsequently regulate NFκB activation (102, 119, 121,

129, 162, 176). In addition to NLRPs, most inflammasome complexes include the important adapter protein, apoptosis speck-like containing protein (ASC) to regulate pro-inflammatory events, in particular, caspase-1 activation. Most NLR proteins are targeted by various pathogen-associated molecular pattern (PAMP) ligands (e.g., bacterial flagellin or DNA/ RNA fragments) that accumulate during microbial infection, but NLRP3 (nucleotide-binding domain and leucine rich repeat containing pyrin-protein 3, also known as cyropyrin) is indirectly targeted by multiple danger-associated molecular pattern (DAMP) ligands (such as extracelluar ATP or phagocytosed monosodium urate [MSU] crystals). DAMPS are released from dead or damaged mammalian cells within tissue sites of injury and inflammation (22). Recently, several gain-of-function mutations in NLRP3 that result in the protein being constitutively active, have been associated with

16 autoimmune diseases like Muckle-Wells syndrome and Familial Fever Syndrome

(59, 130).

Activation and induction of inflammasome assembly is incompletely understood. Current hypotheses include lysosomal release of cathepsins (in particular, but not in all cases, cathepsin B.), increased mitochondrial

+ + dysfunction, enhanced ROS production, and a reduction in intracellular K (K i)

(Fig. 1.9). Not only does inflammasome activation seem to require activating stimuli which are known to produce ROS, but some groups indicate mitochondrial stress and damage may additionally provide DAMPS, such as mitochondrial DNA

(mtDNA) and mitochondrial ATP (mtATP), though these results are somewhat controversial (139, 202, 214). Mitochondria also play a vital role in regulating

2+ 2+ 2+ intracellular Ca (Ca I) homeostasis and perturbations in Ca I could additionally regulate inflammatory events. The link between ROS production and caspase-1 activation may involve free TXNIP (Thioredoxin-Interacting-Protein), which accumulates in response to multiple types of metabolic and oxidative stress (Fig. 1.10). Interestingly, hyperglycemia has been shown to both directly upregulate TXNIP expression through MONDO/MLX transcription factors, as well as to increase free TXNIP levels due to hyperglycemia-induced oxidative stress and consequent dissociation of pre-existing TXNIP/thioredoxin complexes (Fig.

1.9) (20, 26, 27, 228, 232). TXNIP may directly regulate caspase-1 activation through induction of NLRP3 inflammasomes in non-hematopoietic cells, such as pancreatic β-cells and adipocytes (103). Additionally, free TXNIP is known to regulate apoptosis signal-regulating kinase 1 (ASK1) phosphorylation which

17 modulates cytochrome C release from mitochondria thereby regulating apoptosis of damaged or stressed cells. The role of TXNIP in mediating retinal inflammation has been the focus of recent studies making the mitochondrial stress/TXNIP signaling axis a particularly attractive hypothesis in the context of high glucose- induced caspase-1 activation in retinal Müller cells (4, 156, 157, 161)

Regardless of the particular inducers, caspase-1 activation via the NLRP3 inflammasome in hematopoietic monocyte-lineage cells is well-described as a two-step process requiring NFκB-dependent upregulation of NLRP3 expression followed by DAMP/PAMP-dependent formation of the NLRP3/ASC/procaspase-1 inflammasome complexes (Fig. 1.10) (184). Recent emphasis has focused on describing regulation NLRP3 inflammasomes as a distinct mechanism of caspase-1 activation in multiple inflammatory diseases (103, 127, 195, 205).

Notably, purinergic signaling in response to stimulation of cells with exogenous

ATP or following endogenous ATP release from damaged cells (leading to autocrine/paracrine activation of purinergic receptors, specifically P2X7 receptors

(P2X7R)), has garnered much attention with regard to its role in the activation of

NLRP3/caspase-1 signaling in hematopoietic cells such as macrophages (12, 58,

94, 123).

Of particular relevance to the research which comprises this dissertation are recent reports that indicate the NLRP3 inflammasome is also capable of integrating ionic and metabolic stress and breaches in membrane integrity through P2X7R dependent- and independent- mechanisms, converging on caspase-1 activation in non-hematopoietic cell types such as pancreatic β-islet

18 cells and adipocytes (232). A distinct link between NLRP3 and caspase-1 activation and metabolism has been described (Fig. 1.11) (187). Under acute metabolic stress, NLRP3 regulates glucose metabolism through activation of the glycolytic enzymes aldolase and pyruvate kinase (195, 205). NLRP3 inflammasomes have also been shown to play an important role in regulating lipid metabolism. However, under chronic metabolic stress, as is the case in DM,

NRLP3-induced caspase-1/IL-1β signaling leads to pancreatic β-islet cell dysfunction and death resulting in reduced insulin secretion. In addition, increased insulin resistance in fat, muscle and liver seem to be regulated by chronic NLRP3 mediated inflammation (184, 195, 201, 205). Knockdown of

NLRP3 has been shown to reduce inflammation and increase insulin sensitivity in rodents (205).

Several inflammasome-independent mechanisms of caspase-1 activation by caspase-11 and p53 have been suggested (96, 210). Receptor interacting protein 2 (RIP2, aka RICK), another scaffolding and adaptor protein which is not a member of the NLR super-family has also been implicated in caspase-1 activation (35). RIP2, a CARD-containing kinase and scaffolding protein, has been shown to activate both pro-survival and pro-death pathways, though the interactions between these functions are unclear (129). NFκB activation by RIP2 has also been demonstrated and can promote either pro-survival or pro-death pathways by regulating gene transcription (116, 129, 179). It is possible that differential complex formation may determine which pathway RIP2 will initiate

(121, 122, 129, 179). A possible link between hyperglycemia, increased RIP2

19 expression and elevated NFκB activation will be discussed more in depth in

Chapter 5. Interestingly, RIP2 has been shown to activate caspase-1 in neuronal cells in response to a variety of stimuli such as TNFα and in Huntington’s disease

(211). As retinal Müller cells are of neuronal origin, the role of RIP2 in neuronal cell death is of particular interest to this project.

1.5.2 Interleukin-1-beta

IL-1β production is mediated via the activation of caspase-1 (30, 35, 45,

200). Expressed as a precursor (pro-IL-1β) at relatively low levels basally, pro-IL-

1β expression is tightly controlled in myeloid leukocytes. After proteolytic cleavage of pro-IL-1β by caspase-1, mature active IL-1β is secreted via microvesicles and exosomes into the extracellular space where it can then activate its receptor, IL-1R1, in an autocrine or paracrine manner. Complex regulation of IL-1β secretion is a critical step in inflammatory events and has been the focus of many recent studies (167, 168, 170, 171, 206).

The IL-1R1 receptor most closely resembles the structure and function of the toll-like receptor (TLRs) family of receptors (53, 149). Activation requires binding of its ligand, either IL-1β or IL-1α (which does not need to be cleaved to be active), and subsequent dimerization of the receptor. Once dimerized, MyD88 is recruited by the toll-IL-1R (TIR) domain to activate the IL-1R associated kinase

(IRAK). IRAK is responsible for recruitment of TNF receptor associated factor 6

(TRAF6) which is necessary for activation of TAB1 and TAK1. Next, TAB1/TAK1 triggers either MEK1 or NFκB inducing kinase (NIK) to activate the IKK (I-kappa-

20 kinase) complex. Phosphorylation of IκB marks it for ubiquitination and rapid proteosomal degradation, thereby removing the inhibitory effect of IκB on NFκB.

Activation of NFκB facilitates its translocation into the nucleus wherein it regulates expression of many pro-inflammatory proteins including pro-IL-1β, inducible cyclooxygenase, adhesion molecules, chemokines, tissue degrading enzymes, and serum amyloid A. Physiological expression of an IL-1 receptor antagonist (IL-1RA) negatively regulates this signaling pathway by competitively binding to the IL-1R1 making engineered forms of IL-1RA attractive therapies for many inflammatory conditions. Indeed, several autoinflammatory diseases such as rheumatoid arthritis and DMT2 have benefited from a recombinant IL-1RA therapeutic called anakinra (158). Furthermore, knockdown and inhibition of IL-

1R1 prevented acellular capillary formation in the retinas of diabetic mice, but the exact mechanism of this protection was not elucidated in the initial study (207).

Many observations suggest that sustained low-grade IL-1β production is a major contributor to chronic tissue inflammation and tissue damage. Although the exact mechanisms of continuous IL-1β production under chronic inflammatory conditions are poorly understood, the concept of sterile auto-inflammation has emerged to begin to explain this phenomenon (22, 173). Sterile auto- inflammation is characterized by localized, chronic tissue inflammation in the absence of obvious microbial infection (173). In a recent review, Charles

Dinarello defined “auto-inflammatory” as “the release of the active form of IL-1β driven by endogenous molecules (44).” Recent studies have indicated that following initial IL-1β production, this cytokine can activate its own producer

21 caspase-1 in an autocrine fashion thereby providing feedback amplification of the inflammatory response (12).

1.6 Purinergic Signaling

Adenosine triphosphate or “ATP” has been labeled the “energy currency” of the cell and is produced via many processes including oxidative phosphorylation of glucose, pyruvate and NADH by the electron transport chain in the mitochondria. Aside from its pivotal role as a cellular energy source, ATP participates in nucleic acid synthesis and, of particular interest to this study, acts as an extracellular signaling molecule in a diverse array of circumstances.

Importantly, nucleotides other than ATP also play important roles in autocrine and paracrine signaling events through nucleotide receptors. Adenylate cyclase utilizes ATP in the production of cyclic AMP (cAMP), an important intracellular signaling molecule known to participate in regulating protein kinase A (PKA) activation and subsequently, many events including gene transcription. Not only can ATP participate in intracellular signaling, but once outside of the cell ATP can act as a “find me” signal following cell death so that immune cells are recruited to clear out cellular debris (56, 235). In cancer treatment with certain chemotherapeutics, the immune response to ATP released by dead or dying cancer cells plays an integral role in cancer cell destruction (6, 118).

22 1.6.1 ATP/Adenosine Metabolism

Metabolism of ATP is one mechanism of regulating extracellular purine concentrations and subsequently, terminating purinergic receptor activation. Four types of ecto-nucleotidases are responsible for the sequential degradation of

ATP into ADP, AMP, adenosine, and either Pi, or PPi. These include (1) 5’- triphosphate diphosphydrolases (ENTPDases, such as CD39 ), (2) ecto- nucleotide pyrophosphates/phosphodiesterases (ENPPs), (3) glycosyl phosphatidylinositol (GPI-) anchored ectonucleotidases (like CD73), and (4) GPI- anchored alkaline phosphatases (233). Upon generation of adenosine, adenosine reuptake transporters recycle adenosine back into the cell so that intracellular ATP can be regenerated. In addition, adenosine deaminase (ADA) enzymatically degrades adenosine into inosine providing an additional pathway for regulating extracellular adenosine concentrations. Interestingly, some data indicate that facilitated transport of adenosine into the extracellular environment is the main source for Müller adenosine rather than through the degradation of

AMP derived from hydrolysis of released ATP (83).

Signaling by nucleotides is a well-described process whereby purinergic receptors are activated by their nucleotide ligand (29, 31, 38, 76, 82, 86, 101,

146, 197). Purinergic receptors are classified into two distinct classes- P1 receptors whose ligand is adenosine (including the A1, A2A/B, and A3 receptor subtypes) and P2 receptors whose ligands, depending on P2 receptor subtype include ATP, ADP, UTP, UDP, or UDP-sugars with varying degrees of selectivity

(Table 1.3). These receptors are widely expressed amongst all tissues

23 contributing to their integral role in cellular homeostasis and intercellular communication.

1.6.2 P1 Adenosine Receptors

Adenosine receptors are G protein-coupled receptors (GPCR) which include the A1 subtype which is coupled to Go or Gi proteins, A2A and A2B subtypes which are coupled to Gs proteins and the A3 subtype which is coupled to Gq or Go proteins. These receptors have EC50 values of approximately 300nM for adenosine. The Gi/o–coupled A1 receptors inhibit adenylate cyclase activity thereby reducing cytosolic cAMP levels. Specifically, A1 receptors are well- characterized for their contribution to cardiac function by modulating the rate and force of contraction. In the brain, activation of A1 receptors decreases neurotransmitter release and neuronal firing of neurons located within the cortex, hippocampus, spinal cord and cerebellum by decreasing cAMP levels, or through

+ activation of K channels (38, 68, 101). In contrast, A2A/B receptors are Gs- coupled which, once activated, increase cAMP levels and PKA activation which is an important process in glucose and lipid metabolism as well as regulation of gene transcription (5, 188).

A feed-forward mechanism increases adenosine concentration following an increase in cAMP levels. Adenosine concentration is tightly controlled both intracellularly and extracellularly. ATP metabolism is the predominant source of extracellular adenosine in most tissues, but not in the retina. Transport of adenosine across the plasma membrane into the extracellular environment

24 requires activity of equilibrative nucleoside transporters that can also regulate reuptake of extracellular adenosine. Another class of concentrative, Na+- dependent nucleoside transporters also contributes to adenosine uptake when extracellular levels are lower than intracellular concentrations. Once in the extracellular space, adenosine can be quickly converted into inosine by endogenous adenosine deaminase (ADA) or conversion into AMP by adenosine kinase (101).

1.6.3 P2 Receptor Family

The P2 family of purinergic receptors includes the G protein- coupled metabotropic P2Y receptors and the ionotropic P2X receptors. The P2Y family encompasses eight subtypes while the P2X family contains seven subtypes. P2Y receptor function is closely related to osmotic cell volume regulation, especially within the retina, as well as neuroprotective events by increasing neurotransmitter reuptake. Five P2Y receptor subtypes (P2Y1, P2Y2, P2Y4,

P2Y6 and P2Y11) are predominantly Gq-coupled receptors that control phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis by phospholipase C

(PLC) into DAG and inositol triphosphate (IP3). PKC activation by DAG then

2+ occurs and is known to be elevated in diabetics (3), while IP3 activates Ca release from the endoplasmic reticulum. Other P2Y receptor subtypes (PY12,

P2Y13, P2Y14) are coupled with Gi and decrease cytosolic cAMP levels or increase G-protein regulated K+ channels. The P2Y11 subtype uniquely couples to Gs in addition to Gq. The P2Y12, P2Y13, and P2Y14 subtypes are Gi-coupled receptors and can thereby negatively regulate cAMP levels while positively

25 + regulating Gi-regulated K channels and several MAP kinase or Akt kinase cascades.

P2X receptors function as non-selective cation channels following activation (51). P2X receptors play important roles in nociception, control of blood flow, lipid metabolism and in regulating inflammatory events (29, 75). To differentiate between the seven distinct P2X subtypes, one can experimentally take advantage of the fact that prolonged activation of P2X7R results in the formation of a large, non-selective pore through which large molecules such as dyes like ethidium bromide (EtBr) and Yo-Pro can pass as an indices of P2X7R activation (Fig. 1.12) (124, 147, 182).

While P2Y receptors respond to a variety of nucleotides in the nanomolar to micromolar range, ATP is the predominant ligand for P2X receptors (Table

1.3). The relative sensitivity of P2Y receptors to nucleotides is higher than for

P2X receptors and especially for P2X7R whose EC50 for ATP is an exceptionally high 300 µM making its activation under physiological conditions somewhat of a conundrum (146, 197). Recent evidence has begun to explain how P2X7R are activated under physiological context. The discovery of two splice variants,

P2X7RA and P2X7Rk where the K variant is activated by lower concentrations of

ATP than the A variant, began to elucidate this mystery (2, 145). In addition, studies by our labs and others have demonstrated that ADP-ribosylation P2X7R by extracellular NAD and ecto-ADP-ribosyl transferases increases P2X7R sensitivity to ATP (79).

26 1.6.4 Purinergic signaling within the retina

P1 and P2 receptors play integral roles in maintaining retinal homeostasis and mediating communication between the retina and the RPE across the subretinal space (Fig. 1.12). These receptors are localized to photoreceptor outer segments and the apical RPE membrane (34). ATP is released by photoreceptors into the subretinal space in response to glutamate (the predominant neurotransmitter involved in visual processing) and in cases of ischemia where it is then dephosphorylated into adenosine (134). Release of

VEGF under disease states is regulated by ischemia-induced ATP release (196).

2+ 2+ Synergistically, ATP and adenosine act to increase cytosolic Ca (Ca i) in RPE cells and under chronic activation (as is the case in disease) can cause retinal edema and cell death.

P2Y1 and P2Y2 are the most widely expressed P2 receptor subtypes within the retina and primarily regulate cell volume even under physiological conditions (34). Purines are tonically released in the dark in a Ca2+-dependent manner in response to an increase in neuronal activity (142). However, chronic activation of these receptors has been shown to be deleterious and may play an integral role in the development of proliferative retinopathies and retinal detachment (31, 218, 219).

Retinal Müller cells are characterized by both autocrine and paracrine purinergic signaling via A1, A2A/B and P2Y receptors making them especially susceptible to damage due to chronic purinergic activation (Fig. 1.13) (57, 65, 69,

27 70, 97, 133, 143, 217-220, 226). In particular, Müller cells regulate synaptic activity between rods and cones and retinal ganglion by releasing ATP and adenosine in response to glutamate receptor activation and osmotic membrane stretching (97, 217, 220). Under pathological conditions, sustained P2Y2R activation results in a prolonged increase in Müller cell cytosolic Ca2+ that is correlated with increased oxidative stress and inflammation (65, 111, 155). This increased cytosolic Ca2+ results in hyperpolarization of adjacent ganglion cells.

ATP released from Müller cells under these conditions is rapidly converted to

+ adenosine and activates A1 receptors of ganglion cells and subsequent K channel opening resulting in prolonged hyperpolarization. Stimulation of A2B receptors also amplifies the ATP-induced Ca2+ response (37, 133, 143, 144). The ratio of ATP and adenosine levels ultimately determines the level of ganglion cell death due to glial cell dysfunction (218, 219). Interestingly, it has been shown that extracellular ATP concentration in the retina can be altered either by increased production or decreased metabolism under hyperglycemic conditions

(32).

Whether P2X7R expression and function is important in retinal Müller cells under physiological or pathological states is somewhat controversial (17, 33, 49,

85). Some data suggest that P2X7R plays an integral role in mediating glial cell dysfunction (67, 113, 131, 225). P2X7R expression and function has been shown to be upregulated in diabetic fibroblasts, but the effects of hyperglycemia on

P2X7R function in other cell types is not entirely clear (193). Data from

Bringmann and colleagues suggest that an increase in Müller cell P2X7R

28 activation mediates proliferative retinopathies despite overwhelming evidence that the predominant source of ATP-induced Ca2+ transients is release from internal stores of Ca2+ rather than through Ca2+ influx from the extracellular environment (17, 65, 220). Several studies with Müller cells from different mammalian species demonstrate an absence of P2X7R expression suggesting any role for P2X7R might be species-specific (85, 153). However, prolonged activation of P2X7R in human Müller cells did not result in ethidium influx and my studies have demonstrated an absence of P2X7R protein expression and function in the rat retinal Müller cell line rMC-1 which has been widely used as a model for Müller cell signaling responses to hyperglycemic stress (220, 226).

Whether Müller cell caspase-1 activation is regulated by P2X7R activation or other purinergic receptors under hyperglycemic conditions was a major issue addressed in this dissertation.

29 1.7 Summary

During the development of diabetic retinopathy, a chronic inflammatory response characterized by increased levels of pro-inflammatory cytokines such as IL-1β and increased caspase-1 activity is initiated. Intervention in this signaling cascade ameliorates retinal Müller cell dysfunction both in the mouse model and in transformed cell models (rat retinal Müller cell line, rMC-1) (207). In the animal model, this inflammatory response is characterized by increased caspase-1 activity and increased levels of pro-inflammatory cytokines such as IL-

1β and NO (16, 138, 159, 198). Vitreous of diabetic patients has been shown to contain elevated levels of IL-1β, IL-6, and TNFα. Interestingly, the severity of hyperglycemia in diabetic and galactosemic mice directly correlates to increases in caspase-1 activity (138). Treatment with anti-inflammatory agents such as aspirin and TNFα-neutralizing antibodies seems to prevent the progression of the disease (91, 198, 207). IL-1 receptor knockout mice were also protected from the development of the disease (198). These data indicate that activation of caspase-1 and subsequent IL-1β production play integral roles in the development of diabetic retinopathy. Previous work in the Mohr lab was focused predominantly on the downstream effects of caspase-1 activation. However, how high glucose activates caspase-1 remained unclear.

Significant work on mechanisms of caspase-1 activation has focused predominantly on acute activation of inflammatory processes via bacterial or viral infection. Exactly how caspase-1 activation and IL-1β production and release are regulated under chronic, sterile conditions is not well understood. The ATP-gated

30 cation channel, P2X7R, which is known to participate in inflammasome formation and caspase-1 activation in hematopoietic cell types, has been shown to be upregulated in fibroblasts of diabetic patients. Whether this puringeric receptor plays a role in inducing inflammasome formation and caspase-1 activation in retinal Müller cells under hyperglycemic conditions remained to be determined.

P2X7R-mediated activation of IL-1β secretion has predominantly been characterized in LPS-primed, ATP stimulated macrophages and monocytes but not in cases of sterile inflammation such as diabetic retinopathy. Studies in monocyte/macrophages indicate this is a two-step process where LPS priming first serves to promote transcription/translation of both pro-IL1β (the downstream substrate of caspase-1) and the NLRP3 adapter (a key upstream regulator of caspase-1) while ATP stimulation of P2X7R induces assembly of the

NLRP3/ASC/procaspase-1 inflammasome complexes required for caspase-1 activation. As such, my studies sought to determine what effect high glucose has on inflammasome components, like NLRP3, NLRP1 (another NLR family member linked to inflammasome regulation in astrocytes (1, 183)), and ASC, known to be required for caspase-1 activation and IL-1β production and release.

Identifying mechanisms involved in this IL-1β induced IL-1β release has major implications not only on the health of Müller cells themselves due to their autocrine activation by IL-1β, but also directly on the health of the local retinal vasculature because of potential IL-1β-dependent paracrine effects that could mobilize leukocytes and macrophages to the retinal vasculature.

31 Therefore, the focus of this dissertation was to determine the mechanism(s) by which high glucose initiates caspase-1 activation in retinal Müller cells.

Objective: To determine mechanisms of high glucose-induced caspase-1 activation with a specific focus on purinergic receptor function and autoinflammatory signaling in retinal Müller cells under hyperglycemic conditions

(Fig. 1.14).

Specific Aims:

Specific Aim 1: To determine the role of the purinergic receptor activation in high

glucose-induced caspase-1 activity in retinal Müller cells.

Specific Aim 2: To identify the role of an IL-1β autocrine signaling mechanism in

high glucose induced caspase-1 activity in retinal Müller cells.

32 A)

Figure 1.1

33 Figure 1.1: (A) A drawing of a section through the human eye with a schematic enlargement of the retina (B) Simple diagram of the organization of the retina.

Courtesy of: Webvision, Available online at: http://webvision.med.utah.edu, retrieved 8.18.2011. Used under the Attribution, Non-commercial, No Derivative

Creative Commons License.

Figure 1.2

35 Figure 1.2: Retinal trypsin digest from diabetic human retina demonstrating acellular capillary and micro-aneurysm formation.

Courtesy of: T. Kern Lab

Figure 1.3

37 Figure 1.3: Biochemical events during diabetic retinopathy (from rodent studies)

Reference: Yego, E.C, Mechanisms for the regulation of pro-death glyceraldehyde-3-phosphate-dehydrogenase nuclear accumulation in retinal

Müller cells under high glucose conditions, CWRU, May 2010.

Hyperglycemi a

Non-proliferative diabetic retinopathy Pericyte loss, basement membrane thickening, vascular leaking, hypoxia

Pre-proliferative diabetic retinopathy Continued Hypoxia, microaneurysms

Proliferative diabetic retinopathy Angiogenesis, retinal detachment blindness

Figure 1.4

Figure 1.4: Morphological changes during the stages of diabetic retinopathy.

A) B)

Normal Retina Mild Non-Proliferative Stage (micro anneurisms)

C) D)

Severe Non-Proliferative Stage (hemorrhage) Severe Proliferative Stage (massive hemorrhage, growth of new leaky vessels)

Figure 1.5

41 Figure 1.5: (A) Retinalgrams from normal and (B-D) diabetic patients in varying stages of diabetic retinopathy

Courtesy of: Vitreoretinal Dieseases and Surgery Service for NEEC. Available online http://www.NEEC.com/pages/services/vitreoretinal/diabetic_retinopathy

Retrieved 8/18/2011. Used with permission (Appendix 1).

Figure 1.6

43 Figure 1.6: (A) Normal Vision (B) Obstructed vision in Diabetic patient.

Courtesy of: National Eye Institute, National Institutes of Health.

Figure 1.7

45 Figure 1.7: (A) Immunohistochemistry staining of rat retina. Green:

Photoreceptors, Red: Müller cells (B) Schematic representation of Müller cells

(green) and their relationship with the retinal vasculature (red) (Mohr Lab files).

