THE ROLE OF INTERLEUKIN-6 SIGNALLING MOLECULES IN … · ii The role of Interleukin-6 signalling molecules in murine models of inherited photoreceptor degeneration Doctor of Philosophy,

THE ROLE OF INTERLEUKIN-6 SIGNALLING

MOLECULES IN MURINE MODELS OF INHERITED

PHOTORECEPTOR DEGENERATION

by

Michael Joseph Szego

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Molecular Genetics University of Toronto

© Copyright by Michael Joseph Szego 2009

ii

The role of Interleukin-6 signalling molecules in murine models of inherited photoreceptor degeneration

Doctor of Philosophy, 2009 Michael Joseph Szego

Department of Molecular Genetics University of Toronto

Abstract

We previously reported that in inherited photoreceptor degenerations (IPDs), the mutant photoreceptors

(PRs) are at a constant risk of death (Pacione, Szego et al, 2003). Using microarrays to identify genes that

may mediate the constant risk, I identified 145 differentially expressed transcripts in the Rds+/- mouse model

of IPD at a time when 90% of the PRs were alive. A major finding was the up-regulation, quantified by

qPCR, of four components of a putative IL-6 cytokine signaling pathway: Oncostatin M (Osm) (2-fold

increased) → Oncostatin M receptor (Osmr)(2.6-fold increased) → Stat-3 (2.3-fold increased) → C/EBPδ (3.2-

fold increased). Similarly, I found increases in the cognate proteins Osmr (3-fold), Stat-3 (2.6-fold), and the

phosphorylated, transcriptionally active form of Stat-3, pStat-3 (5.8-fold)(all p<0.01). Other Il-6 cytokine

signaling molecules were largely unchanged, but the mRNA of leukemia inhibitory factor (Lif), was

increased (3.0-fold). Comparable increases of most transcripts were also present in the Rd1-/- and mutant

rhodopsin P347S transgenic (P347S) IPD models. The increases in cytokine signaling molecules occurred

predominantly in Müller glia, although C/EBPδ transcript was increased in PRs. Because exogenous IL-6

cytokine treatment slows PR death in IPDs, I asked whether the endogenous increases in IL-6 pathway

proteins in IPD retinas were a survival response, and generated IPD models with Osmr, Lif or C/EBPδ loss-

of-function (LOF) mutations. Osmr LOF decreased PR survival in the retinas of Rds+/-;Osmr-/- mice, which

had 12.5% fewer PRs than those of Rds+/-;Osmr+/+ mice (n=9, p<0.05) at 4 month of age, and Tg-

RHO(P347S);Osmr-/- mice had 13.5% fewer PRs (n=6, p<0.01) at 31 days of age. Unexpectedly, Osmr LOF

had no effect on pStat3 levels in Rds+/-;Osmr-/- retinas, indicating that retinal Stat3 activation may be

predominantly regulated by other molecules. In contrast to the Osmr LOF, Lif or C/EBPδ LOF

unexpectedly increased mutant PR survival. Rd1-/-;Lif -/- mice at 13 days had 14% more PRs than Rd1-/-

;Lif+/+ mice (n=6, p<0.003) and a 1.7 fold decrease in pStat-3 (n=4, p<.05). Similarly, 8 month-old Rds+/-;

C/EBPδ-/- mice had 18% more PRs than Rds+/-; C/EBPδ+/+ mice (n=5, p<0.005). These findings suggest

that in mutant PRs: 1) up-regulation of the Osmr receptor is protective; 2) the presence of Lif or C/EBPδ

is pathogenic, and therefore 3) Osmr, Lif and C/EBPδ act either in different pathways or different cells, to

account for the differing effects of their LOF on PR cell death; and 4) the partial effects of Osmr, Lif and

C/EBPδ LOF indicate that other genes also mediate the constant risk of death of mutant PRs in IPDs.

iii

Acknowledgments

I have met some incredible people while performing the work summarized in these pages-

many of whom will remain life long friends. I would like to thank Rod for his enthusiasm, ability to

ask key questions and his support of my career pursuit. I would also like to thank my committee

members, Michael Salter and Brenda Andrews for their guidance over the years. Jayne Danska and

Freda Miller were also very generous with their time.

A former student once described the lab culture as being temporally organized into

dynasties. I have been lucky enough to be apart of two such dynasties. In the first dynasty I am

indebted to Jonathan, Rachel, Lynda and Laura, otherwise known as members of the Commie Bay.

Jonathan, your strong self-identity and ability to say exactly what was on your mind certainly inspired

me. You have my utmost respect and love. Rachel, you were always there to support me and spent

many a late night helping me edit papers for my various graduate courses. Finally, to Laura, who

kept plugging away at our project and kept things moving even when I was feeling discouraged. I

still laugh at the thought of the late night we spend working on the review together. Lynda, you have

certainly made the lab a fun place to work. I thank for all of your help particularly over this last year.

To the second, most recent, dynasty I am especially thankful to Alexa, who spent many

hours helping me complete experiments while I wrote this thesis and entered into fatherhood. It has

been fun watching you grow as a scientist and your comments over the years were always insightful

and certainly helped shape the direction of this project. However, your penchant for motorcycle

racing has always concerned me; please drive safely!

My Dad has always described my educational trajectory as being an “S” shaped curve. Thus,

he was not surprised when I told him I would be doing a Master’s degree after my PhD. Mom and

Dad you have always supported me through all of my endeavors. You are model parents that I

attempt to emulate.

After living for several months in the frigid, dark, and condemned old Graduate House, I

moved into a cozy apartment with Arvin and Pleasie and we quickly became like family. Our kitchen

was a communal one and I have fond memories of the time we spent together.

The best thing that ever happened to me while in graduate school was meeting my wife,

Joby. Her intelligence, confidence and infectious laugh drew me right in and I have never looked

back. I am so very lucky we met. Finally, I would like to thank Quincie, who was born April 4, 2008,

for enabling me to finally stop doing experiments. She has brought learning back to the basics and

has renewed my love of developmental biology.

iv

Table of Contents

Abstract ...............................................................................................................ii

Acknowledgments..............................................................................................iii

Table of Contents ............................................................................................... iv

List of Tables....................................................................................................viii

List of Figures .................................................................................................... ix

List of Abbreviations ...........................................................................................x

1. Introduction ..................................................................................................... 1

1.1 Structure of the mammalian retina .......................................................................................................1

1.1.1 Neuronal cell types in the retina ............................................................................................................1

1.1.2 Retinal pigment epithelium....................................................................................................................5

1.1.3 Müller Glia .........................................................................................................................................5

1.2 Types of Photoreceptor Degeneration ................................................................................................7

1.3 Genes Implicated in Retinal Degeneration .........................................................................................8

1.3.1 Structural proteins- Rds and Rom1 ......................................................................................................8

1.3.2 Phagocytosis- Mertk .......................................................................................................................... 10

1.3.3 Cilia maintenance/Trafficking of intracellular proteins- BBS genes .................................................... 10

1.3.4 Phototransduction- Rhodopsin and Rd1............................................................................................. 11

1.3.5 Signalling, cell-cell interaction, or synaptic interaction- Sema4A.......................................................... 12

1.3.6 Vitamin A metabolism- Rpe65 ........................................................................................................ 12

1.3.7 Transporters/channels- ABCR (rod photoreceptor ABC transporter) or ABCA4............................. 13

1.3.8 Transcription factors- Beta2/NeuroD1 ............................................................................................. 14

1.3.9 RNA intron-splicing factors .............................................................................................................. 14

1.3.10 Enzymes- IMPDH1...................................................................................................................... 15

v

1.4 Common Features of IPD.................................................................................................................. 16

1.4.1 Altered calcium homeostasis............................................................................................................... 16

1.4.2 Cell death by apoptosis ...................................................................................................................... 17

1.4.3 Common kinetics of cell death ............................................................................................................ 18

1.4.4 Implications of these common features occurring in IPD ...................................................................... 18

1.5 Microarrays and IPD ........................................................................................................................... 19

1.5.1 Identification of photoreceptor-enriched genes ....................................................................................... 19

1.5.2 Identification of regulatory networks in the retina................................................................................ 21

1.5.3 Gene expression changes in response to IPD....................................................................................... 22

1.6 Neurotrophic factors protect photoreceptors ................................................................................. 26

1.6.1 Fibroblast growth factor ..................................................................................................................... 26 1.6.1.1 Fibroblast growth factor signalling in photoreceptors.................................................................................... 27

1.6.2 Brain-derived neurotrophic factor and nerve growth factor .................................................................... 27

1.6.3 The Interleukin-6 family of cytokines ................................................................................................. 28 1.6.3.1 Pathways activated by IL-6 cytokines................................................................................................................ 29

1.7 Thesis objectives .................................................................................................................................. 33

2. Characterization of the retinal transcriptomes of the Rom1-/- and Rds+/-

murine models of inherited photoreceptor degeneration ................................. 34

2.1 Abstract ................................................................................................................................................. 35

2.2 Introduction.......................................................................................................................................... 35

2.3 Materials and Methods ........................................................................................................................ 38

2.3.1 Animals used in this study ................................................................................................................ 38

2.3.2 Preparation of retinal RNA for the microarray experiment ................................................................ 38

2.3.3 Preparation of retinal RNA for cDNA synthesis .............................................................................. 40

2.3.4 Microarray hybridization experiments ................................................................................................ 40

2.3.5 Microarray data analysis ................................................................................................................... 40

vi

2.3.6 cDNA synthesis ............................................................................................................................... 41

2.3.7 Quantitative real-time PCR .............................................................................................................. 41 2.3.7.1 Design and evaluation of primers ...................................................................................................................... 41 2.7.2 Analysis of gene expression ................................................................................................................................... 45 2.7.3 Statistical analysis..................................................................................................................................................... 45

2.4 Results.................................................................................................................................................... 46

2.4.1 Characterization of the Rom1-/- retinal transcriptome ......................................................................... 46

2.4.2 Characterization of the Rds+/- retinal transcriptome ........................................................................... 52

2.5 Discussion ............................................................................................................................................. 57

Appendix 2.1: Up-regulated transcripts in the retinas of Rom1-/- mice............................................... 63

Appendix 2.2: Down-regulated transcripts in the retinas of Rom1 -/- mice ........................................ 64

Appendix 2.3: Up-regulated transcripts in the retinas of Rds+/- mice ................................................. 65

Appendix 2.4: Down-regulated transcripts in the retinas of Rds+/- mice............................................ 71

Appendix 2.5: Up-regulated network involved in cell and immune signalling.................................. 72

Appendix 2.6: Up-regulated network involved in molecular transport, cancer and cell death....... 73

Appendix 2.7: Up-regulated network involved in viral function, immune response and injury..... 74

Appendix 2.8: Up-regulated network involved in cell-cell signalling.................................................. 75

Appendix 2.9: Up-regulated network involved immunological and inflammatory disease ............. 76

Appendix 2.10: Up-regulated network involved in cellular growth and proliferation ..................... 77

3 A role for IL-6 signalling molecules in inherited photoreceptor degenerations

........................................................................................................................... 78

3.1 Abstract ................................................................................................................................................. 79

3.2 Introduction.......................................................................................................................................... 80

vii

3.3 Material and Methods.......................................................................................................................... 84

3.3.1 Quantitative real-time PCR .............................................................................................................. 84

3.3.2 Histology and outer nuclear layer measurements.................................................................................. 84

3.3.3 Retinal protein isolation..................................................................................................................... 84

3.3.4 Immunoblot analysis.......................................................................................................................... 86

3.3.5 Immunofluorescence staining............................................................................................................... 87

3.3.6 In situ hybridization.......................................................................................................................... 88

3.3.7 Laser capture microdissection ............................................................................................................. 89

3.4 Results.................................................................................................................................................... 89

4 General discussion and future directions ....................................................108

4.1 Significance of findings ..................................................................................................................... 108

4.2 Identifying the mechanism(s) of IL-6 mediated photoreceptor protection .............................. 110

4.3 The mode of IL-6 signalling: direct or indirect action on photoreceptors................................ 112

4.4 Is Stat-3 signalling a protective response? ...................................................................................... 114

4.5 Determining whether C/EBPδ expression in photoreceptors is pathogenic ........................... 115

4.6 Reconciling the pathogenic and protective affects of IL-6 signalling ........................................ 117

4.7 Stat-3 may be involved in retinal regeneration in lower vertebrates........................................... 117

References ........................................................................................................ 119

viii

List of Tables

Table 1.1: Retinal responses to IL-6 cytokine activation .......................................................................... 31

Table 2.1: List of qPCR primers ................................................................................................................... 43

Table 2.2: Comparison of transcript fold-differences as determined by microarray and qPCR

analyses in Rom1-/- and wild-type retinas............................................................................................. 51

Table 2.3: Comparison of transcript fold-differences as determined by microarray and qPCR

analyses in Rds+/- and wild-type retinas ............................................................................................... 55

Table 3.1: List of primers used for qPCR.................................................................................................... 85

Table 3.2: Quantification of selected differentially expressed transcripts in three different mouse

models of IPD........................................................................................................................................ 90

ix

List of Figures

Figure 1.1: Structure of the vertebrate retina. ................................................................................................3

Figure 1.2: Structure of a rod photoreceptor. ................................................................................................4

Figure 2.1: Summary of the Rom1-/- normalized microarray data............................................................. 47

Figure 2.2: Poor concordance between qPCR and microarray analyses in the Rom1-/- model of IPD50

Figure 2.3: Summary of the Rds-+/- normalized microarray data .............................................................. 53

Figure 2.4: Moderate concordance between qPCR and microarray analyses in the Rds+/- model of

IPD........................................................................................................................................................... 56

Figure 2.5: Up-regulation of a putative IL-6 cytokine pathway in Rds+/- retinas.................................... 62

Figure 3.1: Several IL-6 cytokines and receptors are up-regulated in the retinas of Rds+/- mice......... 92

Figure 3.2: Rds+/- retinas exhibit increased IL-6/Jak-Stat signaling ......................................................... 93

Figure 3.3: Increased Osmr, Stat3 and pStat3 staining in Müller glia of Rds+/- retinas ......................... 95

Figure 3.4: Lif and C/EBPδ are up-regulated in several cell types in the retinas of 7-week old Rds+/-

mice .......................................................................................................................................................... 97

Figure 3.5: Lif up-regulation occurs predominantly in the inner nuclear layer (INL), while C/EBPδ

up-regulation is observed mainly in the outer nuclear layer (ONL)............................................... 98

Figure 3.6: In two models of PR degeneration, Osmr is protective in a pStat-3 independent

mechanism ............................................................................................................................................ 100

Figure 3.7: Lif is pathogenic in the Rd1 -/- model of PR degeneration.................................................. 101

Figure 3.8: C/EBPδ is pathogenic in the Rds+/- model of PR degeneration ........................................ 103

x

List of Abbreviations

α2M Alpha-2-macroglobulin bFGF Basic fibroblast growth factor C/EBPδ CCAAT Enhancer binding protein delta Clc Cardiotrophin-like cytokine Cntf Ciliary neurotrophic factor Cntfr Ciliary neurotrophic factor receptor Egf Epidermal growth factor ERG Electroretinogram Erk Extracellular signal-regulated kinase gp130 Glycoprotein 130 IL-11 Interleukin-11 IL-11r Interleukin-11 receptor IL-6 Interleukin-6 IL-6r Interleukin-6 receptor INL Inner nuclear layer IPD Inherited photoreceptor degeneration Jak Janus kinase Lif Leukemia inhibitory factor Lifr Leukemia inhibitory factor receptor Ngf Nerve growth factor Np Neuropoetin Nt-3 Neurotrophin-3 ONL Outer nuclear layer OS Outer segment Osm Oncostatin Osmr Oncostatin receptor P347S Transgenic mouse line expressing a human mutant rhodopsin

transgene (P347S) pERK phosphorylated extracellular signal-regulated kinase PR Photoreceptor pStat-1 phosphorylated signal transducer and activator of transcription-1 pStat-3 phosphorylated signal transducer and activator of transcription-3 qPCR Quantitative polymerase chain reaction rAAV Recombinant adeno-associated virus Rd1 Retinal degeneration-1 Rds Retinal degeneration slow Rom1 Rod outer segment membrane protein-1 RP Retinitis pigmentosa RPE Retinal pigment epithelium Socs-3 Suppressor of cytokine signalling-3 Stat-1 signal transducer and activator of transcription-1 Stat-3 signal transducer and activator of transcription-3 Tyk Tyrosine kinase

1

1. Introduction

1.1 Structure of the mammalian retina

The mammalian retina is a highly ordered tissue, composed of four cellular layers

that line the back of the eye (Figure 1.1) (Stryer, 1996). The primary function of the retina is to

convert visual cues into a neural signal, which then travels to the visual cortex for further processing

resulting in sight (Stryer, 1996). As the most accessible organ of the central nervous system, the

retina has been extensively characterized (Hubel, 1995; Masland, 2001). Santiago Ramon y Cajal

published a series of very accurate anatomical drawings of the vertebrate retina in 1900 (Ramón y

Cajal and Bresler, 1900). Since then, the neuroanatomy of the retina has been further refined and the

functional relationship between many of the cell types in the retina have been characterized (Hubel,

1995). In addition, the biochemical basis of vision has been well characterized and many genes

required for normal retinal development to occur have been identified (Molday, 1998; Livesey and

Cepko, 2001). The retina is also a non-essential organ. Mutations that affect retinal function or

viability are compatible with life. Consequently, mutations in virtually any gene necessary for eye

development alone, or for ocular function alone, can occur and be studied. In the context of retinal

degeneration, for example, 193 genes or loci associated with inherited retinal degenerations have

been identified (Daiger et al., 2008), making the retina an ideal system for studying neuropathological

mechanisms.

1.1.1 Neuronal cell types in the retina

Photoreceptors (PRs), which constitute 70% of the cells in the mammalian retina (Young,

1985), absorb light and transduce it into a neural signal. Rod PRs are most sensitive to low intensity

monochromatic light, while cone PRs function predominantly at higher light intensity where they are

2

involved in detecting colour (Hubel, 1995). In humans, rods and cones exhibit a non-uniform

distribution in the retina (Curcio et al., 1990). The fovea is a region near the center of the retina that

has a high concentration of cones and is most sensitive to detail. The concentration of cones rapidly

decreases with increasing distance from the fovea, while the concentration of rods increases (Curcio

et al., 1990). Retinas of mice, in contrast, exhibit an almost uniform distribution of rods and cones

(Carter-Dawson and LaVail, 1979). Thus, mice do not have an analogous structure similar to the

fovea.

Rod PRs are polarized neurons that consist of three major compartments: the outer segment

(OS), the inner segment (IS), and the synaptic region (Figure 1.2). The synaptic region is the most

proximal region of the PR and transmits signals to the interneurons of the inner nuclear layer. The

inner segment contains the PR nucleus and most of the cell’s biosynthetic machinery (Kandel, 2008).

Many of the proteins made in the inner segment, are actively transported through a structure called

the connecting cilium to the OS (Kandel, 2008). The OS is a modified cilium, containing about 1000

membranous disks in primates (Young, 1971), stacked in an array oriented perpendicular to the long

axis of the OS. Disk morphogenesis occurs as evaginations of the connecting cilium (Steinberg et al.,

1980). OS disk renewal occurs at a rate of 10% per day in primates (Young, 1971), with the older

disks being phagocytosed by the retinal pigment epithelium at the distal tip of the OS in a process

that is calibrated such that the OSs maintain their overall length and structure (Kandel, 2008). The

high renewal rate may be critical to prevent photodamage from occurring on PR molecules (Young,

1976).

Mature OS disks are covered with rhodopsin and are the site of light absorption and

phototransduction. Rhodopsin, which accounts for greater than 70% of all protein in PRs (Molday,

1998), consists of the protein moiety opsin, bound to the light sensitive chromophore, retinal

(Stryer, 1996). Upon light exposure, 11-cis retinal isomerizes into 11-trans retinal which causes a

conformational change in opsin, resulting in activated transducin, which biochemically amplifies the

signal further; the end result of this process is a drop in cGMP concentration, that results in closed

cGMP-gated Ca2+ channels and a decrease in the release of glutamate from the PR axon terminal

(Molday, 1998; Hargrave, 2001). PRs transmit signals to bipolar cells in the inner nuclear layer (INL),

which in turn send a signal to ganglion cells (Figure 1.1). Axons from the ganglion cells then

converge to form the optic nerve, which sends visual signals into the visual cortex for higher

processing (Hubel, 1995).

3

Figure 1.1: Structure of the vertebrate retina. The neural retina has three nuclear layers, each composed of specific cell types. The nuclear layers are separated by the outer and inner plexiform layers, which are composed of synaptic terminals. On the left is a cartoon of the retina with corresponding structures indicated on a toluidine-stained section of a wild-type adult murine retina. A=amacrine cell, B=biplolar cell, C=cone photoreceptor, G=ganglion cell, H=horizontal cell, M=Müller cell, R=Rod photoreceptor.

4

Figure 1.2: Structure of a rod photoreceptor. The rod PR is composed of the following structures: the outer segment (OS), the inner segment (IS), the cell body and the synaptic terminal. A single outer segment disk has been enlarged on the right. Three OS disk proteins relevant to the research in this thesis include the photopigment rhodopsin, an intrinsic membrane protein of the disk membrane; Peripherin or Rds, which is localized to the disk rims and forms a mixture of disulfide-linked dimers (indicated), octomers and higher-order oligomers. Rds can also form noncovalent homo- and hetero tetramers with Rom-1, another disk rim protein. Similar to Rds, Rom1 can also form homodimers but is not usually observed in homomeric higher order structures (reference Biochemistry. 2008 Jan 29;47(4):1144-56. and Loewen and Molday 2000).

5

Other neuronal cell types in the retina function to modulate the neural signals originating from PRs.

A single horizontal cell is capable of providing feedback to multiple PR cells. Horizontal cells are

thought to enhance the contrast between adjacent light and dark visual fields by allowing the PR at

the center of a visual field to excite the ganglion cell below it, while inhibiting neighboring PRs

(Masland, 2001). In contrast, amacrine cells can modulate the activity of bipolar and ganglion cells

(Kolb et al., 1992; Masland, 2001).

1.1.2 Retinal pigment epithelium

The retinal pigment epithelium (RPE) is a monolayer of cells situated immediately behind

the PRs. The RPE is critical to the function and survival of PRs. For example, if the RPE is

separated from PR layer, the PRs will die (Anderson et al., 1983; Berglin et al., 1997). One of the

main roles of the RPE is in vitamin A metabolism in the retina. As previously discussed, all-trans

retinaldehyde results from light-activated rhodsopsin, which is subsequently isomerized to all-trans

retinol by another enzyme located in the PR OS, and transported to the RPE, where it is converted

back to 11-cis retinaldehyde (Stryer, 1996). The RPE65 gene encodes an isomerase required for this

conversion, and mutations in RPE65 have been associated with PR degeneration (Marlhens et al.,

1997). The RPE is also involved in the phagocytosis of disks at the distal tip of the OS. The RPE is

also thought to provide trophic support to PRs. For example, an in vitro study showed that PR

cultures survived longer in the presence of RPE-conditioned media, suggesting the conditioned

media contained trophic factors that improved PR survival (Gaur et al., 1992).

1.1.3 Müller Glia

Müller cells are radial glia and the principal glial cell type in the vertebrate retina (Bringmann

et al., 2006). Müller glia span from the outer limiting membrane, which is formed by the junction

between Müller cells and PR inner segments to the inner limiting membrane at the base of ganglion

cells, ensheathing all retinal neurons in between (Bringmann et al., 2006). This morphological

relationship allows Müller cells to carry out their many functions. For example, Müller cells play an

important role in neurotransmitter recycling. Glutamate is the dominant excitatory neurotransmitter

in the retina and its timely clearance is required so that excitatory synapses can continue to function

normally; neurotoxicity can also occur in the presence of excessive glutamate (Barnett and Pow,

6

2000). Müller cells express the glutamate/aspartate transporter GLAST, which rapidly removes

glutamate from synapses; glutamine synthetase then catalyzes the conversion of glutamate into

glutamine within Müller cells (Derouiche and Rauen, 1995). Finally, glutamine is transported back to

neurons as a precursor required for glutamate synthesis (Bringmann and Reichenbach, 2001).

The maintenance of retinal potassium ion (K+) homeostasis is another critical activity of

Müller cells. Müller cells express a specific class of K+ channel called inward-rectifying K+ channels

which make the cell highly permeable to K+ (Newman, 1993), an ion released by active neurons.

Excess K+ is absorbed by Müller cells from the neuronal extracellular space, and is distributed to the

blood, the vitreous and the subretinal space (Bringmann and Reichenbach, 2001).

Several other important functions of Müller glia have been identified. The retina produces a

lot of water as a result of its high metabolic rate (Bringmann and Reichenbach, 2001). In addition,

water influx also occurs as a result of metabolic substrate uptake (Bringmann and Reichenbach,

2001). Aquaporins present on Müller glia facilitate the redistribution of water out of the inner retinal

tissue (the RPE facilitates the redistribution of water out of the subretinal space) (Bringmann and

Reichenbach, 2001). Müller cells also provide reduced glutathione to neurons, which reduces the

number of free radicals and reactive oxygen species (Schutte and Werner, 1998; Bringmann and

Reichenbach, 2001). Finally, Müller glia can phagocytose debris from dead neurons or RPE cells

(Francke et al., 2001). While Müller glia play a role in maintaining homeostasis in the healthy retina,

there is increasing evidence that they also play a neuroprotective role in retinas under stress (Zack,

2000; Bringmann and Reichenbach, 2001).

Müller glia also appear to be important in PR degeneration (Zack, 2000). PR cell death can

be caused by pathogenic mutations or injury (LaVail et al., 1992; Pacione et al., 2003). Several

neurotrophic factors, including basic fibroblast growth factor (bFGF), ciliary neurotrophic factor

(Cntf), leukemia inhibitory factor (Lif), brain-derived neurotrophic factor (Bdnf), and neurotrophin-

3 (NT-3) have been shown to slow the rate of PR degeneration in light-damaged retinas (LaVail et

al., 1992); Müller cells may be central to this protection. Upon neurotrophin treatment, Müller cells

exhibit a dramatic up-regulation of pathways thought to be involved in PR protection. For example,

the phosphorylation of signal transducer and activator of transcription-3 (pStat-3) was observed in

Müller glia and ganglion cells after Cntf treatment (Peterson et al., 2000; Wang et al., 2002), while

extracellular signal-regulated kinase-1,2 (pErk1,2) was observed in Müller glia after Cntf, bFGF, or

Bdnf treatment (Peterson et al., 2000; Wahlin et al., 2000). Interestingly, these downstream pathways

are not activated in PRs in most cases (Peterson et al., 2000); treatment with Lif, however, results in

7

pStat-3 staining in a small subset of PRs and all Müller glia (Ueki et al., 2008). The lack of pStat-3 or

pErk1,2 activation in PRs upon treatment with most neurotrophins suggests that Müller glia may be

the primary cells responding to the neurotrophin treatment, which then results in PR protection by

an unknown mechanism. These results support the Müller cell hypothesis which states that cytokine-

mediated PR protection in IPD and light-damaged retinas is dependent on Müller glia (Zack, 2000).