(A) Reference: Bringmann, A, et al; Müller cells in the healthy and diseased retina, Prog. In Ret. And Eye Research, 25 (2006) 397–424. Used with permission (Appendix 1).

Table 1.1

47 Table 1.1: Selective functions of retinal Müller cells.

Adapted from: Bringmann, A, et al; Müller cells in the healthy and diseased retina, Prog. In Ret. And Eye Research, 25 (2006) 397–424.

Table 1.2

49 Table 1.2: Characteristics of reactive Müller cells in various diseases.

Adapated from: Bringmann, A, et al; Müller cells in the healthy and diseased retina, Prog. In Ret. And Eye Research, 25 (2006) 397–424.

Figure 1.8

51 Figure 1.8: Extrinsic (pathway is in color) and intrinsic (pathway is in black) activation of the canonical apoptosis cascade. Extrinsic (receptor mediated) activation of apoptosis requires activation of the Fas receptor by its ligand

FasLigand. Following receptor dimerization, pro-caspase-8 is recruited to the receptor complex via its CARD domain. Following cleavage of the pro-form, caspase-8 is activated and can now cleave and activate the executioner caspases, caspase-3 and -7 which are responsible for mediating apoptotic events such as DNA cleavage. The intrinsic pathway converges on caspase-3 and -7 mediated apoptosis, but does not require extracellular receptor activation.

DNA damage due to a variety of stimuli (such as UV ray exposure, genotoxic drug treatment, etc), p53 activation mediates changes in mitochondrial membrane integrity which allows for cytochrome C release. Cytochrome C induces formation of the apoptosome, a multi-meric protein complex which is responsible for activation of pro-caspase-9. Activated caspase-9 cleaves the executioner caspases-3 and -7, initiating apoptotic events.

Figure 1.9

53 Figure 1.9: High glucose conditions modulate TXNIP expression.

Thioredoxin-interacting-protein (TXNIP) is in complex with thioredoxin under basal conditions. Following oxidative stress, the complex dissociates, thus increasing free TXNIP which has been suggested to directly regulate NLRP3 inflammasome induced caspase-1 activation. In addition, metabolic stress, can directly regulate TXNIP expression in addition to providing a source of oxidative stress thereby increasing TXNIP in two ways- increasing free TXNIP, and increasing transcription of TXNIP.

Figure 1.10

55 Figure 1.10: Activation of the inflammasome is a two-step process requiring a priming step (Signal 1) and an activating step (Signal 2) for efficient production of mature, biologically active IL-1β. Priming by TLR ligands and other NFκB activators, induces gene transcription of rate limiting pro-inflammatory proteins such as NLRs and pro-IL-1β. Inflammasome assembly and pro-caspase-1 cleavage into its active form is induced by a variety of signal 2’s which either induce K+ efflux and/or perturbations in lysosomal integrity.

Figure 1.11

57 Figure 1.11: Mitochondria: integrators of metabolic stress and activators of the

NRLP3 inflammasome. DAMPs and PAMPs, directly or via the alteration of cellular metabolism, induce partial mitochondrial dysfunction resulting in the elevated production of ROS. Via an unknown mechanism downstream of ROS, the NLRP3 inflammasome Is activated and triggers (1) an extracellular inflammatory response through processing of the cytokines IL-1β and IL-18, and

(2) an intracellular response that alters cellular glucose and lipid metabolism.

Reference: Tschopp, J, et Mitochondria: Sovereign of Inflammation? Eur. J.

Immunol. 2011. 41: 1196–1202 Used with permission (Appendix 1).

Figure 1.12

59 Figure 1.12: Structure and Function of P2X7R channels and Intracellular

Signaling Cascade (Dubyak Lab Files).

Table 1.3

61 C)

Table 1.3 cont’d

62 Table 1.3: (A) Relative binding affinity for the various P1, Adenosine receptors and their G-protein coupling. (B) P2X Receptor relative binding affinities for ATP.

Figure 1.17

Figure 1.13

64 Figure 1.13: Schematic illustration of the key components of purinergic signaling in the subretinal microenvironment. Stimulation of P2 receptors on the RPE can enhance transepithelial fluid absorption while P1 receptors can modulate

− phagocytosis. ATP released through CFTR and other Cl channels can stimulate

P2 receptors or be converted to ADP, AMP, and adenosine (Ado) by a series of ectonucleotidases present on the apical membrane of the RPE. By controlling the balance of extracellular purines available to stimulate these receptors these mechanisms can control levels of endogenous purines available to activate the receptors. While theoretically possible, it remains to be determined whether these subretinal purines can actually stimulate photoreceptors.

Reference: Mitchell, CH, and Reigada, David; Purinergic Signalling in the subretinal space, Purinergic Signalling (2008) 4:101–107 Used with permission

(Appendix 1).

Figure 1.14

66 Figure 1.14: Scheme of the autocrine glutamatergic-purinergic signaling cascade involved in the VEGF-induced inhibition of Müller cell swelling. Activation of

KDR/flk-1 by VEGF evokes a calcium-, phospholipase C (PLC)-, protein kinase C

(PKC)-, and Src kinase-dependent exocytotic release of glutamate from Müller cells. Voltage-gated sodium channels mediate rapid fluctuations of the membrane potential required for the activation of voltage-gated calcium channels implicated in the exocytosis of glutamate-containing vesicles. Glutamate activates metabotropic glutamate receptors (mGluRs) that results in a calcium- independent release of ATP from Müller cells. ATP is extracellularly converted by the nucleoside triphosphate diphosphohydrolase-2 (NTPDase2) to ADP that activates P2Y1, resulting in nucleoside transporter-mediated release of adenosine. Activation of A1 adenosine receptors causes a cAMP-, protein kinase

A (PKA)-, and phosphatidylinositol-3 kinase (PI3K)-dependent opening of potassium and chloride channels; the ion efflux equalizes the osmotic gradient across the plasma membrane and thus prevents water influx and cellular swelling under hypo-osmotic stress conditions. In swollen cells, the ion efflux is associated with a water efflux, resulting in decreased cell volume. While the release of glutamate from Müller cells is calcium-dependent, all steps of the cascade after activation of mGluRs are calcium-independent. Neuron-derived glutamate and ATP may activate the volume-regulatory signaling cascade in dependence on the neuronal activity. In the murine retina, activation of P2Y4, 6 by UTP and UDP, respectively, might result in a release of neuronal ATP that activates glial P2Y1. Müller cell-derived glutamate and adenosine may also

67 activate neuronal a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainate

(AMPA/KA) and A1 receptors, resulting in stimulation and inhibition, respectively, of neuronal activity. In the retinal parenchyma of diabetic rats, NTPDase1 (which hydrolyzes ATP and ADP about equally well) is upregulated, and extracellular formation of adenosine contributes to the swelling- inhibitory effect of glutamate and ATP.

Reference: Wurm, Antje, et al, Purinergic Signaling involved in Müller cell function in the mammalian retina Progress in Retinal and Eye Research Volume

30, Issue 5, September 2011, Pages 324-342 Used with permission (Appendix

1).

Figure 1.15

69 Figure 1.15: Proposed model for high glucose -induced caspase-1 activation in retinal Müller cells.

CHAPTER 2

MATERIALS AND METHODS

71 2.1 Materials

Caspase-1 antibody was purchased from Invitrogen (Carlsbad, CA). TXNIP (Anti-

VDUP-1/TXNIP) antibody was purchased from MBL (Woburn, MA). Anti-Glut1 antibody was from AbCam (Cambridge, MA). Primary antibodies against rat

P2X7R (APR-004 and APR-008) were from Alamone Labs (Jerusalem, Israel).

Goat Anti-Mouse IgG and Goat Anti-Rabbit IgG conjugated to HRP, and anti-beta actin primary antibody were purchased from Santa Cruz Biotechnology (Santa

Cruz, CA). Anti-RIP2/RICK primary antibody was from ProSci, Inc. (Poway, CA).

Goat anti-rabbit IgG conjugated to FITC and fura2-AM were from Invitrogen

(Carlsbad, CA). Goat Anti-Rabbit IgG conjugated to horseradish peroxidase

(HRP) and 4-20% Gradient Tris- SDS-PAGE gels were purchased from BioRad

(Hercules, CA). Nucleotides, adenosine deaminase (ADA), apyrase (grade I), forskolin, dipyridimole, TRIzol, 7-amino-4-trifluoro-methylcoumarin (AFC), ethidium bromide and minocycline were obtained from Sigma-Aldrich Chemical

Co. (St. Louis, MO). The caspase-1 inhibitors YVAD-fmk, caspase peptide substrates (AFC-coupled), elastase and mouse anti-vimentin were purchased from Calbiochem (San Diego, CA). IL-1 receptor antagonist (IL-1ra) and recombinant rat IL-1β were purchased from R&D Systems (Minneapolis, MN). IL-

1β ELISAs were from Pierce Thermo Scientific (Rockford, IL). RT2 SYBR

Green/ROX qPCR Master Mix (PA-012) and predesigned qPCR primers for rat

IL-1β (PPR06480A-200), rat NLRP3 (PPR56639A-200), rat NLRP1

(PPR45641A-200), rat ASC/Pycard (PPR06566A-200), rat TXNIP (PPR42837A-

200), rat GAPDH (PPR06557A), rat caspase-1 (PPR06427A), rat P2X7R

(PPR44893A), rat P2Y2R (PPR45235B), and rat 18S ribosomal RNA

72 (PPR57734E) were purchased from SA BioSciences (Frederick, MD). RNeasy

Mini Kits were from Qiagen (Valencia, CA). Transcriptor first strand cDNA synthesis kit was from Roche (Indianapolis, IN). 5’-N-ethylcarboxamido- adenosine (NECA), P2X7 R antagonist A4387079, P2X7R antagonist

AZ10606120, suramin, 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX), MRS1754, and SCH442416 were from TOCRIS (Ellisville, MO). S-(4-nitrobenzyl)-6- thioguanosine (NBTG) was from RBI (Natick, MA). Bradford protein assay reagent was from BioRad (Hercules, CA). Sulfo-NHS-Biotin and enhanced chemiluminescence (ECL) detection reagent were from Pierce (Rockford, IL).

73 2.2 Methods

Animal Models: Caspase-1-/- (Cas-1-/-) (gift from Dr. T. McCormick, Case

Western Reserve University: in a C57BL/6 background) and IL-1R1-/- (Jackson

Laboratories; strain name: B6.129S7-Il1r1tm1jm in a C57BL/6 background) mice were bred using homozygous breeding pairs. Male mice weighing 20 g from wildtype C57BL/6 the Cas-1-/- strain, or the IL-1R-/- strain were randomly assigned to be either diabetic or normal controls. Streptozotocin injections

(60mg/kg body wt i.p. on 5 consecutive days) were utilized to induce diabetes

(138). Galactosemia was induced by feeding normal mice a diet enriched with

30% galactose as previously described (138). Diabetic animals were maintained with insulin injections (0.1-0.2 units of Neutral Protamine Hagedorn (NPH) insulin subcutaneously, two to three times a week) as needed. Animals had free access to food and water, were caged in pairs, and were maintained under a 14 h on/10 h off light cycle. Treatment of animals conforms to the Association for Research in Vision and Ophthalmology Resolution on Treatment of Animals in Research and euthanasia was performed via peritoneal injection of Fatal-Plus. Animals with fasted blood glucose levels 300 mg/dl were used in these studies. Fasted blood glucose levels were estimated by measuring non-enzymatically glycated hemoglobin (GHb) levels using affinity chromatography (Glyc-Affin; Pierce,

Rockford, IL).

Histologic assessment of retinal vascular pathology. Isolated retinas were washed in running water overnight and then digested in elastase solution

74 (0.4U/ml in 100 mM sodium phosphate buffer, 150 mM sodium chloride, 5mM

EDTA) for 30-45 min at 37°C. The retinal tissue was transferred into 100mM Tris-

Hydrochloric Acid buffer (pH 8.5) and left overnight at RT. The cleaned vessel network was dried onto a mounting slide, stained with hematoxylin and PAS, dehydrated and covered with coverslip. Acellular capillaries were quantitated in

4-7 field areas in the mid-retina (20X magnification) in blinded fashion as previously described (207). The number of acellular capillaries is reported per mm2 retina.

Tissue Culture Models:

rMC-1: The rat retinal Müller cell line (rMC-1) utilized in these experiments was originally obtained from Dr. V.R. Sarthy (Northwestern University, Evanston

IL). These cells have been well characterized and established by both our laboratory and others as an excellent tool for retinal Müller studies (112, 138,

180). rMC-1 cells were grown in normal (5 mM glucose) DMEM growth medium containing 10% FBS and 1% penicillin/streptomycin (P/S) at 37°C and 5% CO2 in a humidified incubator. Experiments were done with passages lower than 37.

Human retinal Müller cells (hMC): Handling of human tissue conformed to the tenets of Declaration of Helsinki for research involving human tissue. Human

Müller cells were isolated from retinal tissue of healthy donors with no history of diabetes and characterized as previously described (112, 165, 224). Briefly, isolation of hMC was performed by successive trypsin splits (0.25% trypsin) and

75 grown in normal glucose (5 mM) DMEM/Ham’s F12 (1:1 ratio) containing 10%

FBS and 1% P/S. Following the third trypsin split, cultures were 95% pure for

Müller cells as determined by vimentin and CRALBP as positive stains for Müller cell identification and GFAP as a negative stain. Experiments were only performed with cells from passages 3-9.

HEK-rP2X7R: HEK293 cells stably transfected with the wild-type rat

P2X7R (HEK-rP2X7 cells) were selected for, and maintained in, DMEM (25mM

Glucose, 10% CS, 1% l-glutamine, 1% P/S) supplemented with 400 µg/ml G418 at 37°C and 5% CO2 in a humidified incubator. Expression of rP2X7R was verified using western blot and FACS analysis. Parental HEK293 cells that lack native P2X7R expression were used as negative controls.

Cell Culture Treatment Protocols

High glucose treatment: rMC-1 (1x106) and hMC (5x106) were treated with either normal (5mM) glucose DMEM or high (25mM) glucose DMEM supplemented with 2% FBS, 1% P/S for 12, 24, 48, 72, or 96 hours. For experiments extending beyond 24 hours, medium was changed every day to maintain glucose levels to the end of the respective experiments. Cells treated with normal (5 mM) glucose DMEM supplemented with 2% FBS served as controls.

76 Cytokine Treatment: rMC-1 (1x106 / 10 cm dish) were incubated in normal

(5mM) DMEM supplemented with 2% FBS, 1% P/S or normal (5mM) DMEM plus exogenous rat IL-1β (2ng/ml) for 24 hours.

Drug washout studies: rMC-1 (1x106 / 10 cm dish) were incubated in high glucose media plus either 100µM YVAD-fmk (an irreversible caspase-1 inhibitor), or 100µM minocycline for 28 hours. Drug-containing media was washed out and replaced with high glucose media for the remainder of the experiments up to 96 hours. For IL-1ra studies, rMC-1 were pretreated in high glucose DMEM, 0.1%

BSA, and 1% penicillin/streptomycin) plus 50ng/ml IL-1ra for one hour. Following pretreatment with IL-1ra, rMC-1 were incubated in high glucose medium for up to

96 hours as described above.

High glucose washout studies: rMC-1 (1x106 / 10 cm dish) were incubated in high glucose media for 24 hours or 48 hours, respectively. After this initial high glucose medium treatment, high glucose media was washed out and replaced with normal glucose medium. At 72 hours, cells were collected and lyzed.

Treatment with exogenous nucleotides or NECA: rMC-1 (1x106/ 10 cm dish) were transferred to either normal glucose or high glucose DMEM with 2%

FBS, 1% P/S. The DMEM was then supplemented as indicated in particular experiments with various nucleotides (ATP, ADP, or UTP at 10 µM -10 mM final concentration), or the adenosine analog NECA (at 10-100 µM final concentration) and the treated cells were incubated for 24 hours prior to experimental assays.

77 Treatment with nucleotide/nucleoside scavenging enzymes, P1/P2 receptor antagonists, or nucleoside transport inhibitors: rMC-1 (1x106/ 10 cm plate) were transferred to either normal or high glucose DMEM supplemented with 2% FBS, 1% P/S and then incubated in the presence or absence of either apyrase (5 units/ml), adenosine deaminase (ADA) (2U/ml), suramin (10 µM),

A438079 (10-20 µM), AZ 10606120 (10 µM), MRS1754 (10 µM), SCH442416 (10

µM), DPCPX (10 µM), NBTG (80 µM) or dipyridamole (5 µM) for 24 hours prior to experimental analysis.

Preparation of cell lysates/samples:

Whole cell lysate preparation: Following the incubations/ treatments in normal or high glucose DMEM described above, rMC-1 (1x106) were lysed in

200µl of CHAPS lysate buffer (100mM HEPES, pH 7.5, 10% sucrose, 0.1%

CHAPS, 1mmol/l EDTA, 1 mM PMSF and leupeptin (10µg/ml)]. Cell extracts remained on ice for at least an hour prior to sonication and centrifugation at

10,000 RPM for 10 minutes at 4°C (137). Supernatants were retained, assayed for protein content by the Bradford method, and then used for caspase-1 activity measurements.

RIPA Lysis (for TXNIP Western Blots): RIPA Lysates were prepared as previously described (167). Briefly, after treatment with either normal or high glucose for the time points indicated, rMC-1 (250,000) were washed once with ice cold PBS. Cells were then lyzed with 150 µl of radioimmunoprecipitation

78 assay (RIPA) extraction buffer [RIPA Buffer (1% Nonidet P-40, 0.5% Na- deoxycholate, 0.1% SDS (pH 7.4) in PBS) containing DTT (1mM), PMSF (1 µM), leupeptin (2µg/ml), and aprotinin (2µg/ml)] for 5 minutes and then gently scraped from 6-well (2 cm wells) plates. Cells were incubated in RIPA lysis buffer for 15 minutes on ice before centrifugation at 10,000 RPMs for 10 minutes at 4°C.

Supernatants were retained to be used in experiments and subjected to Bradford

Protein measurement.

Membrane protein extraction and preparation: Following treatment, rMC-1,

HEK-rP2X7R, or control HEK293 cells were washed with ice cold PBS, permeabilized with 20µl of digitonin (5mg/ml) for 5 min at 37ºC to release soluble cytosolic proteins, and washed again with ice cold PBS. The permeabilized cells were lysed with 150 µl of radioimmunoprecipitation assay extraction buffer

(RIPA) (1% Nonidet P-40, 0.5% Na-deoxycholate, 0.1% SDS (pH 7.4) in PBS) containing PMSF, leupeptin, and aprotinin for 15 min. on ice. Lysates were centrifuged at 10,000x g for 10 min at 4ºC. Supernatants were retained, assayed for protein content by the Bradford method, and then prepared Western

Blot Analysis.

Biotinylation of membrane surface proteins: rMC-1 or HEK-rP2X7R cells were subjected to cell surface biotinylation and streptavidin pull down as described (77). Briefly, adherent cells were trypsinized and resuspended

(~25x106 cells/ml) in 1 ml of ice cold PBS (pH 8.0). 1mg/ml of Sulfo-NHS-Biotin reagent was added and the suspensions incubated at 4ºC with continuous rotation. Biotinylation was quenched with PBS (pH 8.0) containing 50mM glycine.

79 Samples were then subjected to RIPA lysis as described above prior to streptavidin pull down. 50µl of unprecipitated RIPA lysate was retained as a whole cell lysate sample and the remaining 100µl was processed for incubation with streptavidin beads (200µl of 50% bead slurry per sample). The sample/ bead slurries were incubated at 4ºC for 2 hours with continuous rotation prior to centrifugation at 10,000 x g for 10 min. The bead pellets were washed 3 times by resuspension (in 500 µL PBS) and recentrifugation prior to addition of 20 µl SDS sample buffer and boiling for 5 min to elute bound proteins. Samples were then subjected to SDS-PAGE and Western blot analysis for P2X7R.

Caspase-1 Activity Assay: Caspase-1 activities were measured as described previously (112, 114, 207, 212, 221). Briefly, equal amounts of sample (15 µg) were incubated in the presence of fluorogenic peptide substrates (2.5 µM) for caspase-1 at 32°C for one hour while shaking. The cleavage of substrate by active caspase-1 in the samples causes release of a fluorescent product (AFC) which was measured by a Tecan Spectra FluorPlus fluorescence plate reader

(excitation: 400 nm, emission: 510 nm). Caspase-1 activities were calculated against an AFC standard curve and expressed as pmol/AFC/mg protein/min.

FACS analysis for P2X7R expression: rMC-1 and HEK-rP2X7R were stained for P2X7R following treatment with anti-P2X7 extracellular domain primary antibody (10 µg/ml, Alamone, APR-008) for 30 min at 4°C and then Alexa Fluor

80 594-conjugated secondary Ab for 30 min at 4°C (39). Cells were washed, fixed

(1% paraformaldehyde) and analyzed on an EPIC flow cytometer (79).

Western Blot Analysis:

For Caspase-1: Under reduced conditions, equal amounts of samples (15

µg/ml of protein as determined by the Bradford Protein Assay) were loaded onto

4-20% Gradient Tris- SDS-PAGE gels (BioRad) and separated (90 min., 40 V) followed by transfer to nitrocellulose (90 min. at 100 V) at 4ºC. Following transfer, membranes were blocked in 5% non-fat dry milk made with TBS-T for at least 1 hour. Membranes were incubated with primary antibody anti-caspase-1 (1:800) or anti-β actin (1:10,000) overnight at 4ºC. Membranes were washed 5 times for

20 minutes with TBS-T and incubated with the secondary HRP-conjugated antibodies (1:3,000) for one hour at room temperature. Membranes were washed

5 times for 20 minutes with TBS-T and developed using enhanced chemiluminescence HRP detection reagent. Bands were quantified via densitometry analysis using Biorad quantity one program and expressed as the ratio of caspase-1/ actin band densities.

For TXNIP and P2X7R: Under reduced conditions, equal amounts of samples (50µg/ml of protein as determined by the Bradford Protein Assay) were loaded onto 15% Tris-SDS-PAGE gels and separated (180 min. 35 mA) followed by transfer to Polyvinylidene Fluoride (PVDF) (54 min. at 24 V). Following transfer, membranes were blocked in 2% non-fat dry milk made with TBS-T for at

81 least 1 hour. Membranes were incubated with primary antibody anti-TXNIP

(1µg/ml), anti-rP2X7R (APR-004, 0.3µg/ml or APR-008, 0.9 µg/ml), anti-GLUT1

(1µg/ml), or anti-β actin (1:10,000) overnight at 4ºC. Membranes were washed 3 times for 5 minutes with TBS-T and incubated with the appropriate secondary

HRP conjugated antibodies (1:5,000) for one hour at room temperature.

Membranes were washed 3 times for 5 minutes with TBS-T and developed using chemiluminescence HRP detection reagent.

For RIP2: Under reduced conditions, 10 µg/ml of each sample was loaded onto 10% SDS-PAGE gels. SDS-PAGE gels were blotted onto nitrocellulose membranes which were then exposed to primary anti-RICK/RIP2 antibody

(1:1,000) or anti-β actin (1:500) overnight at 4ºC. Membranes were washed 5 times for 5 minutes with TBS-T and incubated with appropriate secondary HRP conjugated antibodies (1:3,000) for one hour at room temperature. Following detection, band intensities were measured using densitometry analysis. Data are expressed as arbitrary band intensity units. As a control, a blocking peptide against RIP2 was used to determine specificity of the Anti-RICK/RIP2 antibody.

Cytokine Assays: IL-1β ELISAs were performed according to the manufacturer’s directions. Briefly, equal amounts of medium (75µl) from rMC-1 and hMC treated with normal or high glucose containing medium as described above were assayed for active, secreted IL-1β. Samples were added to pre- coated 96 well plates followed by the addition of 50µl biotinylated antibody

82 reagent. Plates were sealed tightly and incubated for 2 hours at room temperature. Plates were washed and 100µl Streptavidin-HRP solution was added. Plates were sealed tightly again and incubated for 30 min at room temperature. Plates were washed again. 100µl TMB substrate solution was added and enzymatic reaction was allowed to develop for 30 min at room temperature in the dark. 100µl of stopping solution was added. Absorbance was measured at 450 nm and 550 nm using a Tecan Spectra FluorPlus plate reader.

Values were compared to a standard curve and normalized to milligrams of total protein.

qPCR analysis of mRNA transcripts encoding inflammasome proteins,

TXNIP, rP2X7R, rP2Y2R and TXNIP. Total RNA was extracted by TRIZol reagent from rMC-1 cells incubated for 24 h in normal or high glucose DMEM ± various agonists and antagonists as described previously (163). RNA was purified and then concentrated using an RNeasy Mini Kit. A Transcriptor First

Strand cDNA Synthesis kit was utilized for synthesis of first-strand cDNA from purified RNA. Quantitative PCR (qPCR) analysis of IL-1β, NLRP3, NLRP1, ASC,

P2X7R, P2Y2R, TXNIP, caspase-1, GAPDH or 18S ribosomal RNA was performed using a StepOne-Plus Real-Time PCR System (Applied Biosystems).