This topic will be discussed in greater detail in the section dedicated to neurotrophins and IPD.

Interactions between Müller glia and PRs were suggested by a study examining the effect of

bright light on the expression of certain neurotrophin receptors (Harada et al., 2000). The high-

affinity TrkC and low-affinity p75NTR receptors displayed increased expression in Müller glia upon

light exposure (Harada et al., 2000). Blocking the TrkC receptor resulted in an 80% increase in PR

death upon light exposure; interestingly, blocking the p75NTR receptor elicited the opposite response,

a 26% decrease in cell death (Harada et al., 2000). In vitro experiments suggested that bFGF secretion

by Müller glia could account for these results. Activation of the TrkC receptor resulted in increased

bFGF secretion, while activation of the p75NTR receptor resulted in decreased bFGF secretion

(Harada et al., 2000). Since bFGF has been shown to protect PRs (LaVail et al., 1992; LaVail et al.,

1998), increased bFGF secretion by Müller cells upon TrkC activation could account for the

neuroprotection, while decreased bFGF signalling upon p75NTR activation may explain the increased

PR cell death. Thus, PR degeneration modulates a neurotrophic pathway that involves PRs

communicating to Müller glia, which are capable of responding by sending a neuroprotective signal

back to PRs.

1.2 Types of Photoreceptor Degeneration

IPDs are a genetically and phenotypically diverse groups of diseases that lead to visual

impairment (Pacione et al., 2003). Different types of PR degeneration can be distinguished by the

following characteristics: the mode of inheritance, pattern of visual loss, and by the mutant gene

involved in the disorder (Pacione et al., 2003). In the majority of IPDs both the cones and rods die,

but the degree to which each cell type is affected differs between the various disorders. For example,

retinitis pigmentosa (RP) is characterized by the initial loss of rods, leading to night blindness and

loss of peripheral vision, followed by the loss of cones, leading to a loss of central vision and

blindness (Hartong et al., 2006). Most cases of RP are monogenic (Hartong et al., 2006). A second

8

group of IPDs, the macular degenerations, involve both rods and cones since both cell types are

present in the macula, a region near the center of the eye responsible for high acuity vision (Hubel,

1995). These macular degenerations typically result in poor visual acuity, impaired colour vision and

photophobia (Rattner and Nathans, 2006). Age-related macular degeneration (AMD) is the most

common form of macular degeneration, with disease incidence rising sharply after age 70 (Rattner

and Nathans, 2006). Results from twin and family studies have pointed to a strong genetic

component for AMD and several susceptibility loci have been identified (Rattner and Nathans,

2006). There are also strong environmental risk factors, including a history of smoking, obesity, a

high intake of vegetable fat and a low intake of antioxidants (Jager et al., 2008). Monogenic forms of

macular degeneration also exist and tend to affect younger individuals (Rattner and Nathans, 2006).

A unique feature of IPDs is the functional diversity of the proteins encoded by the genes

associated with these diseases (Pacione et al., 2003). To date, 193 human IPD loci have been

identified and the genes at 142 of these loci have been determined (Daiger et al., 2008). Much

research has focused on characterizing the normal function of proteins encoded by IPD genes. Of

all the proteins involved in the monogenic forms of IPD examined to date, twelve functional

categories have been identified. These include proteins involved in structure, phagocytosis, cilia

maintenance, phototransduction, signalling, vitamin A metabolism, transport, transcription, RNA

splicing, and enzyme catalysis (Pacione et al., 2003; Hartong et al., 2006). In the next section, I will

describe at least one gene in each functional category and focus on how the mutant protein may be

contributing to cell death. It is important to note, however, that the precise relationship between the

altered function of an IPD protein and the death of the mutant PR (years to decades after the birth

of the mutant neuron) is not clear in any instance except perhaps the Rd1 mutant. In the following

discussion, I give particular attention to the Rom1-/-, Rds+/-, Rd1-/- mutants, and to mice carrying the

human mutant Rhodopsin P347S transgene, because experiments employing these mouse IPD

models constitute the bulk of the research reported in this thesis.

1.3 Genes Implicated in Retinal Degeneration

1.3.1 Structural proteins- Rds and Rom1

The retinal degeneration slow (Rds) mouse is one of the oldest mouse models of PR

degeneration (Sanyal et al., 1980; Sanyal and Jansen, 1981). Rds mutants exhibit a semidominant

9

pattern of inheritance (Ma et al., 1995). The mutant allele responsible for this spontaneous model

was identified as a 9.2 kb repetitive element insertion into exon two of the Rds gene (Travis et al.,

1989; Ma et al., 1995). Although the structure of PRs in Rds-/- mice is normal at birth, they fail to

develop outer segments and exhibit rod and cone PR loss that begins at two to three weeks of age

and is complete by 9-12 months of age (Sanyal et al., 1980; Sanyal and Jansen, 1981). The PRs of

Rds+/- mice degenerate at a slower rate than in Rds-/- mice, and rods are more quickly affected than

cones; by 18 months of age, 50% of the PRs have already degenerated in Rds+/- mice (Sanyal et al.,

1980). While outer segments do form in Rds+/- retinas, their development is delayed, they achieve

only half the normal length, and the disks within the OS form irregular whorls and remain

disorganized throughout life (Hawkins et al., 1985). The importance of the Rds protein is not unique

to rodents; 34 different mutations in the Rds gene has been associated with human RP.

The Rds protein is thought to have only a structural role in PRs (Figure 1.2). The Rds

protein is an integral membrane protein localized to the rim regions of rod and cone OS disks and

lamellae (Boon et al., 2008). Rds can form homo-tetramers or hetero-tetramers with another protein

called rod outer segment membrane protein-1 (Rom1); higher order structures can form from these

core tetramers (Loewen and Molday, 2000). The C-terminus of Rds has been associated with

membrane fusion, which is important for disk renewal (Boon et al., 2008). Interestingly, while Rds-/-

mice contain no OS disks, cone lamellae (the equivalent of disks in cones) are able to form,

suggesting that Rds is less important in cone lamellae genesis or stability than it is in the biogenesis

of rod OS disks.

Rom1 is also an integral membrane protein localized to the rims of PR OS disks where it can

form oligomers with Rds (Figure 1.2) (Bascom et al., 1992; Moritz and Molday, 1996). The

importance of Rom1 in PR biology was suggested by the identification of patients with digenic RP.

In this genetically complex disorder, affected individuals are heterozygous for mutations in both the

Rds and Rom1 genes (Dryja et al., 1997). A Rom1 complete loss-of-function mouse mutant was

created, which exhibited disorganized OS disks and PR degeneration (Clarke et al., 2000b). Rom1-/-

mice are able to form outer segments, indicating that the Rds protein alone, in the absence of

Rds:Rom1 oligomers, is sufficient for disk and OS morphogenesis. However, the disks produced by

Rom1-/- mice are large and disorganized, indicating that Rom1 is required for normal disk formation

(Clarke et al., 2000b). Furthermore, the loss of Rom1 results in slow PR degeneration; by 18 months

42% of PRs had died by apoptosis (Clarke et al., 2000b). Thus, Rom1, like Rds, is required for

normal disk morphogenesis and is essential for PR viability. The mechanism through which

10

dysmorphic OS disks, or some other consequence of the absence of Rom1, leads to PR

degeneration is not currently understood.

1.3.2 Phagocytosis- Mertk

PRs are dependent on the underlying RPE to phagocytose the oldest OS disks. The

autosomal recessive Royal College of Surgeons (RCS) rat model of IPD has a defect in this process,

resulting in debris accumulation in the subretinal space (Dowling and Sidman, 1962) and a fast rate

of PR degeneration; only a few PRs remaining by post-natal day 60 (Vollrath et al., 2001). Chimeric

studies first suggested that mutant RPE cells and not PRs were the cause of PR death in the RCS

model (Mullen and LaVail, 1976). The receptor tyrosine kinase Mertk was later identified as the

mutant gene in the RCS rat; gene transfer of Mertk into the RCS retina corrected the phagocytosis

defect and preserved PRs (Vollrath et al., 2001). Moreover, Mertk co-localizes with outer segment

material during the early stages of phagocytosis into the RPE (Feng et al., 2002). Thus, Mertk is

required for the receptor-mediated phagocytosis of outer segments by the RPE, an essential process

for maintaining the appropriate balance between synthesis and degradation of the outer segment

membranes (Pacione et al., 2003).

1.3.3 Cilia maintenance/Trafficking of intracellular proteins- BBS genes

The PR OS is connected to the rest of the cell by a connecting cilium called the inner

segment (Figure 1.2). Proteins synthesized in the inner segment must be transported through the

connecting cilium to reach the outer segment; for example, it is estimated that ~2000 opsin

molecules a minute are required to maintain the rod outer segment (Rosenbaum and Witman, 2002).

Mutations in genes encoding proteins involved in cilia maintenance or transport lead to the Bardet-

Biedl syndrome (BBS), a rare pleiotropic genetic disorder with retinal degeneration as one of its

primary features (Mykytyn and Sheffield, 2004). Most forms of BBS are autosomal recessive, but

triallelic inheritance has been well documented in some instances (Katsanis, 2004). Eleven BBS

genes have been identified to date, and current data suggest that BBS proteins are required for cilia

formation, maintenance and function (Badano et al., 2006). Recently, it has been recognized that

cilia function in many processes, including protein degradation, neuronal migration, axonal guidance,

vesicular transport, and signalling (Badano et al., 2006). Impairment of one or several of these

processes may account for the pleiotropy observed in this syndrome.

11

Studies in mice and zebrafish have elucidated some BBS protein functions. Knockdown of

each of the six zebrafish BBS orthologs led to the loss of cilia in Kupffer’s vesicle, a ciliated organ

which plays a role in left-right patterning and retrograde melanosome transport (Yen et al., 2006).

Multiple BBS proteins have been localized to the connecting cilium of the retina (Nishimura et al.,

2004), including the BBS8 protein which was also shown to interact with PCM1, a protein thought

to be involved in ciliogenesis (Ansley et al., 2003). Mice lacking the Bbs4 gene display defects in the

transport of phototransduction proteins from the inner segment to the OS (Abd-El-Barr et al.,

2007). BBS4 is also capable of binding dynactin, suggesting it may play a role in microtubule

transport (Badano et al., 2006). Recently, BBS6 was shown to be involved in the planar cell polarity

pathway, a non-canonical Wnt pathway (Ross et al., 2005), suggesting BBS proteins may also play a

role in cell signalling. In conclusion, BBS proteins play a role in cilia formation, maintenance and

function. The importance of cilia participatation in a broad range of biological processes, including

signalling is emerging. Disruption of normal ciliary function can affect many sytems, including PR

survival.

1.3.4 Phototransduction- Rhodopsin and Rd1

Mutations in genes encoding many of the proteins involved in phototransduction,

the biochemical process that transduces light, are associated with PR degeneration. For example,

over 100 mutations in rhodopsin (RHO), the molecule responsible for photon capture, are

associated with PR degeneration (Hartong et al., 2006). Most RHO mutations are autosomal

dominant (Pacione et al., 2003). Interestingly, RHO mutations are quite prevalent; approximately

25% of autosomal dominant retinitis pigmentosa patients have a RHO mutation (Hartong et al.,

2006). The mechanism by which mutant or absent rhodopsin leads to PR death is unknown

(Pacione et al., 2003). Proposed mechanisms include interference with metabolism, the formation of

toxic aggregates, the prevention of normal intracellular transport, and the disruption of PR outer

segment biogenesis (Hartong et al., 2006).

Many mouse models have been created to examine the role of mutant rhodopsin in PR

biology. One strategy involves the creation of transgenic mice expressing a mutant human RHO

transgene to study the role of a specific mutation in PR degeneration. Li et al. created a mouse

expressing human RHO with a proline 347 to serine mutation (P347S), a mutation associated with

autosomal dominant RP in humans (Li et al., 1996). The PRs in P347S mice develop normally and

the mutant rhodopsin protein is predominantly found in the outer segments; the remaining mutant

12

rhodopsin was found to accumulate in small vesicles near the junction between inner and outer

segments, suggesting a partial defect in rhodopsin transport. The outer segments were also shorter in

length and phtotoreceptor degeneration was observed (Li et al., 1996).

The Rd1-/- mouse is a spontaneous mutant that lacks a functional β cGMP

phosphodiesterase subunit (Pde6b) and exhibits rapid PR degeneration (Chang et al., 2002). In the

normal PR, light-activated rhodopsin activates Pde6b through transducin; active Pde6b then

hydrolyzes cGMP to 5’-GMP, resulting in a mean decrease in the concentration of cGMP, which

causes cGMP-gated ion channels to close, resulting in decreased PR Ca2+ contentration (Stryer,

1996). The PR is then hyperpolarized, and ceases neurotransmitter release at its synaptic terminal

(Stryer, 1996). Pde6b null mice have very high PR cGMP concentrations, which in turn opens an

increased proportion of cGMP-gated channels in the plasma membrane. Rod cell death is thought to

result from the excessive Ca2+ influx through the cGMP-gated channels (Hartong et al., 2006).

Calcium channel blockers have been successful at slowing the rate of PR death in Rd-/- mice,

supporting the notion of Ca2+ toxicity as the mechanism of cell death in this model (Frasson et al.,

1999).

1.3.5 Signalling, cell-cell interaction, or synaptic interaction- Sema4A

Semaphorins are a large conserved family of secreted or membrane-bound ligands with roles

as either attractive or repulsive cues involved in axon outgrowth (Fiore and Puschel, 2003). Recent

biochemical evidence suggests that semaphorins may also play a role in the organization of the

cytoskeletal network (Pasterkamp and Kolodkin, 2003). For example, Semaphorin 4A null mice

produce PRs that fail to produce elongated outer segments and die at a rapid rate, with few PRs

remaining after the first month of life (Rice et al., 2004). Sema4a has been shown to be involved in

cell-cell interaction in the immune system (Kumanogoh et al., 2002). The finding that Sema4a was

expressed in the RPE at a time when contacts are usually made between PRs and the RPE suggests

that Sema4a may function as a transmembrane ligand for a receptor present on PRs, an interaction

that is required for OS formation and PR survival (Rice et al., 2004).

1.3.6 Vitamin A metabolism- Rpe65

The activation of rhodopsin requires the conversion of 11-cis retinal, a product of vitamin A

metabolism, to all-trans retinal. The RPE is involved in the uptake of all-trans retinol (all-trans retinal

is converted into all-trans retinol in PRs) and the regeneration of 11-cis retinal through the action of

13

the RPE65 enzyme (Moiseyev et al., 2005). This process is critical for the visual cycle and PR

survival since mutations in the RPE65 gene have been associated with PR degeneration in mice,

humans and dogs (Bok, 2005a).

A lack of Rpe65 activity effectively deprives PRs of 11-cis retinal, resulting in a lack of

rhodopsin (opsin coupled with 11-cis retinal) but a high concentration of opsin (Fan et al., 2005).

Opsin alone is “noisy” and can occasionally trigger the activation of transducin (Fan et al., 2005).

When Rpe65-/- mice were crossed with animals lacking transducin, a molecule required for the

activation of the phototransduction cascade, no degeneration was observed in the double mutant

animal (Woodruff et al., 2003). These results suggest that the high concentration of opsin in the

Rpe65-/- retina results in the continuous activation of the phototransduction cascade, a toxic event

leading to PR degeneration (Redmond et al., 1998). A chronic decrease in intracellular Ca2+ has been

proposed as a possible trigger for PR death in Rpe65-/- mice (Fain, 2006).

1.3.7 Transporters/channels- ABCR (rod photoreceptor ABC transporter) or ABCA4

Mutations in the rod PR ABC transporter (ABCR) gene were first associated with a juvenile

form of macular degeneration called Stargardt disease and, more recently, with some cases of cone-

rod dystrophy, retinitis pigmentosa, and age-related macular degeneration (Westerfeld and Mukai,

2008). The ABCR gene encodes an ATP-binding transmembrane protein localized to the rim of rod

OS disks. It is a member of a highly conserved protein family that uses the energy from ATP to

translocate molecules across membranes, suggesting that ABCR is involved in the active transport of

protein in the outer segment (Westerfeld and Mukai, 2008).

The generation of Abcr -/- mice provided insight into the function of the Abcr protein (Weng

et al., 1999). Mice lacking Abcr displayed increased all-trans-retinal in PRs following light exposure,

suggesting that Abcr is involved in transporting all-trans-retinal from the OS disk interior to the PR

cytoplasm, where it available for uptake by the RPE (Weng et al., 1999). Abcr -/- mice also displayed

an accumulation of visual cycle byproducts which form in PRs due to the presence of increased all-

trans-retinal concentration (Mata et al., 2000; Bok, 2005b). These byproducts are then taken up and

processed by the RPE cells, resulting in the production of epoxides, which is highly toxic to the

RPE (Mata et al., 2000; Bok, 2005b). In this model, loss of the RPE is the primary cause of PR

degeneration (Bok, 2005b).

14

1.3.8 Transcription factors- Beta2/NeuroD1

Several transcription factors involved in retinal development have also been associated with

retinal disease (Phelan and Bok, 2000; Pennesi et al., 2003). For example, BETA2/NeuroD1 is a

bHLH transcription factor involved in the regulation of neuron versus glial cell fate and in amacrine

and bipolar cell specification in the developing retina (Morrow et al., 1999). Surprisingly,

BETA2/NeuroD1 null mice also exhibited PR death, suggesting that BETA2/NeuroD1 is also

required for PR maintenance in addition to the development of non-PR neurons (Pennesi et al.,

2003). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) staining suggested

two phases of PR death in this model: an early-phase peaking at postnatal day 3 and a late-phase,

characterized by a slow rate of cell death that continued into adulthood (Pennesi et al., 2003). This

late phase could be due to improperly differentiated rod cells dying by apoptosis. Electroretinograms

of BETA2/NeuroD1 null mice indicated that surviving PRs are functional; why PR death occurs in

the adult BETA2/NeuroD1 null mouse is not understood (Pennesi et al., 2003).

1.3.9 RNA intron-splicing factors

Three ubiquitously expressed genes encoding proteins involved in pre-mRNA splicing have

recently been associated with autosomal dominant RP (Pacione et al., 2003). The majority of pre-

mRNA introns are sliced by the major spliceosome, a complex of five uridine-rich

ribonucleoproteins (snRNP) called U1, U2, U4, U5, and U6. The U1 snRNP binds to the 5’ pre-

mRNA splice site, while the U2 snRNP binds to the branch-point sequence, near the 3’ pre-mRNA

splice site (Hastings and Krainer, 2001). A catalytically active spliceosome is created when a tri-

snRNP composed of U5 and U4/U6 binds to the 5’ spice site, allowing a complex series of

rearrangements resulting in intron removal (Staley and Guthrie, 1998). Three genes associated with

dominant RP encode proteins that are a part of this U5-U4/U6 complex and are essential for

splicing to occur in vitro : PRPF31 is required for U5-U4/U6 tri-snRNP complex formation in

human cells possibly through the tethering of the U5 snRNP to the U4/U6 snRNP (Makarova et al.,

2002); similarly, PRPF8 contibutes to 5’ and 3’ splice site recognition and formation of the catalytic

core in yeast, likely by facilitating the tethering of the U5 snRNP to the spliceosome (Collins and

Guthrie, 2000); finally, PRPF3 is required for the stable formation of the U4/U6 snRNP in yeast

(Anthony et al., 1997).

Several hypotheses have been suggested to explain why mutations in these three ubiquitously

expressed splicing factor genes cause adRP, a PR specific phenotype, yet do not seem to affect the

15

viability of other cell types. The first hypothesis is that splicing may be required to occur at

maximum efficiency in rods to supply them with their unusually high protein synthesis requirements

(Pacione et al., 2003); approximately 10% of the total rod OS disks are replaced each day (Young,

1976), with rhodopsin accounting for greater than 70% of the total protein found in the OS disks

(Hargrave, 2001). If the production of rhodopsin is reduced by 50%, as in mice who are

heterozygous for the rhodopsin null allele, rod PR degeneration occurs (Humphries et al., 1997). In

vitro experiments demonstrated that primary retinal explant cultures transfected with a PRPF31

mutant construct displayed approximately 50% less rhodopsin compared to controls (Yuan et al.,

2005). This experiment suggests that splicing factor mutations might act by limiting the protein

synthesis of rhodopsin accounting for the death of PRs.

A second hypothesis is that PR death may result from a PR-specific effect of mutant splice

factor gene expression (Pacione et al., 2003). This hypothesis is supported by the finding that PRs

transfected with mutant PRPF3 exhibited mutant protein aggregates, while aggregation did not occur

in fibroblasts or epithelial cells (Comitato et al., 2007). Thus, the mutant protein might have a cell-

specific dominant effect in PRs but not other cells (Comitato et al., 2007). The origin of this

differential effect is not known.

1.3.10 Enzymes- IMPDH1

Inosine monophosphate dehydrogenase type 1 (IMPDH1) was first identified as a candidate

IPD loci by experiments indicating that IMPDH1 was enriched in PRs (Bowne et al., 2002; Kennan

et al., 2002), a characteristic of many IPD genes (Blackshaw et al., 2001). Mutations in IMPDH1

were found in affected individuals from four families: the Asp226Asn missense mutation was found

in three families and the Arg224Pro substitution was found in the fourth family (Bowne et al., 2002;

Kennan et al., 2002). Unaffected individuals in the families did not have either mutation, suggesting

the identified mutations were associated with RP (Bowne et al., 2002; Kennan et al., 2002).

IMPDH1 catalyzes the rate-limiting step in de novo synthesis of guanine nucleotides. Similar

to the splicing factors, IMPDH1 is ubiquitously expressed but mutations result only in PR

degeneration. Its high level of expression in the retina and the critical role of guanine nucleotides in

phototransduction could explain the retina-specific nature of IMPDH1 mutations (Kennan et al.,

2002). Upon light activation, rhodopsin activates transducin through the conversion of GDP to

GTP, activating an enzyme which cleaves cGMP, closing cGMP gated ion channels (Stryer, 1996).

16

Thus, the loss of one allele of IMPDH1 may prevent adequate guanine nucleotide synthesis;

however, it is not known how this could lead to PR degeneration (Kennan et al., 2002).

1.4 Common Features of IPD

Mutations in genes encoding proteins from many functional categories result in IPD. This

diversity is unique to PR degenerations; degenerations that affect other parts of the nervous system

usually occur in a group of functionally related proteins (as in Alzheimer’s disease) or in proteins

with common structural features (such as the CAG triplet repeat expansion diseases). Disruption of

PR structure, protein trafficking, calcium homeostasis, the RPE, and PR metabolism all lead to PR

death; however, it remains unclear why disruptions in these diverse classes of proteins lead to PR

death. Several common features of IPD have been reported in the literature and are discussed

below.

1.4.1 Altered calcium homeostasis

A defect in calcium homeostasis has been proposed as one mechanism accounting for

mutant PR death. As previously discussed, the Rd1-/- model of IPD exhibit higher intracellular

calcium levels due to a defect in βPde, which results in higher concentrations of cGMP and more

open cGMP-gated calcium channels. Calpain activity has been proposed as the link between elevated

calcium levels and PR death. Active in Rd1-/- retinas (Doonan et al., 2005), calpains are calcium-

activated proteinases that can trigger cell death by breaking down molecules in the cytoskeleton and

plasma membrane (Paquet-Durand et al., 2007). Treatment with calcium-channel inhibitors slowed

the rate of degeneration in Rd1-/- (Frasson et al., 1999; Takano et al., 2004), as did the use of calpain

inhibitors in vivo (Sanges et al., 2006), suggesting that calcium toxicity is playing a role in PR death.

Several mouse models of IPD were treated with calcium-channel blockers to determine if

decreasing cellular calcium concentrations could be a general strategy to protect PRs (Takeuchi et al.,

2008). Rds+/- retinas treated with calcium channel blockers functioned better, as indicated by

electroretinography, and displayed more organized OS disks; however, no difference in PR number

was observed between treated and un-treated retinas (Takeuchi et al., 2008). Furthermore, increased

expression of Cntf and FGF2, known survival factors, were observed upon calcium-channel blocker

treatement (Takeuchi et al., 2008). Thus, it was not clear whether decreased cell calcium or a

17

secondary effect of drug treatment was responsible for increased Rds+/- PR function. In a similar

experiment, RCS rats were treated with calcium-channel blockers and displayed increased PR

function and a slower rate of PR degeneration. In contrast, a mutant rhodopsin rat model treated

with calcium-channel blockers failed to show any PR rescue. In summary, while calcium may have a

critical role in PR cell death in some models, its role would not appear universal in IPDs.

Interestingly, these results directly refutes the hypothesis that decreased PR calcium can also be toxic

to PRs (Fain, 2006), since none of the treated retinas displayed increased PR death.

1.4.2 Cell death by apoptosis

A common feature of PR death appears to be the initiation of a cascade that ultimately leads

to apoptosis. Until recently, caspase-dependent apoptosis was thought be the only cell death

pathway in PR degenerations (Travis, 1998; Wenzel et al., 2005). Caspases are endopeptidases that

cleave proteins at aspartate residues and were considered to be central executioners of the apoptotic

program, cleaving a variety of intracellular substrates (Beere, 2005). Activation of caspase-3 was

detected in rats expressing a mutant Rhodopsin transgene (Liu et al., 1999). The role of caspase-3 was

directly tested by examining the rate of PR death in Rd1-/- mice with a caspase-3 loss-of-function

mutation; PR death was delayed, but not prevented in double mutant animals (Zeiss et al., 2004).

Similarly, treatment with a pan-caspase inhibitor did not abolish apoptosis but reduced the number

of apoptotic Rd1-/- PRs (Sanges et al., 2006). The overexpression of Bcl-2, an anti-apoptotic protein

that prevents caspase-3 activation, resulted in PR rescue in Rds-/- retinas (Nir et al., 2000). Altogether,

these studies indicate that caspase-dependent apoptosis has a major, but not absolutely essential, role

in PR death.