Reactions were performed in 25 µl reaction volumes containing RT2 SYBR

Green/ROX qPCR Master Mix (12.5µl), 1:100 dilution of RT product, and 1µM

PCR primer pair stock and run in duplicate. Amplification cycle conditions were

95°C for 10 minutes followed by 40 cycles of (95°C, 15 sec; 55°C, 30-40 sec; and

83 72°C, 30 sec.). Melt curves were performed at the end of the reaction with all products demonstrating one predominant peak. Relative expression was calculated using the ΔΔCt method using StepOne software v. 2.1 with values normalized to the reference genes GAPDH or 18S rRNA.

Measurement of nucleotide-induced Ca2+ mobilization: HEK-rP2X7R or rMC-

1 were trypsinized to generate cell suspensions (106/ml) and then incubated with

1µM fura2-AM for 30-60 min. Cytosolic [Ca2+] in the fura2-loaded cell suspensions before and after stimulation with ATP, ADP, or UTP (1µM-1mM) was measured and calibrated using a fluorimeter as previously described (79).

Because rMC-1 and HEK293 cells express Gq-coupled, Ca2+-mobilizing P2Y2 receptors, analysis of possible P2X7R-mediated Ca2+ influx was assayed by first treating the cells with 30 µM UTP to activate and desensitize the P2Y2 receptors before stimulation of P2X7 by the indicated concentrations of ATP.

Measurement of P2X7R-mediated ethidium dye uptake: HEK-rP2X7R and rMC-1 were assayed for P2X7R-mediated changes in ethidium accumulation and were corrected for background dye fluorescence as previously described (79).

Data are expressed as percentages of the maximal fluorescence observed when the cells were permeabilized with 0.003% digitonin.

84 Measurement of ATP-stimulated K+ efflux by atomic absorbance spectrophotometry: rMC-1 or HEK293-rP2X7 cells were plated in 12-well dishes and incubated in either normal or high glucose DMEM for 24 h. The cell monolayers were washed once with PBS and then bathed in 1 ml of a basal salt solution BSS (130 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.5 mM CaCl2, 25 mM

NaHEPES, pH 7.5, 5 mM glucose, 0.1% bovine serum albumin) for 5 min at

37ºC. Cells were then stimulated without or with ATP (1-5mM) for 10 min. at

37ºC. ATP containing medium was rapidly aspirated to terminate K+ efflux reactions and replaced with 1 ml of 10% nitric acid at room temperature for 3-4 hrs to extract K+ from the cell monolayers. K+ content was quantified using atomic absorbance spectroscopy (93).

Data Processing and Statistical Analysis: For caspase-1 experiments, either the average activity or the fold change in caspase-1 activity were calculated by normalizing caspase-1 activity in treated samples (high glucose and a variety of agonists/antagonists) to the paired controls (normal glucose treated samples) for each experiment and graphed as mean± SEM. For qPCR experiments, ΔΔCt values were calculated compared to normal glucose controls and graphed as mean± SEM. In atomic absorbance spectrophotometry experiments, intracellular

K+ concentrations were calculated compared to a K+ standard curve and data were expressed as mean ± SEM. Statistical significance was determined by

ANOVA One-Way analysis by comparing values in treated groups to paired normal glucose control values (correlated samples, p<0.05), followed by Tukey’s

85 post analysis test. For ordinal data, the Kruskal-Wallis test (p<0.05) followed by

Dunn’s post analysis was used to determine statistical significance among groups. For details in statistical analysis see VasserStats Statistical Computation

Web Site. Statistical assistance provided by Case Statistical Consulting Center,

Department of Statistics, Case Western Reserve University.

CHAPTER 3

PURINERGIC REGULATION OF HIGH GLUCOSE- INDUCED CASPASE-1 ACTIVATION IN RAT RETINAL MüLLER CELLS (rMC-1)

Portions of this chapter have been published in AJP-Cell:

Trueblood, K.E., S. Mohr, and G.R. Dubyak, Purinergic regulation of high glucose-induced caspase-1 activation in the rMC-1 rat retinal Müller cell line. Am J Physiol Cell Physiol.

87 3.1 Introduction

Chronic activation of pro-inflammatory caspase-1 in the retinas of diabetic animals and patients in vivo and retinal Müller cells in vitro is well-documented

(16, 18, 57, 100, 111, 138, 189, 207, 223, 224). We previously reported that inhibition of caspase-1/interleukin-1β signaling prevents degeneration of retinal capillaries in diabetes (41, 66, 99, 138, 207). However, the exact mechanism(s) by which hyperglycemia initiates the beginning stages of chronic caspase-1 activation in Müller cells was not elucidated. In this study we characterized how elevated glucose and extracellular purines contribute to the activation of caspase-1 in a cultured rat Müller cell model (rMC-1).

The expression of P2X7R and other purinergic receptors has been described in Müller cells, from some but not all species, and in other retinal cell types with proposed roles in osmotic regulation or metabolism of retinal neurotransmitters (220). Given the established role of P2X7R in driving

NLRP3/caspase-1 activation in hematopoietic cells (as described in chapter 1), and the reported ability of hyperglycemia to upregulate NLRP3/caspase-1 signaling in non-hematopoietic cells, we designed experiments to characterize the potential contribution of purinergic signaling to high glucose induced caspase-1 activation in the rat retinal Müller cell line (rMC-1). Although upregulation of P2X7R expression and function occurs in various diabetic tissues, it was important to consider whether metabolic stress induction of caspase-1 inflammasomes may alternatively involve P2X7R-independent but purine-dependent pathways (153, 193). Indeed, our findings suggest that

88 autocrine stimulation of adenosine-sensing P1 receptors and P2 receptor subtypes other than P2X7R mediate the caspase-1 activation response of Müller cells to high glucose. Moreover, the ability of forskolin to mimic the stimulatory effects of high glucose or purinergic agonists is consistent with a likely role for cAMP signaling in the regulation of caspase-1 in this non-hematopoietic glial cell model.

89 3.2 Results

3.2.1 Treatment of rMC-1 cells with extracellular apyrase or adenosine deaminase suppresses high glucose-induced caspase-1 activation

Purinergic regulation of caspase-1 activation in myeloid hematopoietic cells has been well documented (94). To determine whether extracellular purines are involved in high glucose-induced caspase-1 activation in non-hematopoietic

Müller glial cells, we cultured the rat retinal Müller cell line, rMC-1, in high glucose medium supplemented with potato apyrase, which is a broad spectrum

ATP scavenger that serially metabolizes ATP to ADP and then AMP. Apyrase treatment significantly reduced high glucose induced caspase-1 activation (Fig.

3.1A). In addition to ATP, extracellular adenosine can regulate pro-inflammatory processes (92). Scavenging of endogenous extracellular adenosine by treatment with adenosine deaminase (ADA), which rapidly converts adenosine to inosine, also significantly reduced high glucose-induced caspase-1 activation (Fig. 3.1B).

The ability of ATP- and adenosine-scavenging enzymes to suppress high glucose-induced caspase-1 activation suggested that autocrine stimulation of both adenosine-sensing P1 receptors and ATP-sensing P2 receptors may cooperatively mediate this response of rMC-1 cells to elevated glucose.

3.2.2 Treatment of rMC-1 cells with exogenous ATP, exogenous adenosine analog, or adenosine uptake inhibitors mimics high glucose-induced caspase-1 activation

Because apyrase reduced high glucose-induced caspase-1 activation, we tested whether exogenous ATP (the primary agonist for P2X7R and other P2XR

90 and P2YR subtypes), might directly induce caspase-1 activation in retinal Müller cells incubated under normal glucose conditions. The caspase-1 activation response of rMC-1 cells treated with 5 mM ATP in normal glucose medium was similar in magnitude to that induced by high glucose medium in the absence of exogenous ATP (Fig. 3.2A); inclusion of 5 mM ATP in the high glucose medium did not produce an additive or synergistic effect. Caspase-1 activation was increased by extracellular ATP in a concentration dependent manner under normal glucose conditions (Fig. 3.2B) but required supra-millimolar ATP. This may reflect a requirement for sustained elevation of extracellular ATP -- averaged over the 24 hour test period -- in the 10 – 100 µM concentration range that facilitates maximal activation of most P2 receptor subtypes (75, 82, 86).

Indeed, a 5 mM pulse of exogenous ATP was rapidly (90% clearance within 6 h) metabolized by robust ecto-ATPase activities associated with both the rMC-1 cell surface and the 2% FBS component of the DMEM; rMC-1 cultured in normal or high glucose medium exhibited similar rates of exogenous ATP clearance (Fig.

3.2E).

A requirement for supra-millimolar ATP could also indicate a role for sustained accumulation of extracellular ADP which is the preferred agonist for the P2Y1, P2Y12, and P2Y13 receptor subtypes (209) However, inclusion of 5 mM ADP in the normal glucose DMEM did not significantly increase caspase-1 activity (Fig 3.2A). Moreover, co-incubation with both high glucose and 5 mM exogenous ADP partially attenuated the caspase-1 activation to high glucose alone (Fig. 3.3A). Supra-millimolar ATP could also drive the ecto-nucleotide

91 diphosphokinase-mediated transphosphorylation of released UDP to generate extracellular UTP, a potent agonist for the P2Y2 and P2Y4 receptor subtypes.

Like ADP but unlike ATP, 5 mM UTP alone did not increase caspase-1 activation in rMC-1 cells incubated in control glucose medium but did attenuate the increase in caspase-1 elicited under high glucose conditions (Fig. 3.2A, 3.3B).

Given that adenosine deaminase attenuated high glucose-induced caspase-1 activation (Fig. 3.1B), we tested whether exogenous adenosine also stimulates caspase-1 under control glucose conditions. Because extracellular adenosine per se can be rapidly cleared by endogenously expressed adenosine deaminase or nucleoside transporters, rMC-1 cells were treated with varying concentrations (10-100 µM) of 5'-N-ethylcarboxamido-adenosine (NECA), a stable adenosine analog and non-selective agonist for all four adenosine receptor subtypes (68). At 10 or 30 µM, NECA significantly increased caspase-1 activity compared to normal glucose controls but not to the magnitude observed with high glucose treatment (Fig. 3.2D). Notably, the efficacy of NECA did not increase, but modestly decreased, as the test concentration was elevated to 100

µM. Co-incubation of NECA with UTP did not induce an increase in caspase-1 activity significantly differently than NECA alone (Fig. 3.2F)

To determine whether increased levels of endogenously produced extracellular adenosine also stimulate caspase-1 activity, we treated rMC-1 cells with dipyridamole (Dipyr) or S-(4-nitrobenzyl)-6-thioguanosine (NBTG) which inhibit the nucleoside transporters that can facilitate reuptake of extracellular adenosine (68). Both reagents stimulated caspase-1 activation in normal glucose

92 treated rMC-1 and this effect was attenuated when the cells were treated with either reagent in combination with ADA (Fig. 3.2D).

3.2.3 Treatment of rMC-1 cells with suramin or adenosine receptor antagonists suppresses high glucose-induced caspase-1 activation

Given the stimulatory effects of exogenous ATP, NECA, and the adenosine reuptake inhibitors, we hypothesized that antagonism of either P2 nucleotide receptors or P1 adenosine receptors would reduce high glucose- induced caspase-1 activation. The non-selective P2 receptor inhibitor suramin, which targets the widely expressed P2Y2 receptor and other P2 subtypes (209), reduced high glucose-induced caspase-1 activation and also basal caspase-1 activity (Fig. 3.4A). 8-cyclopentyl-1, 3-dipropylxanthine (DPCPX), which acts as a non-selective adenosine receptor antagonist at high (>1 µM) concentrations

(68) , also markedly reduced high glucose caspase-1 activation (Fig. 3.4B). The effect of 10 µM DPCPX was mimicked by the A2B antagonist MRS1754 (MRS)

(also at 10 µM). The A2A selective antagonist SCH442416 (SCH) (at 10 µM) also significantly reduced high glucose-induced caspase-1 activation but with reduced efficacy relative to DPCPX or MRS1754.

3.2.4 No significant role for P2X7R receptors in purinergic regulation of high glucose-induced caspase-1 activation

Supra-millimolar concentrations of extracellular ATP were required for activation of caspase-1 in rMC-1 cells under normal glucose conditions (Fig.

93 3.2B). P2X7R is the P2 subtype with the lowest ATP affinity and is also the purinergic receptor unequivocally linked to regulation of caspase-1 activation in hematopoietic cells (51, 58). Thus, we characterized P2X7R expression and its possible contribution to high glucose-induced caspase-1 activation in rMC-1 cells using two selective antagonists (29). Although treatment with the P2X7R selective antagonist A438079 significantly reduced high glucose induced caspase-1 activation, a structurally distinct selective P2X7R antagonist,

AZ10606120 (AZ) had no effect (Fig. 3.4C). This suggested that the ability of

A438079 to attenuate caspase-1 activation may reflect an off-target effect rather than specific antagonism of P2X7R.

We performed qPCR analysis for P2X7R, P2Y2R and TXNIP mRNA expression in normal and high glucose treated cells. TXNIP was assayed as a positive control gene with a known glucose-responsive promoter and because of its link to hyperglycemia-induced mitochondrial dysfunction and caspase-1 activation (4, 27, 74, 103, 156, 161, 181, 204, 232). Figure 3.5A shows that high glucose elicited several-fold increases in TXNIP and P2X7R transcript levels.

Western blot analysis of whole cell lysates indicated that the increased TXNIP mRNA levels were matched by increased TXNIP protein expression confirming the glucose responsiveness of rMC-1 cells (Fig. 3.5B, lower right panel). In contrast, western blot analysis of either whole cell lysates (not shown) or membrane-enriched extracts from digitonin-treated samples (Fig 3.5B, left panel) failed to reveal measurable levels of P2X7R protein in control or high glucose- treated rMC-1 cells. These analyses used two different P2X7R antibodies: one

94 which recognizes intracellular C-terminal residues from the full-length 595-amino acid rat P2X7R and another raised against an amino acid sequence from the extracellular loop of P2X7R (Fig. 3.5B, upper right panel). The latter antibody can recognize both full-length P2X7R and recently identified P2X7R splice variants with truncated intracellular C-termini (2). Positive and negative control samples verified that both antibodies detected a 78 kDa band in extracts from HEK293 cells stably transfected with rat P2X7R cDNA (HEK-P2X7R), but not from wildtype HEK293 cells. Membrane extracts from both control and high glucose- treated rMC-1 cells were positive for the Glut1 glucose transporter, an intrinsic membrane protein expressed by Müller cells (Fig. 3.5B). To enrich for possible low abundance cell surface P2X7R proteins, control or high glucose-treated cells were incubated with a membrane-impermeant biotinylating crosslinker prior to detergent extraction and precipitation with streptavidin-agarose beads. However, anti-P2X7R western blots of the streptavidin precipitates again failed to detect immunoreactive protein in the rMC-1 samples. (Fig. 3.5D).

As a well-characterized functional readout of expressed P2X7R, we compared the ability of supra-millimolar ATP to induce K+ efflux from the rMC-1 cells versus HEK-rP2X7R cells (93). While HEK-rP2X7R cells released >50% of their intracellular K+ pool during a 10-min stimulation with 1, 2.5, or 5 mM ATP, no significant decreases in intracellular K+ were observed in either control or high-glucose treated rMC-1 cells challenged with 5 mM ATP (Fig. 3.5C). As an additional index of P2X7R function, we assayed ATP-triggered induction of non- selective pores permeable to organic molecules up to 900 Da in mass (51, 58,

95 79). No ATP-induced accumulation of fluorescent ethidium-DNA complexes, indicative of increased influx of the otherwise size-excluded ethidium+ dye (390

Da), was observed in control or high glucose-treated rMC-1cells; this contrasted with the ATP-triggered dye uptake in positive control murine bone marrow- derived macrophages (3.5E).

3.2.5 Activation of caspase-1 by high glucose, NECA, or ATP is mimicked by forskolin and correlated with increased gene expression of caspase-1 and TXNIP

The antagonistic effects of MRS1754 and SCH442416 (Fig. 3.4B) indicated that the adenosinergic component of high-glucose induced caspase-1 activation may reflect agonistic input from A2 receptor subtypes coupled to

Gs/adenylyl cyclase/cAMP signaling (68) . Notably, treatment of rMC-1 cells with forskolin (FSK), a direct activator of adenylyl cyclase, mimicked the ability of high glucose treatment or NECA to increase caspase-1 activation (Fig. 3.6A).

Almeida et al. recently reported that treatment of osteoblasts with FSK or Gs- coupled prostaglandin E receptor agonists increased caspase-1 activity secondary to the cAMP-dependent stimulation of caspase-1 gene expression (5).

We observed that forskolin similarly induced a marked up-regulation of caspase-

1 expression in rMC-1 cells (Fig. 3.6B). Caspase-1 mRNA transcripts were also elevated in cells treated with high glucose, NECA (30 µM), or ATP (5 mM), but not to the magnitude induced by forskolin. Treatment with forskolin, NECA, or

ATP also induced accumulation of TXNIP mRNA but to levels lower than in high glucose-stimulated cells (Fig. 3.6B).

96 3.2.6 rMC-1 cells express Gq-coupled, Ca2+-mobilizing P2Y2/P2Y4 and P2Y1 receptors and high glucose treatment increases the efficacy of Ca2+ mobilization

The antagonistic action of suramin (Fig. 3.4A) suggested that one or more suramin-sensitive P2Y or P2X receptor subtypes may mediate the contribution of autocrine ATP to high-glucose induced caspase-1 activation in rMC-1 cells. A general response to activation of most P2Y and P2X receptor subtypes is marked perturbation in ion homeostasis, including rapid mobilization or influx of

Ca2+ (29, 79). Control rMC-1 cells exhibited robust and equivalent Ca2+ mobilization responses to 100 µM ATP or UTP while 100 µM ADP was less efficacious (Figs. 3.7A). With [ATP] < 100 µM, the ATP concentration-response relationship was characterized by an EC50 of ~ 3 µM and a plateau phase at 10-

100 µM; increasing ATP to 1 mM produced a further increase in peak [Ca2+] (Fig.

3.7B). Treatment with high glucose increased the efficacy of ATP as a Ca2+ mobilizing agonist at concentrations >10 µM. The UTP concentration-response in control rMC-1 cells was similar to that of ATP but without the secondary increase in peak Ca2+ mobilization as UTP was increased to 1 mM (Fig. 3.7C). As for ATP, growth in high glucose potentiated the efficacy of UTP at concentrations

>10 µM. The ADP concentration-responses were quite different with little response at <10 µM and no plateau at 100 µM in cells cultured in normal or high glucose medium (Fig. 3.7D); these results were consistent low-level expression of the ADP-selective P2Y1 subtype. Notably, initial stimulation of rMC-1 cells with UTP prevented Ca2+ mobilization responses to a secondary stimulation with

ATP (and vice versa) indicating cross-desensitization of a common P2Y subtype

97 that recognizes both ATP and UTP as agonists (not shown). This is consistent with both the pharmacology of the P2Y2 and/or (for this rat-derived cell type)

P2Y4 subtypes (16, 114) and the expression of P2Y2 mRNA in rMC-1 cells (Fig.

3.5A).

98 3.3 Discussion

Glial cell dysfunction and death have been identified as hallmarks of several neurodegenerative diseases including Alzheimer’s disease, Huntington’s disease and diabetic retinopathy (16, 99, 211, 229, 234). Such diseases are also characterized by aberrant caspase-1 activation and consequent IL-1β production similar to that observed in other chronic, sterile inflammatory conditions such as arthritis and diabetes mellitus type 1 (22, 130, 159, 173). Inflammasome-based mechanisms of caspase-1 activation in myeloid hematopoietic cells in response to microbial infections or stimulation of purinergic P2X7 ATP-gated ion channel receptors have been well characterized (58, 184). Whether similar pathways underlie sterile caspase-1 activation in non-hematopoietic inflammatory cells, such as the Müller glia which contribute to diabetic retinopathy, had not been characterized. This present study indicates that a novel purinergic signaling cascade mediates in part the ability of high glucose stress to activate caspase-1 in the rMC-1 cell line. In contrast to the P2X7R-dependent signaling that defines purinergic regulation of caspase-1 in myeloid hematopoietic cells, this new pathway includes a key role for autocrine activation of G protein-coupled A2 adenosine receptors and cyclic AMP-dependent signaling. Moreover, both the purinergic- and the high glucose-stimulated components of caspase-1 regulation are correlated with increased gene expression of caspase-1 and TXNIP, an adapter protein linked to caspase-1 regulation in other non-hematopoietic cells.

Finally, G protein-coupled P2Y receptors appear to drive additional signals that – depending on context – positively or negatively modulate the caspase-1 response to high glucose. Fig. 3.9 illustrates a schematic model that

99 encompasses the signaling pathways and network implicated by our observations.

Most cells release ATP at low rates which are counteracted by multiple ecto-nucleotidases that serially metabolize extracellular ATP into ADP, AMP, and, ultimately, adenosine for re-accumulation and regeneration of cytosolic ATP

(31). Our experiments suggest that high glucose comprises a metabolic stress signal which changes the relative rates of ATP release/hydrolysis and adenosine clearance by rMC-1 cells to facilitate: 1) steady-state concentrations of extracellular ATP, ADP, and adenosine sufficient for autocrine stimulation of various P2 nucleotide receptors and P1 adenosine receptors; and 2) the consequent induction of P2 and P1 receptor-dependent signaling cascades that converge on activation of caspase-1. Pharmacological analyses indicated that autocrine purinergic mediation of the high glucose-induced caspase-1 response in rMC-1 cells involves signals generated both by A2 receptors for adenosine and by apyrase- and suramin-sensitive P2 receptors for ATP. Inclusion of exogenous ATP in the control (5 mM glucose) culture medium elicited caspase-1 activation similar in magnitude to that stimulated by high glucose. Exogenous

ATP will directly stimulate P2 receptors and – due to metabolism by ecto- nucleotidases – will generate adenosine for stimulation of adenosine receptors.

This indicates that co- activation of P2 receptors and A2 adenosine receptors can entrain signaling pathways that lead to accumulation of active caspase-1 even in absence of high glucose metabolic stress.

100 The ability of exogenous NECA to stimulate caspase-1 in the absence of high glucose stress demonstrates that adenosine receptor signaling is sufficient for this proinflammatory response in the rMC-1 cell model, albeit at reduced efficacy in comparison to high glucose (Fig. 3.2 C) or 5 mM ATP (Fig. 3.2A).

Multiple reports have described the expression of functional adenosine receptors in retinal Müller cells (111, 218). Our findings represent the first linkage of these

G protein-coupled receptors to caspase-1 based inflammation and support a model wherein A2 adenosine receptors, coupled to cAMP signaling, induce transcriptional up-regulation of caspase-1 and TXNIP, a glucose-sensitive adapter protein linked to caspase-1 activation. All of the stimuli (high glucose,

NECA, ATP, forskolin) that increased caspase-1 activity also increased caspase-

1 mRNA levels in rMC-1 cells, albeit to different magnitudes (Fig. 3.6B). Myeloid hematopoietic inflammatory cells, including macrophages and microglia, constitutively express high levels of caspase-1 mRNA and protein in its zymogenic procaspase-1 form (12, 123, 184). Accumulation of active caspase-1 at functionally significant levels in myeloid leukocytes is not limited by the amount of procaspase-1 protein, but by the levels of key inflammasome adapter proteins, such as NLRP3, and/or the PAMP and DAMP ligands which allosterically regulate assembly of the inflammasome platforms (12, 94, 123, 184). In contrast, procaspase-1 is expressed at lower limiting levels within non- hematopoietic cells, such as pancreatic β cells (184, 232), adipocytes (103, 195), keratinocytes (230) and osteoblasts (5), known to secrete IL-1β in response to sterile proinflammatory stimuli.

101 Recent studies have indicated up-regulation of caspase-1 gene expression per se as a critical component of the inflammasome activation response in such cells. Fibroblastoid cells from the bone tumors of patients with

NOMID (neonatal-onset multisystem inflammatory disease) exhibit hyper- expression of caspase-1 due to a cAMP/PKA-driven increase in the Ets-1 transcription factor which positively modulates caspase-1 gene expression (5).

The same study reported that forskolin or prostaglandin E2 increased caspase-1 gene expression in MC3T3-E1 murine pro-osteoblasts similar to our findings with rMC-1 glial cells. Undifferentiated 3T3-L1 preadipocytes lack basal expression of caspase-1 but accumulate caspase-1 mRNA and protein during induction of adipocyte differentiation (195). Notably, the classic 3T3-L1 “differentiation cocktail” includes isobutyl-methyl-xanthine (IBMX) which increases cAMP via inhibition of cAMP-phosphodiesterases. Thus, cAMP-dependent regulation of caspase-1 expression to non-limiting levels may be a general pathway for inflammasome regulation and production of bioactive IL-1β by non-myeloid cell types exposed to diverse metabolic stress signals.