Proteolytic degradation is one of the hallmarks of apoptosis. Recent evidence suggests that

the cleavage of proteins in mutant PRs also occur through caspase-independent mechanisms

(Doonan et al., 2003; Wenzel et al., 2005). For example, cathepsins, calpains, granzymes A and B,

and serine proteases-like AP24 are all proteases linked to apoptosis; these enzymes can act together

with caspases or function independently as executioners of apoptosis (Wenzel et al., 2005).

Autophagy is a catabolic process through which a cell digests its own cellular components and is

another caspase-independent cell death process active in the retina (Kunchithapautham and Rohrer,

2007). In autophagy, the primary proteases are cathepsins and proteasomal proteins (Klionsky and

Emr, 2000). Significantly, Rd1-/- and Rds-/- retinas exhibit markers of autophagy, including active

cathepsin B and lysozymal enzymes and increased lipidated LC3, a complex associated with active

18

lysosomes (Kunchithapautham and Rohrer, 2007). The emerging picture of cell death in IPD retinas

is one of multiple parallel pathways, including caspase-dependent apoptosis, contributing to cell

death.

1.4.3 Common kinetics of cell death

Several years ago, Clarke et al. made the important observation that many, if not all, IPD

models share the same kinetics of cell death (Clarke et al., 2000a). In all 11 animal models of IPD

examined, the risk of cell death averaged across the cell population was found to be constant

throughout the life of the animal (Clarke et al., 2000a). The kinetics of cell death were initially

described as being exponential; however, this model has since been revised slightly with the

observation that stretched exponential kinetics is a more accurate description of the cell death

kinetics (Clarke and Lumsden, 2005). The stretched exponential model accounts for the fact that the

rate of cell death slows slightly over time for a population of mutant neurons, although the risk of

cell death for an individual mutant neuron remains constant (Clarke and Lumsden, 2005). In the

stretched exponential model, each neuron still has a constant risk of cell death, but the risk differs

from that of other mutant cells in the population (Clarke and Lumsden, 2005). For example, spatial

differences are known to influence the rate of cell death (Pacione et al., 2003). Thus, a pool of

mutant neurons in the same region of the retina might have the same constant risk of cell death,

which may differ from another population in a different region of the same retina. Alternatively,

there may be endogenous differences between PRs in a population that can influence the constant

risk.

Stretched exponential decay implies that the system is characterized by at least four features.

In the case of mutant PRs, these features are as follows: (i) the risk of cell death is constant

throughout the lifetime of each mutant PR; (ii) each mutant neuron has a different risk of cell death,

but that risk remains constant over time; (iii) the time at which cell death occurs is random; and (iv)

the death of each neuron is independent of other neurons.

1.4.4 Implications of these common features occurring in IPD

Although IPD-associated proteins display great functional diversity, there are some common

features as well. Apoptosis seems to be one of the final pathways resulting in cell death. While it has

been suggested that processes such as autophagy also play a role in PR death, the pervasiveness of

autophagy in different IPD models and the degree to which it contributes to cell death remains to be

19

determined. Stretched exponential kinetics also seems to be another common feature of IPD. These

commonalities argue against the possibility that there are 190 private IPD gene-specific preapoptotic

pathways, all of which can individually trigger cell death. A more parsimonious model is that there

one, or very few pre-death pathways (or gateways) that are shared by all IPD loci, that all converge

toward apoptosis. The identification of critical molecules that participate in the pre-death pathways

should provide substantial insight into the pathogenesis of IPD and could suggest new therapeutic

avenues to prevent or slow PR death.

1.5 Microarrays and IPD

Global gene expression analysis facilitated by microarray technology and serial analysis of

gene expression (SAGE) have been used to further the understanding of IPD in two predominant

ways. First, global gene expression analysis has been employed to identify PR-specific or PR-

enhanced transcripts. This strategy is useful because it has been established that transcripts encoded

by IPD loci tend to be enriched in PRs (Blackshaw et al., 2001); novel PR-enriched transcripts

represent candidate IPD loci. Secondly, the identification of differences between IPD and wild-type

retinal transcriptomes is a promising approach to gain insights into why PRs die in IPDs. Analysis of

global gene expression can identify biochemical responses shared by many or all IPD retinas, as well

as reponses unique to individual IPD genes. The next sections review the use of arrays in IPDs.

1.5.1 Identification of photoreceptor-enriched genes

The mouse genome project and the use of high-throughput gene expression profiling tools

have facilitated the identification of PR-specific transcripts in the retina (Swaroop and Zack, 2002).

The crucial role of PR-enriched genes is highlighted by the estimation that approximately half of all

retinal degeneration-associated genes have transcripts which are preferentially expressed in the retina

(Blackshaw et al., 2001). A recent study integrating numerous microarray and direct sequencing

approaches estimated that the retinal transcriptome contains between 19,000-20,000 transcripts from

embryonic to adult stages; importantly, 39.1% of these transcripts were not a part of the 60,770

RIKEN set of full-length cDNAs, which are derived from multiple extra-ocular tissues, suggesting

that a high proportion of retinal transcripts are also retina-specific (Zhang et al., 2005). A similar

study estimated that the adult retinal transcriptome is comprised of 13,037 transcripts; however, the

authors acknowledged their estimation may only represent up to 90% of the actual total number due

to the lack of several published PR-specific transcripts in their final list (Schulz et al., 2004).

20

One strategy employed to identify retina-enriched genes has been to create microarrays from

retinal cDNA libraries and compare the hybridization intensity of retinal and non-ocular samples

(Chowers et al., 2003; Hackam et al., 2004a); transcripts with higher expression levels in the retinal

sample are candidate retina-enriched genes. Since PRs constitute 70% of the retina, retina-enriched

transcripts are likely to be PR transcript enriched. In one study, a custom microarray containing a

murine retinal cDNA library representing 5376 genes was hybridized with labeled RNA derived

from mouse brain, retina and liver; 1134 genes were identified as up-regulated in the retina from the

brain-retina and liver-retina comparisons (Hackam et al., 2004a). Of the retina-enriched genes, 205

mapped to known retinal disease regions (Hackam et al., 2004a). Over a third of the retina-enriched

genes were previously uncharacterized ESTs, representing potentially novel IPD candidate genes.

In a conceptually similar study, a cDNA microarray representing 10,034 human genes was

hybridized with post-mortem human retina, cortex and liver tissue from at least two donors

(Chowers et al., 2003). Of the genes identified in the retina-cortex and retina-liver comparisons, 211

were identified as up-regulated in the retina in both comparisons (Chowers et al., 2003).

Interestingly, the authors of both studies noted the limited overlap of genes identified as retinal-

specific by different groups using different high-throughput technologies such as SAGE, microarray

analysis and EST data mining (Chowers et al., 2003; Hackam et al., 2004a). This lack of overlap

could be explained by the following possibilities: (i) each methodology is efficient at sampling only a

limited and partially overlapping dataset; (ii) the existence of a large number of unknown retina-

enriched genes; (iii) significant error rates in at least some of the individual studies. An experimental

approach that includes multiple methodologies may be most effective in cataloguing the complete

library of retina-enriched genes.

Several studies have used mutant mice to identify transcripts specifically enriched in PRs

(Bowne et al., 2002; Kennan et al., 2002). In one study, microarray analysis was performed to

examine gene expression in the Rho-/- retina, at a time-point when all PRs had degenerated,

compared to wild-type (Kennan et al., 2002). Of the 77 significantly differentially expressed genes

identified, 44 were down-regulated in Rho-/- retinas and represent candidate PR-enriched genes; in

fact, nine of the down-regulated genes identified in this study map to chromosomal regions that

have been implicated in IPD (Kennan et al., 2002). Interestingly, one of the identified genes, inosine

monophosphate dehydrogenase type 1 (IMPDH1), was also identified during SAGE analysis

comparing the retinas of Crx-/- mice, which lack differentiated PRs, to wild-type (Blackshaw et al.,

2001; Bowne et al., 2002). IMPDH1 maps to 7q31.1 and was considered to be a strong retinitis

21

pigmentosa-10 (RP10) candidate gene. Further examination of the link between IMPDH1 and

disease revealed that a single mutation (Asp226Asn) in a conserved region of the predicted coding

region was found in all three RP10 families examined, suggesting this substitution is the allele

responsible for PR degeneration at the RP10 locus (Bowne et al., 2002; Kennan et al., 2002). It is

notable that the IMPDH1 gene is expressed in a wide variety of tissues; it became a strong IPD

candidate gene after it was identified as a PR-enriched gene by global gene expression analysis,

validating the approach of using global gene expression analysis to identify PR-enriched genes as

candidates for IPD. On the other hand, apart from the identification of IMPDH1 as a candidate

IPD locus, that was later shown to actually be an IPD gene (Blackshaw et al., 2001; Bowne et al.,

2002), I know of no other IPD genes that have been identified via global gene expression

methodologies.

1.5.2 Identification of regulatory networks in the retina

Microarray analysis performed to examine the genetic networks controlled by PR-specific

transcription factors has been an effective strategy to identify additional PR-enriched genes and

establish relationships between previously characterized genes. Many studies have focused on Crx, a

transcription factor involved in rod and cone differentiation during retinal development (Chen et al.,

1997; Furukawa et al., 1997). Livesey et al. identified 18 differentially expressed transcripts when

comparing Crx-/- and wild-type murine retinas at PN10, a time just prior to terminal PR

differentiation (Livesey et al., 2000). In addition to several known PR-specific genes, four novel up-

regulated genes were detected that are potential IPD-associated genes (Livesey et al., 2000). Livesey

et al. then went on to compare the proximal promoter (250bp upstream of the transcriptional start

site) of all up-regulated genes and discovered a novel Crx binding-motif (Livesey et al., 2000). The

identification of this motif is another step toward the elucidation of the transcriptional network

controlled by Crx, which could be used as a bioinformatics tool to identify other Crx-regulated, PR-

specific genes.

The identification of PR-specific transcripts has also facilitated research on the regulatory

networks controlling retinal gene expression. Novel regulatory targets of CRX, NRL, and NR2E3,

three transcription factors predominantly expressed in PRs, have been identified after a

compendium of PR-specific genes was assembled from the studies mentioned above, combined with

a search for phylogenetically conserved CRX, NRL and NR2E3 consensus binding sites in the

promoter region of each PR-specific gene (Qian et al., 2005). Of the 169, 166 and 97 putative targets

22

of CRX, NRL, and NR2E3, respectively, there was significant overlap between the CRX and NRL

dataset (148 genes); this overlap is consistent with the fact that many PR genes are co-regulated by

Crx and Nrl (Hsiau et al., 2007). Interestingly, 89% of the 58 mouse orthologs of human retinal

disease genes that were examined were dysregulated in Crx-/-, Nrl-/-, or Nr2e3-/- retinas, suggesting the

majority of IPD genes are regulated by Crx, Nrl and/or Nr2e3 (Hsiau et al., 2007). Thus, novel

targets of any of these three transcription factors should be considered to be IPD candidate genes.

In the study of Hsiau et al. that examined the transcriptomes regulated by Crx, Nrl, and

Nr2e3, the 15kb of sequence flanking the transcriptional start site of all dysregulated genes were

examined and genes ranked according to the number, affinity and clustering of phylogenetically

conserved Crx, Nrl, and Nr2e3 binding sites (Hsiau et al., 2007). Novel elements were identified and

tested in vivo by electroporating the novel elements linked to a reporter into retinal explant cultures

and then examined for PR staining (Hsiau et al., 2007). Importantly, 19 novel PR-specific cis-

regulatory elements were identified all of which map near retinal disease genes (Hsiau et al., 2007).

These novel elements represent a bioinformatic tool to identify additional PR-specific genes but also

provide insight into the regulatory networks at work in the retina.

A catalogue of PR-specific genes would be a valuable contribution to the study of the

regulatory networks that control retinal gene expression. Recently, 3000 functional cis-regulatory

elements were identified in the human genome by searching for blocks of extreme phylogenetic

conservation between non-coding regions of Fugu, human, mouse and rat genomes (Pennacchio et

al., 2006). Remarkably, none of the 19 novel PR cis-regulatory elements identified by Hsiau et al. were

present within the 3000 elements, suggesting that approaches that rely solely on identifying regions

of phylogenetic conservation may fail to recognize cell-type specific cis-regulatory elements (Hsiau et

al., 2007). A strategy that accompanies phylogenetic conservation with tissue-specific information

may be more fruitful in identifying tissue specific cis-regulatory regions.

The studies cited above have identified several novel PR genes. The documentation of PR-

specific transcripts has also provided some insight into the transcriptional networks operating in

PRs, a critical step in the understanding of retinal development and maintenance and retinal disease.

1.5.3 Gene expression changes in response to IPD

The molecular events leading to cell death in IPDs are complex. Although we have a detailed

understanding of the normal function of many IPD-associated proteins, the cellular pathways

leading from the primary mutation to cell death are not well understood (Pacione et al., 2003).

23

Microarray analyses comparing IPD to wild-type retinas, at an early time in the life of the mutant

IPD retina, at which the majority of PRs are still present, would in principle appear to be an

important strategy to gain insight into the molecular events preceding PR death. However, only a

few such studies have been performed.

The transcriptome of the Rd1-/- murine retina, which has a deficiency in the cGMP

phosphodiesterase β-subunit (βPde), is one example of an IPD model that has been characterized

by microarray analysis (Jones et al., 2000; Hackam et al., 2004b). Mutations in βPde cause a peak in

rod PR death by PN12-13, followed by cone cell death (Doonan et al., 2003; Punzo and Cepko,

2007). Hackam et al. used a custom microarray incorporating a retinal cDNA library to identify

changes in the Rd1-/- retinal transcriptome compared to wild-type at post-natal day 14 (PN14), PN35

and PN50, time-points which represented rod cell death, the beginning of cone cell death and during

cone cell death, respectively (Hackam et al., 2004b). Several transcripts whose protein products are

known to provide protection against oxidative damage were up-regulated during cone and rod

degeneration, suggesting oxidative stress may be an important factor during PR death (Hackam et

al., 2004b). A group of genes usually associated with tissue growth and differentiation showed

increased expression during rod and cone degeneration (Hackam et al., 2004b). In particular, Dickopf

3 (Dkk3), a member of the Wnt signalling pathway, was up-regulated in Rd1-/- retinas at all time-

points (Hackam et al., 2004b). Interestingly, increased expression of frizzled-related protein-2

(SFRP2), another Wnt-related gene was also identified (Jones et al., 2000) and displayed overlapping

expression with Dkk3, as determined by in situ hybridization (Hackam et al., 2004b). Wnt signalling

has been shown to inhibit apoptosis in vitro (Chen et al., 2001) and may represent a pathway that

inhibits rod and/or cone PR apoptosis in the Rd1-/- retina.

Examining the pathways involved in cone cell death in Rd1-/- retinas is complicated by the

fact that cones begin to degenerate at a time-point (~PN35) when all rod PRs, which account for

70% of the total cells in the retina, have already degenerated. Thus, any down-regulated gene

identified when comparing the transcriptome of a PN35 Rd1-/- retina to a wild-type retina would

likely be rod PR-specific. In addition, there will be an over-representation of RNA from non-PR cell

types in Rd1-/- retinas. Punzo et al. circumvented this problem by using microarray analysis as a

preliminary screening tool, followed by localization studies to identify the cell types in which the

expression changes occurred (Punzo and Cepko, 2007). Genes that were expressed in ganglion cells,

bipolar cells, and possibly Müller glia displayed the greatest degree of dysregulation in response to

the loss of Rd1 function, suggesting multiple cell types react to PR death (Punzo and Cepko, 2007).

24

Differentially expressed genes were then organized according to functional classes; genes involved in

cell-cell communication, metabolism and transcriptional regulation represented the largest classes of

differentially expressed genes, suggesting that an increase in cellular communication and a

restructuring of the remaining tissue occur during cone cell death (Punzo and Cepko, 2007).

The lack of functional βPde in Rd1-/- retinas causes a dramatic increase in cytoplasmic cGMP

concentration, resulting in a continual opening of cGMP-gated channels and the excessive entry of

calcium into the PR cytoplasm (Farber and Lolley, 1974). It has been suggested that this increase in

calcium is one of the primary events leading to apoptosis (Chang et al., 1993). Microarray analysis

characterizing the Rd1-/- transcriptome using commercial oligonucleotide microarrays has been

employed to further investigate this possibility (Rohrer et al., 2004). Retinas from Rd1-/- and wild-

type mice were compared at P6, before significant degeneration had occurred, and assayed

extensively during the degeneration process (every 3-4 days up until PN21); significantly, genes

involved in calcium binding were up-regulated starting at P10 (Rohrer et al., 2004), a time-point

when PRs begin dying by apoptosis (Punzo and Cepko, 2007). For example, calbindin, was up-

regulated at all time-points examined (Rohrer et al., 2004; Hartmann and Konnerth, 2005). There

was also an over-representation of genes encoding transcription factors, immune response proteins

and angiogenic factors in Rd1-/- retinas (Rohrer et al., 2004). These classes of up-regulated genes lead

Rohrer et al. to speculate that the primary insult in the Rd1-/- retina is caused by the rise in cGMP

and the accompanying calcium influx and toxicity, followed by secondary affects such as

neuroinflammation and blood-retina barrier breakdown (Rohrer et al., 2004).

Calpain activity has been proposed as a mechanism through which calcium toxicity could

lead to cell death. Calpains are calcium-activated proteinases that have been shown to be active in a

light-induced PR death model (Donovan and Cotter, 2002) and in Rd1-/- retinas (Doonan et al.,

2005). Excessively active calpains are thought to break down molecules in the cytoskeleton such as

actin-associated proteins, microtubule-associated proteins, and cell adhesion molecules, which can

lead to the degradation of the cytoskeleton and plasma membrane (Paquet-Durand et al., 2007).

Microarray analysis aimed at identifying differentially-expressed members of the calpain

system between Rd1-/- and wild-type retinas revealed that calpastatin, an inhibitor of calpain, was

down-regulated 20-fold in Rd1-/- retinas (Paquet-Durand et al., 2006). Although the calpain transcript

was not up-regulated, the down-regulation of calpastatin suggests the mechanism through which

calpain protein activity is increased in Rd1-/- retinas (Paquet-Durand et al., 2006). A recent study

25

determined that blocking calpain activity prevented PR death in Rd1-/- retinas, suggesting calpain

activity may be involved in mediating PR death in this model (Sanges et al., 2006).

Photoreceptor injury or disease often leads to the activation of Müller glia, the dominant glial

cell type in the retina (Bringmann and Reichenbach, 2001; Garcia and Vecino, 2003). A recent

microarray study examining the transcriptome of a light-mediated PR death model and the prCAD-/-

model of IPD identified 281 and 32 differentially expressed genes, respectively (Rattner and

Nathans, 2005). Interestingly, 8 of the 12 transcripts that were highly induced by light damage (>2.5-

fold) were also induced in the genetic model of IPD; five of the 8 commonly differentially expressed

genes were expressed in Müller glia, indicating a prominent Müller cell genomic response to PR

damage (Rattner and Nathans, 2005). The discrepancy in the number of genes identified in the light-

mediated damage and prCAD-/- models suggests there are many model-specific responses as well

(Rattner and Nathans, 2005).

Endothelin-2 (Edn2) was observed to be one of the most highly up-regulated genes by

microarray and northern analysis in both light-mediated and the prCAD-/- model of IPD.

Localization studies revealed that the up-regulation of the End2 mRNA occurred in PRs (Rattner

and Nathans, 2005). The Edn2 receptor, EdnrB, was also up-regulated in the light-mediated damage

model and was localized to Müller glia (Rattner and Nathans, 2005). Taken together, increased PR-

derived Edn2 may function as a stress signal that activates Müller glia by binding to the EdnrB

(Rattner and Nathans, 2005). Studies examining other endothelin family members in the central

nervous system have suggested that endothelins promote active astrocytosis, a process that occurs

during neuroinflammation and cell death (Barbeito et al., 2004). The role of Edn2 in IPD models

remains to be elucidated and further study is required to determine whether the increase in Edn2 is

pathogenic or a response to cell stress.

Global gene expression profiling has been productively used in several aspects of IPD

research. Many candidate IPD loci have been discovered using microarray analysis to identify PR-

specific or enhanced transcripts. Transcriptional networks have been characterized and novel, PR-

specific, cis-regulatory regions have been identified by efforts to find all PR-enriched transcripts.

While a few microarray experiments have been performed to identify transcriptional changes early in

the degeneration process, the examination of other IPD mouse models is likely to provide a better

representation of the full range of molecular responses that occur in mutant retina. In chapter three

of this thesis, I discuss microarray experiments characterizing the transcriptome of two additional

models of IPD.

26

1.6 Neurotrophic factors protect photoreceptors

A variety of neurotrophic factors have been shown to be protective in IPD and light-

damaged retinas (LaVail et al., 1992). The factors that display the greatest degree of PR protection

are ligands for the following receptors: the fibroblast growth factor receptor, the Trk neurotrophin

receptor and the interleukin 6 (IL-6) family of receptors (Chaum, 2003). An understanding of the

mechanism(s) through which these factors protect the mutant retina is likely to provide insight into

the pathogenesis of PR cell death, and could ultimately lead to improved therapies for patients with

IPD.

1.6.1 Fibroblast growth factor

Treatment with basic fibroblast growth factor (bFGF) slows the rate of degeneration in

light-damage and IPD rodent models (Faktorovich et al., 1990; LaVail et al., 1992; Lau et al., 2000).

Interestingly, preconditioning with bright light evokes a protective response against further light

damage that is concomitant with endogenous FGF up-regulation, suggesting endogenous

neurotrophic factor expression might mediate the protection (Liu et al., 1998; O'Driscoll et al.,

2008); localization studies determined that bFGF transcripts increased in several cell types, including

the RPE, PR inner segments and Müller cells (Cao et al., 1997; Wen et al., 1998). Cultured Müller

cells were shown to secrete bFGF into the surrounding media (Harada et al., 2000). For bFGF to

act, it must bind to its receptor, which has been identified in the retina and which may act directly on

PRs (Tcheng et al., 1994). For instance, bFGF can bind directly to rod outer segments, presumably

through the bFGF receptor (Fayein et al., 1990). In another study, isolated PRs grown in culture

(>99.5% purified) normally die, but the number of dead cells was attenuated by bFGF, suggesting

that PRs can respond directly to bFGF (Fontaine et al., 1998). While these experiments strongly

suggest bFGF can act on PRs, further experiments are required to definitively establish the presence

of Fgf receptors on these cells.

Basic FGF may play a broader role in the retina, as a trophic support molecule for the

maintenance of PRs (Chaum, 2003). Wild-type rats expressing a dominant negative FGF receptor in

PRs exhibited PR degeneration, suggesting that FGF signalling is required for PR maintenance

(Campochiaro et al., 1996). Thus, bFGF might provide critical trophic support to normal PRs, as

well as providing neuroprotection in conditions of stress.

27

1.6.1.1 Fibroblast growth factor signalling in photoreceptors

The PI3-K signalling pathway is activated by ligand-bound FGF receptors (Chaum, 2003).

The anti-apoptotic effects of the PI3-K cascade are mediated through the phosphorylation of Akt

(protein kinase B), a serine threonine kinase (Chaum, 2003). Once phosphorylated, Akt directly

represses caspase-9, one of the caspases that initiates apoptosis, and inhibits several pro-apoptotic

proteins including Bad, which represses the anti-apoptotic activity of BCL-xL and BCL-2 (Beere,

2005). Over-expression of PI3-K or Akt are sufficient to prevent neuronal apoptosis when trophic

factor support is withdrawn from a primary culture of cortical neurons (Philpott et al., 1997).

Activation of the PI3-K pathway has also been shown to be protective in a model of glaucoma and

after optic nerve transection (Zhou et al., 2007). In PRs, Jomary et al. observed a decrease in Akt

signalling in the period leading up to the peak of PR death in the Rd1-/- model, suggesting that the

down-regulation of this survival pathway might contribute to the PR death in this model (Jomary et

al., 2006). Thus, FGF-mediated inhibition of apoptosis may be the consequence of activation of the

PI3-K pathway.

1.6.2 Brain-derived neurotrophic factor and nerve growth factor

Brain-derived neurotrophic factor (Bdnf) was first shown to be neuroprotective in a light-

damage model of PR degeneration (LaVail et al., 1992; Hojo et al., 2004; Gauthier et al., 2005). The

effectiveness of Bdnf treatment was then examined in mouse models of IPD; surprisingly, although

Bdnf treatment rescued PRs in the Nr-/- model, it was not effective in Rd1-/- retinas or in the Q334ter

mutant rhodopsin transgenic mice (LaVail et al., 1998). Similarly, in a feline model of PR

degeneration, Bdnf was ineffective in slowing the rate of degeneration (Chong et al., 1999). Since

these studies typically involve one or a few Bdnf injections, Okoye et al. proposed that the length of

time that PRs were exposed to Bdnf was a critical factor for PR protection to occur (Okoye et al.,

2003). To test this theory, mice expressing an inducible Bdnf transgene were created and crossed

with mice expressing the Q334ter mutant rhodopsin transgene; double transgenic animals exhibited a

slower rate of degeneration when the Bdnf transgene was induced, indicating that continuous Bdnf

expression does protect PRs in Q334ter mice (Okoye et al., 2003). Continuous Bdnf expression

might also be effective in delaying PR death in other models.

Bdnf treatment may rescue mutant PRs through its effects of other retinal cell types. Bdnf

binds to the tyrosine kinase B (TrkB) receptor, which is present on Müller cells, the RPE, ganglion

cells and cone but not rod PRs (Chaum, 2003). Thus, other cell types mediate the effect of Bdnf on

28

PR survival. One hypothesis is that Bdnf causes an increase in bFGF and Cntf (another protective

cytokine that will be discussed in the next section) production and release from Müller cells, which

enhances PR survival (Harada et al., 2002).

Müller cells may also mediate PR protection by nerve growth factor (NGF) treatment

(Harada et al., 2000). NGF has been shown to protect PRs in light-damage and the RCS rat model

of IPD (Harada et al., 2000; Lenzi et al., 2005). It binds to the high affinity tyrosine kinase A (TrkA)

receptor, leading to increased bFGF production and PR protection (Harada et al., 2000).

Interestingly, NGF can also bind to the low-affinity neurotrophin receptor p75, resulting in lower

bFGF production from Müller cells (Harada et al., 2000). Thus, the abundance of each receptor on

the surface of Müller cells is correlated with the level of bFGF released, which determines the

neurotrophic response.

1.6.3 The Interleukin-6 family of cytokines

The interleukin-6 family consists of 8 cytokines, several of which have been found to delay

PR death in IPD and light-damage models. The IL-6 family includes Cardiotrophin-1 (Ct-1),

Cardiotrophin-like cytokine (Clc), ciliary neurotrophic factor (Cntf), interleukin-6 (IL-6), interleukin-

11 (IL-11), oncostatin M (Osm), leukemia inhibitory factor (Lif), and neuropoetin (Np) (Heinrich et

al., 1998). Some cytokine receptors can bind multiple cytokines. For example, Lif, Ct-1 and Clc bind

to the Lif receptor (Lifr), while Cntf and Np bind to the Cntf receptor (Cntfr). The Cntfr is unique

in its ability to form a tripartite receptor complex with Lifr and gp130 upon ligand binding (Heinrich

et al., 1998). All the other cytokines bind to a single receptor (Heinrich et al., 1998). Once an IL-6

ligand binds to its specific receptor, the complex associates with common gp130 co-receptor

(Heinrich et al., 1998). Although IL-6 cytokine receptors and gp130 have no intrinsic kinase activity,

they associate with Janus kinase 1 (Jak1), Janus kinase 2 (Jak2) and Tyrosine kinase 2 (Tyk2) on their

cytoplasmic side; upon ligand binding and association with gp130, the relevant kinases are

transphosphorylated and therby activated. These events trigger two main signalling pathways, by the

phosphorylation of extracellular signal-regulated kinases 1,2 (Erk1,2) and of signal transducers and

activators of transcription (STAT) proteins (Kamimura et al., 2003).

Several members of the interleukin-6 family of cytokines (IL-6) have neuroprotective

properties in IPD and light-damage models of PR degeneration (LaVail et al., 1992; LaVail et al.,

1998). For example, Cntf, first identified as a trophic factor required to culture chick ciliary ganglion

cells, is capable of slowing PR death in retinas exposed to bright light (Barbin et al., 1984; LaVail et

29

al., 1992). Recent experiments have demonstrated that Lif is also protective during light damage

(Ueki et al., 2008). The Nr -/-, Rd1-/- and Q334ter mutant rhodopsin mouse models of IPD display PR

protection when treated with a single intravitreal injection of Cntf; Lif and Ct-1 were also shown to

be protective in the Q334ter mutant rhodopsin model (LaVail et al., 1998; Song et al., 2003).

Surprisingly, Lif was not protective in the Nr-/- or Rd1-/- retinas; however, the authors suggested that

technical difficulties or the need for multiple doses of the cytokines could account for the negative

results (LaVail et al., 1998). Lif has also been reported to be neuroprotective in the RCS and P23H

transgenic rhodopsin rat models of IPD (Matthew Lavail, personal communication).

Therapeutic strategies have been developed based on the neuroprotective properties of Cntf.

The intravitreal injection of an adeno-associated virus (AAV) expressing the Cntf gene resulted in a

slower rate of PR death in Rds-/-, P23H, and S334ter rat models of IPD (Liang et al., 2001). This

result was independently confirmed by another group examining a mouse Rds-/- model (Bok et al.,

2002). The phototransduction of PRs was then examined with the expectation that retinas treated

with AAV-Cntf would have more functional PRs. Surprisingly, electroretinagrams (ERGs) (which

measure the electrical activity in the retina) had lower amplitudes in AAV-Cntf treated retinas

compared to those treated with AAV-GFP (a negative control), indicating that Cntf negatively

affects PR function (Liang et al., 2001; Bok et al., 2002). However, these studies were both

confounded by the fact that the cyctomegalovirus promoter was used in the AAV construct and this

strong promoter may have led to the production of toxic amounts of Cntf. McGill et al. tested

several doses of Cntf and noted that low doses did not impair ERG amplitudes, yet still delayed the

rate of PR death in light-damaged retinas (McGill et al., 2007). The use of encapsulated cells that

secrete Cntf is another mode of neurotrophin delivery. This system was used to deliver a dose of

Cntf that did not adversely affect the ERG amplitudes of wild-type rabbits (Bush et al., 2004). When

implanted into the phosphodiesterase 6B-deficient rcd1 canine model of IPD, increased PR survival

and an increase in ERG amplitudes were observed (Tao et al., 2002). This technology is now being

tested in humans to determine if encapsulated cells secreting Cntf can slow cell death and preserve

PR function in patients with retinitis pigmentosa (Sieving et al., 2006).

1.6.3.1 Pathways activated by IL-6 cytokines

The observation that IL-6 cytokines are neuroprotective in the retina led investigators to

study which cell types respond to cytokine treatment and which downstream pathways are activated

(Table 1.1). Axokine, an analog of Cntf, was injected into the vitreous of wild-type rats resulting in

30

the phosphorylation of Stat-3 and Erk1,2 and a delayed up-regulation of total Stat-1 and Stat-3

proteins; Stat-1 was only marginally phosphorylated 16 hours after axokine treatment (Peterson et

al., 2000). Activated Stat-3 was predominantly localized to Müller and ganglion cell nuclei and

astrocyte nuclei, while activated Erk1,2 was localized in Müller cell nuclei and astrocyte nuclei; PRs

did not exhibit Stat-3 or Erk1,2 staining (Peterson et al., 2000). These results suggest that Cntf

strongly activates both Stat-3 and Erk1,2 in Müller glia and astrocytes, and that these effects may be

responsible for the PR protection observed with Cntf.

Lif has also been shown to activate the Stat-3 and Erk1,2 pathways (Ueki et al., 2008). In a

recent study, Lif was injected into the vitreous of wild-type mice followed by exposure to bright light

two days later. The goal of the experiment was to determine if Lif protects PRs from apoptosis

caused by light exposure and to identify the pathways up-regulated in response to Lif treatment.

Ueki et al. showed that Lif protected PRs from light-mediated damage; low doses of Lif protected

the structure and function of PRs, but high doses inhibited PR function (Ueki et al., 2008). An

immediate up-regulation of both pStat-3 and pErk1,2 was observed with Lif treatment; however,

only pStat-3 up-regulation was sustained two days post-injection, when the animals were exposed to

the bright light, suggesting that the Stat-3 pathway is responsible for PR protection in Lif-treated

animals (Ueki et al., 2008). Up-regulation of pStat-3 was observed in Müller glia, ganglion cells and

PRs (Ueki et al., 2008). Although the staining was predominantly observed in Müller glia and

ganglion cells, consistent with the cell types that respond to Cntf treatment (Peterson et al., 2000),

the occasional PR exhibited intense pStat-3 staining, suggesting that Lif may have an direct effect on

PRs in addition to an indirect effect through Müller cells (Ueki et al., 2008).

Increased endogenous cytokine signalling can also occur in retinas under stressful conditions.

Preconditioning animals with a 12-48 hour exposure to bright light evokes a protective response

against future light damage in the rat retina (Liu et al., 1998; O'Driscoll et al., 2008). The

31

Table 1.1: Retinal responses to IL-6 cytokine activation

Mode of IL-6 pathway activation

↑pStat-3

pStat-3 localization

↑pStat-1

pStat-1 localization

↑pERK1,2

pERK1,2 localization

Light treatment (low dose)1,2,3

yes3 - yes3 - yes1,2 Müller glia2

Light treatment (high dose)4

Yes - yes - yes -

Rd1-/- retinas4

Yes - no - no -

Vpp+/- retinas4

Yes - no - no -

Axokine (Cntf analog) treatment3

Yes Müller glia, Astrocytes, ganglion cells

marginal - yes Müller glia

Cntf treatment5

Yes Müller glia, ganglion cells

- - - -

Lif treatment6

Yes Müller glia, some PRs, ganglion cells

- - yes Müller glia, ganglion cells

Ct-1 treatment7 Yes Müller glia, ganglion cells,

- - yes -

References in table: 1O'Driscoll, 2008; 2Liu, 1998; 3Peterson, 2000; 4Samardzija, 2006; 5Wang, 2002; 6Ueki, 2008; 7Song, 2003.

32

identification of increased bFGF, Cntf and pErk1,2 in pre-treated retinas suggests a possible

mechanism of neuroprotection (Liu et al., 1998; O'Driscoll et al., 2008). Other studies have

examined the retinal response to toxic levels of light. Bright light results in the up-regulation of

pStat-1, pStat-3, pErk1,2, total Stat-1 and total Stat-3 proteins (Peterson et al., 2000; Samardzija et

al., 2006); Clc, Cntf, Lif and bFGF transcripts were also up-regulated (Samardzija et al., 2006). Given

that treatment with IL-6 cytokines protects PRs from a variety of insults, the endogenous up-

regulation of IL-6 cytokines and their downstream pathways by bright light suggests that these

responses are protective.

The Stat pathway was also noted to be up-regulated in the Rd1-/- and Vpp+/- models of IPD

(Samardzija et al., 2006). As in the light-damaged model, Lif and Clc transcripts were increased, as

was pStat-3 and total Stat-3; however, the levels of pStat-1 and pErk1,2 remained unchanged,

suggesting their up-regulation may be specific to the light damage models (Samardzija et al., 2006).

In an attempt to determine if the activation of Stats plays a role in PR death, a Jak2 inhibitor,

AG490, was injected into retinas of IPD models or mice exposed to bright light and the rate of cell

death and the phosphorylation state of Stat-1 and Stat-3 determined (Samardzija et al., 2006). With

light damage, AG490 dramatically reduced both pStat-1 and pStat-3, and increased PR survival

(Samardzija et al., 2006). This result was somewhat unexpected, given the apparent protective

increase in pStat-3 seen with the administration of IL-6 cytokines, described above; however, the

level of pStat-1, a pro-apoptotic factor, was also reduced, which could explain this result (Samardzija

et al., 2006). Interestingly, the abundance of pStat-3 were unchanged in IPD models treated with

AG490, as was the rate of cell death upon AG490 treatment (Samardzija et al., 2006). Thus, the role

of Stat-3 in PR protection in IPD models still remains uncertain.

In summary, a variety of neurotrophic factors have been shown to protect PRs in light-

damage and IPD models. Basic FGF appears to act directly on PRs, whereas Bdnf, Ngf and IL-6

cytokines seem to act predominantly through other cell types, particularly Müller cells. The up-

regulation of pStat-3 and pERK1,2 in response to IL-6 cytokines suggests that these transcription

factors initiate downstream survival responses, although direct evidence to support this hypothesis is

presently lacking. In chapter three, I will report genetic studies that demonstrate that Lif, Osmr and

C/EBPδ (a downstream target of IL-6 signalling) all make significant contributions to the retinal

responses occurring in IPDs.

33

1.7 Thesis objectives

1) The study of IPD proteins has not has not led to an understanding of why PRs die in IPD retinas.

In order to identify transcripts involved in PR death, I performed a whole-genome analysis

employing microarray and quantitative PCR experiments examining transcript levels in IPD models

compared to wild-type retinas. These data are described in Chapter 2. From these experiments, I

identified several up-regulated transcripts in the retinas of Rds+/- mice that I hypothesized were

participating in a common IL-6 pathway. These results in combination with data reported in the

literature (LaVail et al., 1992; LaVail et al., 1998) led to further investigations into the role of IL-6

signalling in the IPD retina as a protective response.

2) In Chapter 3, I further characterized the role of the putative IL-6 pathway members in protecting

PRs in models of IPD. I determined that most IL-6 pathway members that were up-regulated in the

retinas of Rds+/- mice were also up-regulated in the retinas of Rd1-/- and P347S murine models of

IPD, suggesting a specific IL-6 response in many models of IPD. I also compared the protein levels

and cell localization of key members of the identified putative IL-6 pathway in the retinas of Rds+/-

and wild-type mice. I then employed a genetic analysis to functionally determine whether several

members of the putative IL-6 pathway influence the rate of cell death in IPD models.

34

2. Characterization of the retinal transcriptomes of the Rom1-/- and

Rds+/- murine models of inherited photoreceptor degeneration

The majority of the work presented in this chapter is my own, with the following exceptions:

1) The Montréal Genome Institute performed the microarray hybridizations

2) Laura Pacione performed the majority of the qPCR analysis

35

2.1 Abstract

A central question in inherited photoreceptor (PR) degeneration (IPD) is why mutant

neurons die after functioning for many years, even though the mutation is present from birth. The

survival of mutant neurons suggests that the pathogenic mutation cannot be the direct cause of cell

death. Rather, a cascade of event(s) might be initiated by the mutation that eventually leads to the

death of the cell. To identify molecules that may be involved in IPD, the retinal transcriptomes of

Rom1-/- and Rds+/- mouse models of IPD were characterized using microarray and quantitative PCR

(qPCR) analysis. RNA was harvested from each model and age-matched wild-type controls at a

time-point when >90% of mutant PRs were still present. A triplicate repeat microarray experiment

was then performed. The data was normalized using robust multi-array average (RMA) followed by

significance analysis of microarrays (SAM) to identify significantly differentially expressed genes.

Using a false discovery rate (to control for multiple testing) of 4.7%, 29 transcripts and 145

transcripts were identified as differentially expressed in Rom1-/- and Rds+/- retinas, respectively.

Interestingly, none of the differentially expressed genes were common in both models. Due to the

large number of differentially expressed genes in the Rds+/- model, Ingenuity pathway finding

software was used and identified a putative IL-6 pathway (p<10-50), which included 20 genes of

the 145 originally identified by SAM. This IL-6 pathway has been implicated in PR cell survival and

may represent a protective response to PR cell stress.

2.2 Introduction

Inherited photoreceptor degenerations (IPD) are a genetically and phenotypically

heterogeneous group of disorders causing visual impairment. The individual types of PR

degeneration can be distinguished by their mode of inheritance, pattern of vision loss, and the

defective gene involved (Pacione et al., 2003). In the majority of IPDs, both cone and rod PRs die,

but the degree to which these cell types are affected differs depending on the disorder (Pacione et

al., 2003). For example, Retinitis pigmentosa (RP), the leading monogenic cause of blindness, is

characterized by an initial loss of rods, resulting in night blindness and peripheral vision loss,

followed by the loss of cones and total vision loss (Hartong et al., 2006). Since the initial discovery in

1990 that a point mutation in the rhodopsin gene can cause autosomal dominant retinitis

36

pigmentosa (adRP) (Dryja et al., 1990), the rate of IPD gene identification has progressed rapidly;

currently, there are 193 different loci, including 142 cloned genes implicated in IPD (SP Daiger,

2008).

Genes implicated in IPD encode proteins with great functional diversity. Mutations in genes

involved in the following processes lead to PR cell death: the phototransduction cascade, vitamin A

metabolism, structural or cytoskeletal, signalling, cell-cell interaction, synaptic interaction, RNA

splicing factors, trafficking of intracellular proteins, maintenance of cilia/ciliated cells, pH regulation,

and phagocytosis (Pacione et al., 2003; Hartong et al., 2006). Despite the rapidly increasing

knowledge of the genes and proteins whose functions are disrupted in IPD, the fundamental

mechanisms underlying PR death remains to be elucidated.

Inherited PR degenerations share some common features. For example, caspase-dependent

apoptosis is thought to play a role in mediating cell death in many IPD models; however, the events

that trigger the apoptotic cascade in mutant PRs is poorly understood (Chang et al., 1993). Secondly,

IPDs share the same kinetics of cell death (Clarke et al., 2001; Clarke and Lumsden, 2005). In all 11

examples of IPD that were examined, the population of mutant PRs exhibit exponential cell death

kinetics (Clarke et al., 2000a). This observation was later refined to stretched exponential kinetics,

which is characterized by four important features: (i) the risk of mutant cell death is constant, (ii)

each mutant PR has the same risk of death (other factors being equal) (iii) mutant cell death occurs

randomly, and (iv) the death of each neuron is independent of other neurons (Clarke and Lumsden,

2005). The common features shared by IPDs, despite the great functional diversity of IPD-

associated proteins, has at least one important implication for understanding mechanisms by which

these mutations lead to PR death: it is unlikely that there are 192 private gene-specific pre-apoptotic

pathways (Pacione et al., 2003). Rather, there is likely to be a one or a few pathways or gateways that

converge to trigger cell death. Additionally, there may also be a limited number of pathways that

function to resist cell death in IPD. The identification of pathogenic and protective molecules will

provide substantial insight into the pathogenesis of IPDs and may provide opportunities for novel

therapies.

Microarray technology has been used to identify global gene expression differences between

IPD and wild-type retinas (Swaroop and Zack, 2002). Initial studies focused on comparing the

transcriptomes of wild-type retinas to IPD models at a time-point when most PRs had already

degenerated (Kennan et al., 2002; Hackam et al., 2004b). This approach was useful for the

identification of PR-specific genes but little insight into the mechanism(s) leading to cell death in

37

IPD models was gained. In a limited number of studies, microarray experiments have been

performed comparing the retinal transcriptomes of wild-type and IPD models at a time-point when

few mutant PRs have died; the presence of mutant PRs are important for the identification of any

differentially expressed genes/pathways that may be playing a causal role in PR death (Hackam et al.,

2004b; Rattner and Nathans, 2005). For example, an analysis of the Rd1-/- mouse retina, a well-

studied model of IPD, which has a deficiency in the cGMP phosphodiesterase β-subunit (βPde)

(Chang et al., 2002), revealed the up-regulation of several genes implicated in the Wnt pathway

(Jones et al., 2000; Hackam et al., 2004b). Since Wnt signalling has been shown to inhibit apoptosis

in vitro (Chen et al., 2001), this putatively protective pathway may inhibit rod and/or cone PR

apoptosis in the Rd1-/- retina. Transcripts whose protein products are known to mitigate oxidative

damage were also up-regulated, suggesting oxidative stress may also be playing a role in Rd1--/-

retinas (Hackam et al., 2004b).

The lack of functional βPde in Rd1-/- retinas causes a dramatic increase in cytoplasmic cGMP

concentration, resulting in the permanent opening of cGMP-gated channels and the excessive entry

of calcium (Farber and Lolley, 1974). It has been suggested that this increase in calcium is one of the

primary events leading to apoptosis (Chang et al., 1993). Microarray analysis characterizing the Rd1-/-

transcriptome throughout the degeneration process revealed that genes involved in calcium

homeostasis, such as calbindin, were up-regulated (Rohrer et al., 2004; Hartmann and Konnerth,

2005), suggesting that Rd1-/- retinas are responding to high calcium levels.

Calpain activity has been proposed as a mechanism through which calcium toxicity could

lead to cell death (Sharma and Rohrer, 2004; Paquet-Durand et al., 2007). Excessively active calpains

are thought to break down molecules in the cytoskeleton such as actin-associated proteins,

microtubule-associated proteins, and cell adhesion molecules, which can lead to the degradation of

the cytoskeleton and plasma membrane (Paquet-Durand et al., 2007). Calpains have been shown to

be active in Rd1-/- retinas (Doonan et al., 2005) and microarray analysis revealed that calpastatin, an

inhibitor of calpain, was down-regulated 20 fold in Rd1-/- retinas (Paquet-Durand et al., 2006).

Although the calpain transcript itself was not up-regulated, the down-regulation of calpastatin

suggests that reduced inhibition may be the mechanism through which calpain protein activity is

increased in Rd1-/- retinas (Paquet-Durand et al., 2006). In vivo experiments have demonstrated that

blocking calpain activity prevented PR death in Rd1-/- retinas, suggesting calpain activity may be

involved in mediating PR cell death in this model (Sanges et al., 2006).

38

Photoreceptor injury or disease often leads to the activation of Müller glia, the dominant glial

cell type in the retina (Bringmann and Reichenbach, 2001; Garcia and Vecino, 2003). A recent

microarray study examining the transcriptome of a light-mediated PR death model and the prCAD-/-

model of IPD revealed that 8 of the 12 transcripts that were highly induced by light damage were

also up-regulated in the genetic model of IPD; in situ hybridization analysis of five of the 8

commonly differentially expressed genes revealed a staining pattern consistent with Müller cell

localization, suggesting a prominent Müller cell genomic response to PR damage (Rattner and

Nathans, 2005). One such response involved Endothelin-2 (Edn2), the most highly up-regulated gene

by microarray and northern analysis in both light-mediated and the prCAD-/- model of IPD (Rattner

and Nathans, 2005). Localization studies revealed that End2 mRNA was expressed in the PR layer

(Rattner and Nathans, 2005). The Edn2 receptor, EdnrB, was also found to be up-regulated in the

light-damage model and was localized to Müller glia (Rattner and Nathans, 2005). The role of Edn2

in IPD models remains to be elucidated, but PR-derived Edn2 may function as a stress signal and

activate Müller glia by binding to the EdnrB in light damage (Rattner and Nathans, 2005). Studies

examining other endothelin family members in the central nervous system have suggested that

endothelins promote active astrocytosis, a process which occurs during neuroinflammation and is

associated with cell death (Barbeito et al., 2004).

Microarrays can be an effective tool to identify differentially expressed genes in IPD models,

which may be playing a role in cell death. In this study, we identified differentially expressed

transcripts in two models of IPD using both microarray and qPCR analysis.

2.3 Materials and Methods

2.3.1 Animals used in this study

Mouse strains used in this study include Rom1-/-, in a mixed 129/CD1 (outbred) background

(Clarke et al., 2000b) and Rds+/-, which is in the C3H background (Jackson Laboratory, Bar Harbor,

Maine, USA) (Sanyal and Jansen, 1981). Animals were raised under conditions of twelve-hour cycles

of light and dark and treated in accordance with the policies of the Hospital for Sick Children.

2.3.2 Preparation of retinal RNA for the microarray experiment

Retinal tissue was obtained from male and female Rom1-/- and Rds+/- animals at three months,

and seven weeks of age, respectively, along with age-matched wild-type controls of the appropriate

39

background strain. Animals were sacrificed by cervical dislocation and the retinas were dissected

away from the eyecup and RPE in ice-cold phosphate buffered saline (PBS) (137mM NaCl, 19mM

Na2HPO4, 3mM KCl, 2mM KH2PO4 [pH7]). Retinas from single animals were homogenized in a

dounce homogenizer in 1mL of filter-sterilized GIT buffer (4M guanidine isothiocyanate, 25mM

sodium acetate [pH6] with 120mM β-mercaptoethanol added immediately before use). To prevent

ribonuclease (RNase) contamination, all solutions used for RNA isolation were made with

diethylpyrocarbonate (DEPC)-treated distilled-deionized water (ddH2O) (0.1% (v/v) DEPC in

ddH2O dissolved by stirring overnight at 370C, followed by autoclaving). Retinal RNA was isolated

by cesium chloride density gradient centrifugation. To shear genomic DNA, retinal homogenates

were passed 5 times through a 10cc syringe with a 23-gauge needle. GIT buffer was added to

3.33mL final volume, and this was layered over 1.65mL of 5.7M CsCl, 25mM sodium acetate [pH6]

in an RNase-free SW55 5mL centrifuge tube. Tubes were balanced and centrifuged using an SW55

rotor at 116,000xg for 16 hours at 20–250C. The supernatant was removed, leaving approximately

1cm of liquid behind, and tubes were cut just above the liquid level to facilitate washing. The

remaining buffer was removed, and RNA pellets were rinsed 3 times in 70% (v/v) ethanol and

resuspended in 300µL DEPC-treated H2O. RNA was precipitated by the addition of 30µL of 3M

sodium acetate [pH5] and 750µL of absolute ethanol, followed by incubation overnight at -200C.

Tubes were then centrifuged at 21,000xg for 30 minutes at 4°C, washed in 70% ethanol, re-

centriguged and resuspended in DEPC-treated water. To examine the quality of RNA, the

absorbance of 1µL (diluted 1:49) was measured spectrophotometrically at 260 and 280nm.

Additionally, 1µL of sample was run on an agarose gel to examine RNA quality on the basis of the

18S to 28S ribosomal RNA ratio (2:3 ideally). RNA isolated from three animals of the same

genotype were pooled and was bound to an RNeasy® column (Qiagen, Mississauga, Ontario)

according to the manufacturer’s protocol for further purification. RNA was extracted from the

column with two 50µL elutions of DEPC-treated H2O; 2µL was used for gel and

spectrophotometric analysis as previously described. RNA was precipitated by the addition of 10µL

of 3M sodium acetate [pH5] and 220µL of absolute ethanol, followed by incubation overnight at -

200C. RNA was pelleted by centrifugation at 21,000xg for 10min at 40C, and the pellet was washed

with 1mL of 75% ethanol, and centrifuged for a further 5min at 7,500xg at 40C. The remaining

ethanol was removed and the pellet was air-dried for 5–10min and resuspended in 10µL DEPC-

treated ddH2O (approximate yield of 15µg).

40

2.3.3 Preparation of retinal RNA for cDNA synthesis

Retinas were harvested as previously described. Retinas from single animals were

homogenized in a dounce homogenizer with 1mL of Trizol reagent (Invitrogen, Burlington,

Ontario), transferred to a microfuge tube and RNA was isolated according to the manufacturer’s

instructions. Briefly, 0.2 ml of chloroform was added to each tube containing 1 ml of the Trizol

reagent, mixed well, then centrifuged at 12,000xg for 15 minutes at 40C. The top clear phase was

isolated and 0.5ml of isopropanol was added to precipitate the RNA followed by centrifugation at

12,000xg for 15 minutes at 40C. The pellet was then washed in 70% ethanol, re-centrifuged at

7,500xg for 5 minutes at 40C and resusended in 12µl of DEPC-treated water. The RNA was then

DNase treated to remove any genomic DNA contamination. A solution of 1mM dithiothreitol, 1x

first strand buffer (Invitrogen, Burlington, Ontario), 1ul RNase inhibitor (Roche, Laval, Québec)

and DEPC treated water to a volume of 49µl followed by 1µl of RNase free DNaseI (Invitrogen,

Burlington, Ontario). Samples were incubated at 37°C for 45 minutes then repurified using the

Trizol reagent (Invitrogen, Burlington, Ontario) employing the same protocol as described above.

The typical yield of RNA from 2 retinas is 4-7µg.

2.3.4 Microarray hybridization experiments

Isolated RNA was shipped on dry ice to Montreal where the staff at the Montreal Genome

Centre (Montreal, Québec) performed the microarray experiments. Each sample represented retinas

pooled from 3 individual mice of the same genotype. Briefly, labeled cRNA, produced from mutant

and wt RNA samples, was hybridized, using three separate samples per genotype, to the Affymetrix

Murine Genome U74A array (Affymetrix, Santa Clara, California) representing approximately 12,000

genes and ESTs from the UniGene database (build 74) (Rom1-/-model), or to the Affymetrix Murine

Expression Set 430A array (Affymetrix, Santa Clara, California) representing approximately 22,500

genes and ESTs from the UniGene database (build 107) (Rds+/- model).