Most of the stimuli (high glucose, NECA, ATP) that increased caspase-1 activity in rMC-1 cells also up-regulated expression of TXNIP, but to different levels (Fig. 3.6B). Defining how TXNIP is involved in inflammasome regulation and IL-1β secretion has been an active but contentious area of recent investigation. Zhou et al. reported that TXNIP binds to NLRP3 and is critical for activation of the NLRP3/caspase-1 inflammasome in human THP1 monocytes, primary murine macrophages, and murine pancreatic β cells (232). That study

102 proposed that increased reactive oxygen species (ROS) production after mitochondrial damage leads to dissociation of TXNIP/thioredoxin complexes, thus freeing thioredoxin to scavenge ROS and TXNIP to interact with NLRP3 and drive inflammasome activation. However, another report which compared

P2X7R- or NLRP3 inflammasome activation in macrophages from wildtype versus TXNIP-knockout mice failed to demonstrate a critical role for TXNIP (127).

Conversely, analysis of hyperglycemia-induced IL-1β secretion by human adipocytes confirmed a stimulatory role for TXNIP in mature IL-1β production, but at the level of increased IL-1β gene expression rather than caspase-1 activation

(103, 195). Thus, the precise role of TXNIP in caspase-1/IL-1β cascades may vary with cell type and/or the relative expression levels of TXNIP and other inflammasome components. It is relevant to note that elevated expression of

TXNIP in rat models of retinal inflammation is associated with increased neurotoxicity, endothelial dysfunction, glial activation, and enhanced release of

IL-1β and TNFα (4, 156, 157). Moreover, knockdown of TXNIP in a rat model of diabetic retinopathy protected against capillary dropout and retinal gliosis, though this effect was not specifically linked to protection of Müller cells from high glucose stress (4, 157) .

Regardless of the precise role(s) for TXNIP in regulating caspase-1 and other high glucose-induced inflammatory responses in Müller cells, our observations confirm and extend reports regarding the ability of extracellular adenosine and other purines to positively modulate TXNIP gene expression under control and hyperglycemic conditions (228). TXNIP expression is

103 stringently linked to extracellular glucose concentration via glycolytic flux-induced activation of the MondoA and Max-like protein X (MLX) transcription factors which interact with a carbohydrate response element in the TXNIP gene promoter (25, 227, 228). However, incubation of a broad range of cell types with extracellular adenosine or other adenosine-containing compounds, including

ATP, also increased transcription of TXNIP by facilitating nuclear translocation of

Mondo A (228). Increased cAMP can also decrease TXNIP protein stability

(188). As noted previously, high glucose also activates NFκB-based gene expression in rMC-1 cells (99, 175) and NLRP3 expression in myeloid cells is markedly up-regulated by stimuli coupled to NFκB (12, 94). In other studies described in Chapter 4, we determined that high glucose treatment induced upregulation of NLRP3 and proIL-1β mRNA in rMC-1 cells. Thus, Müller cells exposed to high glucose may coordinately up-regulate NLRP3 via NFκB transcriptional signaling, TXNIP via MondoA/MLX transcriptional signaling, and caspase-1 via PKA/CREB/Ets-1 transcriptional signaling (Fig. 3.9).

Although adenosine deaminase significantly reduced high glucose- induced caspase-1 activation implicating a specific role for adenosine in this regulatory process, treatment with apyrase similarly reduced caspase-1 activation (Fig. 3.1A). These data indicated that ATP per se, and not merely its metabolite adenosine, also participates in regulating caspase-1 activation (Fig.

3.9). Surprisingly, our experiments demonstrated no significant expression of functional P2X7Rs in rMC-1cells or contribution of P2X7R signaling to high glucose-induced caspase-1 activation (Fig. 3.5). We explored whether P2X4R,

104 another P2XR subtype recently linked to caspase-1 activation, might be involved by testing whether ivermectin, which allosterically potentiates ATP-gating of

P2X4 channels, would increase ATP-induced Ca2+ influx in control or high glucose-treated rMC-1 cells (186). However, these experiments failed to identify

P2X4-like function in this cell model (Fig. 3.8). Rather, several lines of experimental evidence (Fig. 3.4A, 3.7) suggested that Gq-coupled, Ca2+- mobilizing P2Y2 receptors may mediate the stimulatory action of exogenous

ATP. How ATP, presumably acting via P2Y2R, might increase caspase-1 expression (Fig. 3.6B) and activity (Fig. 3.2A, B) remains to be defined. Plausible mechanisms include Ca2+-dependent induction of mitochondrial reactive oxygen species implicated in NLRP3- and TXNIP-dependent regulation of inflammasome assembly (Fig. 3.9). Additionally, activation of P2Y receptors has been shown to regulate osmotic changes within the diseased and damaged retina and may be

2+ linked to fluxes of Ca or other ions that modulate caspase activity (19, 65, 111,

217, 219). However, exogenous UTP, which also triggered robust Ca2+ mobilization (Fig. 3.7B), failed to mimic the ability of exogenous ATP to significantly increase caspase-1 activity (Fig. 3.2A, 3.3A and B). This could reflect the inability of UTP to additionally support generation of adenosine and activation of A2 receptor signaling. We tested whether combined treatment with

UTP, as a non-adenine nucleotide agonist of P2Y2 receptors, plus NECA, as a pan-P1 receptor agonist, would produce additive increases in caspase-1 activation. However, these experiments did not demonstrate any significant differences between treatments with NECA alone or NECA plus UTP (Fig. 3.3E).

105 Finally, we noted that exogenous UTP and ADP – in contrast to ATP – attenuated high glucose activation of caspase-1 (Fig. 3.2A) even when added at submillimolar levels (Fig. 3.4A and B). This suggests an additional inhibitory role of P2Y family receptors in the complex multi-component signaling network that links high glucose stress to activation of the caspase-1/IL-1β response in Müller cells. Activation of Gi-coupled P2Y receptors that counteract Gs-coupled A2 receptors or drive activation of cAMP phosphodiesterases are possible pathways that should be considered.

In summary, we demonstrate convergence of multiple signaling pathways in a tissue culture model of retinal Müller cells whereby high glucose conditions regulate caspase-1 activity via a purinergic-dependent but P2X7R-independent mechanism (Fig. 3.9). Whether similar purinergic cascades regulate caspase-1

Müller cells within the complex in vivo environment of the intact retina remains an important question for future experiments. If so, inhibition of the relevant adenosine or ATP receptors may protect retinal Müller cells from hyperglycemia- induced inflammation and thereby attenuate the progression of diabetic retinopathy.

106

Figure 3.1

107 Figure 3.1: Treatment of rMC-1 cells with extracellular apyrase or adenosine deaminase suppresses high glucose-induced caspase-1 activation rMC-1 cells were incubated in control (5mM) or high (25 mM) glucose medium ± apyrase (5units/ml) in panel (A) or ± adenosine deaminase (2units/ml) in panel

(B). Caspase-1 activity was measured (as described in Methods) after 24 h, normalized to the control activity, and is expressed as mean ± SEM (n=5). (*) p<0.02 compared to control, (#) p<0.004 compared to high glucose (ANOVA).

108 E. F.

A 2 6&7*8$9*:*6&7*;<= 7 46&7*8$9*:*6&7*;<= & ?

* 6 =

; 5 * . ) $ 1

" $ $

% 4 9 )

. ( / C B 0

0 1 2 3 /4 /> 10 <,&%*?@A

Figure 3.2 !"##$%&%'()$*+,-".%*/

109 Figure 3.2: Treatment of rMC-1 cells with exogenous ATP, exogenous adenosine analog, or adenosine uptake inhibitors mimics high glucose- induced caspase-1 activation (A) rMC-1 cells were incubated in control or high glucose medium ±ATP (5mM), ADP (5mM), or UTP (5mM). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (n=5). (*) p<0.0002 as compared to control, (#) p<0.05 as compared to + high glucose samples (ANOVA). (B) rMC-1 cells were incubated in control (5 mM glucose) medium supplemented with the indicated concentrations of ATP (1-10mM). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (n=5). (*) p<0.0002 compared to control (ANOVA). (C) rMC-1 cells were incubated in control medium, high glucose medium, or control medium + NECA (10, 30,

100µM). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (n=5). (*) p <0.03 compared to control

(ANOVA). (D) MC-1 cells were incubated in control (5 mM glucose) medium supplemented with 80µM dipyridamole (Dipyr) or 5µM NBTG, with and without

ADA (2U/ml). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM. (n=5) (*) p<0.05 compared to control. (#) p<0.05 compared to +Dipyr or +NBTG alone (ANOVA). (E) rMC-1 cells were incubated in control medium, high glucose medium, or control medium

+ NECA (30 or 100 µM) in the presence or absence of 100 or 1000 µM UTP.

Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (n=9). (F) rMC-1 cells were incubated in control

110 medium or high glucose medium for 24 h prior to supplementation with 5 mM

ATP. Extracellular medium samples were taken at the indicated times, heated to

100ºC to inactivate soluble nucleotidases, and diluted for ATP analysis. ATP levels were quantified by the luciferase-based assay described in Methods. Data are mean + SE (N=3).

111

Figure. 3.3

112

Figure 3.3: Treatment of rMC-1 cells with exogenous ADP or exogenous

UTP attenuates high glucose-induced caspase-1 activation (A, B) rMC-1 cells were incubated in control medium or in high glucose medium ± the indicated concentrations of ADP (A) or UTP (B). Caspase-1 activity was measured after

24 h, normalized to the control activity, and expressed as mean ± SEM (n=3-10).

(*) p<0.0002 as compared to control, (#) p<0.05 as compared to + high glucose samples (ANOVA).

113

Figure 3.4

114 Figure 3.4: Treatment of rMC-1 cells with suramin or adenosine receptor antagonists, but not P2X7R antagonists, suppresses high glucose-induced caspase-1 activation (A) rMC-1 cells were incubated in control medium ± suramin (10 µM) or high glucose medium ± suramin (10 µM). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (n=5). (*) p<0.02 compared to controls, (#) p<0.004 compared to high gluocse (ANOVA). (B) rMC-1 cells were incubated in control medium or in high glucose medium ± DPCPX (10µM), MRS1754 (10µM), or

SCH442416 (10µM). Caspase-1 activity was measured after 24 h, normalized to the control activity, expressed as mean ± SEM. (*) p<0.004 compared to 5mM glucose, (#) p<0.05 compared to 25mM glucose (n=5). (C) rMC-1 cells were incubated in control medium or in high glucose medium ± A438079 (10 µM) (n=8) or AZ10606120 (10µM) (n=5). Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM (*) p <0.0001 compared to control, (#) <0.02 compared to high glucose (ANOVA).

115

Figure 3.5

116

Figure 3.5 (cont’d)

117 F)

Figure 3.5 (cont’d)

118 Figure 3.5: rMC-1 cells do not express immunoreactive P2X7R receptor protein or functional P2X7R receptors (A) RNA was isolated from rMC-1 cells after incubation in control or high glucose media for 24 h and was then reverse- transcribed to cDNA for qPCR analysis of P2X7R, TXNIP and P2Y2R gene expression. Target gene expression was quantified relative to the housekeeping

GAPDH gene product and the fold-increases in gene expression between high glucose versus control media incubation were calculated by the ΔΔCt method and graphed as mean ± SEM (n=13), (*) = P <0.001 compared to control. (B) Left panel. Whole cell lysates were prepared from digitonin-treated rMC-1 cells after incubation in control medium, high glucose medium, or control medium + 5 mM

ATP for 24 h. Lysates were similarly prepared from digitonin-treated HEK-

P2X7R cells or wildtype HEK293 cells. Samples were resolved by SDS-PAGE and analyzed by serial western blots using antibodies against β-actin, TXNIP,

P2X7R-C terminus (CT), P2X7R-extracellular loop (EC), and Glut1.

Representative blots from 1 of 4 independent experiments are shown. (B) Right panel, upper: epitope domains of the P2X7R antibodies. (B) Right panel, lower:

Whole cell lysates were prepared from intact MC-1 cells after incubation in control or high glucose medium and for TXNIP and β-actin expression as described for left panel. (C) rMC-1 cells in 6-well plates were incubated in control or high glucose medium for 24 h. After washing and transfer to basal salt solution, parallel wells of adherent cells were incubated ± 5 mM ATP for 15 min prior to nitric acid extraction and quantification of K content by atomic absorbance spectrophotometry. HEK-P2X7R cell cultures were subjected to the

119 identical stimulation protocol and K analysis. (D) rMC-1 or HEK-P2X7R positive control cell suspensions were subjected to cell surface biotinylation, extraction, precipitation with immobilized streptavidin, SDS-PAGE, and western blot analysis for P2X7R protein as described in Methods and the Figure 4B legend. Blot is representative of data from 2 experiments. (E) rMC-1 cells or positive control cell suspensions of murine bone marrow-derived macrophages were incubated in basal salt solution supplemented with 10 mg/ml ethidium bromide, stimulated by addition of 5 mM ATP, and finally permeabilized by addition of digitonin. Traces illustrate continuous readout of fluorescence (360 nm ex/ 540 nm em). Data are expressed as mean ± SEM (n=5). (*) p<0.002 compared to unstimulated (0mM

ATP HEK-rP2X7R) samples (ANOVA). (F) As a positive control, HEK-rP2X7R cells were stained for P2X7R and subjected to FACS analysis. The left panel is representative of the negative control (Primary antibody only). The right panel is representative of the positive control as indicated by the right-ward shift in MFU

(mean fluorescence units). (G) rMC-1s were incubated in either 5mM (left) or

25mM (right) glucose containing DMEM for 24 hours and stained for P2X7R and subjected to FACS analysis.

120

!" ()*+,-. $()*+,-. '

C 4 4 A = : B = = + A

< & . ; ! : + 9 @ / 8 <

7 % > + 9 6 - ? 5 > 9 0

7 $ +

# /+012 /+012 3+012

I" GJEKH 79>?9>

> 4 1 > @## ' < D @ + ? 5 M A + F 6 + < < O : = - < 9

, )* + ) @# < D )* +,+0 B 5 = 4 +,-. 4 )* A E N 9 +,-( - <

L 4

- 5 -. 2 ! H AD , 1 7 G : + 0 F ! 5 * E 7 +) ( $

Figure!"#$ 3.6%&'(

121 Figure 3.6: Activation of caspase-1 by high glucose, NECA, or ATP is mimicked by forskolin and correlated with increased gene expression of caspase-1 and TXNIP (A) rMC-1 cells were incubated in control medium ± 10

µM forskolin (FSK) or high glucose medium. Caspase-1 activity was measured after 24 h, normalized to the control activity, and expressed as mean ± SEM

(n=5). (*) p<0.02 compared to control. (ANOVA). (B) RNA was isolated from rMC-1 cells after incubation in control medium, high glucose medium, control medium plus 10 µM forskolin, control medium plus 30 µM NECA, or control medium plus 5 mM ATP for 24 h and then reverse-transcribed to cDNA for qPCR analysis of TXNIP and caspase-1 gene expression. Target gene expression was quantified relative to the housekeeping 18S rRNA gene product and the fold- increases in gene expression between the various treatments versus control media incubation were calculated by the ΔΔCt method and graphed as mean ±

SEM (n=4), ANOVA (*) = P <0.05 compared to control.

122

Figure 3.7

123 Figure 3.7: rMC-1 cells express Gq-coupled, Ca2+-mobilizing P2Y2/P2Y4 and P2Y1 receptors and high glucose treatment increases the efficacy of

Ca2+ mobilization (A-D) rMC-1 cells were incubated in control medium or high glucose medium for 24 h and then trypsinized to generate single cell suspensions for fura2 loading and fluorimetric analysis of cytosolic [Ca2+] as described in methods. (A) Representative time courses of Ca2+ mobilization triggered by 100 mM pulses of ATP, UTP, or ADP. (B-D) Concentration- response relationships describing peak changes in cytosolic Ca2+ elicited by the indicated [ATP], [UTP], or [ADP] in rMC-1 cells isolated from the control (■) versus high glucose (▲) culture conditions. Data are the means ± SEM (N=5).

(*) =P <0.05 compared to control incubated cells (Students’ T- Test).

124

3µM Ivermectin, 30µM UTP, µM ATP

Ca2+-free extracellular buffer

Figure 3.8

125 Figure 3.8: ATP induced Ca2+ mobilization is from intracellular stores and is unchanged following ivermectin treatment (A-B) rMC-1 cells were incubated in control medium or high glucose medium for 24 h and then trypsinized to generate single cell suspensions for fura2 loading and fluorimetric analysis of cytosolic [Ca2+] as described in methods. (A) Cells were fura2 loaded in Ca2+ containing extracellular buffer. Panel 1: To first desensitize native P2Y receptors, rMC-1 were stimulated with 30µM UTP. To observe changes in Ca2+ mobilization due to P2X receptor activation, cells were subsequently stimulated with 30µM ATP. Panel 2: cells were pretreated with 3µM Ivermectin in order to promote P2X4R activation prior to UTP, ATP stimulation. (B) Cells were fura2 loaded in Ca2+ free extracellular buffer. Panel 1: To first desensitize native P2Y receptors, rMC-1 were stimulated with 30µM UTP. To observe changes in Ca2+ mobilization due to P2X receptor activation, cells were subsequently stimulated with 30µM ATP. Panel 2: cells were pretreated with 3µM Ivermectin in order to promote P2X4R activation prior to UTP, ATP stimulation.

126

Figure 3.9

127 Figure 3.9: Proposed model for high glucose-induced caspase-1 activation via purinergic signaling

128

CHAPTER 4

DIABETES-INDUCED CASPASE-1 ACTIVITY IS REGULATED BY AN AUTOINFLAMMATORY FEED-BACK MECHANISM

129 4.1 Introduction

Sustained low grade IL-1β production is a major contributor to chronic tissue inflammation and damage (43, 44, 46, 150). Although the exact mechanisms of continuous IL-1β production under chronic inflammatory conditions are poorly understood, the concept of sterile auto-inflammation has emerged as a possible explanation of this phenomenon (22, 173). Sterile auto- inflammation is characterized by localized, chronic tissue inflammation in the absence of obvious microbial infection (44, 173). Recent studies have indicated that following initial IL-1β production, this cytokine can activate its own producer caspase-1 in an autocrine fashion thereby providing feedback amplification of the inflammatory response (12, 150). Normally, caspase-1 activation and IL-1β production are tightly regulated by gene transcription and induction of multi- protein complexes called inflammasomes (126, 183). Recent studies have suggested that inflammasome formation; especially those based on the NLRP3 regulatory protein, might be responsible for caspase-1 activation and prolonged

IL-1β production in sterile inflammatory diseases, such as diabetes (59, 126, 130,

150, 184). However, little is known about how caspase-1 activation is initiated and sustained under diabetic inflammatory conditions.

Increased levels of caspase-1 activity have been observed in retinas of diabetic and galactosemic mice beginning at 8 weeks and 24 weeks duration of diabetes (138). Inhibition of caspase-1 activation using minocycline, an antibiotic that has been speculated to inhibit caspase-1 activation, prevented degeneration of the retinal vasculature in diabetic and galactosemic mice (207). Exactly how

130 the activation of caspase-1 is induced, regulated, and sustained in the retina and retinal cells under hyperglycemic conditions has not been determined. Therefore, identifying potential autoinflammatory feedback mechanisms regulating sustained caspase-1 activation and IL-1β production under hyperglycemic conditions in vivo and in vitro was the focus of the present study.

131 4.2 Results

4.2.1 Inhibition of diabetes-induced vascular damage in caspase-1 knock-out mice

The development of diabetic retinopathy has previously been linked to the activation of the IL-1 receptor-signaling pathway (41, 99, 110, 138, 207).

Caspase-1 is the major but not the sole processing protease for production of mature, biologically active IL-1β (128). To demonstrate the importance of upstream caspase-1 in the development of diabetic retinopathy, we initiated a long-term study using 4 experimental groups (normal and diabetic wild type (wt), and normal and diabetic cas-1-/- mice). After 6 months of diabetes, retinal vasculatures were isolated, and the number of acellular capillaries/mm2 retina were counted. Figure 4.1A shows that diabetes significantly increased the number of acellular capillaries/mm2 in diabetic wt mice 2.3 ± 0.5 fold compared to normal wt mice. Caspase-1 knock-out inhibited the diabetes-induced increase in numbers of acellular capillaries/mm2 by 65±19%. Severity of diabetes was not affected by caspase-1 knock-down (Table 4.1).

4.2.2 Hyperglycemia-induced caspase-1 activity pattern in the retinas of diabetic mice

Given this demonstration that caspase-1 is important for the development of diabetic retinopathy, we next investigated the caspase-1 activation pattern throughout the development of the disease. Previous data have shown that caspase-1 is active in the retinas of diabetic mice at 8 weeks and at 24 weeks duration of diabetes compared to control (138). However, whether caspase-1

132 remains constitutively active throughout the duration of diabetes has not been previously determined. Our studies demonstrate that caspase-1 activity is significantly increased at 10 weeks duration of diabetes compared to non- diabetic control animals. Interestingly, at 13 weeks duration of diabetes caspase-

1 activity was no longer elevated and caspase-1 activity levels were similar to controls. At 20 weeks duration of diabetes, caspase-1 activity was again increased compared to controls indicating the hyperglycemia-induced caspase-1 activity pattern is indicative of a cyclic inflammatory condition (Fig. 4.2A).

To test the potential role for IL-1β feedback signaling in retinal auto- inflammation, we examined the caspase-1 activity pattern in diabetic IL-1R1-/- mice. As caspase-1 activity is upstream of IL-1β production, caspase-1 activity was elevated by at 10 weeks duration of diabetes as expected. As seen in the wild type animals, at 13 weeks duration of diabetes, there was no significant increase in caspase-1 activity compared to controls. Interestingly, at 20 weeks duration of diabetes, hyperglycemia did not induce later activation of caspase-1 in the retinas of IL-1R1-/- mice, unlike in retinas of wild type mice at 20 weeks of diabetes (Fig. 4.2B). Knock-out of the IL-1R1 did not influence the severity of diabetes (Table 4.2). These data strongly indicate that the late, sustained phase of caspase-1 activity is dependent on functional IL-1β signaling supporting the hypothesis of an auto-inflammatory feedback mechanism in the development of diabetic retinopathy (Fig. 4.2B).

133 4.2.3 High glucose-induced caspase-1 activity patterns in rMC-1

To study the role of caspase-1 activation in hyperglycemia-induced auto- inflammation under stringently controlled conditions, we used cultured retinal

Müller cells as the cellular model for mechanistic in vitro studies. We have previously identified that Müller cells are a source for active caspase-1 and IL-1β under hyperglycemic conditions (16, 223, 224). First, we determined the consequence of long-term exposure to high glucose on the caspase-1 activity pattern in these cells. Consistent with experiments described in chapter 3, caspase-1 activity was significantly increased in rMC-1s (Fig. 4.3A) cultured in high glucose conditions with an initial peak at 24 hours compared to controls.

However, caspase-1 activity decreased to control levels at 48 hours of high glucose treatment in either rMC-1 or human Müller cells. Starting at 72 hours, another phase of caspase-1 activity occurred which was sustained through 96 hours in high glucose conditions (Fig. 4.3A). To confirm these activity data, we determined caspase-1 activation via western blot analysis of the p20 subunit of active caspase-1. Results demonstrate that cleavage of caspase-1 is multi- phasic consistent with the caspase-1 activity pattern (Fig. 4.3B). Long-term exposure of retinal Müller cells to elevated glucose levels induced a multi-phasic caspase-1 activation pattern as observed in vivo.

To demonstrate that high glucose-induced activation of caspase-1 leads to the production of active IL-1β, the release of IL-1β was determined by ELISA.

High glucose induced a similar IL-1β release pattern as seen for the caspase-1

134 activity pattern with an early peak at 24 hours and another phase of IL-1β release starting at 72 hours in rMC-1 (Fig. 4.3C).

4.2.4 High glucose-induced caspase-1 activity and IL-1β production patterns in hMC

We repeated the caspase-1 activity pattern experiments in isolated human

Müller cells (hMC) to avoid any pitfalls from using the transformed rMC-1 cell line, in addition to verifying that the high glucose-induced caspase-1 signaling pathway was operative in human as in rodent Müller cells. An initial increase in high glucose-induced caspase-1 activity was observed at the 24 hour time point

(Fig. 4.4A). Caspase-1 activity decreased to control levels at 48 hours in hMCs.

Beginning at 72 hours, another phase of increased in caspase-1 activity was observed in high glucose treated hMCs (Fig. 4.4A). The high glucose-induced IL-

1β release pattern closely followed the pattern seen for high glucose-induced caspase-1 activity although IL-1β release in the later phase seemed to be delayed (Fig. 4.4B).