2.3.5 Microarray data analysis

To analyze the microarray data from the Rom1-/- model, the raw data was first normalized

using the robust multi-array average (RMA), a normalization procedure within the ArrayAssist®

(Statagene, La Jolla, California) software package (Irizarry et al., 2003). Normalized and log2

transformed data was then subject to significance analysis of microarrays (SAM) that employs a user-

41

defined input (delta) to select significantly differentially expressed genes; the software calculates the

false discovery rate for each delta employed (Tusher et al., 2001). The Rds+/- microarray data was also

normalized using RMA followed by SAM as described above. In addition, microarray data from the

Rds+/-model was also analyzed using Affymetrix Microarray Suite 5.0 (MAS5) software (Affymetrix,

Santa Clara, California). Since MAS5 can only perform pair-wise analysis, significantly differentially

expressed genes in all 9 possible pair-wise comparisons were identified. Genes identified by the all

nine pair-wise comparisons and SAM were selected for further analysis. Concordance correlation

coefficients were calculated using software available at:

www.niwa.cri.nz/services/free/statistical/concordance. Ingenuity pathway analysis was performed

using the Ingenuity software package (Stratagene, La Jolla, CA).

2.3.6 cDNA synthesis

Samples were reverse transcribed directly following DNase treatment. 2µL of 50ng/µL

random primers (Invitrogen, Burlington, Ontario) were added to 1µg RNA in a total volume of

13µl. Tubes were heated to 700C for 10min, quickly chilled on ice and the contents were collected by

brief centrifugation. The following was then added to each tube: 4µl of 5X first strand buffer

(Invitrogen, Burlington, Ontario), 1µl 4mM dNTPs, 1µl of 0.1M dithiothreitol (Invitrogen,

Burlington, Ontario) and 1µL RNase inhibitor (Roche, Laval, Québec). After mixing and incubation

at 420C for 2min, 1µL of Superscript® II reverse transcriptase (Invitrogen, Burlington, Ontario) was

then added to each tube, with the exception of the reverse transcriptase-free (-RT) control. One -

RT control reaction was performed for each sample. All samples were incubated for 50min at 420C,

followed by heating at 700C for 15min to inactivate the reverse transcriptase prior to storage at -

200C.

2.3.7 Quantitative real-time PCR

2.3.7.1 Design and evaluation of primers

Whenever possible, primers were designed to span genomic intron/exon boundaries to

avoid quantitative real-time PCR (qPCR) amplification of any contaminating genomic DNA.

Primer3 software (available at http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) was

used to design primers against a rodent mispriming library according to the following specifications:

70–120bp product size, 20–80% primer GC content, 57–630C melting temperature (600C optimally),

42

maximum 3 base pairs of primer self-complementarity and no primer 3’ self-complementarity (to

minimize primer-dimer product formation) (Table 1.1).

Primer pairs were then evaluated for qPCR suitability. All qPCR reactions were carried out

in ABI Prism™ 96-well optical reaction plates (Applied Biosystems, Foster City, California), using

plugged pipette tips. cDNA, typically from a C3H animal, along with a corresponding -RT control

was diluted 1:50 in sterile distilled deionized water (ddH2O). For each primer set to be evaluated,

8µL of diluted stock cDNA, -RT control (to control for contaminating genomic DNA) or ddH2O

(to control for primer-dimer product formation) was added to individual wells. The following was

then added to each well as part of a master mix: 10µL SYBR® green PCR master mix (Applied

Biosystems, Foster City, California), 0.6µL of a 50ng/µL solution of each primer and 0.8µL

sddH2O. Wells were sealed using optical caps (Applied Biosystems, Foster City, California) and

contents were collected in the bottom of the wells by brief centrifugation. qPCR was carried out

using an ABI Prism® 7900HT sequence detection system (Applied Biosystems, Foster City,

California) at 500C for 2min, 950C for 10min followed by 45 cycles of 950C for 15sec and 590C for

1min. In order to construct dissociation curves, samples were further incubated at 950C for 15sec

and then 600C for 15sec with a slow (2%) ramp speed up to 950C.

Quantitative PCR analysis was carried out using Sequence Detector System 2.2.1 software

(Applied Biosystems, Foster City, California) with threshold 0.02 and cycles 3–15 typically used to

calculate the baseline fluorescence. A specific product was identified based on single, narrow peak

for the PCR-product dissociation curve, with a melting temperature >750C for the reaction

containing cDNA and no peak for the -RT or sddH2O controls.

43

Table 2.1: List of qPCR primers All primers have an annealing temperature of 58–630C and produce amplicons in the range of 75–120bp.

Gene Name Accession

Number

Sequence (5' to 3') Crosses Intron/ Exon Boundary

Housekeeping Genes

glyceraldehyde-3-phosphate dehydrogenase (Gapdh)

NM_008084 atcttcttgtgcagtgccagc tgaccaggcgcccaat

No

succinate dehydrogenase complex, subunit A, flavoprotein (Fp) (Sdha)

XM_127445 tttgttcagttccaccccaca cccccctctccacgacac

Yes

Genes interrogated in the Rom1-/- model of IPD by qPCR

transglutaminase 3, E polypeptide (Tgm3) NM_009374 tggaaggactcagccacaa No ttgtctgccttcaggtatctctc transmembrane protein 14C (Tmem14c) NM_025387 tcctagctacatctgggacctt Yes ttccaactttggcaaccatc mitochondrial ribosomal protein L16 (Mrpl16)

NM_017840 ccatccctgaaaggtccaa ggagataaccaccgccaag

Yes

peptidylglycine α-amidating monooxygenase (Pam)

U79523 agttggaggggaaactggaa tgtaaggacacaccggaaca

Yes

protein tyrosine phosphatase receptor type A (Ptpra)

M36033 gtggacaagctggaagagga ggataggacaagcagggaga

Yes

solute carrier family 17 (sodium phosphate), member 1 (Slc17a1)

NM_009198 agcaccgtcattttcctgac tctgggagcaatatccaagg

Yes

Genes interrogated in the Rds+/- microarray experiment by qPCR

transferrin (Trf) AF440692.1 Yes

gctgacagggaccaatatgaa gagccacaacagcatgagaa

signal transducer and activator of transcription 3 (Stat3)

AI325183 gaggggtcactttcacttgg gctgcttggtgtatggctct

Yes

leucine-rich repeat-containing 2 protein (Lrrc2)

AJ428068.1 ctgagtgacctgccacaaga ctgacaaccagcaaggtgag

Yes

thrombospondin 1 (Thbs1) AV026492 agaccaaagcctgcaagaaa Yes

44

tctctgcactcctcctccac glial fibrillary acidic protein (Gfap) BB183081 Yes

ctggaggcagagaacaacct ctccagcgattcaacctttc

α-2-macroglobulin (α2m) BB185854 ggttcctgaacgtggaaaga Yes cacctgttggacaaagcaaa procollagen C-proteinase enhancer protein (Pcolce)

BB250811 gaaattttgcggagacaagg ggtcctgtaggaggctgagaa

Yes

CCAAT/enhancer binding protein (Cebpδ) BB831146 gttccgcctttgctatgtct

ttccctccttcctgtttgtg No

FBJ osteosarcoma oncogene B (Fosb) BG076079 Yes

ccataaaagtttccccagtcc tgtggctcctttctttggtt

transthyretin (Ttr) BG141874 Yes

ggacaccaaatcgtactggaa agagtcgttggctgtgaaaa

MHC class I H-2D1 (H-2D1) M69068.1 catggtgatcgttgctgttc Yes gcatagtcccctccttttcc complement component 1, q subcomponent, α polypeptide (C1qa)

NM_007572.1 ccgggtctcaaaggagagag ccccacattgccaggttt

Yes

endothelin-2 (Edn2) NM_007902.1 ctggcttgacaaggaatgtg Yes gccgtagggagctgtctgt serine/cysteine proteinases inhibitor, clade G, member 1 (SerpinG1)

NM_009776.1 tcgtccttctcaatgctgtct aaaggcgccatcatcttttt

Yes

dopachrome tautomerase (Dct) NM_010024.1 gtttgacagccctcccttct Yes agtccagtgttccgtctgct Keratin complex 1 acidic gene 12 (Krt1-12) NM_010661.1 gcctggagatggagattgag Yes actgatggtgtctgcggttt oncostatin receptor (Osmr) NM_011019.1 attcgcatcacaaccaacaa Yes tccttgacctcttcgtgtcc

45

2.7.2 Analysis of gene expression

Gene expression was quantified by the standard curve method (ABI, 2001). Typically, the

expression of two genes was interrogated in a single 96-well qPCR reaction plate. To compare the

expression level of a gene between samples, primers were used in reactions as previously described:

four standard curve dilutions performed in triplicate, six separate cDNA samples from three mutant

and three wild-type individuals also run in triplicate. A standard curve was performed for each

reaction as recommended for accurate quantification (Bustin, 2002). Additionally, three ddH2O

controls and one –RT control for each sample is also present on each plate. Typically, 100µl of a

1:50 cDNA sample dilution was prepared and the amount of cDNA starting material was

normalized using either Gapdh or Sdha housekeeping genes. Samples were kept at 40C for up to 5

days before they are re-normalized so the expression levels of multiple query genes could be

determined. Normally, 8µl of a 1:50 cDNA dilution was added to each well.

2.7.3 Statistical analysis

Analysis was carried out, and dissociation curves examined, using Sequence Detector System

2.1 software (Applied Biosystems, Foster City, California), as previously described. The standard

deviation for triplicate repeat wells was typically <10%. To normalize for the starting amount of

cDNA, the mean raw control gene expression was standardized by dividing the value for each

sample by the lowest mean i.e. the standardized control gene expression of the animal with the

lowest expression was defined at ‘1’. A correction value was calculated for each animal by geometric

averaging of the standardized expression values of two internal control genes (Vandesompele et al.,

2002). To normalize gene expression for each animal, the mean raw expression value of the query

gene was divided by the correction value calculated for that animal. A two-tailed Student’s t-test was

used to determine if differences in mutant and wt expression were statistically significant. Statistically

significant (p<0.05) differences in gene expression were expressed as fold-differences by dividing the

mean mutant expression by the mean wt expression. For fold-differences <1 the inverse of the

fold-difference was taken and assigned a negative value to represent down-regulation.

46

2.4 Results

2.4.1 Characterization of the Rom1-/- retinal transcriptome

The retinal transcriptome of Rom1-/- was compared to Rom1+/+ mice in order to identify

differentially expressed genes that may be pathogenic, protective or bystanders to PR death. A

triplicate repeat microarray experiment was performed using pooled retinal RNA from three mice

per sample; the RNA was harvested from three month-old mutant and wild-type animals at a time-

point when 93% of PRs were still present in mutant retinas (Clarke et al., 2000a). This early time-

point was chosen because any secondary effects due to cell death in the mutant retina would be

minimal. RNA samples were then labeled and hybridized onto separate Affymetrix U74A

microarrays, scanned, and gene expression values were assigned based on hybridization intensity.

In order to compare gene expression values between Rom1-/-and wild-type retinas, raw

expression data from all six microarray experiments (three Rom1-/- and three Rom1+/+) was

normalized. The robust multi-array average (RMA) normalization method was used, which employs

background correction and quantile normalization to estimate gene expression values (Irizarry et al.,

2003). After normalization, a scatter plot was generated comparing mean log2 Rom1-/- and Rom1+/+

hybridization signals with the log2ratio of Rom1-/- to Rom1+/+ (Figure 2.1). This type of plot is also

called an M versus A plot and is a convenient way to visualize the expression values from multiple

arrays representing Rom1-/- and wild-type samples; the majority of data points are predicted to be

centered around a y-axis value of zero, indicating that the majority of genes are not differentially

expressed (Yuen et al., 2002). As expected, most data points were centered around this value,

suggesting most genes have similar expression values in both Rom1-/- and Rom1+/+ retinas. Data

points representing Rom1 had the lowest log2ratio, indicating, as expected, that it was the most

down-regulated gene in mutant retinas (Figure 2.1, arrows).

To identify significantly differentially expressed genes, significance analysis of microarrays

(SAM) was performed using the RMA normalized microarray data. The SAM algorithm incorporates

the variability of replicate arrays to calculate the likelihood that observed expression level changes

are real differences rather than artifacts (Tusher et al., 2001). The false discovery rate (FDR) is also

incorporated into the SAM algorithm to control for multiple testing. The problem of multiple

testing arises because the differentially expressed status of each gene on the array is determined by a

separate test; using the standard p<0.05 level of significance would yield

47

Rom1

Rom

1-/- lo

g 2-R

om1+

/+lo

g 2

2Rom1-/-

log2+ Rom1+/+

log2

Rom1 Rom1

Rom

1-/- lo

g 2-R

om1+

/+lo

g 2R

om1-

/- log 2

-Rom

1+/+

log 2

2Rom1-/-

log2+ Rom1+/+

log2Rom1-/-

log2+ Rom1+/+

log2

Rom1

Figure 2.1: Summary of the Rom1-/- normalized microarray data M versus A scatter plot representing gene expression in three month old wild-type compared to the Rom1-/- murine model of PR degeneration. Each data point represents mean hybridization signals of Rom1-/- and wild type compared to the mean ratio of Rom1-/- to Rom1+/+ hybridization signals from three independent microarray experiments for each gene. Genes represented by data points centered around a y-axis value of zero are not differentially expressed, while data points with a mean ratio greater than or less than zero represent genes that are up-regulated or down-regulated, respectively. Arrows point to data points representing Rom1.

48

approximately 5% false positives. Microarray analysis using a platform capable of detecting 20 000

genes would identify 1000 false positives amongst the differentially expressed genes if the p<0.05

level of significance were employed. The FDR reduces the number of false positives; it is calculated

after comparing the expression ratios of arrays hybridized with mutant RNA compared to wild-type

with the ratios of randomly permutated mutant and wild-type groups from the same experiment.

These permutations allow for the estimation of the number of genes that would be identified by

chance given a user-defined threshold of significant gene expression; the choice of FDR is a balance

between wanting to identify as many differentially expressed genes as possible, while minimizing the

number of predicted false positives within the group (Tusher et al., 2001).

The data-set from the microarray comparing Rom1-/- to wild-type retinas was analyzed using

SAM at a user-defined threshold such that 4.7% of the genes identified as differentially expressed

were predicted to be false positives. Twenty-nine genes were identified as significantly up-regulated

(Appendix 1), while 14 genes were identified as significantly down-regulated (Appendix 1). As

expected, Rom1 was identified as the most significantly down-regulated gene in Rom1-/- retinas

(Appendix 1). Most of the other identified genes displayed less than a two-fold difference in

expression levels. In fact, of the 29 up-regulated genes, only six were greater than 2-fold and only

one, an expressed sequence tag (EST), was greater than 3-fold up-regulated (Appendix 1). Similarly,

only one gene was down-regulated by 2-fold (Appendix 1). In addition to an estimated fold

difference, a SAM score was also assigned to each differentially expressed transcript on the basis of

the change in gene expression between microarrays hybridized with Rom1-/- compared to wild-type

retinal RNA, relative to the standard deviation of repeated measurements of random permutations

of the array data (Appendix 1). Up-regulated genes with a high score and down-regulated genes with

a low score have a greater probability of being truly differentially expressed, as opposed to being

identified by chance (Tusher et al., 2001).

To validate the microarray results, independent retinal RNA samples from Rom1-/- and

Rom1+/+ animals were reverse transcribed and used as a template in quantitative real-time polymerase

chain reactions (qPCR). All qPCR experiments were performed in triplicate; retinal cDNA samples

from three separate 3-month old Rom1-/- animals were compared to cDNA samples obtained from

three separate age-matched wild-type animals. Using independent RNA samples to validate

microarray experiments is advantageous since the biological reproducibility of each differentially

expressed transcript can be directly examined. Known genes with a high fold difference and SAM

score relative to the group were selected for qPCR analysis. Of the six genes examined, three

49

transcripts were confirmed to be differentially expressed by qPCR, while two were not confirmed,

including Tmem14c, which was identified as up-regulated by SAM and down-regulated by qPCR

(Table 2.2). The lack of correlation between the microarray and qPCR results is shown on a scatter

plot comparing the log2 qPCR fold difference to log2 microarray fold-difference (Figure 2.2).

Although the slope of the trend line was positive, indicating a positive correlation between qPCR

and microarray results, the large degree of scatter suggests this microarray analysis provides only a

rough estimate of transcriptional changes that occur in Rom1-/- compared to wild-type retinas (Figure

2.2).

To quantify the degree of correlation between the qPCR and microarray results, the

concordance correlation coefficient (CCC) was employed. The CCC is an established method to test

the validity of microarray results after qPCR analysis has been performed on a subset of genes

identified as differentially expressed; a CCC of 1 indicates perfect concordance between the

microarray and qPCR fold differences, while a CCC of -1 indicates perfect reversed agreement

(Miron et al., 2006). A comparison of microarray and qPCR fold differences for the Rom1-/-

experiment yielded a CCC of 0.15 (Figure 2.2), confirming the lack of strong correlation between the

microarray and qPCR results.

In summary, the Rom1-/- transcriptome was characterized using Affymetrix oligonucleotide

microarrays, followed by RMA normalization and SAM to identify significantly differentially

expressed genes. A sample of differentially expressed genes identified by microarray analysis was

selected for confirmation by qPCR. Only three of the six selected genes were confirmed to be

differentially expressed. The 50% false positive rate is almost 10-fold higher than the predicted 4.7%

false discovery rate predicted by the SAM, resulting in a low concordance correlation coefficient of

0.15. Despite the high false positive rate, the microarray analysis suggests that the transcriptome of

the Rom1-/- retinas are minimally perturbed; 84% of genes had less than a 2-fold difference in

transcript levels between Rom1-/- and Rom1+/+ retinas (excluding Rom1 itself). The most differentially

expressed transcript was an EST (AI037577), which was up-regulated by 3.5 fold. The Rom1-/- model

of IPD displays a comparatively slow rate of PR degeneration compared to other models (Clarke et

al., 2000a). The transcriptome of a faster model of IPD, the Rds+/- mouse (Sanyal and Jansen, 1981),

was also examined to determine if a faster model of IPD displays greater differences in gene

expression when compared to Rom1-/-.

50

Figure 2.2: Poor concordance between qPCR and microarray analyses in the Rom1-/- model of IPD Scatter plot comparing fold differences estimated by SAM analysis and qPCR in three-month old Rom1-/- and wild-type retinas. Each data point represents the mean log2fold-change following a triplicate repeat microarray experiment followed by SAM (x-axis) compared to the mean log2fold change following a triplicate repeat qPCR analysis (y-axis). The positive trend-line indicates a positive correlation between the microarray and qPCR analyses; the concordance correlation coefficient of 0.15 indicates that the correlation between qPCR and microarray, while positive, is poor.

51

Table 2.2: Comparison of transcript fold-differences as determined by microarray and qPCR analyses in Rom1-/- and wild-type retinas.

Gene Name

Microarray fold

difference

qPCR fold

difference p value Mitochondrial ribosomal protein L16 -1.8 -1.3 0.012

Transmembrane protein 14C 2.7 -2.6 0.00021

Protein tyrosine phosphatase, receptor type, A 2.2 1.4 0.010

Transglutaminase 3, E polypeptide 2.5 3.2 4.0x10-5

Solute carrier family 17 (sodium phosphate), member 1

1.7 1 0.33

Peptidylglycine α -amidating monooxygenase -1.65 1 0.22

52

2.4.2 Characterization of the Rds+/- retinal transcriptome

The retinal transcriptome of the Rds+/- model of IPD was compared to Rds+/+ retinas to

identify differentially expressed genes that may be pathogenic, protective or bystanders to PR death.

A triplicate repeat microarray experiment was performed using pooled retinal RNA from 3 mice per

sample; the RNA was harvested from 7 week-old wild-type and mutant animals, a time-point when

90% of mutant PRs were still present (Clarke et al., 2000a). The isolated RNA was labeled, and each

sample was hybridized onto separate Affymetrix 430A microarrays. The arrays were then scanned,

and each gene on each array was assigned an expression value based on hybridization intensity. This

microarray experiment differs from the previously described Rom1-/- experiment in two ways: the

Rds+/- model of IPD displays faster cell death kinetics compared to Rom1-/- and the updated

Affymetrix 430A microarray platform was used compared to the U74A array.

Significantly differentially expressed genes were identified using two independent data

analysis tools. First, data analysis software provided by Affymetrix (microarray analysis suite 5 or

MAS5) was used to analyze the raw microarray data and identify differentially expressed genes. A

limitation of MAS5 is its ability to identify differences between only two individual microarrays at a

time. Since three microarrays were hybridized with Rds+/+ RNA while another three were hybridized

with Rds+/- RNA, all 9 possible pair-wise comparisons were performed using MAS5; 18 genes were

identified as differentially expressed in all comparisons (Appendix 2). The second method of data

analysis involved RMA normalization followed by SAM. An M versus A scatter plot was generated

comparing RMA normalized mean log2 Rds+/- and Rds+/+ hybridization signals with RMA normalized

log2ratio of Rds+/- to Rds+/+; data points centered around a y-axis value of zero exhibit the same level

of expression in Rds+/- and Rds+/+ retinas (Figure 2.3). While most data points were centered around

this value, several putatively differentially expressed genes were identified with log2ratios above and

below this value (Figure 2.3). The SAM algorithm was performed with the false discovery rate

adjusted to 4.7%; 128 transcripts were identified as up-regulated (Appendix 2), while 17 transcripts

were identified as down-regulated in Rds+/- retinas (Appendix 2). Significantly, all 18 genes identified

by the 9 pair-wise comparisons were also identified by SAM.

To independently validate the microarray results, RNA samples extracted from mouse

retinas not used in the microarray experiments were reverse transcribed and tested by qPCR.

53

Figure 2.3: Summary of the Rds-+/- normalized microarray data M versus A scatter plot representing gene expression in 7-week old wild-type compared to the Rds+/-

murine model of PR degeneration. Each data point represents mean hybridization signals of Rds+/- and wild type compared to the mean ratio of Rds+/- to Rds+/+ hybridization signals from three independent microarray experiments for each gene. Genes represented by data points centered around a y-axis value of zero are not differentially expressed, while data points with a mean ratio greater than or less than zero represent genes that are up-regulated or down-regulated, respectively.

54

Genes identified by both the 9 pair-wise comparisons and SAM were selected for further analysis. In

addition to these 18 genes, Ahrr and Socs-3 were also included in the qPCR analysis since Ahrr had a

high SAM score and Socs-3 is known to be regulated by Stat-3 (Heinrich et al., 2003), one of the 18

genes identified by the 9 pair-wise comparison and SAM. All qPCR experiments were performed

examining transcripts levels in three independent 7 week-old Rds+/- retinas compared to three

independent 7 week-old Rds+/+ retinas. A moderate degree of concordance was observed between

the qPCR and microarray results. Of the 20 examined genes, the differentially expressed status of 12

transcripts was confirmed by qPCR (Table 2.3). To directly compare the microarray and qPCR

results, a scatter plot was generated comparing the log2 qPCR fold differences to the log2 microarray

fold differences (Figure 2.4). The positive trend line indicates that the microarray and qPCR data are

positively correlated and CCC analysis revealed a value of 0.49, although this value may be slightly

inflated due to a single data point (Figure 2.4, arrow), representing a highly differentially expressed

gene identified by both microarray and qPCR. Microarray analysis tends to underestimate the actual

fold difference as determined by qPCR (Yuen et al., 2002; Chowers et al., 2003). Consistent with this

tendency, the fold differences determined by qPCR were higher than the predicted microarray fold

difference in 11 out of 12 examples.

Due to the large number of significantly differentially expressed genes, the Ingenuity

pathway analysis software was used to identify any relationships between the identified genes. The

Ingenuity pathway finding algorithm functions by searching for relationships within a user-defined

set of molecules- in this case, the differentially expressed genes identified by SAM- using a database

of known interactors. Networks that maximize the interconnectedness of the user-defined molecules

of interest with all the molecules in the ingenuity database are then created. Additional molecules

from the Ingenuity database are used to specifically connect two or more smaller networks by

merging them into a larger one. The 145 significantly differentially expressed genes identified by

SAM were analyzed by the Ingenuity pathway analysis algorithm and 6 networks were identified

among the differentially expressed genes that function in a variety of different processes, including:

cell and immune signalling (Appendix 2.5); molecular transport, cancer and cell death (Appendix

2.6); immune and injury response (Appendix 2.7); intercellular signalling and interaction (Appendix

2.8); immune and inflammatory disease (Appendix 2.9), and cellular growth proliferation (Appendix

2.10).

55

Table 2.3: Comparison of transcript fold-differences as determined by microarray and qPCR analyses in Rds+/- and wild-type retinas. Genes represented on the microarray with multiple probe sets were assigned multiple fold differences.

Gene symbol Microarray

Fold Difference

qPCR Fold Difference

p-value (qPCR)

Endothelin-2 8.34 32.00 0.00056 Oncostatin M receptor 2.30

1.25 2.30 3.0x10-5

Solute carrier family 29 (nucleoside transporters), member 1

1.73 5.20 0.00020

CCAAT enhancer binding protein δ 3.29 1.16

3.20 0.011

Α-2-macroglobulin 1.80 3.70 0.0035 SerpinG1 1.77 3.40 0.0043 Stat3 1.84

1.50 1.14

2.30 0.012

Histocompatibility 2, D region locus 1

1.16 2.10 0.049

Complement component 1, q subcomponent, α polypeptide

1.50 1.80 0.024

Dopachrome tautomerase 1.61 1.00 0.13 Glial fibrillary acidic protein 2.10 3.10 0.11 Procollagen C-proteinase enhancer protein

1.67 1.12

1.00 0.078

Transthyretin 1.55 1.88

1.00 0.11

Keratin complex 1 acidic gene 12 2.42 1.00 0.50 Transferrin 1.93 1.00 0.11 Fosb 1.10 1.00 0.45 Leucine-rich repeat-containing 2 protein

1.84 1.00 0.47

Thrombospondin 1 1.57 1.17

1.00 0.40

Suppressor of cytokine signalling 3 1.56 1.17 1.26

2.80 2.80

0.001

Aryl hydrocarbon receptor 2.30 1.36 0.025

56

Figure 2.4: Moderate concordance between qPCR and microarray analyses in the Rds+/- model of IPD. Scatter plot comparing fold differences estimated by SAM analysis and qPCR in 7-week old Rds+/- and wild-type retinas. Each data point represents the mean log2fold-change following a triplicate repeat microarray experiment followed by SAM (x-axis) compared to the mean log2fold change following a triplicate repeat qPCR analysis (y-axis). The positive trend-line indicates a positive correlation between the microarray and qPCR analyses; the concordance correlation coefficient of 0.49 indicates that the correlation between qPCR and microarray is quite good. Arrow indicates Endothelin-2, the most highly differentially expressed gene by microarray and qPCR analyses.

57

The highly significant (p<10-50) up-regulated network involved in cell and immune signalling

(a putative IL-6 pathway) was the most statistically significant and was comprised of transcripts with

some of the highest estimated fold differences, suggesting this network is playing an important role

in Rds+/- retinas (Appendix 2.5). One highly differentially expressed gene not included in this

network is endothelin-2 (32-fold up-regulated), which was identified in a network involved in cellular

growth and proliferation (Appendix 2.10).

In conclusion, the Rds transcriptome was characterized using Affymetrix oligonucleotide

microarrays, followed by two different data analysis strategies. Of the 20 genes selected for qPCR

validation, 12 were confirmed to be significantly up-regulated in Rds+/- retinas. This false positive rate

of 40% (8/20) is significantly greater than the 4.7% false discovery rate predicted by the SAM. The

Ingenuity pathway finding algorithm was then employed to determine if any of the 145 differentially

expressed genes identified by SAM participated in any common networks. Of the 6 putative

pathways identified, a pathway involved in immune and cell signalling was identified that

encompassed a large number of significantly differentially expressed transcripts at comparatively

high fold differences, suggesting it is playing an important role in Rds+/- retinas.

2.5 Discussion

Significantly differentially expressed transcripts were identified in both the Rom1-/- and Rds+/-

retinas. Remarkably, none of the identified genes were differentially expressed in both models. This

lack of overlap was surprising because Rom1 and Rds are homologs and both encode proteins that

associate to form oligomers thought to play a structural role in the outer segment of the retina

(Connell et al., 1991; Bascom et al., 1992; Clarke et al., 2000b). The Rds+/- model of IPD displays

much faster cell death kinetics compared to the Rom1-/- model, which exhibits one of the slowest

rates of degeneration of any IPD model (Clarke et al., 2001). One possible interpretation of the lack

of overlapping differentially expressed genes is that the rate of degeneration may be an important

factor in determining the magnitude, number and type of transcriptional changes occurring in IPD

retinas; the functional category of the protein encoded by the IPD gene may be less important in

determining the transcriptional response. The observation that several common transcriptional

responses have been observed in different models, all displaying cell death kinetics faster than the

58

Rom1-/- model, supports this hypothesis. For example, members of the IL-6 pathway were up-

regulated in response to light damage, mechanical stress, and the Rd1-/- model of IPD (Chen et al.,

2004; Vazquez-Chona et al., 2004; Rattner and Nathans, 2005; Zacks et al., 2006).

Differentially expressed genes identified in Rds+/- retinas by microarray analysis were more

likely to be biologically reproducible when compared to the genes identified in Rom1-/- retinas; the

Rom1-/- microarray-qPCR comparison yielded more false positives and a lower CCC compared to the

Rds+/- microarray-qPCR comparison. The different microarray platforms used for each experiment

could be one source of the differing reproducibility. The Rom1-/- microarray experiment employed

Affymetrix U74A arrays, which contained probes designed to interrogate the ~12 000 genes and

ESTs present in the unigene database build 71 (March, 2000)(Affymetrix, 2004). In comparison, the

Rds+/- microarray experiment was performed using the more advanced 430A array, containing

probes representing ~14 000 genes and ESTs, many of which were redesigned based on the draft

sequence of the mouse genome and the updated sequences contained within the unigene database

build 107 (June, 2002)(Affymetrix, 2004). The microarray platform itself may have also been

improved between the U74A and 430A, although no specific details were released by the company.

Thus, the use of the 430A array may have facilitated the identification of more validated

differentially expressed genes in Rds+/- retinas.

Both the MAS5 9 pair-wise comparison and SAM identified an overlapping set of 18

differentially expressed genes in the Rds+/- model of IPD. This overlap is significant because each

method of data analysis is quite different. For example, MAS5 employs a normalization technique

based on scaling the hybridization intensities from each microarray being compared, so that the

average is the same (Affymetrix, 2004). In contrast, RMA involves a quantile normalization method

such that only genes with similar hybridization intensities are used in the normalization process

(Irizarry et al., 2003). The MAS5 algorithm computes the hybridization intensity for each gene by

identifying the degree to which RNA binds to its perfect match probe compared to the mismatch

probe- each perfect match probe has a mismatch negative control- on the microarray (Affymetrix,

2004). In contrast, the SAM algorithm ignores the mismatch data and identifies differentially

expressed genes based on the perfect match hybridization intensities only. The fact that two such

different data analysis tools were able to identify an overlapping set of differentially expressed genes,

suggests that the identified genes are likely to be truly differentially expressed. The fact that only

10/18 genes identified by both methods were verified by qPCR suggests an unknown bias in either

59

the microarray platform itself, or a difference in the RNA samples used for the microarray analysis

compared to qPCR.

The genetic background of the Rom1-/- and Rds+/- animals may have also contributed to the

variability of the microarray results. While Rds+/- and Rds+/+ mice are inbred animals (Schalken et al.,

1990; jackson-laboratories, 2008), Rom1-/- and its wild-type control are outbred (CD1 and 129 mix)

(Clarke et al., 2000b). The Rom1-/-microarray experiments were performed using pooled samples,

which is known to decrease variability by mitigating the possible effect of a single outlier (Allison et

al., 2006). However, genetic modifiers are known to affect the rate of cell death in many models of

IPD (Pacione et al., 2003); therefore, the lack of genetic homogeneity within and between Rom1-/-

and Rom1+/+ animals may have increased the variability in the Rom1-/- microarray experiment.

Increased variability would manifest itself as a decrease in the number of identified significantly

differentially expressed genes; increased variability would make it less likely that a given transcript

would be identified as significantly differentially expressed. The Rds+/- microarray analysis yielded

more reproducible results, a greater number of significantly differentially expressed genes with

higher fold differences. Additionally, six networks were identified among the differentially expressed

genes identified in Rds+/- retinas; no pathways were identified among the differentially expressed

genes identified in the Rom1-/- model. As such we decided to focus on the results obtained from the

Rds+/- model.

Six networks were identified as up-regulated in Rds+/- retinas (Appendix 2.5), including: 1) a

network involved in cell and immune signalling (a putative IL-6 pathway), featuring signal transducer

and activator of transcription (STAT) signalling, which is known to be activated when IPD retinas

are treated with neurotrophic factors that slow the rate of cell death (Chaum, 2003); 2) a network

involved in cancer and cell death including the proto-oncogene B-cell CLL/lymphoma 3 (BCL3)

and transforming growth factor beta (Tgf-b), which is known to mediate cell death in the retina are

members of this pathway (Duenker, 2005). Neuroprotective proteins were also identified in this

pathway, including GADD45, a DNA-damage inducible gene thought to play a protective role in

damaged neurons and platelet derived growth factor (Pgdf), known the protective PI 3-kinase/Akt

pathway in the retina were both (Torp et al., 1998; Biswas et al., 2008). Thus, this pathway consists

of both neuroprotective and pathogenic components; 3) a pathway involved in viral function,

immune response and injury was also up-regulated. Tumor necrosis factor is a major node in this

pathway and is a known pro-apoptotic factor involved in systemic inflammation and known to be

involved in neuronal cell death (Venters et al., 2000); 4) a cell signalling pathway featuring beta-

60

estradiol, which has recently been shown to exhibit neuroprotective properties on retinal ganglion

cells (Zhou et al., 2007); 5) an additional inflammatory pathway featuring STAT and mitogen

activated protein kinase signalling, generally thought be protective in the retina (Wahlin et al., 2001;

Chaum, 2003); 6) finally, the pathway involved in cellular growth and differentiation includes tumour

protein 53 (p53), which is involved in regulating the cell cycle, especially during DNA damage. It has

been suggested that p53 controls the activity of C/EBPδ, a transcription factor involved in

apoptosis during mammary gland development (Thangaraju et al., 2005).

The putative IL-6 pathway involved in cell and immune signalling (pathway #1 above) was

the most significant network identified. The pathway included more genes compared to the other

five pathways; 24 of the 145 significantly differentially expressed transcripts identified by the

microarray analysis were present in this pathway. As such, it was the most statistically significant of

all the pathways (p<10-50), since the likelihood that 24 genes that make up a common pathway could

randomly be present in a group of 145 genes is very low.

IL-6 pathway activation (Figure 2.5) may represent a near universal response to cell stress

and is probably involved in PR survival (Ueki et al., 2008). Our report represents one of the few

studies that have recognized increased IL-6 signalling in IPD retinas. For example, Rattner et al.

listed the Oncostatin M receptor (Osmr), a member of the IL-6 family of receptors, Stat-3, and

C/Ebpδ, a gene known to be regulated by Stat-3, as up-regulated in prCAD-/-retinas; however, they

were never identified as being a part of the same pathway (Rattner and Nathans, 2005). In contrast, a

recent report did identify increased IL-6 pathway activation in IPD retinas; cardiotrophin-like

cytokine (clc), and leukemia inhibitory factor (lif), both IL-6 cytokines, were up-activated

concomitant with increased phosphorylated Stat-3, an indicator of IL-6 signalling (Samardzija et al.,

2006). IL-6 pathway activation has also been reported in response to mechanical stress (Vazquez-

Chona et al., 2004; Zacks et al., 2006) and light mediated damage (Chen et al., 2004; Rattner and

Nathans, 2005). Several studies have shown that treatment with IL-6 cytokines, including ciliary

neurotrophic factor and cardiotrophin-1, have resulted in Stat-3 phosphorylation and increased PR

survival, suggesting endogenous IL-6 signalling may be protective in the retina (LaVail et al., 1998;

Song et al., 2003).

We selected the putative IL-6 pathway for additional analysis for several reasons. First, the

pathway the most significant compared to the other pathways identified by Ingenuity analysis.

Secondly, there is evidence in the literature to suggest that the IL-6 pathway is protective; its up-

regulation in a variety of contexts also makes IL-6 signalling an important therapeutic target. Finally,

61

loss-of-function mouse mutants were available for several genes in the IL-6 pathway, facilitating

genetic experiments to test the role of pathway members in IPD. The subject of the next chapter is

the use of genetic analysis to determine whether the up-regulation of several IL-6 pathway members

is mechanistically linked to cell survival.

62

Figure 2.5: Up-regulation of a putative IL-6 cytokine pathway in Rds+/- retinas Lif and Osm bind to Lifr and Osmr, respectively, which both share gp130 as a common co-receptor. Upon ligand binding, tyrosine kinases (Jak1,Jak2 and Tyk2) associated with gp130, Osmr and Lifr become activated. Once active, these kinases can phosphorylate Stat-3, which then dimerizes and is transported to the nucleus, where it acts as a transcriptional activator of several genes, including: Glial fibrillary acidic protein (Gfap), α-2 macroglobulin (α-2m), Suppressor of cytokine signalling (Socs3), CAAAT enhancer binding protein δ (C/EBPδ). Socs-3 is a negative regulator of Stat-3 phosphorylation. All fold differences are qPCR values.

63

Appendix 2.1: Up-regulated transcripts in the retinas of Rom1-/- mice

Accession Number

Gene Name SAM Score

SAM Numerator

Fold Change

96222_at AI037577 7.97 1.83 3.54 96353_at Transmembrane protein 14C

(Tmem14c) 6.86 1.44 2.73

100908_at Protein tyrosine phosphatase, receptor type A (Ptpra)

6.50 1.13 2.19

93180_at AI506816 4.58 1.19 2.29 103424_at RIKEN cDNA 6330442E10 4.40 0.48 1.39 96566_at Transglutaminase 3, E polypeptide

(Tgm3) 4.29 1.34 2.53

93351_at Hydroxyprostaglandin dehyogenase 15 (NAD) (Hpgd)

4.23 0.80 1.74

102413_at LIM domain only 1 (Lmo1) 3.65 0.41 1.33 99180_at GTP binding protein 4 (Gtpbp4) 3.51 0.57 1.48 99469_at Peroxisomal biogenesis factor 6 (Pex6) 3.48 0.40 1.32 93268_at Glyoxalase 1 (Glo1) 3.48 1.08 2.15 95052_at RIKEN cDNA 1110035L05 3.44 0.49 1.41 103744_at SH3 domain binding glutamic acid-rich

protein like 2 (Sh3bgrl2) 3.42 0.58 1.49

95508_at NCK-associated protein 1 3.23 0.58 1.49 104212_at Leucine-rich PPR motif containing

(Lrpprc) 3.19 0.58 1.49

94407_at Beta-1,3-glucuronyltransferase 3 (glucuronosyltransferase I) (B3gat3)

3.09 0.51 1.43

99494_at AJ001700 3.01 0.38 1.30 93328_at Histidine decarboxylase (Hdc) 2.88 0.67 1.59 104206_at RIKEN cDNA 5730557B15 2.86 0.89 1.86 160965_at RAS p21 protein activator 4 (Rasa4) 2.77 0.36 1.29 96077_at Solute carrier family 17 (sodium

phosphate), member 1 (Slc17a1) 2.71 0.79 1.70

95061_at Breast carcinoma amplified sequence 2 (Bcas2)

2.66 0.49 1.40

96672_at Homeobox only domain (Hod) 2.59 0.40 1.32 98596_s_at Sialyltransferase 9 (Siat9) 2.58 0.32 1.25 99840_at Prodynorphin (Pdyn) 2.54 0.30 1.23 96020_at Complement component 1, q

subcomponent, beta polypeptide (C1qb)2.52 0.36 1.29

95105_at RIKEN cDNA 2010110M21 2.48 0.32 1.25 93008_at LSM4 homolog, U6 small nuclear RNA

associated (Lsm4) 2.48 0.33 1.26

104516_at Claudin 5 (Cldn5) 2.47 0.42 1.34

64

Appendix 2.2: Down-regulated transcripts in the retinas of Rom1 -/- mice

Accession Number

Gene Name SAM Score

SAM Numerator

Fold Change

93453_at Rod outer membrane protein 1 (Rom1) -21.27 -5.88 -59.25 99140_at Mitochondrial ribosomal protein L16

(Mrpl16) -4.64 -0.83 -1.78

160799_at AW060549 -4.25 -0.56 -1.48 100496_at Peptidylglycine alpha-amidating

monooxygenase (Pam) -3.93 -0.73 -1.65

104314_at RIKEN cDNA 1110032A03 -3.51 -0.64 -1.56 103590_at Gamma-glutamyl carboxylase (Ggcx) -3.31 -0.80 -1.72 93464_at A kinase (PRKA) anchor ptotein 9

(Akap9) -3.09 -0.99 -2.00

95379_at Mab-21-like 2 (Mab21) -3.05 -0.57 -1.49 96646_at Ubiquitin specific protease 39 (Usp39) -2.99 -0.40 -1.32 99823_r_at DNA segment, Chr 18, ERATO Doi

232 -2.85 -0.48 -1.40

104000_at RIKEN cDNA 2210023G05 -2.86 -0.39 -1.31 95355_at Angiotensin II, type I receptor-

associated protein (Agtrap) -2.81 -0.41 -1.33

92202_g_at Zinc finger protein 145 (Zfp145) -2.72 -0.70 -1.63 96003_at Metastasis-associated gene family,

member 2 (Mta2) -2.72 -0.36 -1.28

65

Appendix 2.3: Up-regulated transcripts in the retinas of Rds+/- mice

Accession Number

Gene Name Gene symbol

Identified by MAS 5

SAM Score

SAM Numerator

Fold Change

1449161_at Endothelin 2 ET-2 Yes 16.32 3.08 8.34 1451683_x_at Histocompatibility 2,

K1, K region H2-K1 Yes 16.24 1.69 3.23

1437165_a_at Procollagen C-proteinase enhancer protein

Pcpe Yes 12.68 1.44 2.71

1423233_at CCAAT enhancer binding protein delta

C/Ebpδ Yes 10.86 1.71 3.29

1427388_at Leucine-rich repeat-containing 2

Lrrc2 Yes 9.46 0.87 1.84

1420796_at Aryl-hydrocarbon receptor repressor

Ahrr 8.66 1.20 2.31

1418021_at Complement component 4 (within H-2S)

C4 Yes 8.54 0.96 1.94

1455393_at Ceruloplasmin Cp 8.53 0.78 1.71 1435541_at Betacellulin, epidermal

growth factor family member

Btc 8.37 0.78 1.71

1454849_x_at Clusterin Clu 8.28 0.53 1.45 1448433_a_at Procollagen C-

proteinase enhancer protein

Pcolce 7.92 0.74 1.67

1426587_a_at Signal transducer and activator of transcription-3

Stat-3 Yes 7.88 0.88 1.84

1417381_at Complement component 1, q subcomponent, alpha polypeptide

C1qa 7.52 0.59 1.50

1449453_at Bone marrow stromal cell antigen 1

Bst1 7.09 0.45 1.37

1425546_a_at Transferrin Trf Yes 7.01 0.95 1.93 1418449_at Adinin Lad1 6.98 0.78 1.72 1418332_a_at ATP/GTP binding

protein 1 Agtpbp1 6.90 0.60 1.52

1416625_at Serine (or cysteine) proteinase inhibitor, clade G, member 1

Serping1 Yes 6.46 0.83 1.77

1423909_at RIKEN cDNA 0610011I04 gene

6.33 0.42 1.34

1415899_at Jun-B oncogene JunB 6.08 0.50 1.42

66

1417332_at Regulatory factor X, 2 (influences HLA class II expression)

Rfx2 6.01 0.51 1.42

1418726_a_at troponin T2 Tnnt2 5.83 0.58 1.49 1434719_at Alpha-2-macroglobulin α2m Yes 5.74 0.85 1.80 1426508_at Glial fibrillary acidic

protein Gfap Yes 5.72 1.08 2.10

1428942_at Metallothionein 2 Mt2 5.70 0.58 1.50 1451782_a_at Solute carrier family 29

(nucleoside transporters), member 1

Slc29a1 Yes 5.67 0.78 1.73

1418674_at Oncostatin M receptor Osmr Yes 5.67 1.21 2.31 1419309_at Glycoprotein 38 Gp38 5.66 0.41 1.33 1426708_at Anthrax toxin receptor

2 Antxr2 5.62 0.78 1.72

1449773_s_at Growth arrest and DNA-damage-inducible 45 beta

Gadd45b 5.51 0.74 1.68

1450650_at Myosin X Myo10 5.47 0.71 1.64 1416776_at Crystallin, mu Crym 5.41 0.55 1.47 1417928_at PDZ and LIM domain

4 Pdlim4 5.36 0.43 1.35

1424133_at RIKEN cDNA 6530411B15

5.29 0.51 1.42

1423754_at Interferon-induced transmembrane protein 3

fitm3 5.20 0.87 1.84

1437458_x_at Clusterin Clu 5.13 0.46 1.37 1418004_a_at RIKEN cDNA

1810009M01 5.12 0.37 1.30

1449289_a_at Beta-2 microglobulin Β2m 5.12 0.62 1.54 1449363_at Activating transcription

factor 3 Atf3 5.09 0.46 1.38

1421374_a_at FXYD domain-containing ion transport regulator 1

Fxyd1 4.94 0.44 1.35

1417460_at Interferon induced transmembrane protein 2

fitm2 4.94 0.46 1.37

1451132_at Pre-B-cell leukemia transcription factor interacting protein 1

Pbxip1 4.86 0.52 1.43

1419231_s_at Keratin complex 1, acidic, gene 12

Krt1-12 Yes 4.81 1.01 2.04

67

1416961_at Budding uninhibited by benzimidazoles 1 homolog, beta

Bub1b 4.76 0.73 1.66

1460344_at RIKEN cDNA 2310033F14

4.71 0.41 1.33

1423860_at Prostaglandin D2 synthase (brain)

Ptgds 4.64 0.63 1.55

1421322_a_at Interferon dependent positive acting transcription factor 3 gamma

Isgf3g 4.61 0.32 1.25

1416340_a_at Mannosidase 2, alpha B1

Man2b1 4.61 0.24 1.18

1427893_a_at phosphomevalonate kinase

Pmvk 4.56 0.43 1.35

1418181_at Protein tyrosine phosphatase 4a3

Ptp4a3 4.52 0.30 1.23

1427703_at platelet-activating factor acetylhydrolase, isoform 1b, beta1 subunit

Pafah1b1 4.45 0.32 1.25

1420549_at Guanylate nucleotide binding protein 1

Gbp1 4.43 0.59 1.50

1424080_at RIKEN cDNA 1700001E16

4.36 0.45 1.37

1449159_at Guanine nucleotide binding protein, beta 3

Gnb3 4.35 0.51 1.42

1418536_at Histocompatibility 2, Q region locus 7

H2-Q7 4.27 0.68 1.60

1448673_at Poliovirus receptor-related 3

Pvrl3 4.27 0.48 1.40

1422603_at Ribonuclase, RNase A family 4\!Rnase4

Rnase4 4.24 0.38 1.30

1452428_a_at Beta-2 microglobulin B2m 4.23 0.45 1.36 1418028_at Dopachrome

tautomerase Dct Yes 4.20 0.69 1.61

1419100_at Serine (or cysteine) proteinase inhibitor, clade A, member 3N

Serpina3n 4.18 1.01 1.98

1422631_at Aryl-hydrocarbon receptor

Ahr 4.16 0.39 1.31

1450582_at Histocompatibility 2, Q region locus 5

H2-Q5 4.15 0.38 1.30

1455899_x_at Suppressor of cytokine signalling 3

Socs3 4.12 0.63 1.56

1417750_a_at Mitochondrial solute carrier protein

Mscp 4.10 0.36 1.28

68

1417063_at Complement component 1, q subcomponent, beta

C1qb 4.07 0.40 1.32

1423653_at ATPase, Na+/K+ transporting, alpha 1 polypeptide

Atp1a1 4.04 0.42 1.34

1427221_at X transporter protein 3 similar 1 gene

Xtrp3s1 4.02 0.34 1.27

1419488_at TNFAIP3 interacting protein 2

Tnip2 3.99 0.31 1.24

1425560_a_at S100 calcium binding protein A16

S100a16 3.98 0.32 1.25

1449059_a_at 3-oxoacid CoA transferase 1

Oxct1 3.94 0.30 1.23

1455562_at SRY-box containing gene 12

Sox12 3.90 0.30 1.23

1419814_s_at S100 calcium binding protein A1

S100a1 3.85 0.31 1.24

1417977_at RIKEN cDNA 1300018P11

3.73 0.33 1.25

1423859_a_at Prostaglandin D2 synthase

Ptgds 3.70 0.51 1.42

1439240_x_at Lin 7 homolog b Lin7b 3.64 0.21 1.15 1451580_a_at Transthyretin Ttr Yes 3.62 0.63 1.55 1420915_at Signal transducer and

activator of transcription 1

Stat1 3.61 0.27 1.20

1439409_x_at Tyrosinase-related protein 1

Tyrp1 3.56 0.42 1.34

1449454_at bone marrow stromal cell antigen 1

Bst1 3.56 0.31 1.24

1424123_at cDNA sequence BC011209

3.54 0.33 1.26

1419877_x_at Phosphatidylinositol glycan, class M

Pigm 3.50 0.21 1.16

1422557_s_at Metallothionein 1 Mt1 3.50 0.50 1.42 1418440_at Procollagen, type VIII,

alpha 1 Col8a1 3.47 0.34 1.27

1448591_at Cathepsin S Ctss 3.46 0.57 1.49 1428088_at RIKEN cDNA

2410002I01 3.44 0.22 1.17

1428988_at ATP-binding cassette, sub-family C (CFTR/MRP), member 3

Abcc3 3.43 0.25 1.19

1415810_at ubiquitin-like, containing PHD and RING finger domains

Uhrf1 3.42 0.20 1.15

69

1432385_a_at ATP/GTP binding protein 1

Agtpbp1 3.40 0.52 1.44

1423294_at Mesoderm specific transcript

Mest 3.40 0.25 1.19

1416695_at Benzodiazepine receptor, peripheral

Bzrp 3.40 0.25 1.19

1418097_a_at Thymic stromal-derived lymphopoietin, receptor

Tslpr 3.40 0.20 1.15

1449556_at Histocompatibility 2, T region locus 23

H2-T23 3.39 0.39 1.31

1451186_at RIKEN cDNA 2700083B06

3.36 0.17 1.13

1422591_at Transcription elongation factor B (SIII), polypeptide 3

Tceb3 3.33 0.16 1.12

1449401_at Complement component 1, q subcomponent, gamma polypeptide

C1qg 3.32 0.38 1.31

1419420_at Sialyltransferase 7 Siat7e 3.27 0.21 1.16 1420989_at RIKEN cDNA

4933411K20 3.27 0.25 1.19

1427713_x_at POU domain, class 2, transcription factor 2

Pou2f2 3.26 0.24 1.18

1456212_x_at Suppressor of cytokine signalling 3

Socs3 3.24 0.22 1.17

1426221_at RIKEN cDNA 5830475I06

3.21 0.30 1.23

1427881_at RIKEN cDNA 4930588M11

3.20 0.14 1.11

1438676_at AI595338 3.19 0.14 1.10 1417630_at MAP kinase-interacting

serine/threonine kinase 1

Mknk1 3.18 0.18 1.13

1416612_at Cytochrome P450, family 1, subfamily b, polypeptide 1

Cyp1b1 3.18 0.27 1.20

1418589_a_at Myeloid leukemia factor 1

Mlf1 3.18 0.44 1.35

1428740_a_at Phosphatidylinositol glycan, class T

Pigt 3.17 0.30 1.23

1449172_a_at Lin 7 homolog b Lin7b 3.17 0.28 1.21 1437689_x_at Clusterin Clu 3.15 0.33 1.26 1418133_at B-cell

leukemia/lymphoma 3 Bcl3 3.13 0.39 1.31

1421811_at Thrombospondin 1 Thbs1 Yes 3.13 0.63 1.57

70

1434121_at Leucine-rich repeat LGI family, member 4

Lgi4 3.12 0.31 1.24

1423762_at AarF domain containing kinase 1

Adck1 3.11 0.18 1.13

1418396_at G-protein signalling modulator 3

Gpsm3 3.10 0.28 1.21

1419376_at RIKEN cDNA 1110018M03

3.10 0.30 1.23

1427434_at Baculoviral IAP repeat-containing 1

Birc1f 3.10 0.19 1.14

1419873_s_at Colony stimulating factor 1 receptor

Csf1r 3.06 0.32 1.24

1422544_at Myosin X Myo10 3.05 0.37 1.29 1450716_at Disintegrin-like and

metalloprotease (reprolysin type) with thrombospondin type 1 motif, 1

Adamts1 3.05 0.50 1.42

1416619_at RIKEN cDNA 4632428N05

3.04 0.22 1.16

1445367_at Mm.346510 3.03 0.27 1.20 1419230_at Keratin complex 1,

acidic, gene 12 Krt1-12 3.03 1.23 2.42

1422903_at Lymphocyte antigen 86 Ly86 3.02 0.40 1.32 1425078_x_at RIKEN cDNA

5830484A20 3.02 0.26 1.20

1418392_a_at Guanylate nucleotide binding protein 3

Gbp3 3.01 0.23 1.17

1449178_at PDZ and LIM domain 3

Pdlim3 3.01 0.29 1.22

1418240_at Guanylate nucleotide binding protein 2

Gbp2 3.01 0.34 1.27

1449164_at CD68 antigen Cd68 3.00 0.23 1.17 1427363_at RIKEN cDNA

9930007B02 2.99 0.30 1.23

71

Appendix 2.4: Down-regulated transcripts in the retinas of Rds+/- mice

Accession number

Gene Name Symbol Identified by MAS5

SAM Score

SAM Numer-ator

Fold Change

1449588_at ATP-binding cassette, sub-family A (ABC1)

Abca4 -7.89 -0.70 -1.6

1426819_at FBJ osteosarcoma oncogene B

Fosb Yes -6.97 -0.95 -1.9

1449956_at Protein kinase C, epsilon

Prkce -5.11 -0.57 -1.5

1426921_at ATP-binding cassette, sub-family F (GCN20), member 1

Abcf1 -4.98 -0.25 -1.2

1418780_at Cytochrome P450, family 39, subfamily a, polypeptide 1

Cyp39a1 -4.90 -0.34 -1.3

1436372_a_at AA415817 -4.72 -0.34 -1.3 1431326_a_at Tropomodulin 2 Tmod2 -4.65 -0.51 -1.4 1420682_at Cholinergic receptor,

nicotinic, beta polypeptide 1 (muscle)

Chrnb1 -4.58 -0.50 -1.4

1451529_at Small glutamine-rich tetratricopeptide repeat (TPR)-containing, beta

Sgtb -4.57 -0.30 -1.2

1452090_a_at Olfactomedin 3 Olfm3 -4.40 -0.40 -1.3 1436547_at Diacylglycerol kinase,

epsilon Dgke -4.34 -0.31 -1.2

1455053_a_at Mm.298486 -4.34 -0.23 -1.2 1450504_a_at 1-acylglycerol-3-

phosphateOacyltransferase 3

Agpat3 -4.27 -0.29 -1.2

1425898_x_at Olfactomedin 3 Olfm3 -4.25 -0.21 -1.2 1415802_at Solute carrier family 16

(monocarboxylic acid transporters), member 1

Slc16a1 -4.21 -0.33 -1.3

1437855_at Microtubule-associated protein 4

Mtap4 -4.15 -0.30 -1.2

1436912_at RIKEN cDNA 3110038O15

-4.03 -0.30 -1.2

72

Appendix 2.5: Up-regulated network involved in cell and immune signalling

Twenty-four differentially expressed genes identified from the microarray analysis are included in this network (p<10-50). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: aryl hydrocarbon receptor (AHR), aryl-hydrocarbon receptor repressor (AHRR), ATPase, Na+/K+ transporting, alpha 1 polypeptide (ATP1A1), betacellulin (BTC), CCAAT/enhancer binding protein (C/EBP), delta (CEBPD), colony stimulating factor 1 receptor (CSF1R), cathepsin S (CTSS), cytochrome P450, family 1, subfamily B, polypeptide 1 (CYP1B1), dopachrome tautomerase (DCT), FXYD domain containing ion transport regulator 1 (FXYD1), glial fibrillary acidic protein (GFAP), interferon induced transmembrane protein 2,3,9 (IFITM2,3,9), janus kinase (JAK), jun B proto-oncogene (JUNB), MAP kinase interacting serine/threonine kinase 1 (MKNK1), oncostatin M receptor (OSMR), S100 calcium binding protein A1 (S100A1), serine (or cysteine) peptidase inhibitor, clade A, member 3N (SERPINA3N), suppressor of cytokine signalling 3 (SOCS3), signal transducer and activator of transcription 1 (STAT1), TNFAIP3 interacting protein 2 (TNIP2), tyrosinase-related protein 1 (TYRP1).