4.2.5 Exogenous interleukin-1β-induced caspase-1 activity in retinal Müller cells

The multi-phasic activation of caspase-1 by hyperglycemia was indicative of a potential autocrine feedback pathway. Therefore, we attempted to identify the mechanism for the later caspase-1 activation. Recent reports have introduced the concept of IL-1β induced IL-1β production (39, 55, 148), and we therefore designed experiments to test whether IL-1β might be responsible for

135 regulating the second, sustained phase of caspase-1 activity in our model. rMC-1 were treated with either normal glucose medium or normal glucose medium plus

2 ng/ml IL-1β. Results show that within 24 hours, IL-1β was able to significantly induce capase-1 activity (Fig. 4.5A) demonstrating that IL-1β is indeed capable of activating its own producer, caspase-1, in retinal Müller cells.

To determine if the later phase of caspase-1 activity was dependent on the production of IL-1β in the early phase and subsequent IL-1β signaling, studies using an IL-1 receptor antagonist were performed. Using an IL-1 receptor antagonist (IL-1ra), which prevents downstream signaling of IL-1β, we determined that caspase-1 activity was significantly decreased at 72 hours compared to high glucose alone (Fig. 4.5B). Utilizing YVAD-fmk, a caspase-1 specific activity inhibitor, activity was decreased significantly at 72 hours (Fig.

4.5C). Minocyline, a second generation tetracycline derivative suggested to be an inhibitor of caspase-1 activation, also reduced caspase-1 activity at 72 hours as compared to high glucose alone indicating that interfering in the early high glucose-induced caspase-1/IL-1β signaling cascade prevents the second, sustained phase of caspase-1 activity (Fig. 4.5C).

4.2.6 The effect of inhibition of early phase caspase-1/IL-1 β signaling on Müller cell viability

To determine whether inhibition of early high glucose-induced caspase-1 and IL-1β action not only prevents another phase of caspase-1 activation but also increases survival of Müller cells, Trypan Blue viability assays were

136 performed. High glucose induced a 3.30 ± 0.09 fold increase in cell death compared to controls. Data show that treatment with IL-1ra for only 24 hours decreased high glucose-induced cell death by 55.6 ± 2.3%. YVAD-fmk application for only 24 hours was capable of preventing Müller cell death by 50.0

± 1.9%. Finally, minocycline treatment for 24 hours also prevented high glucose- induced cell death by 53.5 ± 3.3% at the 96 hour time point (Fig. 4.6). These results suggest that activation of the second, sustained phase of caspase-1 activity is detrimental to Müller cell viability.

4.2.7 Glucose washout effect on late phase caspase-1 activity

Glucose washout experiments were completed to determine whether glucose removal prevents late phase caspase-1 activity. Removal of high glucose at 24 hours and replacement with normal glucose media reduced caspase-1 activity at 72 hours by 31.8 ± 10.8% (Fig. 4.7). Interestingly, removal of high glucose media at 48 hours when caspase-1 activity and IL-1β levels were low was no longer protective (Fig. 4.7) suggesting that after the first phase of high glucose-induced caspase-1 activity is completed, later caspase-1 activity becomes apparently independent of glucose levels, but rather dependent on autocrine IL-1β feed-back signaling.

4.2.8 Hyperglycemia-induced induction of inflammasome component mRNA

It is well established that caspase-1 activation is tightly regulated by formation of inflammasomes, multi-protein complexes which enable

137 autoproteolytic activation of caspase-1. One major type of inflammasome for caspase-1 activation involves the NOD-family protein NLRP3/NALP3 in complexes with ASC and caspase-1 (126, 183). To determine possible effects of high glucose treatment on the expression of mRNAs encoding components of the

NLRP3 inflammasome, cDNA was isolated from normal and high glucose-treated

Müller cells. qPCR results demonstrate that both NLRP3 and NLRP1 mRNAs were significantly increased, 11.8 ± 6 fold and 5.3±1.9 fold, respectively, following 24 hours of high glucose treatment consistent with a hyperglycemia- induced induction of NLRP3-based inflammasomes (Fig. 4.8A). Elevated glucose levels did not affect mRNA levels of the adapter protein, ASC. qPCR results also confirmed that IL-1β mRNA is increased by 30.6 ± 12.1 fold within 24 hours of high glucose treatment consistent with the our findings showing high glucose- stimulated accumulation of IL-1β protein (Fig. 4.8A).

To determine whether IL-1β itself is capable of inducing caspase-1 activation via NLRP3 inflammasomes, rMC-1 were exposed to IL-1β. IL-1β treatment induced a significant upregulation in NLRP3 (1.9 ± 0.7 fold) and

NLRP1 (1.7 ± 0.2 fold) mRNA levels and a significant 2.3 ± 0.7 fold increase in

IL-1β transcripts following treatment with IL-1β alone (Fig. 4.8B). These data demonstrate that IL-1β is able to sustain caspase-1 activity by sustaining the upregulation of NLRP3 inflammasome components.

rMC-1 lysates were also subjected to Western blot analysis for RIP2 expression. RIP2 is a CARD-containing, protein kinase and scaffolding protein that has been shown to activate both pro-survival and pro-death pathways (129).

138 The ability of RIP2 to stimulate NFκB, a process required for priming cells for caspase-1 activation, has also been demonstrated (116, 129, 179). Recently,

RIP2 has been linked to caspase-1 activation in neuronal cells in response to a variety of stimuli, such as TNFα, and in Huntington’s Disease (211). Because retinal Müller cells are of neuronal origin, the role of RIP2 is of particular interest to this project. Results from western blot analysis demonstrate that RIP2 protein expression is upregulated by high glucose conditions (Fig. 4.8C).

139 4.3 Discussion

Increased inflammatory cytokine production is a characteristic of diabetic retinopathy and many other neurodegenerative and chronic inflammatory conditions (141). Many investigators have sought to understand the effects of pro-inflammatory signaling pathways, such as the caspase-1/IL-1β pathway, on tissue destruction but these studies have primarily focused on acute activation of inflammatory processes via bacterial or viral infection (28, 192, 208). How caspase-1 activity and IL-1β production are regulated under chronic sterile conditions, and more specifically, under chronic hyperglycemic conditions, is not well understood and was the focus of this study.

This current study provides data demonstrating that the activation of caspase-1 is crucial for the development of diabetic retinopathy. Caspase-1 activation in this chronic disease model is surprisingly multi-phasic and requires functional signaling of its product, IL-1β, in the retinas of diabetic mice. Using retinal Müller cells as a model system for studying the effect of elevated glucose levels on caspase-1 activation, our results show that prolonged treatment of

Müller cells with high glucose also induces a multi-phasic activation pattern of caspase-1 that is dependent on functional IL-1β signaling. Sustained caspase-1 activation appears to be independently regulated by an autocrine feedback mechanism rather than the initial glucose insult responsible for the early phase of inflammation (21, 107). This might explain why diabetic retinopathy still progresses even when blood glucose levels are tightly controlled (21, 107).

Finally, for the first time, we have demonstrated that high glucose conditions and

IL-1β itself can act as a priming stimulus for caspase-1 activation by inducing

140 expression of both the limiting NLRP3 inflammasome component and pro-IL-1β.

Taken together, our data support a model wherein, (1) hyperglycemia induces initial caspase-1 activation and IL-1β production possibly through the NLRP3 inflammasome leading to (2) an IL-1β driven, autoinflammatory feedback signal that maintains continuous caspase-1 activation and IL-1β production. Based on our in vivo studies, the late, sustained phase of caspase-1 activation is most detrimental to the health of the retinal vasculature (Figure 4.1). Our previous studies demonstrated that IL-1R1-/- mice are protected from hyperglycemia- induced capillary degeneration which we propose is due to decreased late phase caspase-1 activation within IL-1R1-/- retinas as observed in the present study

(Fig. 4.2B). These data further support the hypothesis that sustained activation of caspase-1 in the diabetic retina potentially participates in the process of vascular drop-out (207).

The activation of the NLRP3 inflammasome has recently been suggested to act as a sensor for metabolic stress (184, 232). However, the mechanism of this “sensing” of metabolic stress is only poorly understood and requires further investigation (184, 203). The results described in chapter 3 indicate that accumulation of extracellular purines provides at least one mechanism for linking metabolic stress to inflammasome activation during high glucose stimulation of

Müller cells. Whether the NLRP3 inflammasome is required for caspase-1 activation within the context of diabetic retinopathy in vivo and in vitro remains to be determined. Interestingly, IL-1β itself is able to induce mRNA for the NLRP3 inflammasome component and its own production via caspase-1. These effects

141 may primarily be due to both hyperglycemia and IL-1β’s ability to activate NFκB, which is known to be increased in diabetic retinas (148, 175, 231). One potential pathway for hyperglycemia induced NFκB activation that we have identified is through the upregulation and activation of RIP2 (receptor interacting protein 2), which also follows a multi-phasic expression pattern (Fig. 4.8) (137). Although the upregulation and production of pro-IL-1β is an integral part of the caspase-

1/IL-1β signaling pathway, this study was primarily focused on the caspase-1 activation pathway and the regulation of caspase-1 activity over time. More studies are required to fully characterize the regulation of pro-IL-1β synthesis under diabetic conditions.

It seems that diabetic retinopathy progresses via auto-inflammatory responses, especially the activation of the caspase-1/IL-1β signaling pathway.

This phenomenon has been observed in other chronic diseases, although the function of such a response is somewhat controversial and not entirely clear (28,

36, 42, 192). It has been suggested for some chronic inflammatory diseases like cryopyrinopathies, Crohn’s disease, osteoarthritis, brain injury, and even ischemic stroke, that the initial, acute inflammatory response may actually be protective (28, 36, 141, 208). In the long term, however, this response incompletely shuts off making the switch from being protective and beneficial to causing major tissue damage. Consistent with the presence of a multiphasic caspase-1 activity pattern, we show that secretion of active, mature IL-1β (a product of caspase-1) occurs in a multiphasic pattern in high glucose-treated

Müller cells. It is thought that early IL-1β release may be intended to mount

142 forces against the initial insult but is then potentiated resulting in a vicious cycle of recurrent inflammation ultimately leading to cell death (208). As summarized in a review by Viviani et al., early IL-1β release from glia occurs before any neuronal cell death indicating that early activation of glia is an important step in the neurodegenerative process (208). Although neuronal cell loss in the diabetic retina has been documented, the extent to which retinal glia are responsible for directly initiating this cell death via IL-1β release remains to be shown.

Although we provide evidence for a diabetes-induced IL-1β driven autoinflammatory feedback mechanism, the pathways leading to the downregulation of caspase-1 activity and IL-1β production after the initial caspase-1 activation peak remain to be determined. Future studies must address the question of whether this multiphasic pattern is due to: 1) a defect in the down-regulating machinery in the presence of a maintained triggering signal; 2) to degradation of previously activated caspase-1; or 3) the possible translocation of active caspase-1 to the nucleus which would sequester caspase-1 from its pro-IL-1 β substrate that the same time nuclear caspase-1 may contribute cell death signaling (211).

It must be pointed out that these autocrine inflammatory responses in a hyperglycemic environment are independent of any bacterial infection which are often the initiating stimulus for caspase-1 activation in other pathological inflammatory conditions. The new concepts of sterile auto-inflammation based on amplifying IL-1 production seem applicable to the development of diabetic retinopathy (22, 173). Based on our results, we propose an autocrine signaling

143 mechanism whereby IL-1β signals to the cell via the IL-1R1 to induce additional caspase-1 activation and tissue damage (Fig. 4.9). Interfering in this second, sustained phase of caspase-1 activity, therefore, represents a potential therapeutic target with major clinical implications. Thus, the observations described in this study may have significant implications in the treatment of not only diabetic retinopathy but perhaps for other chronic inflammatory diseases. It will be interesting to learn whether the current clinical trials of anakinra will be translatable into treatments for diabetic retinopathy patients (46, 95).

144

Figure 4.1

145 Figure 4.1: Retinal capillary degeneration in wild type and caspase-1 knockout mice Retinas of wild type normal (n=6) and diabetic mice (n=8), and normal

(n=6) and diabetic cas1-/- mice (n=5) were isolated at 7 months of diabetes.

Diabetes-induced formation of acellular capillaries in wild type mice was significantly different from normal p<0.05. Number of acellular capillaries in diabetic cas-1-/- animals was significantly reduced from number seen in diabetic animals with p<0.05, and did not significantly differ from normal cas-1-/- or wild type animals. Kruskal-Wallis test (p<0.05) followed by Dunn’s post analysis was used to determine statistical significance among groups. (Data from the Mohr

Lab)

146 Table 4.1

147 Table 4.1: Summary of Animal Data Severity of diabetes was estimated by measuring non-enzymatically glycated hemoglobin (GHb) levels using affinity chromatography (Glyc-Affin; Pierce, Rockford, IL). (*) = p< 0.05 significantly different from non-diabetic wild type animals compared to diabetic wild type animals and non-diabetic IL-1R1 -/- animals compared to diabetic IL-1R1 -/- animals.

148

Figure 4.2

149 Figure 4.2: Hyperglycemia-induced caspase-1 activity pattern in the retinas of diabetic mice wild type and IL-1R1-/-mice (A) Retinas of wild type normal

(ND) (n=7) and diabetic (SD) mice (n=7), were isolated at 10, 13, and 20 weeks of diabetes. (B) Retinas of normal (n=8) and diabetic (n=5) IL-1R1-/- mice were isolated at 10, 13, and 20 weeks of diabetes. At time points indicated, caspase-1 activity was measured and expressed as mean ± SEM with (*) = p< 0.05 significantly different from normal wild type animals compared to diabetic wild type animals and normal IL-1R1 -/- animals compared to diabetic IL-1R1 -/- animals (#) = p<0.05 significantly different from 20 week normal wild type animals compared to 10 week normal wild type animals and 20 week normal IL-

1R1 -/- animals compared to 10 week normal IL-1R1 -/- animals. Kruskal-Wallis test (p<0.05) followed by Dunn’s post analysis was used to determine statistical significance among groups.

150 A)

Figure 4.3

151 Figure 4.3: High glucose-induced caspase-1 activation, activity, and IL-1β release patterns in rMC-1 rMC-1 were incubated in medium containing 5mM glucose or 25mM glucose media. At time points indicated, (A) caspase-1 activity was measured using a caspase-1 specific fluorescence substrate (YVAD-AFC).

Caspase-1 activity is expressed as mean ± SEM (n=15). (B) Caspase-1 activation was determined by Western Blot analysis. Band intensities of caspase-

1 Western Blots were quantified using densitometry analysis, normalized to actin, and graphed as mean ± SEM (n=10). (C) IL-1β release was measured by ELISA.

IL-1β levels are presented as the mean ± SEM (n=5). (*) = p< 0.05 compared to

5mM glucose (ANOVA, followed by Tukey’s Test).

152

Figure 4.4

153 Figure 4.4: High glucose-induced caspase-1 activity and IL-1β production in human Müller cells (hMC) hMC were incubated in medium containing 5 mM glucose or 25 mM glucose media. (A) At time points indicated, caspase-1 activity was measured and expressed as mean ± SEM (n=11). (*) = p< 0.05 compared to

5mM glucose. (B) IL-1β release was measured by ELISA. IL-1β levels are presented as the mean ± SEM (n=4). (*) = p< 0.05 compared to 5mM glucose.

(Data from E. Chepchumba K. Yego) (ANOVA, followed by Tukey’s Test).

154

Figure 4.5

155 C)

Figure 4.5 cont’d

156 Figure 4.5: Exogenous IL-1β induced caspase-1 activation in retinal Müller cells (A) rMC-1 were treated in 5mM glucose or 5mM glucose + IL-1β (2ng/ml) media for 24hrs. Following treatment, caspase-1 activity was measured and expressed as mean ± SEM (n=5). The SEM for controls in this experiment was negligible. (*) = p<0.05 compared to 5mM glucose. (B) rMC-1 were cultured in

5mM glucose media, 25mM glucose media, or 25mM glucose media containing either 50ng/ml IL-1ra for 24 hours at which point drug containing medium was washed out and replaced with 25mM glucose media. At 72 hours, caspase-1 activity was measured and expressed as mean ± SEM (n=15). (C) rMC-1 cells were cultured in 5mM glucose media, 25mM glucose media, or 25mM glucose media containing either 50ng/ml IL-1RA, 100µM YVAD-FMK (an irreversible caspase-1 inhibitor), or 100µM minocycline (Mino.) for 24 hours at which point drug containing medium was washed out and replaced with 25mM glucose media. At 72 hours, caspase-1 activity was measured and expressed as mean ±

SEM (n=15). (*) = p< 0.05 compared to 5mM glucose, (#) = p< 0.05 compared to

25mM glucose (ANOVA, followed by Tukey’s Test).

157

Figure 4.6

158 Figure 4.6: The effect of inhibition of early phase caspase-1/IL-1β signaling on Müller cell viability rMC-1 cells were cultured in 5mM glucose or 25mM glucose or 25mM glucose medium containing IL-1ra (50ng/ml), YVAD-fmk (100

µM), or mino. (100µM). Trypan blue viability assays were performed at 96 hours and expressed as mean ± SEM (n=8). (*)=p<0.05 compared to 5mM glucose, (#)

=p<0.05 compared to 25mM glucose (ANOVA, followed by Tukey’s Test).

159

Figure 4.7

160 Figure 4.7: Glucose washout effect on late phase caspase-1 activity rMC-1 were cultured in either 5mM glucose media, 25mM glucose media, or in 25mM glucose media for 24 or 48 hours followed by washout and replacement of media with 5mM glucose media. At 72 hours, caspase-1 activity was measured and expressed as mean ± SDEV (n=6). The SEM for controls in this experiment were negligible. (*)=p<0.05 compared to 5mM glucose, (#) =p<0.05 compared to

25mM glucose (ANOVA, followed by Tukey’s Test).

161 A)

Figure 4.8

162 Figure 4.8: High glucose-induced NLRP3 inflammasome component mRNA induction (A) cDNA was isolated from rMC-1 cultured in 5mM glucose or 25mM glucose media for 24 hours (n=8) and qPCR analysis of IL-1β, NLRP3, NLRP1 and ASC was performed. Fold changes in relative expression were calculated using the ΔΔCt method using StepOne software v. 2.1 with values normalized to the reference gene, 18S and graphed as mean ± SDEV. (*) = p<0.005 compared to 5mmol/lM glucose. (B) rMC-1 were cultured in 5mM glucose or 5mM glucose media containing 2ng/ml IL-1β for 24. qPCR analysis of IL-1β, NLRP3, NLRP1 and ASC was performed. Fold changes in relative expression were calculated using the ΔΔCt method using StepOne software v. 2.1 with values normalized to the reference gene, 18S and graphed as mean ± SDEV. (*) = p<0.005 compared to 5mM glucose (ANOVA, followed by Tukey’s Test). (C) rMC -1 were treated in

5mM or 25mM or 5mM glucose media +2 ng/ml IL-1β for 24 hours. RIP2 levels were determined by Western Blot analysis. (n=4)

163

Figure 4.9

164 Figure 4.9: Hyperglycemia-induced auto-inflammatory feed-back signaling

165 CHAPTER 5

DISCUSSION AND

FUTURE DIRECTIONS

166 Chapter 5

Discussion and Future Directions

5.1 Summary

Chronic activation of pro-inflammatory caspase-1 in the retinas of diabetic animals and patients in vivo and retinal Müller cells in vitro is well documented.

In this study I characterized: (1) how elevated glucose and extracellular purines contribute to early activation of caspase-1 in a cultured rat Müller cell model

(rMC-1) (Chapter 3); and (2) how sustained high glucose-induced caspase-1 activation and retinal degeneration is regulated in vitro and in vivo (Chapter 4).

The findings from the chapter 3 studies support a novel role for autocrine P1 and

P2 purinergic receptors coupled to cyclic AMP signaling cascades and transcriptional induction of caspase-1 in mediating the high glucose-induced activation of caspase-1 and secretion of IL-1β in non-hematopoietic retinal Müller cells. The data from the chapter 4 studies demonstrate the crucial role of an auto-inflammatory response mediated by IL-1β autocrine signaling that acts to sustain the caspase-1 activation in retinal cells initiated by high glucose stress in vitro and in vivo. These studies further support numerous studies which have suggested, but not unequivocally demonstrated, that chronic caspase-1 activation is crucial for the development of diabetic retinopathy based on the absence of acellular capillary development in caspase-1 knock-out mice. Of course, these studies have raised additional important questions that must be addressed in future research.

167 5.2 Is high glucose-induced caspase-1 activation mediated by inflammasome signaling complexes in retinal Müller cells?

Activation and assembly of several different inflammasomes has been well described in pathogenic models of caspase-1 dependent inflammation and cell death (63, 183, 214). Induction of the NRLP3 inflammasome in particular, has been shown to require priming and activating stimuli in these models. Priming by

DAMPS or PAMPS results in increased NFκB-dependent expression of both

NLRP3, a critical upstream regulator of caspase-1, and pro-IL-1β, the major downstream substrate of caspase-1 (94, 123, 168). Sustained P2X7R activation is just one well-characterized stimulator of NLRP3 inflammasome assembly. It is not entirely clear how P2X7R activation induces assembly, but multiple studies have indicated a central role for increased K+ efflux (62, 93). However, P2X7R activation is not required in all cases of NLRP3-mediated inflammasome activation. Other mechanisms known to induce K+ efflux, including lysosomal membrane disruption and exposure to microbial pore-forming toxins, also induce

NLRP3 inflammasome assembly (11, 62, 64, 73, 203). For example, phagocytosis of particulate crystals (such as cholesterol, asbestos, and alum) which stimulates lysosomal disruption, has been linked to K+ efflux-dependent

NLRP3 inflammasome activation. Interestingly, phagocytosis of IAPP (islet amyloid polypeptide) has also been directly linked to pancreatic β-islet cell inflammation and death through induction of the NLRP3 inflammasome (127). In addition, ROS production has been shown to induce NLRP3 inflammasome activation. Whether high glucose-induced caspase-1 activation unequivocally

168 requires the NLRP3 inflammasome is unclear. Inflammasome independent mechanisms for caspase-1 activation have been described but are not well understood.

Even less understood is the role and regulation of inflammasome signaling cascades within the context of sterile inflammatory diseases such as diabetic retinopathy. Recent evidence demonstrates a specific role for the mutations in

NLRP3 in mediating chronic inflammation in a number of genetic diseases (59,

130). Several studies have also described a role for the NRLP3 inflammasome in

“sensing” metabolic stress. However how NLRP3 senses metabolic disturbances is not entirely clear, and whether sterile inflammatory mechanisms necessarily require autocrine P2X7R activation or alternative K+ efflux mechanisms to mediate NLRP3 inflammasome assembly has not been fully characterized (159,

184, 203).

In this dissertation research, I demonstrate that elevated glucose levels positively regulate mRNA levels of NLRP3 within retinal Müller cells (Chapter 4).

However, whether high glucose-induced caspase-1 activation is indeed dependent on NLRP3 inflammasome assembly remains to be tested. Therefore, siRNA and NLRP3 knockdown studies in rMC-1 and rodent models of diabetes will shed light on this important issue. Knockdown of NLRP3 in vivo or in vitro should reduce retinal caspase-1 activity, IL-1β production, and acellular capillary formation if these processes are specifically dependent on NLRP3-mediated inflammasome signaling. One potential pitfall of these experiments is that other

NLRs can participate in inflammasome mediated caspase-1 activation. Since

169 apoptosis speck-like containing protein (ASC) is required for almost all types of inflammasome assembly – regardless of which NLR acts as the primary PAMP or DAMP sensor -- ASC knockdown experiments might better indicate whether caspase-1 activation require inflammasome signaling platforms under hyperglycemic conditions.

5.2.1 Does P2X7R participate in the development of diabetic retinopathy in vivo?

Although I was unable to detect significant expression levels of P2X7R in rMC-1 cells, it is possible that knockdown of P2X7R expression (P2X7R-/-) on other retinal cell types may still confer protection in vivo. Moreover, it remains formally possible that P2X7R are expressed in primary Müller cell lines or Müller cells in vivo, but that P2X7R expression is suppressed in rMC-1 cells as a consequence of their transformed status. Therefore, studies designed to measure acellular capillary formation and caspase-1 activation within the retinas of P2X7R-/- mice will be informative. Even if P2X7R are not expressed within mouse Müller cells, the receptor may still be expressed on other retinal cell types, such as resident macrophages and retinal vasculature endothelial cells that are known to also be directly affected by hyperglycemia, possibly resulting in protection following knockdown.

170 5.2.2 What are the sources of DAMPS that regulate high glucose-induced caspase-1 activation?

One of the main questions within diabetic research is, “what is the stimulator of inflammation.” Is it glucose itself or byproducts of glucose metabolism? Or, might high glucose-induced ROS or RAGE production be required? Could be hyperglycemia-induced release of cytochrome C act as either a DAMP or an inducer of apoptosis? Additionally, mitochondrial stress could

2+ induce an increase in cytosolic Ca i. Could such mitochondrial stress-triggered

Ca2+ release induce ATP secretion? Inhibition of mitochondrial stress has been shown to prevent acellular capillary generation in rodent models but it is not clear whether this protection is due to inhibition of mitochondrial generated ROS or cytochrome C or Ca2+ release (or a combination of the three) (106, 108-110).