73

Appendix 2.6: Up-regulated network involved in molecular transport, cancer and cell death

Twenty-two differentially expressed genes identified from the microarray analysis are included in this network (p<10-44). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: alpha-2-macroglobulin (A2M), ADAM metallopeptidase with thrombospondin type 1 motif, 1 (ADAMTS1), activating transcription factor 3 (ATF3), beta-2-microglobulin (B2M), B-cell CLL/lymphoma 3 (BCL3), complement component 1, q subcomponent, A,B,C chain (C1QA,B,C), ceruloplasmin (ferroxidase) (CP), growth arrest and DNA-damage-inducible, beta (GADD45B), guanylate binding protein 2, interferon-inducible (GBP2), histocompatibility 2, D region (H2-LD), major histocompatibility complex, class I, C,E (HLA-C,E), metallothionein 1E (MT1E,F), procollagen C-endopeptidase enhancer (PCOLCE), platlet derived growth factor (Pdgf), podoplanin (PDPN), POU class 2 homeobox 2 (POU2F2), serpin peptidase inhibitor, clade G (C1 inhibitor), member 1 (SERPING1), solute carrier family 29 (SLC29A1), transforming growth factor (Tgf), thrombospondin 1 (THBS1)

74

Appendix 2.7: Up-regulated network involved in viral function, immune response and injury

Sixteen significantly differentially expressed genes identified from the microarray analysis are included in this network (p<10-29). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: ATP-binding cassette, sub-family C member 3 (ABCC3), arachidonate 5-lipoxygenase-activating protein (ALOX5AP), asparagine synthetase (ASNS), BUB1,3 budding uninhibited by benzimidazoles 1 homolog beta (yeast) (BUB1,3), chemokine (C-C motif) receptor 2 (CCR2), CCAAT/enhancer binding protein (C/EBP), beta (CEBPB), carboxypeptidase B2 (plasma) (CPB2), deoxynucleotidyltransferase, terminal, interacting protein 2 (DNTTIP2), estrogen receptor 1 (ESR1), Fc fragment of IgG, receptor, transporter, alpha (FCGRT), guanylate nucleotide binding protein 1 (GBP1), guanylate binding protein 2, interferon-inducible (GBP2), guanylate binding protein 4 (GBP4), guanine nucleotide binding protein (G protein), alpha inhibiting activity polypeptide 2 (GNAI2), guanine nucleotide binding protein (G protein), beta polypeptide 3 (GNB3), G-protein signalling modulator 3 (AGS3-like, C. elegans) (GPSM3), hepcidin antimicrobial peptide (HAMP), haptoglobin-related protein (HPR), interferon induced transmembrane protein 3 (1-8U) (IFITM3), interferon, beta 1, fibroblast (IFNB1), interleukin 3 receptor, alpha (low affinity) (IL3RA), ladinin 1 (LAD1), mesoderm specific transcript homolog (mouse) (MEST), macrophage activation 2 like (MPA2L), msh homeobox 2 (MSX2).

75

Appendix 2.8: Up-regulated network involved in cell-cell signalling

Fifteen significantly differentially expressed genes identified from the microarray analysis are included in this network (p<10-27). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: amiloride-sensitive cation channel 3 (ACCN3), ATP/GTP binding protein 1 (AGTPBP1), aryl-hydrocarbon receptor nuclear translocator 2 (ARNT2), bone marrow stromal cell antigen 1 (BST1), chromosome 10 open reading frame 54 (C10ORF54), CD46 molecule, complement regulatory protein (CD46), cytokine receptor-like factor 2 (CRLF2), discs, large homolog 4 (DLG4), enolase 1, (alpha) (ENO1), Fc fragment of IgG, receptor, transporter, alpha (FCGRT), forkhead box P3 (FOXP3), guanylate binding protein 2, interferon-inducible (GBP2), glial fibrillary acidic protein (GFAP), histocompatibility 2, Q region locus 5,8 (H2-Q5,8), major histocompatibility complex, class I, C (HLA-C), interleukin 6 (IL6), lin-7 homolog B (C. elegans) (LIN7B), lymphocyte antigen 86 (LY86), microRNA 21 (MIRN21 (includes EG:406991)), nuclear receptor co-repressor 2 (NCOR2), 3-oxoacid CoA transferase 1 (OXCT1), pre-B-cell leukemia homeobox interacting protein 1 (PBXIP1), platelet-derived growth factor alpha polypeptide (PDGFA), PDZ and LIM domain 4 (PDLIM4), PDZ domain containing 1 (PDZK1), polypyrimidine tract binding protein 2 (PTBP2), protein tyrosine phosphatase, non-receptor type 13 (APO-1/CD95 (Fas)-associated phosphatase) (PTPN13), RD RNA binding protein (RDBP).

76

Appendix 2.9: Up-regulated network involved immunological and inflammatory disease

Twelve significantly differentially expressed genes identified from the microarray analysis are included in this network (p<10-20). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: anthrax toxin receptor 2 (ANTXR2), complement component 4B (C4B), Casein kinase 2, Casein Kinase II, CKII (Ck2), clusterin (CLU), collagen, type VIII, alpha 1 (COL8A1), cAMP response element binding (Creb), DnaJ (Hsp40) homolog, subfamily C, member 11 (DNAJC11), elongation protein 2 homolog (S. cerevisiae) (ELP2), heat shock protein 90 (Hsp90), interleukin 31 (IL31), interleukin 9 receptor (IL9R), keratin 72 (KRT72), low density lipoprotein (LDL), MAP kinase protein (Mapk), oncostatin M receptor (OSMR), platelet-activating factor acetylhydrolase, isoform Ib, alpha subunit (PAFAH1B1), PI3 kinase (PI3K), protein kinase C (Pkc), protein tyrosine phosphatase type IVA, member 3 (PTP4A3), retinol binding protein 4, plasma (RBP4), suppressor of cytokine signalling 4 (SOCS4), signal transducer and activator of transcription 3 (STAT3), transcription elongation factor B3 (TCEB3), transferrin (TF), transthyretin (TTR), ubiquitin-like, containing PHD and RING finger domains, 1 (UHRF1).

77

Appendix 2.10: Up-regulated network involved in cellular growth and proliferation

Twelve significantly differentially expressed genes identified from the microarray analysis are included in this network (p<10-27). Red shading indicated an up-regulated gene as identified by microarray analysis while un-shaded circles are genes included by the ingenuity software package with a known relationship with one of the differentially expressed genes. Solid lines indicate a direct relationship such as a protein-protein interaction while dashed lines indicates an indirect relationship. Abbreviations: aurora kinase B (AURKB), Bruton agammaglobulinemia tyrosine kinase (BTK), cyclin-dependent kinase inhibitor 1B (p27, Kip1) (CDKN1B), CCAAT/enhancer binding protein (C/EBP), delta (CEBPD), crystallin, mu (CRYM), endothelin 2 (EDN2), endothelin receptor type A (EDNRA), etoposide induced 2.4 mRNA (EI24), fibulin 5 (FBLN5), growth arrest-specific 1 (GAS1), glucuronidase, beta (GUSB), hexokinase 2 (HK2), interleukin 31 (IL31), integrin, beta 1 (fibronectin receptor, beta polypeptide, antigen CD29 includes MDF2, MSK12) (ITGB1), mannosidase, alpha, class 2B, member 1 (MAN2B1), mitogen-activated protein kinase 1 (MAPK1), melanoma inhibitory activity (MIA), myeloid leukemia factor 1 (MLF1), v-myc myelocytomatosis viral oncogene homolog (MYC), myosin X (MYO10), phosphomevalonate kinase (PMVK), protein phosphatase 1, regulatory (inhibitor) subunit 15A (PPP1R15A), prostaglandin D2 synthase (PTGDS), prostaglandin F receptor (PTGFR), poliovirus receptor-related (3PVRL3), regulatory factor X, 2,3 (RFX2,3), ribonuclease, RNase A family, 4 (RNASE4).

78

3 A role for IL-6 signalling molecules in inherited photoreceptor degenerations

The majority of the work presented in this chapter is my own, with the following exceptions:

1) Figure 3.2 and Figure 3.7c: Immunoblots were performed under my supervision by Lynda Ploder

2) Figure 3.5: laser capture microdissection and qPCR was performed by myself, Coco Jiang and

Alexa Bramall.

3) Figure 3.7b: Alexa Bramall performed the sectioning and staining of retinas from the Lif+/+ and

Lif-/- mice.

79

3.1 Abstract

We previously reported that in inherited photoreceptor degenerations (IPDs), the mutant

photoreceptors (PRs) are at a constant risk of death (Clarke et al. Nature 2000). Using microarrays

and quantitative PCR (qPCR) to identify genes that mediate the constant risk, we found up-

regulation of a putative IL-6 cytokine pathway in 3 mutant PR mouse models, when 60-90% of PRs

are still alive. For example, in the Rds+/- model, Oncostatin M (2X up by qPCR) → Oncostatin M receptor

(Osmr)(2.6X up) → Stat3 (2.3X up) → the transcription factor C/EBPδ (3.2X up), with increases in

the cognate proteins Osmr (3X up), Stat3 (2.6X up), and the phosphorylated, transcriptionally

active form of Stat-3 (5.8X up)(all p<0.01). Leukemia inhibitory factor (Lif) mRNA, another IL-6

cytokine was also found to be up-regulated (3.0X up). These increases occurred predominantly in

Müller glia, but the increase in C/EBPδ mRNA was in several cell types, including PRs. Since

exogenous IL-6 cytokines have been shown to slow PR death and also increase Müller cell pStat-3,

we asked whether the endogenous increases in IL-6 pathway genes in mutant retinas were a survival

response, and generated mutant PR models with Osmr, Lif or C/EBPδ loss-of-function (LOF)

mutations. Osmr LOF decreased PR survival in the retinas of: 4 month old Rds+/-;Osmr-/- mice had

12.5% fewer PRs than those of Rds+/-;Osmr+/+ mice (n=9, p<0.05), and the retinas of 31 day-old Tg-

RHO(P347S);Osmr-/- mice had 13.5% fewer PRs (n=6, p<0.01). The putative IL-6 response pathway

(above), if predominantly activated through the Osmr, would be expected to exhibit a decline in

pStat3 levels in the absence of Osmr. Unexpectedly, however, Osmr LOF had no effect on pStat3

levels in Rds+/-;Osmr-/- retinas, indicating that Stat3 activation is mediated predominantly through IL-

6 cytokines other than Osm, or other pathways. In contrast to the Osmr LOF, Lif LOF increased

mutant PR survival in the retinas of: PN13 Rd1-/-;Lif -/- mice had 14% more PRs than Rd1-/-;Lif+/+

mice (n=6, p<0.003) and a 1.7 fold decrease in pStat-3 (n=4, p<0.05). Similarly, C/EBPδ LOF

increased mutant PR survival in the retinas of: 8-month old Rds+/-;C/EBPδ-/- mice had 18% more

PRs than Rds+/-;C/EBPδ+/+ mice (n=5, p<0.005). These findings suggest that in mutant PRs 1) up-

regulation of the Osmr receptor is a survival response; 2) the presence of Lif or C/EBPδ is

pathogenic, and therefore 3) Osmr , Lif and C/EBPδ act either in different pathways or different

cells to account for the differing effects of their LOF on PR cell death; 4) the partial effects of Osmr,

Lif and C/EBPδ LOF indicate that other genes also mediate the constant risk of death.

80

3.2 Introduction

Photoreceptors are highly specialized cells in the retina capable of biochemically responding

to visual cues, beginning a cascade of events required for vision (Burns and Arshavsky, 2005;

Fernald, 2006). They are also particularly susceptible to genetic insults. For example, more than 193

different loci, including 142 cloned genes, have been implicated in IPD, a major cause of visual

impairment for which no therapy has been effective, except for the recent gene therapy outcomes

for Leber’s congenital amaurosis (Acland et al., 2001; Dejneka et al., 2004; Hartong et al., 2006;

Daiger et al., 2008; Hauswirth et al., 2008). One promising approach is treatment with neurotrophic

factors, since their exogenous administration promotes PR survival in a variety of animal models

(Faktorovich et al., 1990; LaVail et al., 1992; LaVail et al., 1998).

Photoreceptors are partially protected by treatment with neurotrophic factors in animal

models of PR degeneration (LaVail et al., 1998; Chaum, 2003). PR protection was first demonstrated

in the RCS rat model of IPD, in which a single intravitreal injection of basic fibroblast growth factor

(bFGF) slowed the rate of PR degeneration; two months after treatment, five to seven rows of PR

nuclei remained, compared to one row of nuclei remaining in untreated RCS controls (Faktorovich

et al., 1990). Exposing rodents to bright light also induces PR degeneration. In a light-induced model

of PR degeneration, bFGF also had a significant protective effect (Faktorovich et al., 1992; LaVail et

al., 1992; O'Driscoll et al., 2008). LaVail et al. then used the light-induced model of IPD to

determine if other neurotrophic factors could also provide PR protection; of the 8 factors tested,

four provided a high degree of PR rescue, including interleukin 1β (IL-1β), acidic fibroblast growth

factor (aFgf), ciliary neurotrophic factor (Cntf), and brain derived neurotrophic factor (Bdnf) (LaVail

et al., 1992). Interestingly, when a subset of these neurotrophic factors were tested in three mouse

models of IPD, it was discovered that only intravitreally injected Cntf could rescue PRs in all models

(LaVail et al., 1998). This result was confirmed by studies in a cat model of IPD, in which intravitreal

injections of Cntf but not Bdnf increased PR survival (Chong et al., 1999). One drawback to these

studies was that the neurotrophin being tested was only injected once. The generation of transgenic

mice with an inducible Bdnf transgene in an IPD background demonstrated that continuous Bdnf

expression slowed PR cell death and preserved PR function (Okoye et al., 2003). Taken together,

these results suggest that while a single bolus of Cntf is sufficient to delay PR degeneration, a

constant supply of Bdnf is required for PR rescue.

81

Gene transfer has also been employed to test the effectiveness of a continuous supply of

survival factors in the retinas of mice or rats with IPD. Mutant retinas treated with adenoviral

vectors (AAV) encoding a secretable form of bFGF or Cntf displayed a slower rate of PR

degeneration (Cayouette et al., 1998; Lau et al., 2000; Liang et al., 2001), although a corresponding

preservation of PR function, as measured by electroretinography, was observed in only one study

(Cayouette et al., 1998). Another study concluded that treatment with AAV encoding Cntf rescued

mutant PRs but lowered ERG function, suggesting that the surviving PRs were not functioning

normally (Liang et al., 2001). One explanation for these results is the use of the CMV promoter, a

strong promoter, to drive Cntf expression may have produced toxic amounts of Cntf, which

impaired PR function. A recent study has demonstrated that, at high doses in wild-type rats, Cntf

can inhibit PR function as measured by ERG and visual acuity tests; lower doses were effective at

delaying cell death and did not disrupt PR function in IPD models (McGill et al., 2007).

The intravitreal injection of encapsulated cells constitutively expressing Cntf is another

approach to deliver a constant supply of Cntf. This mode of delivery has been shown to be effective

in delaying cell death in the rcd1 canine model of IPD (Tao et al., 2002); wild-type rabbits exposed to

the same therapeutic dose did not show decreased PR function as measured by ERG (Bush et al.,

2004). These findings support the notion that a low dose of Cntf protects PR morphology and

function. Clinical trials are ongoing to establish whether Cntf is an effective treatment for humans

with IPD (Sieving et al., 2006).

Cntf is a member of the IL-6 family of cytokines (Kamimura et al., 2003). Other members

include: oncostatin M (Osm), leukemia inhibitory factor (Lif), Cardiotrophin-1 (Ct-1), Interleukin-6

(IL-6), Interleukin-11 (IL-11), Cardiotrophin-like cytokine (Clc), and neuropoetin (Np). While Cntf

has been shown to be an effective neuroprotective agent in 13 animal models of IPD (MacDonald et

al., 2007), several other IL-6 family members have also been shown to have neuroprotective effects

in the retina (LaVail et al., 1998; Song et al., 2003). Repeated intravitreal injections of Ct-1 led to

enhanced PR survival in a line of transgenic rats carrying the S334ter rhodopsin mutation (Song et

al., 2003), and Lif has been demonstrated to slow the rate of PR degeneration in the Q344ter mutant

rhodopsin mice, as well as in mice exposed to bright light (LaVail et al., 1998; Ueki et al., 2008). In

conclusion, several IL-6 cytokines have displayed neuroprotective properties in the retina,

prompting my interest in whether endogenous IL-6 cytokines also play a neuroprotective role in

IPDs, and, if so, whether the downstream responses to endogenous cytokines in IPD are uniform.

82

The downstream signalling pathways activated by IL-6 cytokines are well characterized

(Figure 2.5, Chapter 2). Each ligand is capable of binding to an IL-6 type receptor, which then binds

to gp130, the common co-receptor of the IL-6 cytokines (Kamimura et al., 2003). Of the five IL-6

receptors, only two can bind more than one ligand; the Lifr can bind Clc, Ct-1 and Lif, while the

Cntfr can bind Np and Cntf (Kamimura et al., 2003; Derouet et al., 2004). Upon ligand binding, IL-

6 type receptors bind to gp130, resulting in the trans-phosphorylation and activation of associated

janus kinases (JAK) and tyrosine kinases (Tyk), which in turn then trigger two major signalling

pathways by the phosphorylation of extracellular signal-regulated kinases 1,2 (ERK-1,2), and signal

transducers and activators of transcription (STAT) proteins (Kamimura et al., 2003). Suppressor of

cytokine signalling 3 (Socs-3) is up-regulated by Stat-3, functioning as part of a negative feedback

loop that prevents Stat-3 phosphorylation (Fischer et al., 2004). Work performed in non-neuronal

systems has shown that pStat-3 can activate the expression of the transcription factor CCAAT

enhancer binding protein delta (C/EBPδ), alpha-2-macroglobulin (α2M) and glial fibrillary acidic

protein (Gfap) (Hutt et al., 2000; Zhang and Darnell, 2001).

The responses of signalling pathways to the administration of cytokines has been largely

restricted to examination of changes in the abundance and phosphorylation of Stat-3 and ERK-1,2.

Wild-type rat retinas treated with axokine, a Cntf analog, displayed up-regulated phosphorylated

Stat-3 (pStat-3) and phosphorylated ERK-1,2 (pERK-1,2); the increase in pStat-3 was largely

confined to Müller glia and ganglion cells, while up-regulated pERK-1,2 was localized exclusively to

Müller glia (Peterson et al., 2000). Similarly, cardiotrophin-1 mediated PR protection in a rat IPD

model was associated with increased pStat-3 in Müller glia and ganglion cells (Song et al., 2003).

Animals preconditioned with sub-lethal levels of light became resistant to subsequent light damage;

increased Cntf and Fgf transcripts were detected in the whole retina by qPCR (Liu et al., 1998).

Additionally, increased pERK-1,2 localized to both Müller glia and PRs (Liu et al., 1998). Lif

treatment also increased pStat-3 predominantly in Müller glia concomitant with PR protection from

light-mediated damage (Ueki et al., 2008); a subset of PRs also exhibited pStat-3 staining after Lif

treatment, suggesting that Lif has the ability to activate pStat-3 in PRs (Ueki et al., 2008).

Phosphorylated Stat-3 has not been observed in the PRs of Cntf/axokine treated retinas despite the

observation that the Cntfr is present on PRs (Peterson et al., 2000; Beltran et al., 2003). In summary,

the increases in pERK-1,2 and pStat-3 during cytokine-mediated PR rescue suggests these molecules

may play critical roles in protecting mutant PRs from death. Furthermore, these molecules exhibit

83

the greatest activation in Müller glia, suggesting these cells are central to the rescue effect of

cytokines.

Pathogenic mutations, stress or injury to the retina induces the expression of several

neurotrophic factors and down-stream signalling molecules (Table 1, Chapter 1). For example,

mechanical injury and light damage were associated with increased Fgf and Cntf (Wen et al., 1995;

Cao et al., 1997; Liu et al., 1998; Walsh et al., 2001; O'Driscoll et al., 2008). A recent study identified

Cntf, Lif and Clc as up-regulated after acute light exposure and in the Rd1 and Vpp models of IPD;

Stat-3 and phosphorylated Stat-3 were also highly up-regulated in both IPD and light-induced

damage models, while phosphorylated ERK-1,2 was only up-regulated in the light damaged model

(Samardzija et al., 2006). These studies suggest that Stat-3, but not ERK-1,2, activation may be a

general response to PR stress; the observation that Stat-3 is also activated during treatment with

protective neurotrophins indicates that the endogenous up-regulation of pStat-3 may be a protective

response to PR stress.

The inhibition of the Jak kinases is one method that has been used to determine if the

JAK/STAT pathway is pathogenic or protective in the retina (Samardzija et al., 2006). Upon light

exposure, retinas treated with the Jak kinase inhibitor AG-490 exhibited a significant decrease in

Jak2, Stat-3 and Stat-1 phosphorylation; AG-490 treated retinas displayed a decrease in the number

of apoptotic nuclei, suggesting Stat phosphorylation was pathogenic (Samardzija et al., 2006). One

explanation for these results in that the prevention the phosphorylation of Stat-1, a pro-apoptotic

member of the Stat family (Stephanou and Latchman, 2005), may be responsible for the decrease in

cell death (Samardzija et al., 2006). Interestingly, Rd1-/- and Vpp-/- retinas exhibited increased pStat-3

but not pStat-1. In contrast to its effect on light damaged retinas, AG-490 treated mutant retinas did

not exhibit decreased pStat-3 levels or a change in the number of apoptotic nuclei (Samardzija et al.,

2006). Thus, although the JAK/STAT pathway appears to have pro-apoptotic properties in light-

mediated damage, the role of activated Stats in IPD could not be determined because AG-490

treatment was not effective in IPD models.

I previously described (Chapter 2) the up-regulation of a several transcripts belonging to an

IL-6 cytokine pathway in three murine models of IPD. In this study, I further characterized the

protein abundance and localization of several up-regulated IL-6 family members. I also generated

mutant PR models with the Osmr, Lif or C/EBPδ loss-of-function mutations to determine if these

up-regulated genes are involved in the pathogenesis of IPD.

84

3.3 Material and Methods

3.3.1 Quantitative real-time PCR

qPCR was performed as described in the materials and methods section of Chapter 2.

Primers used in qPCR reactions for this chapter are listed in Table 3.1.