Inhibition of the BAX channels which mediate cytochrome C release may begin to delineate between these potential signaling cascades. However, since both cytochrome C release and ROS generation have been linked to elevated TXNIP levels and subsequent caspase-1 activation, the identity of the involved DAMPS is probably more complicated than we realize and not simply an “all or none” situation. It is very likely a case of coordinated signaling between several mechanisms which will make identifying these pathways more difficult.

Data from the Mohr lab has shown that there is an increase in GLUT1 expression and glucose uptake in rMC-1 following high glucose treatment.

Experiments designed to inhibit glucose transport or intracellular metabolism may provide mechanistic insights as to how glucose regulates caspase-1 activation.

171 For example, in lieu of a high glucose stimulus, cells could be treated with various metabolites (i.e., fructose) or non-hydrolyzable forms of glucose (i.e. L- glucose) followed by measurement of caspase-1 activation to determine what specific roles that glucose itself as a ligand, intracellular glucose accumulation, or intracellular glucose metabolism play in mediating this process.

5.2.3 What mediates hyperglycemic induction of NF-κB activation?

Several studies have begun to shed light on how NFκB is regulated by high glucose conditions (48, 175, 190). It is possible that it is glucose itself, which can trigger carbohydrate-responsive elements in the promoters/enhancers of various “glucose-inducible” genes, can also indirectly stimulate NFκB-regulated gene activation, via signaling proteins that couple high glucose stress to the upstream kinase cascades required for NFκB nuclear translocation. Receptor

Interacting Protein 2 (RIP2) has also been shown to regulate NFκB activation and our studies have demonstrated upregulation of this protein in rMC-1 following high glucose treatment (Chapter 4) (47, 102, 119, 129, 179, 211).

However, these correlations raise yet another question: how does glucose regulate RIP2 expression or activation? Activation of carbohydrate responsive elements likely participate in directly regulating RIP2 expression while post- translational modifications like advanced glycation probably affect activation states of RIP2 (20, 27, 156, 175, 189, 190, 213). Following TLR activation

(specifically, TLR2 and TLR4) RIP2 is recruited to the intracellular signaling complex to mediate NFκB activation. Glycolipids are one type of TLR2 ligand that may be especially relevant to our hyperglycemia-induced RIP2 activation studies

172 (102). RIP2/NOD1/2 complex formation is required for NFκB activation. Thus, it is possible that RIP2 activation could be directly influenced by DAMPs or PAMPs sensed by NODs via the LRR-region (14). Additionally, as RIP2 is a protein kinase which can autophosphorylate, it is possible that function could be transmodulated by other protein kinases or phosphatases activated in response to high glucose stress (47, 129).

Chronic NFκB activation is possibly mediated by autocrine IL-1β signaling that also promotes sustained, late caspase-1 activation in vivo and in vitro

(Chapter 4). Western blots for IκB degradation in rMC-1 following treatment with exogenous IL-1β and high glucose would confirm this hypothesis. Inhibiting ROS generation or cytochrome C release or glucose metabolism and running similar

IκB Western blot analysis might also begin to explain how high glucose initially regulates NFκB.

5.3 What regulates autocrine ATP and adenosine release and/or metabolism within the context of hyperglycemia?

The production, release and metabolism of nucleotides are well-controlled processes that maintain intracellular concentrations ranging from 1 to 10 mM while extracellular concentrations are kept at 1 to 10 nM. Local expression of

ATP release channels and transporters, possible exocytosis of ATP-containing secretory vesicles, and regulated activity of the various ecto-nucleotidases and ecto-kinases all contribute to the tight regulation of local purine concentrations

173 (87). Because the total extracellular volume surrounding approximately 106 cells in a tightly packed tissue is about 1 µl, the total amount of ATP release required to activate P2 receptors (except P2X7R) under in vivo conditions is very small making tight regulation of nucleotide concentrations extremely important.

Identifying which enzymes and channels regulate ATP release and adenosine generation within the context of hyperglycemia in retinal cells is clearly important as purines play a vital role in high glucose-induced inflammation (Chapter 3).

5.3.1 ATP/Adenosine Metabolism

Degradation of either ATP by exogenous apyrase or adenosine by ADA inhibited high glucose-induced caspase-1 activation in retinal Müller cells

(Chapter 3). Exactly how high glucose conditions alter either expression or activity of Müller cell ecto-nucleotidases and ADA remains to be defined. It is possible that high glucose conditions negatively regulate these enzymes thereby increasing extracellular purine concentration.

Expression of NTPDase2 and ecto-5’-nucleotidase has been demonstrated in murine Müller cells (83, 172). The presence of NTPDase2 is particularly interesting because its presence results in ADP accumulation and delayed production of adenosine. One potential hypothesis is that in the diseased state, the balance of NTPDase2 expression is altered, shifting the balance away from negative regulation (by producing ADP and delaying adenosine production) towards an accumulation of either ATP (due to reduced expression of

NTPDase2) or adenosine (due to upregulated expression of NTPDase1).

174 However, expression of ecto-nucleotidases on other retinal cells that are in close proximity to Müller cells may also play a role in regulating purine concentrations.

A study by Ricatti et al demonstrated that NTPDase1 was expressed on the retinal vasculature, neurons, and horizontal and ganglion cell processes; all of which are in contact with Müller cells (172). The same study demonstrated that expression of ecto-nucleotidases may be species-specific necessitating further characterization of other models such as in rMC-1 cells. Studies utilizing inhibitors of ecto-apyrases like NTPDase2 (such as ARL-67156) and antagonists of ecto-5’-nucleotidase (like AOPCP) will be necessary to further understand the role of nucleotide metabolism in the context of diabetic retinopathy. Measurement of extracellular ATP and adenosine concentrations in the presence or absence of a high glucose stimulus, and with or without ecto-nucleotidase inhibitors and caspase-1 activity measurements on these samples will further elucidate the role of purines in regulating caspase-1 activation.

5.3.2 Candidate conduits for autocrine ATP release

Despite the electrochemical gradient for ATP efflux, regulated transport of

ATP is required to maintain critically important intracellular ATP levels. Clear characterization of release mechanisms remains elusive as there does not appear to be a centrally defined mechanism as release is largely cell or tissue type specific. For example, in excitatory cells, ATP is often co-packaged with neurotransmitters in synaptic vesicles as well as with dense core granules and is exocytically released following vesicle/granule fusion. Another release mechanism especially important to the immune response is the lytic release of

175 ATP during cellular injury. But several studies have demonstrated the importance of non-lytic release involving candidate channels and/or transporters from non- excitatory cells (7, 31, 54).

Studies have implicated a role for pannexin (Px) and connexin (Cx) hemichannels as ATP release channels; pannexins may or may not associate with

P2X7R to mediate this process (7, 50, 169). Other candidates include the ATP- binding cassette (ABC) transport proteins (15, 185), volume regulated anion channels (VRAC) (78) and maxi-anion channels (177, 178). Characterizing specific mechanisms for regulated ATP release remain the focus of many ongoing studies which will be of particular interest in the context of purine- mediated retinal cell damage during diabetic retinopathy.

My data demonstrate a role for ATP itself in inducing caspase-1 activation under high glucose conditions (Chapter 3). Therefore, identification of release pathways involved in autocrine ATP release from Müller cells is necessary.

Under physiological conditions, Müller cells normally release ATP and adenosine as an important part of their role in regulating retinal synaptic activity (220).

Exactly what is the stimulus for ATP release and whether there are increased concentrations of extracellular ATP under hyperglycemic conditions has not been determined. As previously discussed, even if ATP release is unchanged, ATP metabolism might be diminished in these models thus indirectly increasing ATP concentrations. It has been previously shown that Ca2+ release from intracellular stores is necessary for ATP release from astrocytes (40, 83, 155, 174). We have shown an increase in Ca2+ dynamics largely from intracellular stores following

176 high glucose stimulation (Chapter 3). Whether changes in intracellular Ca2+ homeostasis directly regulate autocrine ATP release in Müller cells remains to be determined. Experiments designed to measure extracellular ATP (using a luciferase based assay) concentrations following high glucose treatment and ±

Ca2+ chelating agents (such as BAPTA) will begin to shed light on this important issue (163, 164). Other experiments using thapsigargin, an inhibitor of the

SERCA pump (endoplasmic reticulum Ca2+) which is involved in clearance of intracellular Ca2+ by the ER, could also indicate a role for Ca2+ regulated ATP release. Data have shown that Müller also cells release ATP in response to osmotic swelling and glutamate receptor activation, especially in cases of retinal disease (217-220). Because we have demonstrated an increase in Ca2+ dynamics typical of P2Y receptors, it is possible that P2Y1,2 or 4R activation could participate in regulating autocrine ATP release in addition to any potential role these receptors may play in directly regulating caspase-1 activation and in mediating osmotic swelling.

Expression of Px and Cx hemichannels on retinal Müller cells remains somewhat unclear (117). Immunohistochemical co-staining for these channels and the Müller cell markers CRALPB (Cellular retinaldehyde-binding protein) and glutamine synthetase or western blot analysis clearly identifying the presence of these channels will allow for further experimentation to determine if they are utilized in Müller cell ATP release. Treatment with carbenoxalone, a non-specific

Px/Cx inhibitor may indicate a role for these channels if their presence is clearly demonstrated. Additionally, Px1 or 3 knockouts or Cx siRNA may also be

177 necessary to demonstrate which Px/Cx are specifically involved. Treatment with probenicid, a broad-spectrum inhibitor of anion channels, will further elucidate release mechanisms in Müller cells (163, 164). Of particular interest, data demonstrate that ATP is released by Cx43 in response to hypoxia in astrocytes

(115). Whether the hypoxic environment of the damaged diabetic retina contributes to ATP release from Müller cells remains has not been elucidated.

Once candidate release channels or transporters have been identified in control cells, it will be interesting to determine if their expression and/or activity is changed by high glucose treatment.

I have shown that high glucose-induced caspase-1 activation involves cAMP and presumably PKA signaling, so it is possible that PKA could also regulate ATP channel/transporter expression levels. Experiments involving treatment with FSK in normal glucose conditions followed by ATP and adenosine extracellular concentration measurements compared to high glucose treated and control cells will be necessary to determine the contribution of PKA activation to

ATP release mechanisms (and may shed light on what role PKA plays in directly regulating caspase-1 activation). Inhibitors of PKA such as H89 or KT5720 could be utilized to further determine whether increased PKA activation under hyperglycemic conditions regulates Müller cell ATP release (or caspase-1 activation). Interestingly, PKC is also known to regulate ATP release and is elevated in diabetic conditions, making this another possible regulator of ATP release mechanisms (3, 84, 199).

178 5.4 What role do increased levels of either cAMP or active PKA play in modulating high glucose-induced caspase-1 activation within the retina?

P1 adenosine receptors are coupled either positively (Gs-coupled protein receptors like A2a and A2b) or negatively (Gi-coupled protein receptors like A1) to cAMP and PKA activation, while P2Y12, P2Y13, and P2Y14 are Gi-coupled protein receptors which negatively regulate cAMP levels. My data indicate a clear contribution for activation of Gs-coupled protein receptors, like A2a or A2b, in mediating high glucose- induced caspase-1 activation. However, our studies did not completely explain what role the coupling of these receptors to cAMP/PKA signaling plays in regulating caspase-1 activation. Significant data have demonstrated a role for increased P2Y activation in several retinal injuries whereby these receptors function to restore ion and fluid homeostasis (to prevent

Müller cell swelling) and to regulate ATP release within the subretinal space (65,

219, 220). However, chronic activation of these receptors, as is seen in the case of chronic diabetes, could ultimately produce deleterious effects. Whether sustained P2Y activation is responsible for increased ATP release from Müller cells, or contributes to severe intracellular ion perturbations (such as changes in

2+ Ca i) which act as a signal for caspase-1 activation, or whether sustained increases in cAMP could directly affect caspase-1 expression and activation is not entirely clear. It is likely a combination of these events as activation of P1 receptors contributed to caspase-1 activation, presumably through cAMP regulation of pro-caspase-1 as was seen in an osteoblast progenitor study

(Chapter 3) (5). Additional areas of regulation by cAMP and PKA are discussed

179 in the previous section and include regulation of ecto-nucleotidase expression/function or ATP release channels.

5.4.1 Explore the mechanism of negative feedback by ADP and UTP in regulating caspase-1 activation in this system.

Surprisingly, treatment with exogenous ADP and UTP reduced high glucose-induced caspase-1 activation (Chapter 4). Originally, we hypothesized that treatment with UTP would have no effect at all since it does not bind to

P2X7R (the purinergic receptor initially hypothesized to mediate high glucose- induced caspase-1 activation) and that ADP would either have no effect or would positively regulate caspase-1 activation. These surprising results indicate the ability for multiple P2Y receptor subtypes to both positively and negatively regulate caspase-1 activity in the context of hyperglycemia. A likely scenario is that the combinatorial effects of multiple P2Y receptor subtypes function as a sort of a “see-saw” with a fine balance between positive and negative input to the caspase-1 signaling cascade. P2Y receptor subtypes that are Gq-coupled receptors could be responsible for positive regulation of caspase-1 by modulating cytosolic Ca2+ (such as P2Y1, P2Y2, P2Y4, P2Y6 and P2Y11) while negative

PKA regulation by Gi-coupled receptor subtypes (P2Y12, P2Y13, and P2Y14) could represent negative feedback. Under pathological conditions, perhaps the system is biased towards positive regulation. It is not clear exactly what processes are involved or why this negative regulation is not functioning in vivo or in vitro when both ADP and UTP are present. It is possible that very low concentrations of these nucleotides are not efficacious or that high

180 concentrations used in these experiments actually desensitized the P2Y receptors that normally positively regulate caspase-1 activation. Experiments to determine which P2Y receptors mediate the positive and negative feedback will be illuminating. Utilizing the P2Y1 inhibitor MRS 2179 will be especially prudent in this case as it is clear that P2Y1/2 receptors are largely responsible for the

2+ dramatic disturbances in Ca I seen in our model.

5.4.2 Which specific adenosine receptor participates in caspase-1 activation in retinal Müller cells?

My studies determined a distinct role for adenosine receptor activation in regulating high glucose-induced caspase-1 activation. Expression of the A1 receptor subtype on Müller cells is suggested by Iandiev et al (83, 218).

Additionally, my FSK studies indicate a distinct role for A2 receptors that positively regulate cAMP levels. However, our DPCPX and MRS1754 studies utilized drug concentrations which are known to inhibit more than one receptor subtype. Experiments using lower concentrations of DPCPX and MRS1754 which will more specifically inhibit either A1 or A2 receptors, as well as another

A2A antagonist, 8-(3-chlorostyryl) caffeine (CSC), will further define which receptors are involved. Several adenosine receptor knockouts are well characterized and available for studies in diabetic mice (222). Additional siRNA experiments in vitro to delete different P1 receptors would indicate which receptor subtype(s) mediate the high glucose-adenosine effect in Müller cells.

181 5.5 How does elevated TXNIP expression participate in regulation of caspase-1 in the retina?

Elevated free TXNIP and increased TXNIP expression have been directly linked to DM pathogenesis and its complications, as well as increased caspase-

1/IL-1β signaling (74, 103, 232). The TXNIP/caspase-1 signaling axis is a particularly interesting hypothesis in the pathogenesis of diabetic retinopathy because glucose, adenosine and ROS all positively regulate TXNIP expression

(20, 25, 27, 181). Several studies have demonstrated that decreased TXNIP expression (either in animal knockout models or in vitro siRNA knockdown) is protective in several diabetes-related complications and my data demonstrate both an increase in mRNA and protein expression of TXNIP in Müller cells

(Chapter 3) (23, 24, 157, 191). However, how TXNIP regulates caspase-1 activation and whether this protein directly complexes with the NLRP3 inflammasome is an area of contentious debate.

To first define whether TXNIP specifically regulates caspase-1 activation in Müller cells, siRNA studies knocking down this protein will be necessary.

Endpoint analysis of caspase-1 activity and IL-1β production following siRNA knockdown and high glucose treatment in vitro should be performed. TXNIP knockout animals are commercially available from Jackson Laboratories and should be used in diabetic animal studies to determine whether TXNIP deletion protects the retina from acellular capillary formation and increased caspase-1 activation in vivo.

182 Studies to define the specific function of TXNIP in mediating caspase-1 activation will certainly be more complicated, but equally interesting. Pull down assays to determine what proteins free TXNIP binds to in the context of metabolic stress will provide some clues as to its potential role in mediating inflammasome formation. It is possible that TXNIP may not directly interact with the inflammasome itself but still regulate caspase-1 activation through chromatin modification, other glucose-dependent modifications of either gene expression or post-translational modification of proteins themselves, as an “activator” of the inflammasome or non-inflammasome dependent mechanisms of caspase-1 activation. Inhibiting mitochondrial stress in studies similar to those performed by the Kowluru group and measuring caspase-1 activation, cytokine production, and

TXNIP expression will prove extremely useful in further answering these questions (108-110).

183 5.6 Conclusion

Currently the only treatments for DR include good control of blood glucose levels with insulin and laser ablation of damaged capillaries. It has proven difficult to maintain consistent blood glucose levels in the long term and this is especially difficult for many patients to achieve due to metabolic memory that many experience in the late stages of diabetes. Laser photocoagulation of bleeding retinal blood vessels usually results in some loss of vision simply from the procedure itself and does not address the underlying chronic inflammatory events associated with diabetes. It is clear then that understanding the mechanism of high glucose induced retinal cell dysfunction is necessary for the development of new therapies. The work presented in this dissertation describes a novel, multi-phasic regulatory process for high glucose-induced caspase-1/IL-

1β signaling in the diabetic mouse retina in vivo and in the rat retinal Müller cell line (rMC-1) in vitro. Initially, ATP and adenosine based purinergic signaling regulate caspase-1 activation in retinal Müller cells while the sustained phase is predominately dependent on functional IL-1β autocrine signaling in vitro and in vivo.

Clinically, the use of purinergic antagonists (such as those that inhibit A2a or A2b or P2Y1/2 receptors) may be especially useful prevent to Müller cell IL-1β production in early stages of diabetes and DR prior to capillary dropout. Inhibition of Müller cell purinergic based inflammatory signaling may prevent retinal degeneration by preventing inflammatory mediated Müller cell dysfunction and death as these cells are vital to retinal homeostasis. Aside from the effects these

184 antagonists may have on Müller cells directly, aberrant purinergic signaling may also participate in mediating hyperglycemia-induced inflammatory events in other retinal cells.

The relevance of anakinra and other IL-1β neutralizing antibodies or receptor antagonists is especially apparent based on data from these studies.

Acellular capillary formation, a hallmark of the disease, was inhibited by knockdown of both the IL-1R1 and caspase-1 in the retinas of diabetic mice.

Therefore, treatment with drugs targeted at preventing sustained caspase-1 activation by inhibiting autocrine IL-1β signaling will prove useful in patients with moderate non-proliferative DR and possibly even in more severe cases of proliferative DR.

In conclusion, data from these studies have identified a novel role for two distinct regulatory pathways in mediating high glucose-induced caspase-1 activation and acellular capillary generation in diabetic retinas and retinal Müller cells.

185

REFERENCES

186

1. Abulafia DP, de Rivero Vaccari JP, Lozano JD, Lotocki G, Keane RW, and Dietrich WD. Inhibition of the inflammasome complex reduces the inflammatory response after thromboembolic stroke in mice. J Cereb Blood Flow Metab 29: 534-544, 2009. 2. Adinolfi E, Cirillo M, Woltersdorf R, Falzoni S, Chiozzi P, Pellegatti P, Callegari MG, Sandona D, Markwardt F, Schmalzing G, and Di Virgilio F. Trophic activity of a naturally occurring truncated isoform of the P2X7 receptor. FASEB J 24: 3393-3404. 3. Aiello LP, Clermont A, Arora V, Davis MD, Sheetz MJ, and Bursell SE. Inhibition of PKC beta by oral administration of ruboxistaurin is well tolerated and ameliorates diabetes-induced retinal hemodynamic abnormalities in patients. Invest Ophthalmol Vis Sci 47: 86-92, 2006. 4. Al-Gayyar MM, Abdelsaid MA, Matragoon S, Pillai BA, and El-Remessy AB. Thioredoxin Interacting Protein Is a Novel Mediator of Retinal Inflammation and Neurotoxicity. Br J Pharmacol. 5. Almeida MQ, Tsang KM, Cheadle C, Watkins T, Grivel JC, Nesterova M, Goldbach-Mansky R, and Stratakis CA. Protein kinase A regulates caspase-1 via Ets-1 in bone stromal cell-derived lesions: a link between cyclic AMP and pro-inflammatory pathways in osteoblast progenitors. Hum Mol Genet 20: 165-175. 6. Aymeric L, Apetoh L, Ghiringhelli F, Tesniere A, Martins I, Kroemer G, Smyth MJ, and Zitvogel L. Tumor cell death and ATP release prime dendritic cells and efficient anticancer immunity. Cancer Res 70: 855-858. 7. Bao L, Locovei S, and Dahl G. Pannexin membrane channels are mechanosensitive conduits for ATP. FEBS Lett 572: 65-68, 2004. 8. Barber AJ, Gardner TW, and Abcouwer SF. The significance of vascular and neural apoptosis to the pathology of diabetic retinopathy. Invest Ophthalmol Vis Sci 52: 1156-1163. 9. Barile GR, Pachydaki SI, Tari SR, Lee SE, Donmoyer CM, Ma W, Rong LL, Buciarelli LG, Wendt T, Horig H, Hudson BI, Qu W, Weinberg AD, Yan SF, and Schmidt AM. The RAGE axis in early diabetic retinopathy. Invest Ophthalmol Vis Sci 46: 2916-2924, 2005. 10. Barlovic DP, Soro-Paavonen A, and Jandeleit-Dahm KA. RAGE biology, atherosclerosis and diabetes. Clin Sci (Lond) 121: 43-55. 11. Bauernfeind F, Ablasser A, Bartok E, Kim S, Schmid-Burgk J, Cavlar T, and Hornung V. Inflammasomes: current understanding and open questions. Cell Mol Life Sci 68: 765-783. 12. Bauernfeind FG, Horvath G, Stutz A, Alnemri ES, MacDonald K, Speert D, Fernandes-Alnemri T, Wu J, Monks BG, Fitzgerald KA, Hornung V, and Latz E. Cutting edge: NF-kappaB activating pattern recognition and cytokine receptors license NLRP3 inflammasome activation by regulating NLRP3 expression. J Immunol 183: 787-791, 2009. 13. Bergsbaken T, Fink SL, den Hartigh AB, Loomis WP, and Cookson BT. Coordinated Host Responses during Pyroptosis: Caspase-1-Dependent Lysosome Exocytosis and Inflammatory Cytokine Maturation. J Immunol. 14. Boyer L, Magoc L, Dejardin S, Cappillino M, Paquette N, Hinault C, Charriere GM, Ip WK, Fracchia S, Hennessy E, Erturk-Hasdemir D, Reichhart JM, Silverman N, Lacy-Hulbert A, and Stuart LM. Pathogen-Derived Effectors Trigger Protective Immunity via Activation of the Rac2 Enzyme and the IMD or Rip Kinase Signaling Pathway. Immunity 35: 536-549. 15. Braunstein GM, Roman RM, Clancy JP, Kudlow BA, Taylor AL, Shylonsky VG, Jovov B, Peter K, Jilling T, Ismailov, II, Benos DJ, Schwiebert LM, Fitz JG, and Schwiebert EM. Cystic fibrosis transmembrane conductance regulator facilitates ATP

187 release by stimulating a separate ATP release channel for autocrine control of cell volume regulation. J Biol Chem 276: 6621-6630, 2001. 16. Bringmann A, Pannicke T, Grosche J, Francke M, Wiedemann P, Skatchkov SN, Osborne NN, and Reichenbach A. Müller cells in the healthy and diseased retina. Prog Retin Eye Res 25: 397-424, 2006. 17. Bringmann A, Pannicke T, Moll V, Milenkovic I, Faude F, Enzmann V, Wolf S, and Reichenbach A. Upregulation of P2X(7) receptor currents in Müller glial cells during proliferative vitreoretinopathy. Invest Ophthalmol Vis Sci 42: 860-867, 2001. 18. Busik JV, Mohr S, and Grant MB. Hyperglycemia-induced reactive oxygen species toxicity to endothelial cells is dependent on paracrine mediators. Diabetes 57: 1952-1965, 2008. 19. Ceruti S, Fumagalli M, Villa G, Verderio C, and Abbracchio MP. Purinoceptor- mediated calcium signaling in primary neuron-glia trigeminal cultures. Cell Calcium 43: 576-590, 2008. 20. Cha-Molstad H, Saxena G, Chen J, and Shalev A. Glucose-stimulated expression of Txnip is mediated by carbohydrate response element-binding protein, p300, and histone H4 acetylation in pancreatic beta cells. J Biol Chem 284: 16898-16905, 2009. 21. Chan PS, Kanwar M, and Kowluru RA. Resistance of retinal inflammatory mediators to suppress after reinstitution of good glycemic control: novel mechanism for metabolic memory. J Diabetes Complications 2008. 22. Chen GY, and Nunez G. Sterile inflammation: sensing and reacting to damage. Nat Rev Immunol 10: 826-837. 23. Chen J, Fontes G, Saxena G, Poitout V, and Shalev A. Lack of TXNIP protects against mitochondria-mediated apoptosis but not against fatty acid-induced ER stress- mediated beta-cell death. Diabetes 59: 440-447. 24. Chen J, Hui ST, Couto FM, Mungrue IN, Davis DB, Attie AD, Lusis AJ, Davis RA, and Shalev A. Thioredoxin-interacting protein deficiency induces Akt/Bcl-xL signaling and pancreatic beta-cell mass and protects against diabetes. FASEB J 22: 3581-3594, 2008. 25. Chen JL, Merl D, Peterson CW, Wu J, Liu PY, Yin H, Muoio DM, Ayer DE, West M, and Chi JT. Lactic acidosis triggers starvation response with paradoxical induction of TXNIP through MondoA. PLoS Genet 6. 26. Chen Z, Lopez-Ramos DA, Yoshihara E, Maeda Y, Masutani H, Sugie K, Maeda M, and Yodoi J. Thioredoxin-binding protein-2 (TBP-2/VDUP1/TXNIP) regulates T-cell sensitivity to glucocorticoid during HTLV-I-induced transformation. Leukemia 25: 440-448. 27. Cheng DW, Jiang Y, Shalev A, Kowluru R, Crook ED, and Singh LP. An analysis of high glucose and glucosamine-induced gene expression and oxidative stress in renal mesangial cells. Arch Physiol Biochem 112: 189-218, 2006. 28. Church LD. Primer: Inflammasomes and interleukin 1β in inflammatory disorders. Nature Clinical Practice: Rheumatology 4: 34-42, 2008. 29. Coddou C, Yan Z, Obsil T, Huidobro-Toro JP, and Stojilkovic SS. Activation and Regulation of Purinergic P2X Receptor Channels. Pharmacol Rev 63: 641-683. 30. Cohen GM. Caspases: the executioners of apoptosis. Biochem J 326 ( Pt 1): 1-16, 1997. 31. Corriden R, and Insel PA. Basal release of ATP: an autocrine-paracrine mechanism for cell regulation. Sci Signal 3: re1. 32. Costa G, Pereira T, Neto AM, Cristovao AJ, Ambrosio AF, and Santos PF. High glucose changes extracellular adenosine triphosphate levels in rat retinal cultures. J Neurosci Res 87: 1375-1380, 2009.

188 33. Coutinho-Silva R, Robson T, Beales PE, and Burnstock G. Changes in expression of P2X7 receptors in NOD mouse pancreas during the development of diabetes. Autoimmunity 40: 108-116, 2007. 34. Cowlen MS, Zhang VZ, Warnock L, Moyer CF, Peterson WM, and Yerxa BR. Localization of ocular P2Y2 receptor gene expression by in situ hybridization. Exp Eye Res 77: 77-84, 2003. 35. Creagh EM, Conroy H, and Martin SJ. Caspase-activation pathways in apoptosis and immunity. Immunol Rev 193: 10-21, 2003. 36. Daheshia M, and Yao JQ. The interleukin 1beta pathway in the pathogenesis of osteoarthritis. J Rheumatol 35: 2306-2312, 2008. 37. Dale N. Purinergic signaling in hypothalamic tanycytes: Potential roles in chemosensing. Semin Cell Dev Biol 22: 237-244. 38. Daly JW, Jacobson KA, and Ukena D. Adenosine receptors: development of selective agonists and antagonists. Prog Clin Biol Res 230: 41-63, 1987. 39. Davis CN, Mann E, Behrens MM, Gaidarova S, Rebek M, Rebek J, Jr., and Bartfai T. MyD88-dependent and -independent signaling by IL-1 in neurons probed by bifunctional Toll/IL-1 receptor domain/BB-loop mimetics. Proc Natl Acad Sci U S A 103: 2953-2958, 2006. 40. De Vuyst E, Decrock E, Cabooter L, Dubyak GR, Naus CC, Evans WH, and Leybaert L. Intracellular calcium changes trigger connexin 32 hemichannel opening. EMBO J 25: 34-44, 2006. 41. Demircan N, Safran BG, Soylu M, Ozcan AA, and Sizmaz S. Determination of vitreous interleukin-1 (IL-1) and tumour necrosis factor (TNF) levels in proliferative diabetic retinopathy. Eye 20: 1366-1369, 2006. 42. Detrick B, Lee MT, Chin MS, Hooper LC, Chan CC, and Hooks JJ. Experimental coronavirus retinopathy (ECOR): retinal degeneration susceptible mice have an augmented interferon and chemokine (CXCL9, CXCL10) response early after virus infection. J Neuroimmunol 193: 28-37, 2008. 43. Dinarello CA. A clinical perspective of IL-1beta as the gatekeeper of inflammation. Eur J Immunol 41: 1203-1217. 44. Dinarello CA. Blocking interleukin-1beta in acute and chronic autoinflammatory diseases. J Intern Med 269: 16-28. 45. Dinarello CA. Interleukin-1 beta, interleukin-18, and the interleukin-1 beta converting enzyme. Ann N Y Acad Sci 856: 1-11, 1998. 46. Dinarello CA, Donath MY, and Mandrup-Poulsen T. Role of IL-1beta in type 2 diabetes. Curr Opin Endocrinol Diabetes Obes 17: 314-321. 47. Dorsch M, Wang A, Cheng H, Lu C, Bielecki A, Charron K, Clauser K, Ren H, Polakiewicz RD, Parsons T, Li P, Ocain T, and Xu Y. Identification of a regulatory autophosphorylation site in the serine-threonine kinase RIP2. Cell Signal 18: 2223-2229, 2006. 48. Du Y, Sarthy VP, and Kern TS. Interaction between NO and COX pathways in retinal cells exposed to elevated glucose and retina of diabetic rats. Am J Physiol Regul Integr Comp Physiol 287: R735-741, 2004. 49. Duarte JM, Oses JP, Rodrigues RJ, and Cunha RA. Modification of purinergic signaling in the hippocampus of streptozotocin-induced diabetic rats. Neuroscience 149: 382-391, 2007. 50. Dubyak GR. Both sides now: multiple interactions of ATP with pannexin-1 hemichannels. Focus on "A permeant regulating its permeation pore: inhibition of pannexin 1 channels by ATP". Am J Physiol Cell Physiol 296: C235-241, 2009.

189 51. Dubyak GR. Go it alone no more--P2X7 joins the society of heteromeric ATP-gated receptor channels. Mol Pharmacol 72: 1402-1405, 2007. 52. Dufner A, and Mak TW. CARD tricks: controlling the interactions of CARD6 with RICK and microtubules. Cell Cycle 5: 797-800, 2006. 53. Dunne A, and O'Neill LA. The interleukin-1 receptor/Toll-like receptor superfamily: signal transduction during inflammation and host defense. Sci STKE 2003: re3, 2003. 54. Ebihara L. New roles for connexons. News Physiol Sci 18: 100-103, 2003. 55. Eisenbarth SC, Colegio OR, O'Connor W, Sutterwala FS, and Flavell RA. Crucial role for the Nalp3 inflammasome in the immunostimulatory properties of aluminium adjuvants. Nature 453: 1122-1126, 2008. 56. Elliott MR, Chekeni FB, Trampont PC, Lazarowski ER, Kadl A, Walk SF, Park D, Woodson RI, Ostankovich M, Sharma P, Lysiak JJ, Harden TK, Leitinger N, and Ravichandran KS. Nucleotides released by apoptotic cells act as a find-me signal to promote phagocytic clearance. Nature 461: 282-286, 2009. 57. Feit-Leichman RA, Kinouchi R, Takeda M, Fan Z, Mohr S, Kern TS, and Chen DF. Vascular damage in a mouse model of diabetic retinopathy: relation to neuronal and glial changes. Invest Ophthalmol Vis Sci 46: 4281-4287, 2005. 58. Ferrari D, Pizzirani C, Adinolfi E, Lemoli RM, Curti A, Idzko M, Panther E, and Di Virgilio F. The P2X7 receptor: a key player in IL-1 processing and release. J Immunol 176: 3877-3883, 2006. 59. Ferrero-Miliani L, Nielsen OH, Andersen PS, and Girardin SE. Chronic inflammation: importance of NOD2 and NALP3 in interleukin-1beta generation. Clin Exp Immunol 147: 227-235, 2007. 60. Fink SL, and Cookson BT. Apoptosis, pyroptosis, and necrosis: mechanistic description of dead and dying eukaryotic cells. Infect Immun 73: 1907-1916, 2005. 61. Fink SL, and Cookson BT. Caspase-1-dependent pore formation during pyroptosis leads to osmotic lysis of infected host macrophages. Cell Microbiol 8: 1812-1825, 2006. 62. Franchi L, Kanneganti TD, Dubyak GR, and Nunez G. Differential requirement of P2X7 receptor and intracellular K+ for caspase-1 activation induced by intracellular and extracellular bacteria. J Biol Chem 282: 18810-18818, 2007. 63. Franchi L, Munoz-Planillo R, Reimer T, Eigenbrod T, and Nunez G. Inflammasomes as microbial sensors. Eur J Immunol 40: 611-615. 64. Franchi L, and Nunez G. The Nlrp3 inflammasome is critical for aluminium hydroxide-mediated IL-1beta secretion but dispensable for adjuvant activity. Eur J Immunol 38: 2085-2089, 2008. 65. Francke M, Uhlmann S, Pannicke T, Goczalik I, Uckermann O, Weick M, Hartig W, Wiedemann P, Reichenbach A, and Bringmann A. Experimental dispase-induced retinopathy causes up-regulation of P2Y receptor-mediated calcium responses in Müller glial cells. Ophthalmic Res 35: 30-41, 2003. 66. Frank RN. Diabetic retinopathy. N Engl J Med 350: 48-58, 2004. 67. Franke H, Gunther A, Grosche J, Schmidt R, Rossner S, Reinhardt R, Faber- Zuschratter H, Schneider D, and Illes P. P2X7 receptor expression after ischemia in the cerebral cortex of rats. J Neuropathol Exp Neurol 63: 686-699, 2004. 68. Fredholm BB, AP IJ, Jacobson KA, Linden J, and Müller CE. International Union of Basic and Clinical Pharmacology. LXXXI. Nomenclature and classification of adenosine receptors--an update. Pharmacol Rev 63: 1-34. 69. Fries JE, Goczalik IM, Wheeler-Schilling TH, Kohler K, Guenther E, Wolf S, Wiedemann P, Bringmann A, Reichenbach A, Francke M, and Pannicke T. Identification

190 of P2Y receptor subtypes in human Müller glial cells by physiology, single cell RT-PCR, and immunohistochemistry. Invest Ophthalmol Vis Sci 46: 3000-3007, 2005. 70. Fries JE, Wheeler-Schilling TH, Kohler K, and Guenther E. Distribution of metabotropic P2Y receptors in the rat retina: a single-cell RT-PCR study. Brain Res Mol Brain Res 130: 1-6, 2004. 71. Gastinger MJ, Singh RS, and Barber AJ. Loss of cholinergic and dopaminergic amacrine cells in streptozotocin-diabetic rat and Ins2Akita-diabetic mouse retinas. Invest Ophthalmol Vis Sci 47: 3143-3150, 2006. 72. Genuth S. Insights from the diabetes control and complications trial/epidemiology of diabetes interventions and complications study on the use of intensive glycemic treatment to reduce the risk of complications of type 1 diabetes. Endocr Pract 12 Suppl 1: 34-41, 2006. 73. Gong YN, Wang X, Wang J, Yang Z, Li S, Yang J, Liu L, Lei X, and Shao F. Chemical probing reveals insights into the signaling mechanism of inflammasome activation. Cell Res 20: 1289-1305. 74. Hamada Y, and Fukagawa M. A possible role of thioredoxin interacting protein in the pathogenesis of streptozotocin-induced diabetic nephropathy. Kobe J Med Sci 53: 53-61, 2007. 75. Harden TK, Sesma JI, Fricks IP, and Lazarowski ER. Signalling and pharmacological properties of the P2Y receptor. Acta Physiol (Oxf) 199: 149-160. 76. Hellman B, Jansson L, Dansk H, and Grapengiesser E. Effects of external ATP on Ca(2+) signalling in endothelial cells isolated from mouse islets. Endocrine 32: 33-40, 2007. 77. Hiken JF, and Steinberg TH. ATP downregulates P2X7 and inhibits osteoclast formation in RAW cells. Am J Physiol Cell Physiol 287: C403-412, 2004. 78. Hisadome K, Koyama T, Kimura C, Droogmans G, Ito Y, and Oike M. Volume- regulated anion channels serve as an auto/paracrine nucleotide release pathway in aortic endothelial cells. J Gen Physiol 119: 511-520, 2002. 79. Hong S, Schwarz N, Brass A, Seman M, Haag F, Koch-Nolte F, Schilling WP, and Dubyak GR. Differential regulation of P2X7 receptor activation by extracellular nicotinamide adenine dinucleotide and ecto-ADP-ribosyltransferases in murine macrophages and T cells. J Immunol 183: 578-592, 2009. 80. Hori S, and Mukai N. Ultrastructural lesions of retinal pericapillary Müller cells in streptozotocin-induced diabetic rats. Albrecht Von Graefes Arch Klin Exp Ophthalmol 213: 1- 9, 1980. 81. Hori S, Nishida T, and Mukai N. Ultrastructural studies on lysosomes in retinal Müller cells of streptozotocin-diabetic rats. Invest Ophthalmol Vis Sci 19: 1295-1300, 1980. 82. Houston D, Costanzi S, Jacobson KA, and Harden TK. Development of selective high affinity antagonists, agonists, and radioligands for the P2Y1 receptor. Comb Chem High Throughput Screen 11: 410-419, 2008. 83. Iandiev I, Wurm A, Pannicke T, Wiedemann P, Reichenbach A, Robson SC, Zimmermann H, and Bringmann A. Ectonucleotidases in Müller glial cells of the rodent retina: Involvement in inhibition of osmotic cell swelling. Purinergic Signal 3: 423-433, 2007. 84. Ishii H, Koya D, and King GL. Protein kinase C activation and its role in the development of vascular complications in diabetes mellitus. J Mol Med (Berl) 76: 21-31, 1998. 85. Jabs R, Guenther E, Marquordt K, and Wheeler-Schilling TH. Evidence for P2X(3), P2X(4), P2X(5) but not for P2X(7) containing purinergic receptors in Müller cells of the rat retina. Brain Res Mol Brain Res 76: 205-210, 2000.

191 86. Jacobson KA, Ivanov AA, de Castro S, Harden TK, and Ko H. Development of selective agonists and antagonists of P2Y receptors. Purinergic Signal 5: 75-89, 2009. 87. Joseph SM, Buchakjian MR, and Dubyak GR. Colocalization of ATP release sites and ecto-ATPase activity at the extracellular surface of human astrocytes. J Biol Chem 278: 23331-23342, 2003. 88. Joussen AM, Huang S, Poulaki V, Camphausen K, Beecken WD, Kirchhof B, and Adamis AP. In vivo retinal gene expression in early diabetes. Invest Ophthalmol Vis Sci 42: 3047-3057, 2001. 89. Joussen AM, Murata T, Tsujikawa A, Kirchhof B, Bursell SE, and Adamis AP. Leukocyte-mediated endothelial cell injury and death in the diabetic retina. Am J Pathol 158: 147-152, 2001. 90. Joussen AM, Poulaki V, Le ML, Koizumi K, Esser C, Janicki H, Schraermeyer U, Kociok N, Fauser S, Kirchhof B, Kern TS, and Adamis AP. A central role for inflammation in the pathogenesis of diabetic retinopathy. Faseb J 18: 1450-1452, 2004. 91. Joussen AM, Poulaki V, Mitsiades N, Kirchhof B, Koizumi K, Dohmen S, and Adamis AP. Nonsteroidal anti-inflammatory drugs prevent early diabetic retinopathy via TNF-alpha suppression. Faseb J 16: 438-440, 2002. 92. Junger WG. Immune cell regulation by autocrine purinergic signalling. Nat Rev Immunol 11: 201-212. 93. Kahlenberg JM, and Dubyak GR. Mechanisms of caspase-1 activation by P2X7 receptor-mediated K+ release. Am J Physiol Cell Physiol 286: C1100-1108, 2004. 94. Kahlenberg JM, Lundberg KC, Kertesy SB, Qu Y, and Dubyak GR. Potentiation of caspase-1 activation by the P2X7 receptor is dependent on TLR signals and requires NF- kappaB-driven protein synthesis. J Immunol 175: 7611-7622, 2005. 95. Kalliolias GD, and Liossis SN. The future of the IL-1 receptor antagonist anakinra: from rheumatoid arthritis to adult-onset Still's disease and systemic-onset juvenile idiopathic arthritis. Expert Opin Investig Drugs 17: 349-359, 2008. 96. Kang SJ, Wang S, Hara H, Peterson EP, Namura S, Amin-Hanjani S, Huang Z, Srinivasan A, Tomaselli KJ, Thornberry NA, Moskowitz MA, and Yuan J. Dual role of caspase-11 in mediating activation of caspase-1 and caspase-3 under pathological conditions. J Cell Biol 149: 613-622, 2000. 97. Karl A, Wurm A, Pannicke T, Krugel K, Obara-Michlewska M, Wiedemann P, Reichenbach A, Albrecht J, and Bringmann A. Synergistic action of hypoosmolarity and glutamine in inducing acute swelling of retinal glial (Müller ) cells. Glia 59: 256-266. 98. Kepp O, Galluzzi L, Zitvogel L, and Kroemer G. Pyroptosis - a cell death modality of its kind? Eur J Immunol 40: 627-630. 99. Kern TS. Contributions of inflammatory processes to the development of the early stages of diabetic retinopathy. Exp Diabetes Res 2007: 95103, 2007. 100. Kern TS, Miller CM, Du Y, Zheng L, Mohr S, Ball SL, Kim M, Jamison JA, and Bingaman DP. Topical administration of nepafenac inhibits diabetes-induced retinal microvascular disease and underlying abnormalities of retinal metabolism and physiology. Diabetes 56: 373-379, 2007. 101. Klotz KN. Adenosine receptors and their ligands. Naunyn Schmiedebergs Arch Pharmacol 362: 382-391, 2000. 102. Kobayashi K, Inohara N, Hernandez LD, Galan JE, Nunez G, Janeway CA, Medzhitov R, and Flavell RA. RICK/Rip2/CARDIAK mediates signalling for receptors of the innate and adaptive immune systems. Nature 416: 194-199, 2002. 103. Koenen TB, Stienstra R, van Tits LJ, de Graaf J, Stalenhoef AF, Joosten LA, Tack CJ, and Netea MG. Hyperglycemia activates caspase-1 and TXNIP-mediated IL-1beta transcription in human adipose tissue. Diabetes 60: 517-524.

192 104. Kolb H. How the Retina Works. American Scientist 91: 8, 2003. 105. Kolb HN, Ralph; Fernandez, Eduardo; Jones, Bryan editor. The organization of the retina and visual system. http://webvision.med.utah.edu/ 1995. 106. Kowluru RA. Diabetic retinopathy: mitochondrial dysfunction and retinal capillary cell death. Antioxid Redox Signal 7: 1581-1587, 2005. 107. Kowluru RA. Effect of reinstitution of good glycemic control on retinal oxidative stress and nitrative stress in diabetic rats. Diabetes 52: 818-823, 2003. 108. Kowluru RA, and Kanwar M. Effects of curcumin on retinal oxidative stress and inflammation in diabetes. Nutr Metab (Lond) 4: 8, 2007. 109. Kowluru RA, Kowluru V, Xiong Y, and Ho YS. Overexpression of mitochondrial superoxide dismutase in mice protects the retina from diabetes-induced oxidative stress. Free Radic Biol Med 41: 1191-1196, 2006. 110. Kowluru RA, and Odenbach S. Role of interleukin-1beta in the development of retinopathy in rats: effect of antioxidants. Invest Ophthalmol Vis Sci 45: 4161-4166, 2004. 111. Krugel K, Wurm A, Pannicke T, Hollborn M, Karl A, Wiedemann P, Reichenbach A, Kohen L, and Bringmann A. Involvement of oxidative stress and mitochondrial dysfunction in the osmotic swelling of retinal glial cells from diabetic rats. Exp Eye Res 92: 87-93. 112. Kusner LL, Sarthy VP, and Mohr S. Nuclear translocation of glyceraldehyde-3- phosphate dehydrogenase: a role in high glucose-induced apoptosis in retinal Müller cells. Invest Ophthalmol Vis Sci 45: 1553-1561, 2004. 113. Le Feuvre R, Brough D, and Rothwell N. Extracellular ATP and P2X7 receptors in neurodegeneration. Eur J Pharmacol 447: 261-269, 2002. 114. Li Q, and Puro DG. Diabetes-induced dysfunction of the glutamate transporter in retinal Müller cells. Invest Ophthalmol Vis Sci 43: 3109-3116, 2002. 115. Lin JH, Lou N, Kang N, Takano T, Hu F, Han X, Xu Q, Lovatt D, Torres A, Willecke K, Yang J, Kang J, and Nedergaard M. A central role of connexin 43 in hypoxic preconditioning. J Neurosci 28: 681-695, 2008. 116. Lin Y, Devin A, Rodriguez Y, and Liu ZG. Cleavage of the death domain kinase RIP by caspase-8 prompts TNF-induced apoptosis. Genes Dev 13: 2514-2526, 1999. 117. Litvin O, Tiunova A, Connell-Alberts Y, Panchin Y, and Baranova A. What is hidden in the pannexin treasure trove: the sneak peek and the guesswork. J Cell Mol Med 10: 613-634, 2006. 118. Locher C, Conforti R, Aymeric L, Ma Y, Yamazaki T, Rusakiewicz S, Tesniere A, Ghiringhelli F, Apetoh L, Morel Y, Girard JP, Kroemer G, and Zitvogel L. Desirable cell death during anticancer chemotherapy. Ann N Y Acad Sci 1209: 99-108. 119. Lu C, Wang A, Dorsch M, Tian J, Nagashima K, Coyle AJ, Jaffee B, Ocain TD, and Xu Y. Participation of Rip2 in lipopolysaccharide signaling is independent of its kinase activity. J Biol Chem 280: 16278-16283, 2005. 120. Mandrup-Poulsen T. Apoptotic signal transduction pathways in diabetes. Biochem Pharmacol 66: 1433-1440, 2003. 121. Manon F, Favier A, Nunez G, Simorre JP, and Cusack S. Solution structure of NOD1 CARD and mutational analysis of its interaction with the CARD of downstream kinase RICK. J Mol Biol 365: 160-174, 2007. 122. Mariathasan S, Newton K, Monack DM, Vucic D, French DM, Lee WP, Roose- Girma M, Erickson S, and Dixit VM. Differential activation of the inflammasome by caspase-1 adaptors ASC and Ipaf. Nature 430: 213-218, 2004. 123. Mariathasan S, Weiss DS, Newton K, McBride J, O'Rourke K, Roose-Girma M, Lee WP, Weinrauch Y, Monack DM, and Dixit VM. Cryopyrin activates the inflammasome in response to toxins and ATP. Nature 440: 228-232, 2006.

193 124. Marques-da-Silva C, Chaves MM, Castro NG, Coutinho-Silva R, and Guimaraes MZ. Colchicine inhibits cationic dye uptake induced by ATP in P2X2 and P2X7 receptor- expressing cells: implications for its therapeutic action. Br J Pharmacol 163: 912-926. 125. Martin PM, Roon P, Van Ells TK, Ganapathy V, and Smith SB. Death of retinal neurons in streptozotocin-induced diabetic mice. Invest Ophthalmol Vis Sci 45: 3330-3336, 2004. 126. Martinon F, Burns K, and Tschopp J. The inflammasome: a molecular platform triggering activation of inflammatory caspases and processing of proIL-beta. Mol Cell 10: 417-426, 2002. 127. Masters SL, Dunne A, Subramanian SL, Hull RL, Tannahill GM, Sharp FA, Becker C, Franchi L, Yoshihara E, Chen Z, Mullooly N, Mielke LA, Harris J, Coll RC, Mills KH, Mok KH, Newsholme P, Nunez G, Yodoi J, Kahn SE, Lavelle EC, and O'Neill LA. Activation of the NLRP3 inflammasome by islet amyloid polypeptide provides a mechanism for enhanced IL-1beta in type 2 diabetes. Nat Immunol 11: 897-904. 128. Mayer-Barber KD, Barber DL, Shenderov K, White SD, Wilson MS, Cheever A, Kugler D, Hieny S, Caspar P, Nunez G, Schlueter D, Flavell RA, Sutterwala FS, and Sher A. Caspase-1 independent IL-1beta production is critical for host resistance to mycobacterium tuberculosis and does not require TLR signaling in vivo. J Immunol 184: 3326-3330. 129. McCarthy JV, Ni J, and Dixit VM. RIP2 is a novel NF-kappaB-activating and cell death-inducing kinase. J Biol Chem 273: 16968-16975, 1998. 130. McDermott MF, and Tschopp J. From inflammasomes to fevers, crystals and hypertension: how basic research explains inflammatory diseases. Trends Mol Med 13: 381- 388, 2007. 131. McLarnon JG, Ryu JK, Walker DG, and Choi HB. Upregulated expression of purinergic P2X(7) receptor in Alzheimer disease and amyloid-beta peptide-treated microglia and in peptide-injected rat hippocampus. J Neuropathol Exp Neurol 65: 1090- 1097, 2006. 132. Meleth AD, Agron E, Chan CC, Reed GF, Arora K, Byrnes G, Csaky KG, Ferris FL, 3rd, and Chew EY. Serum inflammatory markers in diabetic retinopathy. Invest Ophthalmol Vis Sci 46: 4295-4301, 2005. 133. Metea MR, and Newman EA. Calcium signaling in specialized glial cells. Glia 54: 650-655, 2006. 134. Mitchell CH, and Reigada D. Purinergic signalling in the subretinal space: a role in the communication between the retina and the RPE. Purinergic Signal 4: 101-107, 2008. 135. Mizutani M, Kern TS, and Lorenzi M. Accelerated death of retinal microvascular cells in human and experimental diabetic retinopathy. J Clin Invest 97: 2883-2890, 1996. 136. Mohr S. Potential new strategies to prevent the development of diabetic retinopathy. Expert Opin Investig Drugs 13: 189-198, 2004. 137. Mohr S, Trueblood, K., Vincent, J.A. High Glucose-Induced Caspase-1 Activation in Retinal Müller Cells is Bi-Phasic Involving Receptor Interacting Protein 2 (RIP 2) Upregulation. In: American Diabetes Association 67th Scientific Sessions. Chicago, IL: American Diabetes Association, 2007. 138. Mohr S, Xi X, Tang J, and Kern TS. Caspase activation in retinas of diabetic and galactosemic mice and diabetic patients. Diabetes 51: 1172-1179, 2002. 139. Nakahira K, Haspel JA, Rathinam VA, Lee SJ, Dolinay T, Lam HC, Englert JA, Rabinovitch M, Cernadas M, Kim HP, Fitzgerald KA, Ryter SW, and Choi AM. Autophagy proteins regulate innate immune responses by inhibiting the release of mitochondrial DNA mediated by the NALP3 inflammasome. Nat Immunol.

194 140. Nathan DM, Cleary PA, Backlund JY, Genuth SM, Lachin JM, Orchard TJ, Raskin P, and Zinman B. Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes. N Engl J Med 353: 2643-2653, 2005. 141. Neven B, Prieur AM, and Quartier dit Maire P. Cryopyrinopathies: update on pathogenesis and treatment. Nat Clin Pract Rheumatol 4: 481-489, 2008. 142. Newman EA. Calcium increases in retinal glial cells evoked by light-induced neuronal activity. J Neurosci 25: 5502-5510, 2005. 143. Newman EA. Glial cell inhibition of neurons by release of ATP. J Neurosci 23: 1659- 1666, 2003. 144. Newman EA. Glial modulation of synaptic transmission in the retina. Glia 47: 268- 274, 2004. 145. Nicke A, Kuan YH, Masin M, Rettinger J, Marquez-Klaka B, Bender O, Gorecki DC, Murrell-Lagnado RD, and Soto F. A functional P2X7 splice variant with an alternative transmembrane domain 1 escapes gene inactivation in P2X7 knock-out mice. J Biol Chem 284: 25813-25822, 2009. 146. North RA, and Surprenant A. Pharmacology of cloned P2X receptors. Annu Rev Pharmacol Toxicol 40: 563-580, 2000. 147. Nuttle LC, and Dubyak GR. Differential activation of cation channels and non- selective pores by macrophage P2z purinergic receptors expressed in Xenopus oocytes. J Biol Chem 269: 13988-13996, 1994. 148. O'Neill LA. The interleukin-1 receptor/Toll-like receptor superfamily: 10 years of progress. Immunol Rev 226: 10-18, 2008. 149. O'Neill LA. The interleukin-1 receptor/Toll-like receptor superfamily: signal transduction during inflammation and host defense. Sci STKE 2000: re1, 2000. 150. Okamoto M, Liu W, Luo Y, Tanaka A, Cai X, Norris DA, Dinarello CA, and Fujita M. Constitutively active inflammasome in human melanoma cells mediating autoinflammation via caspase-1 processing and secretion of interleukin-1beta. J Biol Chem 285: 6477-6488. 151. Oshitari T, Yamamoto S, Hata N, and Roy S. Mitochondria- and caspase-dependent cell death pathway involved in neuronal degeneration in diabetic retinopathy. Br J Ophthalmol 92: 552-556, 2008. 152. Pachydaki SI, Tari SR, Lee SE, Ma W, Tseng JJ, Sosunov AA, Cataldergirmen G, Scarmeas N, Caspersen C, Chang S, Schiff WM, Schmidt AM, and Barile GR. Upregulation of RAGE and its ligands in proliferative retinal disease. Exp Eye Res 82: 807-815, 2006. 153. Pannicke T, Fischer W, Biedermann B, Schadlich H, Grosche J, Faude F, Wiedemann P, Allgaier C, Illes P, Burnstock G, and Reichenbach A. P2X7 receptors in Müller glial cells from the human retina. J Neurosci 20: 5965-5972, 2000. 154. Park SH, Park JW, Park SJ, Kim KY, Chung JW, Chun MH, and Oh SJ. Apoptotic death of photoreceptors in the streptozotocin-induced diabetic rat retina. Diabetologia 46: 1260-1268, 2003. 155. Pereira Tde O, da Costa GN, Santiago AR, Ambrosio AF, and dos Santos PF. High glucose enhances intracellular Ca2+ responses triggered by purinergic stimulation in retinal neurons and microglia. Brain Res 1316: 129-138. 156. Perrone L, Devi TS, Hosoya K, Terasaki T, and Singh LP. Thioredoxin interacting protein (TXNIP) induces inflammation through chromatin modification in retinal capillary endothelial cells under diabetic conditions. J Cell Physiol 221: 262-272, 2009. 157. Perrone L, Devi TS, Hosoya KI, Terasaki T, and Singh LP. Inhibition of TXNIP expression in vivo blocks early pathologies of diabetic retinopathy. Cell Death Dis 1: e65.

195 158. Petrovski G, Ayna G, Majai G, Hodrea J, Benko S, Madi A, and Fesus L. Phagocytosis of cells dying through autophagy induces inflammasome activation and IL- 1beta release in human macrophages. Autophagy 7. 159. Pickup JC. Inflammation and activated innate immunity in the pathogenesis of type 2 diabetes. Diabetes Care 27: 813-823, 2004. 160. Prevention CfDCa. National Diabetes Fact Sheet: National estimates and general information on diabetes and prediabetes in the United States, 2011. edited by Services USDoHaH. Atlanta, GA: 2011. 161. Price SA, Gardiner NJ, Duran-Jimenez B, Zeef LA, Obrosova IG, and Tomlinson DR. Thioredoxin interacting protein is increased in sensory neurons in experimental diabetes. Brain Res 1116: 206-214, 2006. 162. Proell M, Riedl SJ, Fritz JH, Rojas AM, and Schwarzenbacher R. The Nod-like receptor (NLR) family: a tale of similarities and differences. PLoS One 3: e2119, 2008. 163. Prosdocimo DA, Douglas DC, Romani AM, O'Neill WC, and Dubyak GR. Autocrine ATP release coupled to extracellular pyrophosphate accumulation in vascular smooth muscle cells. Am J Physiol Cell Physiol 296: C828-839, 2009. 164. Prosdocimo DA, Wyler SC, Romani AM, O'Neill WC, and Dubyak GR. Regulation of vascular smooth muscle cell calcification by extracellular pyrophosphate homeostasis: synergistic modulation by cyclic AMP and hyperphosphatemia. Am J Physiol Cell Physiol 298: C702-713. 165. Puro DG, and Mano T. Modulation of calcium channels in human retinal glial cells by basic fibroblast growth factor: a possible role in retinal pathobiology. J Neurosci 11: 1873-1880, 1991. 166. Qaum T, Xu Q, Joussen AM, Clemens MW, Qin W, Miyamoto K, Hassessian H, Wiegand SJ, Rudge J, Yancopoulos GD, and Adamis AP. VEGF-initiated blood-retinal barrier breakdown in early diabetes. Invest Ophthalmol Vis Sci 42: 2408-2413, 2001. 167. Qu Y, and Dubyak GR. P2X7 receptors regulate multiple types of membrane trafficking responses and non-classical secretion pathways. Purinergic Signal 5: 163-173, 2009. 168. Qu Y, Franchi L, Nunez G, and Dubyak GR. Nonclassical IL-1 beta secretion stimulated by P2X7 receptors is dependent on inflammasome activation and correlated with exosome release in murine macrophages. J Immunol 179: 1913-1925, 2007. 169. Qu Y, Misaghi S, Newton K, Gilmour LL, Louie S, Cupp JE, Dubyak GR, Hackos D, and Dixit VM. Pannexin-1 is required for ATP release during apoptosis but not for inflammasome activation. J Immunol 186: 6553-6561. 170. Qu Y, Ramachandra L, Mohr S, Franchi L, Harding CV, Nunez G, and Dubyak GR. P2X7 receptor-stimulated secretion of MHC class II-containing exosomes requires the ASC/NLRP3 inflammasome but is independent of caspase-1. J Immunol 182: 5052-5062, 2009. 171. Ramachandra L, Qu Y, Wang Y, Lewis CJ, Cobb BA, Takatsu K, Boom WH, Dubyak GR, and Harding CV. Mycobacterium tuberculosis synergizes with ATP to induce release of microvesicles and exosomes containing major histocompatibility complex class II molecules capable of antigen presentation. Infect Immun 78: 5116-5125. 172. Ricatti MJ, Alfie LD, Lavoie EG, Sevigny J, Schwarzbaum PJ, and Faillace MP. Immunocytochemical localization of NTPDases1 and 2 in the neural retina of mouse and zebrafish. Synapse 63: 291-307, 2009. 173. Rock KL, Latz E, Ontiveros F, and Kono H. The sterile inflammatory response. Annu Rev Immunol 28: 321-342.

196 174. Romanello M, and D'Andrea P. Dual mechanism of intercellular communication in HOBIT osteoblastic cells: a role for gap-junctional hemichannels. J Bone Miner Res 16: 1465- 1476, 2001. 175. Romeo G, Liu WH, Asnaghi V, Kern TS, and Lorenzi M. Activation of nuclear factor-kappaB induced by diabetes and high glucose regulates a proapoptotic program in retinal pericytes. Diabetes 51: 2241-2248, 2002. 176. Ruefli-Brasse AA, Lee WP, Hurst S, and Dixit VM. Rip2 participates in Bcl10 signaling and T-cell receptor-mediated NF-kappaB activation. J Biol Chem 279: 1570-1574, 2004. 177. Sabirov RZ, and Okada Y. ATP-conducting maxi-anion channel: a new player in stress-sensory transduction. Jpn J Physiol 54: 7-14, 2004. 178. Sabirov RZ, and Okada Y. Wide nanoscopic pore of maxi-anion channel suits its function as an ATP-conductive pathway. Biophys J 87: 1672-1685, 2004. 179. Sarkar A, Duncan M, Hart J, Hertlein E, Guttridge DC, and Wewers MD. ASC directs NF-kappaB activation by regulating receptor interacting protein-2 (RIP2) caspase-1 interactions. J Immunol 176: 4979-4986, 2006. 180. Sarthy VP, Brodjian SJ, Dutt K, Kennedy BN, French RP, and Crabb JW. Establishment and characterization of a retinal Müller cell line. Invest Ophthalmol Vis Sci 39: 212-216, 1998. 181. Sbai O, Devi TS, Melone MA, Feron F, Khrestchatisky M, Singh LP, and Perrone L. RAGE-TXNIP axis is required for S100B-promoted Schwann cell migration, fibronectin expression and cytokine secretion. J Cell Sci 123: 4332-4339. 182. Schilling WP, Wasylyna T, Dubyak GR, Humphreys BD, and Sinkins WG. Maitotoxin and P2Z/P2X(7) purinergic receptor stimulation activate a common cytolytic pore. Am J Physiol 277: C766-776, 1999. 183. Schroder K, and Tschopp J. The inflammasomes. Cell 140: 821-832. 184. Schroder K, Zhou R, and Tschopp J. The NLRP3 inflammasome: a sensor for metabolic danger? Science 327: 296-300. 185. Schwiebert EM. ABC transporter-facilitated ATP conductive transport. Am J Physiol 276: C1-8, 1999. 186. Seil M, El Ouaaliti M, Fontanils U, Etxebarria IG, Pochet S, Dal Moro G, Marino A, and Dehaye JP. Ivermectin-dependent release of IL-1beta in response to ATP by peritoneal macrophages from P2X(7)-KO mice. Purinergic Signal 6: 405-416. 187. Shao W, Yeretssian G, Doiron K, Hussain SN, and Saleh M. The caspase-1 digestome identifies the glycolysis pathway as a target during infection and septic shock. J Biol Chem 282: 36321-36329, 2007. 188. Shao W, Yu Z, Fantus IG, and Jin T. Cyclic AMP signaling stimulates proteasome degradation of thioredoxin interacting protein (TxNIP) in pancreatic beta-cells. Cell Signal 22: 1240-1246. 189. Shelton MD, Distler AM, Kern TS, and Mieyal JJ. Glutaredoxin regulates autocrine and paracrine proinflammatory responses in retinal glial (Müller ) cells. J Biol Chem 284: 4760-4766, 2009. 190. Shelton MD, Kern TS, and Mieyal JJ. Glutaredoxin regulates nuclear factor kappa-B and intercellular adhesion molecule in Müller cells: model of diabetic retinopathy. J Biol Chem 282: 12467-12474, 2007. 191. Shi Y, Ren Y, Zhao L, Du C, Wang Y, Zhang Y, Li Y, Zhao S, and Duan H. Knockdown of thioredoxin interacting protein attenuates high glucose-induced apoptosis and activation of ASK1 in mouse mesangial cells. FEBS Lett. 192. Simi A, Lerouet D, Pinteaux E, and Brough D. Mechanisms of regulation for interleukin-1beta in neurodegenerative disease. Neuropharmacology 52: 1563-1569, 2007.

197 193. Solini A, Chiozzi P, Morelli A, Adinolfi E, Rizzo R, Baricordi OR, and Di Virgilio F. Enhanced P2X7 activity in human fibroblasts from diabetic patients: a possible pathogenetic mechanism for vascular damage in diabetes. Arterioscler Thromb Vasc Biol 24: 1240-1245, 2004. 194. Sparvero LJ, Asafu-Adjei D, Kang R, Tang D, Amin N, Im J, Rutledge R, Lin B, Amoscato AA, Zeh HJ, and Lotze MT. RAGE (Receptor for Advanced Glycation Endproducts), RAGE ligands, and their role in cancer and inflammation. J Transl Med 7: 17, 2009. 195. Stienstra R, Joosten LA, Koenen T, van Tits B, van Diepen JA, van den Berg SA, Rensen PC, Voshol PJ, Fantuzzi G, Hijmans A, Kersten S, Müller M, van den Berg WB, van Rooijen N, Wabitsch M, Kullberg BJ, van der Meer JW, Kanneganti T, Tack CJ, and Netea MG. The inflammasome-mediated caspase-1 activation controls adipocyte differentiation and insulin sensitivity. Cell Metab 12: 593-605. 196. Sueishi K, Hata Y, Murata T, Nakagawa K, Ishibashi T, and Inomata H. Endothelial and glial cell interaction in diabetic retinopathy via the function of vascular endothelial growth factor (VEGF). Pol J Pharmacol 48: 307-316, 1996. 197. Surprenant A, and North RA. Signaling at purinergic P2X receptors. Annu Rev Physiol 71: 333-359, 2009. 198. Swanson MS. Autophagy: eating for good health. J Immunol 177: 4945-4951, 2006. 199. Tang J, Mohr S, Du YD, and Kern TS. Non-uniform distribution of lesions and biochemical abnormalities within the retina of diabetic humans. Curr Eye Res 27: 7-13, 2003. 200. Thornberry NA, and Molineaux SM. Interleukin-1 beta converting enzyme: a novel cysteine protease required for IL-1 beta production and implicated in programmed cell death. Protein Sci 4: 3-12, 1995. 201. Tiganis T. Reactive oxygen species and insulin resistance: the good, the bad and the ugly. Trends Pharmacol Sci 32: 82-89. 202. Tschopp J. Mitochondria: Sovereign of inflammation? Eur J Immunol 41: 1196-1202. 203. Tschopp J, and Schroder K. NLRP3 inflammasome activation: The convergence of multiple signalling pathways on ROS production? Nat Rev Immunol 10: 210-215. 204. Vaira S, Friday E, Scott K, Conrad S, and Turturro F. Wnt/beta-catenin signalling pathway and thioredoxin-interacting protein (TXNIP) mediate the "glucose sensor" mechanism in metastatic breast cancer derived cells MDA-MB-231. J Cell Physiol. 205. Vandanmagsar B, Youm YH, Ravussin A, Galgani JE, Stadler K, Mynatt RL, Ravussin E, Stephens JM, and Dixit VD. The NLRP3 inflammasome instigates obesity- induced inflammation and insulin resistance. Nat Med 17: 179-188. 206. Verhoef PA, Kertesy SB, Lundberg K, Kahlenberg JM, and Dubyak GR. Inhibitory effects of chloride on the activation of caspase-1, IL-1beta secretion, and cytolysis by the P2X7 receptor. J Immunol 175: 7623-7634, 2005. 207. Vincent JA, and Mohr S. Inhibition of caspase-1/interleukin-1beta signaling prevents degeneration of retinal capillaries in diabetes and galactosemia. Diabetes 56: 224- 230, 2007. 208. Viviani B, Bartesaghi S, Corsini E, Galli CL, and Marinovich M. Cytokines role in neurodegenerative events. Toxicol Lett 149: 85-89, 2004. 209. von Kugelgen I. Pharmacological profiles of cloned mammalian P2Y-receptor subtypes. Pharmacol Ther 110: 415-432, 2006. 210. Wang S, Miura M, Jung YK, Zhu H, Li E, and Yuan J. Murine caspase-11, an ICE- interacting protease, is essential for the activation of ICE. Cell 92: 501-509, 1998. 211. Wang X, Wang H, Figueroa BE, Zhang WH, Huo C, Guan Y, Zhang Y, Bruey JM, Reed JC, and Friedlander RM. Dysregulation of receptor interacting protein-2 and caspase

198 recruitment domain only protein mediates aberrant caspase-1 activation in Huntington's disease. J Neurosci 25: 11645-11654, 2005. 212. Ward MM, Jobling AI, Kalloniatis M, and Fletcher EL. Glutamate uptake in retinal glial cells during diabetes. Diabetologia 48: 351-360, 2005. 213. Wendt T, Tanji N, Guo J, Hudson BI, Bierhaus A, Ramasamy R, Arnold B, Nawroth PP, Yan SF, D'Agati V, and Schmidt AM. Glucose, glycation, and RAGE: implications for amplification of cellular dysfunction in diabetic nephropathy. J Am Soc Nephrol 14: 1383-1395, 2003. 214. West AP, Shadel GS, and Ghosh S. Mitochondria in innate immune responses. Nat Rev Immunol 11: 389-402. 215. Westermann D, Van Linthout S, Dhayat S, Dhayat N, Escher F, Bucker-Gartner C, Spillmann F, Noutsias M, Riad A, Schultheiss HP, and Tschope C. Cardioprotective and anti-inflammatory effects of interleukin converting enzyme inhibition in experimental diabetic cardiomyopathy. Diabetes 56: 1834-1841, 2007. 216. Whitmire W, Al-Gayyar MM, Abdelsaid M, Yousufzai BK, and El-Remessy AB. Alteration of growth factors and neuronal death in diabetic retinopathy: what we have learned so far. Mol Vis 17: 300-308. 217. Wurm A, Erdmann I, Bringmann A, Reichenbach A, and Pannicke T. Expression and function of P2Y receptors on Müller cells of the postnatal rat retina. Glia 57: 1680- 1690, 2009. 218. Wurm A, Lipp S, Pannicke T, Linnertz R, Farber K, Wiedemann P, Reichenbach A, and Bringmann A. Involvement of A(1) adenosine receptors in osmotic volume regulation of retinal glial cells in mice. Mol Vis 15: 1858-1867, 2009. 219. Wurm A, Lipp S, Pannicke T, Linnertz R, Krugel U, Schulz A, Farber K, Zahn D, Grosse J, Wiedemann P, Chen J, Schoneberg T, Illes P, Reichenbach A, and Bringmann A. Endogenous purinergic signaling is required for osmotic volume regulation of retinal glial cells. J Neurochem 112: 1261-1272. 220. Wurm A, Pannicke T, Iandiev I, Francke M, Hollborn M, Wiedemann P, Reichenbach A, Osborne NN, and Bringmann A. Purinergic signaling involved in Müller cell function in the mammalian retina. Prog Retin Eye Res. 221. Xi X, Gao L, Hatala DA, Smith DG, Codispoti MC, Gong B, Kern TS, and Zhang JZ. Chronically elevated glucose-induced apoptosis is mediated by inactivation of Akt in cultured Müller cells. Biochem Biophys Res Commun 326: 548-553, 2005. 222. Yaar R, Jones MR, Chen JF, and Ravid K. Animal models for the study of adenosine receptor function. J Cell Physiol 202: 9-20, 2005. 223. Yego EC, and Mohr S. siah-1 Protein is necessary for high glucose-induced glyceraldehyde-3-phosphate dehydrogenase nuclear accumulation and cell death in Müller cells. J Biol Chem 285: 3181-3190. 224. Yego EC, Vincent JA, Sarthy V, Busik JV, and Mohr S. Differential regulation of high glucose-induced glyceraldehyde-3-phosphate dehydrogenase nuclear accumulation in Müller cells by IL-1beta and IL-6. Invest Ophthalmol Vis Sci 50: 1920-1928, 2009. 225. Yiangou Y, Facer P, Durrenberger P, Chessell IP, Naylor A, Bountra C, Banati RR, and Anand P. COX-2, CB2 and P2X7-immunoreactivities are increased in activated microglial cells/macrophages of multiple sclerosis and amyotrophic lateral sclerosis spinal cord. BMC Neurol 6: 12, 2006. 226. Yu D, Buibas M, Chow SK, Lee IY, Singer Z, and Silva GA. Characterization of Calcium-Mediated Intracellular and Intercellular Signaling in the rMC-1 Glial Cell Line. Cell Mol Bioeng 2: 144-155, 2009.

199 227. Yu FX, Chai TF, He H, Hagen T, and Luo Y. Thioredoxin-interacting protein (Txnip) gene expression: sensing oxidative phosphorylation status and glycolytic rate. J Biol Chem 285: 25822-25830. 228. Yu FX, Goh SR, Dai RP, and Luo Y. Adenosine-containing molecules amplify glucose signaling and enhance txnip expression. Mol Endocrinol 23: 932-942, 2009. 229. Yuan J, Lipinski M, and Degterev A. Diversity in the mechanisms of neuronal cell death. Neuron 40: 401-413, 2003. 230. Zepter K, Haffner A, Soohoo LF, De Luca D, Tang HP, Fisher P, Chavinson J, and Elmets CA. Induction of biologically active IL-1 beta-converting enzyme and mature IL-1 beta in human keratinocytes by inflammatory and immunologic stimuli. J Immunol 159: 6203-6208, 1997. 231. Zheng L, Szabo C, and Kern TS. Poly(ADP-ribose) polymerase is involved in the development of diabetic retinopathy via regulation of nuclear factor-kappaB. Diabetes 53: 2960-2967, 2004. 232. Zhou R, Tardivel A, Thorens B, Choi I, and Tschopp J. Thioredoxin-interacting protein links oxidative stress to inflammasome activation. Nat Immunol 11: 136-140. 233. Zimmermann H. Extracellular metabolism of ATP and other nucleotides. Naunyn Schmiedebergs Arch Pharmacol 362: 299-309, 2000. 234. Zipfel GJ, Babcock DJ, Lee JM, and Choi DW. Neuronal apoptosis after CNS injury: the roles of glutamate and calcium. J Neurotrauma 17: 857-869, 2000. 235. Zitvogel L, Kepp O, and Kroemer G. Decoding cell death signals in inflammation and immunity. Cell 140: 798-804.

200