3.3.2 Histology and outer nuclear layer measurements

Eyes were enucleated from the following mice: wild-type, Osmr -/-, Rds+/-;Osmr+/+, Rds+/-;Osmr -/- at 7-

weeks of age; C57/B6 wild-type, P347S; wild-type, P347S;Osmr -/- at post-natal day 31; Rd1+/+, Lif -/-,

Rd1-/-;Lif+/+, Rd1-/-;Lif -/- at post-natal day 13. The dorsal (superior) hemisphere was marked with a

hot needle and placed in 2% gluteraldehyde v/v in 0.1M phosphate buffer, 87.5mM sucrose and

incubated overnight at 4°C. Fixed eyes were then washed in 0.1M phosphate buffer. The dorsal

hemisphere was then isolated and trisected radially through the optic nerve. The sections were then

dehydrated in EtOH and propylene oxide, embedded in Jembed 812 and polymerized at 60°C

overnight. Using a Leica ultramicrotome, 1µm sections were cut and stained with toluidine blue (1%

toluidine blue, 1% sodium borate), and imaged with a Leica DM1000 microscope and digital camera.

Measurements of ONL thickness were made using SigmaScan Pro5 (Systat Software Inc., San Jose,

California). ONL thickness was considered to be the distance between the outer limiting membrane

and the outer plexiform layer, and was measured in sections where the columns of rod nuclei were

apparent, to ensure sections were not oblique. For each sample, five individual measurements were

made in the central retina from at least 3 individuals from each genotype. The data were analyzed

using a two-tailed t-test, to determine if genotype affected ONL thickness.

3.3.3 Retinal protein isolation

Retinal protein was harvested from mice sacrificed by cervical dislocation. Eyes were

immediately removed and placed in ice cold phosphate buffered saline (137mM NaCl, 19mM

Na2HPO4, 3mM KCl, 2mM KH2PO4 [pH7]) and the retina was isolated from the rest of the eye.

Isolated retinas form a single mouse were then placed in a microfuge tube containing 200µl of the

following solution: 10 mM Tris-Hcl, 1% SDS, 1mM sodium ortho-vanidate, 2mM EDTA, 1x

85

Table 3.1: List of primers used for qPCR

Gene Name Accession Number

Sequence (5' to 3') Crosses Intron/ Exon Boundary

IL-6 cytokines cardiotrophin-1 (Ct-1) NM_007795 Accctcttcacggccaac

ctcacccactcgccataga No

cardiotrophin-like cytokine (Clc) NM_019952 Ccagctcttaatcgcacagg

ggattgaagtcaggctcgtt No

ciliary neurotrophic factor (Cntf) NM_170786 Cgactccaagagaacctcca

ggtaggcgaaggcagaaact No

Interleukin-6 (IL-6) NM_031168 Tctctgggaaatcgtggaaa

tccagtttggtagcatccatc Yes

Interleukin-11 (IL-11) NM_008350 Ccgactggaacggctactc

ggggatcacaggttggtct Yes

leukemia inhibitory factor (Lif) NM_008501 Aatgtgctttgccgtctgt caacccaactttttcctttg

No

neuropoietin (Np) AY363390 Cagagcccatcggtcaag

gctgccctggtgttggag Yes

oncostatin M (Osm) NM_001013365 Gctgctccaactcttcctct

caggtcaggtgtgttcaggtt No

IL-6 receptors ciliary neurotrophic factor receptor (Cntfr) NM_016673 Acgcagaaacacagtccaca

ctgtcccgtttaccctcca Yes

glycoprotein 130 (gp130) NM_010560 Caagaaacaaggtgggcaaa

acgggtttaggtggaggtg Yes

Interleukin-6 receptor (IL-6r) NM_010559 Cacgaaggctgtgctgttt

gttgtggctggacttgcttc Yes

Interleukin-11 receptor (IL-11r) NM_010549

Acgtgcctactggatgtgag aaccggaacttgagcagaaa

Yes

leukemia inhibitory factor receptor (Lifr) NM_013584 Tatcctgaacatccccgttt cccaccagtccagttatcct

Yes

oncostatin M receptor (Osmr) NM_011019.1 Attcgcatcacaaccaacaa Yes Tccttgacctcttcgtgtcc

86

Laser capture control genes Chx-10/Vsx2 NM_007701 Aatgaagcccactacccaga

cacttctccctcttcctcca Yes

Rhodopsin (Rho) NM_145383 Tcaccaccaccctctacaca

catcggcttgcagaccac Yes

protease inhibitors (Roche, Laval, Québec) and immediately frozen in liquid Nitrogen. When the

appropriate number of samples were collected for analysis, samples were thawed on ice, briefly

vortexed and boiled for 10 minutes. Samples were vortexed again and then incubated at 60°C for 20

minutes. The lysate was then passed through a 23 gauge needle 20 times followed by centrifugation

at 21 000 x g at 4°C for 20 minutes. The supernatant was then isolated and the protein

concentration was quantified using the Biorad protein assay according to protocol (Biorad,

Missisauga, Ontario).

3.3.4 Immunoblot analysis

Retinal protein was isolated as indicated above and a volume containing 100µg of total

protein was combined with loading buffer (1X final concentration), vortexed briefly, boiled for five

minutes and electrophoresed and transferred to Hybond-C extra membrane (Amersham) according

to established protocols (Maniatis et al., 1987). Membranes were blocked with 1% skim milk powder

dissolved in TBST (25 mM Tris, 140 mM NaCl, 0.05% Tween-20, pH 7.5) for one hour at room

temperature and incubated overnight at 4°C with one of the following antibodies at the indicated

dilutions: Osmr 1:100 dilution (Santa Cruz, Santa Cruz, California), pStat-3 1:1000 (Cell Signalling,

Danvers, Massachusetts), Stat-3 1:1000 (Cell Signalling, Danvers, Massachusetts), pStat-1 1:1000

(Cell Signalling, Danvers, Massachusetts), Stat-1 1:1000 (Cell Signalling, Danvers, Massachusetts),

pERK1,2 1:1000 (Cell Signalling, Danvers, Massachusetts), ERK1,2 (Cell Signalling, Danvers,

Massachusetts), β-tubulin 1:1000 (Cell Signalling, Danvers, Massachusetts). To determine the relative

abundance pStat-3 relative to Stat-3, pStat-1 relative to Stat-1 and pERK1,2 relative to ERK1,2, the

antibody specific to the phosphoprotein was used for the first hybridization followed by incubation

with a secondary antibody conjugated to horseradish peroxidase for detection by enhanced

chemoluminescence and exposed to film. After development, immunoblots were washed with 1x

TBST (25 mM Tris, 140 mM NaCl, 0.05% Tween-20, pH 7.5) and placed in the stripping solution

(100mM β-mercaptoethanol, 2% SDS, 62.5 mM Tris-Hcl pH 6.8) pre-heated to 50°C. The

87

immunoblots were incubated in a 50°C shaking waterbath followed by three 10 minute washes in 1x

TBST. Immunoblots were then blocked and probed with an antibody capable of detecting both the

phosphorylated and un-phosphorylated forms of the protein; the procedure was repeated and an

antibody specific to β-tubulin was used in the last hybridization reaction. Blots were scanned and

densitometry was performed using ImageJ (Abramoff et al., 2004).

3.3.5 Immunofluorescence staining

For Lif, Osmr and Gfap staining, eyes were enucleated from 7-week old wild-type and Rds+/-

mice, fresh frozen in OCT compound (Tissue-Tek, Miles, Elkhart, IN) and 14µm sections were cut

using a Leica cryostat, and allowed to dry overnight. Sections were then fixed in ice-cold acetone for

5 minutes, removed and allowed to air dry for 30 minutes. Sections were then washed in PBS

(137mM NaCl, 19mM Na2HPO4, 3mM KCl, 2mM KH2PO4 [pH7]) and blocked in 1% donkey

serum for 1 hour. Sections were then exposed to one of the following antibodies (or combination of

antibodies in double labeling experiments) at the following dilutions: Gfap 1:50 (Abcam, Cambridge,

Massachusetts), Lif 1:10 (Santa Cruz, Santa Cruz, California), Osmr 1:50 (Santa Cruz, Santa Cruz,

California). Sections were then washed in PBS and incubated with the appropriate secondary

antibody conjugated to a chromofluor. For Lif and Osmr antibodies, 1:200 chicken anti-goat

Alexa488 antibodies were used (Invitrogen, Burlington, Ontario), for GFAP, 1:200 rabbit anti-

mouse TRITC was used (Invitrogen, Burlington, Ontario). Sections were then washed in PBS and

mounted in immuno-mount (Thermo Shandon, Pittsburgh, PA). Sections were visualized using a

Zeiss LSM510 Laser Scanning Confocal Microscope and images were acquired using the LSM510

software package.

For pStat-3, Stat-3 and p27Kip1 staining, eyes were enucleated from 7-week old Rds+/- and

wild type mice, fixed overnight in 4% paraformaldehyde and embedded into paraffin wax and

sectioned. Sections were then subject to antigen retrieval by boiling for 12 min in a 1mM EDTA

solution in the microwave. Sections were then washed in PBS and blocked in 5% goat serum (v/v),

3% bovine serum albumen (w/v) for 1 hour at room temperature. The indicated primary antibody

was then used at the following dilution: Stat3 1:100 (Cell Signalling, Danvers, Massachusetts), pStat3

1:100 (Cell Signalling, Danvers, Massachusetts), p27kip1 1:500 (BD Biosciences, Mississauga, ON) and

incubated overnight at 4°C. Sections were then washed in PBS and incubated with the appropriate

secondary antibody conjugated to a chromofluor. For Stat-3 and pStat-3, 1:100 goat anti rabbit Cy3

(Invitrogen, Burlington, Ontario) was used while for p27kip1, 1:100 goat anti mouse Alexa 488

88

(Invitrogen, Burlington, Ontario) was employed. Sections were then washed in PBS and mounted in

immuno-mount (Thermo Shandon, Pittsburgh, PA). Sections were visualized using a Zeiss LSM510

Laser Scanning Confocal Microscope and images were acquired using the LSM510 software

package.

3.3.6 In situ hybridization

Lif sense and antisense probes were made using a construct available in the National

Institute of Aging 15k murine clone set (accession#: H3041H12) containing 1kb of Lif 3’

untranslated region (UTR). The construct was sequence verified and 20µg of plasmid was linearized

using 2.5µ HindIII (Invitrogen, Burlington, Ontario), 1x Buffer, 1x Bovine serum albumen (BSA) in

a total volume of 50µl. Another 20 µg of plasmid was digested with 2.5 µl EcoR1, 1x Buffer, 1x BSA

in a total volume of 50µl. Each sample was purified using Qiagen PCR columns (Qiagen,

Mississauga, Ontario), precipitated using 100% ethanol and resuspended in a total volume of 10µl

10mM Tris-Hcl such that the concentration of purified DNA was approximately 1µg/ul.

Digoxygenin-labeled riboprobes were then made by adding 1X Roche DIG labeling mix (Roche,

Laval, Quebec), 1X transcription buffer, 37.5 µM Dithiothreitol, 1µl RNase inhibitor (Roche, Laval,

Quebec) to 1µl of purified DNA (1µg/ul); for the HindIII sample, 1µl of T7 RNA polymerase

(Roche, Laval, Quebec) was added, while for the EcoRI sample 1µl of Sp6 RNA polymerase (Roche,

Laval, Quebec) was added. After one hour of incubation at 37°C, another 1µl of the appropriate

RNA polymerase was added and incubated for another hour at 37°C. The sample was ethanol

precipitated and resuspended in 100µl diethylpyrocarbonate-treated water. The quantity of each

probe was assessed by comparing the 260/280 nm ratio using a nano-drop spectrophotometer.

Eyes from 7-week old wild-type and Rds+/- animals were fresh frozen in OCT compound

(Tissue-Tek, Miles, Elkhart, IN) and 14µm sections were cut using a Leica cryostat. Each slide

contained 3-4 sections of each genotype to ensure wild-type and mutant sections would be handled

identically. Sections were post-fixed in PBS (137mM NaCl, 19mM Na2HPO4, 3mM KCl, 2mM

KH2PO4 [pH7]) with 4% paraformaldehyde. In situ hybridization was performed as described in

(Schaeren-Wiemers and Gerfin-Moser, 1993; Rattner and Nathans, 2005). Briefly, 150ng of either

the antisense or sense probe was added to the slide and incubated at 60°C overnight. A probe for

Chx-10 was used as a positive control to ensure equivalent RNA abundance in mutant and wild-type

sections (Horsford et al., 2005). After washing in 2X SSC and .2X SSC sections were blocked with

1% goat serum for 1 hour at room temperature followed by incubation with 1:2000 dilution of anti-

89

Dig antibody (Roche, Laval, Quebec) in blocking solution overnight at 4°C. Following four 10

minute washes in TBS, sections were washed in B3 buffer (0.1M NaCl, 0.1 Tris-HCl pH9.5, 0.05%

MgCl, 1% tween) with 0.02g of levamisole to inhibit endogenous alkaline phosphatase activity.

Sections were then incubated in NBT/BCIP (200µl for every 10 mls in B3) (Roche, Laval, Quebec)

for 1 hour, then sections were washed in PBS, fixed in 3.7% formaldehyde (v/v), 10mM MOPS,

0.02mM EGTA (ethylene glycol tetraacetic acid), 1mM MgSO4 pH7.4 for two hours, dehydrated in

EtOH, washed with xylene and mounted in permount (Sigma-Aldrich, Oakville, Ontario).

3.3.7 Laser capture microdissection

Eyes from 7-week old wild-type and Rds+/- animals were fresh frozen in OCT compound

(Tissue-Tek, Miles, Elkhart, IN) and 14µm sections were cut using a Leica cryostat and placed on

laser capture slides (Molecular Machines and Industries, Haslett, Michigan). Sections were allowed to

dry overnight and stained using a hematoxylin and eosin kit specially designed for laser capture

microdissection (Molecular Machines and Industries, Haslett, Michigan) to visualize all different cell

layers in the retina. The outer nuclear layer and inner nuclear layer were isolated separately according

to manufacturer instructions using a MMI Laser Microdissection/Epifluorescence Microscope Zeiss

Axiovert 200 inverted fluorescence microscope equipped with a Sony 3 chip CCD camera. RNA

was extracted using the Arcturus picopure RNA isolation kit (Arcturus Engineering Inc., Mountain

View, CA) and cDNA synthesis was performed using the Qiagen Quantitect reverse transcription kit

(Qiagen, Mississauga, Ontario). Quantitative PCR was performed as previously described.

3.4 Results

Up-regulated members of a putative IL-6 pathway were initially identified in Rds+/- retinas

following microarray analysis, then confirmed using quantitative real-time PCR (qPCR) (see Chapter

2); specifically, the Osmr (2.6-fold) Stat-3 (2.3-fold) Socs-3 (2.8-fold), α2M (3.7-fold), Gfap (3.1-

fold) and C/EBPδ (3.2-fold) were demonstrated to be up-regulated (Table 3.2). The transcript level

of Osm, the ligand specific to the Osmr, was then examined by qPCR analysis since probes for this

gene were not present on the microarrays employed; Osm mRNA was 2-fold increased in Rds+/-

retinas (Table 3.2). To ascertain whether the up-regulation of Osm and the Osmr was a specific

response among IL-6 cytokines, I examined the levels of the other IL-6 cytokines and receptors

were examined in Rds+/- compared to wild-type retinas. Of the six receptors and 8 cytokines in the

90

Table 3.2: Quantification of selected differentially expressed transcripts in three different mouse models of IPD

All fold-differences were calculated based on the average difference of a triplicate repeat microarray or qPCR experiment; p values were determined using a two-tailed t test. The microarray experiment yielded multiple fold differences for Gfap and Stat-3 due to the presence of additional probes on the microarray. Abbreviations: Oncostatin M (Osm), Leukemia inhibitory factor (Lif), Oncostatin M receptor (Osmr), Leukemia inhibitory factor receptor (Lifr), Glycoprotein 130 (gp130), Signal transducer and activators of transcription-3 (Stat-3), Suppressor of cytokine signaling-3 (Socs-3), Alpha-2-macroglobulin (a2M), CCAAT enhancer binding protein delta (C/EBPδ), Glial fibrillary acidic protein (Gfap).

91

IL-6 family, I found Lif to be 3-fold up-regulated, while gp130 and the IL-6 receptor were up-

regulated 1.3-fold and 1.4-fold, respectively, in Rds+/- retinas (Figure 3.1). Interestingly, the

microarray analysis failed to identify increased Osm, Lif, gp130 or IL-6r, suggesting that the

microarray analysis did not identify all differentially expressed genes. Thus, these results show that a

specific subset of IL-6 cytokines and receptors are up-regulated in Rds+/- retinas. Of particular

interest are Osm, the Osmr, and Lif since they are differentially expressed to the greatest degree. The

oncostatin ligand and its receptor, Osmr, also represent the only up-regulated cytokine-receptor pair.

To establish whether the up-regulation of Lif, Osm, the Osmr and their downstream putative

pathway members are a general response to mutations leading to IPDs, I then interrogated two

additional well-characterized mouse models of IPD, to determine if the same IL-6 cytokine

responses were occurring. The two models include a mouse line expressing a human mutant

rhodopsin trangene with a proline to serine substitution at amino acid 347 (P347S) (Li et al., 1996),

and the Rd1 model, which has a defect in β-phosphodiesterase (Chang et al., 2002). RNA was

harvested from both models at an early time-point when ~60% of PRs were present in the P347S

model while >90% of PRs were present in the Rd1 model. Early time-points were studied to

minimize the secondary effects of apoptosis. All the same IL-6 cytokine-related molecules up-

regulated in the Rds+/- retina were found to be differentially expressed in both P347S and Rd1-/-

retinas (Table 3.2), with two notable exceptions: 1) Osm was up-regulated in the Rds+/- and P347S

models but not the Rd1 -/- model, and 2) the Lifr, which was not differentially expressed in the Rds+/-

or P347S models, was down-regulated in the Rd1-/- model by 1.6-fold (Table 3.2). Interestingly, there

was a positive correlation between the rate of degeneration of the IPD model and the magnitude of

fold change for the examined pathway members. For example, the Osmr was 2.6-fold up-regulated in

the comparatively slow Rds+/- model, while it was up-regulated 3.4-fold and 5.6-fold in P347S and

Rd1-/- retinas, respectively; P347S retinas exhibit an intermediate rate of cell death between Rds+/- and

Rd1-/- retinas. In summary, many components of the IL-6 putative pathway were up-regulated in all

three models of IPD, suggesting that their up-regulation is a general response to the presence of a

PR mutation. Furthermore, the degree to which the pathway members were up-regulated correlated

with the rate of degeneration.

To determine if the changes in the expression of IL-6 cytokines related mRNAs led to

elevations of the cognate proteins in Rds+/- retinas, I performed immunoblot analyses. The Osmr

and Stat-3 exhibited a 3.1-fold up-regulation and 2.6-fold up-regulation in Rds+/- retinas,

92

Figure 3.1: Several IL-6 cytokines and receptors are up-regulated in the retinas of Rds+/- mice Quantitative PCR indicates that Oncostatin M (Osm), Oncostatin M receptor (Osmr), Glycoprotein 130 (Gp130), leukemia inhibitory factor (LIF), IL-6 receptor (IL6r) are up-regulated in Rds+/- retinas; the other IL-6 cytokines and receptors were either not detectable by qPCR or were not differentially expressed. All fold-differences were calculated based on the average difference of a triplicate repeat microarray or qPCR experiment; p values were determined using a two-tailed t test. The Lif receptor (Lifr) can bind Cardiotrophin-1 (Ct-1), Cardiotrophin-like cytokine (Clc), and Lif, while Ciliary neurtrophic factor receptor (Cntfr) can bind Cntf and neuropoietin (Np). All other receptors can bind only a single ligand.

93

Figure 3.2: Rds+/- retinas exhibit increased IL-6/Jak-Stat signaling Retinal proteins from wild-type (lanes 1-3) and Rds+/- mice (lanes 4-6) were extracted from 7 week old mice. Homogenized retinas were subject to immunoblot assays to detect members of the Jak-Stat signaling pathway. Densitometry analysis revealed: a 3.1 fold up-regulation of Osmr in Rds+/- (n=3, 25048 pixels+/- 3439) compared to Rds+/+ retinas (n=3, 8021 pixels +/- 712; p<0.01); ERK-1,2 and pERK-1,2 levels were unchanged in Rds+/- retinas. Stat-1 was 1.9-fold up-regulated in Rds+/- (n=3, 34941.2 pixels+/-6186) compared to Rds+/+ retinas (n=3, 18427.4 pixels+/-8917; p<0.02). A 5.85-fold up-regulation of pStat-3 in Rds+/- (n=3, 17518 pixels +/- 1870) compared to Rds+/+ (n=3, 2993 pixels +/- 639; p<0.002); and Stat-3 displays a 2.6 fold up-regulation in Rds+/- retinas (n=3, 15464 pixels +/- 1730) compared to Rds+/+ retinas (n=3, 5852 pixels +/- 1482; p<0.015). All pixels were followed by +/- standard error of the mean and all p values were calculated using a two-tailed t test.

94

respectively (Fig. 3.2), concordant with the qPCR results. Since the activity of the putative pathway

depends on the phosphorylation state of Stat-3 (Stat-3 must be phosphorylated to function as a

transcription factor (Heinrich et al., 1998; Levy and Lee, 2002)), I also examined the levels of

phosphorylated Stat-3 (pStat-3) in wild-type and Rds+/- retinas. I found a 5.85-fold up-regulation of

pStat-3 in Rds+/- retinas (Fig. 3.2), an increase distinctly higher then the 2.6-fold increase observed

for Stat-3. Interestingly, pStat-1, which is capable of forming heterodimers with pStat-3 (Heinrich et

al., 1998), was below the level of detection in Rds+/- and wild-type retinas, although Stat-1 was up-

regulated by 1.9-fold in the retinas of Rds+/- mice. In conclusion, these results show that protein

levels of key polypeptides of the putative pathway are up-regulated in Rds+/- retinas, and that the

putative pathway that includes pStat-3 is more active in Rds+/- retinas.

The IL-6 family of cytokines are also capable of directly activating the extracellular signal-

regulated kinases 1,2 (ERK-1,2) (Kamimura et al., 2003; Song et al., 2003). To determine if ERK-1,2

are up-regulated in Rds+/- retinas, I again used immunoblot analysis. No difference in total ERK-1,2

abundance or in that of the active phosphorylated form of ERK-1,2 (pERK-1,2) was identified,

suggesting that ERK-1,2 are not activated in this context (Fig. 3.2).

Müller glia have been shown to express Stat-3 and pStat-3 in response to exogenous cytokine

treatment (Peterson et al., 2000; Song et al., 2003; Ueki et al., 2008). To ascertain the cell type(s) in

which up-regulated Osmr, Stat-3 and pStat-3 occurs in IPD, I first performed immunostaining of

Rds+/- retinas. The Osmr was up-regulated in both astrocytes and Müller glia of Rds+/- retinas (Fig 3.3

A,B), as determined by co-localization studies with glial fibrillary acidic protein (Gfap), a marker of

Müller glia projections and astrocytes (Figure 3.3 C,D) (Chen and Weber, 2002). In contrast, to wild-

type and Rds+/- retinas, no Osmr staining was observed in the Osmr-/- retinas, confirming the

specificity of this antibody (Fig 3.3 M). While Stat-3 staining in wild-type retinas was below the level

of detection (Fig. 3.3 E), Stat-3 staining was definite in PRs and strong staining was present in

Müller glia projections and nuclei in Rds+/- retinas (Fig. 3.3 F). Müller glia nuclei were identified using

p27kip1, which co-localized with Stat-3 (Fig. 3.3 G,H). Strong pStat-3 staining was observed in Müller

glia nuclei while PR nuclei were weakly stained. Interestingly, not all Müller glia nuclei in Rds+/-

retinas exhibited pStat-3 staining (Fig. 3.3 J), although all the Müller glia stained for Stat-3; instead

“hotspots” of pStat-3 expression were evident; these pStat-3 “hotspots” co-localized with a subset

of p27kip1 positive Müller glia (Fig. 3.3 K, L) (C. Jiang, McInnes Lab, unpublished). Even when

overexposed, only a few nuclei adjacent to the “hotspots” were weakly stained, but the level of

pStat-3 in the majority of Rds+/- Müller nuclei was below the level of detection (data not shown);

95

Figure 3.3: Increased Osmr, Stat3 and pStat3 staining in Müller glia of Rds+/- retinas Representative confocal micrographs of retinal expression of Osmr, Stat-3 and pStat-3 in 7 week-old Rds+/- (B-D, F-H, J-L) and Rds+/+ animals (A,E,I). A, Antibodies recognizing Osmr show diffused staining in the inner plexiform layer (IPL) of Rds+/+ retinas. B, increased staining in Müller cell projections (arrow head) as identified by Gfap (C). D, Gfap and Osmr are co-localized. E, Stat-3 is below the level of detection in Rds+/+ retinas. F, Increased Stat-3 staining in the inner nuclear layer (INL) and IPL, characteristic of Müller glia nuclei (arrow) and projections (arrowhead); staining was also observed in the outer nuclear layer (ONL). G, A marker of Müller glia nuclei, p27kip1, co-localized with Stat-3 (H). I, pStat-3 is below the level of detection in Rds+/+ retinas. J, pStat-3 expression in a subset of cells in the INL (arrow), identified as Müller glia (K), (L). M, No Osmr staining was observed in Osmr-/- retinas. N, no staining was observed in the secondary only negative control. Scale bars indicate 10µm (A-D).

96

pStat-3 staining was not observed in any part of wild-type retinas (Fig. 3.3 I). No staining was

observed when immunofluorescence was performed omitting the primary antibody (Fig. 3.3 N).

These results demonstrate that the Osmr and Stat-3 proteins are up-regulated in Rds+/- Müller glia

while pStat-3 is up-regulated in a subset of Müller glia nuclei of Rds+/- retinas.

Lif is a secreted cytokine with autocrine and paracrine capabilities (Cheng and Patterson, 1997;

Nakanishi et al., 2007). To determine which cells produce the Lif transcript in wild-type and Rds+/-

retinas, in situ hybridization using a probe specific to ~1 kb of the 3’ UTR of Lif was performed. All

nuclear layers expressed the Lif mRNA in both wild-type and Rds+/- retinas; however, Lif mRNA

expression was increased by the greatest magnitude in the INL of Rds+/- compared to wild-type

retinas (Fig. 3.4 A,B). No staining was observed when the sense negative control probe was used

(Fig. 3.4 C). To quantify the magnitude of the increase in Lif mRNA expression, I used laser capture

microdissection followed by qPCR. Total cDNA derived from the INL and ONL was tested with a

marker specific to each layer to establish the degree of enrichment achieved using laser capture. Chx-

10, a transcript specific to the inner nuclear layer (Liu et al., 1994) exhibited a 21- fold average

enrichment in INL samples compared to the ONL samples (Fig. 3.5 C), whereas rhodopsin, a PR-

specific marker (Molday and Molday, 1979), was 36-fold enriched in the ONL samples compared to

the INL samples (Fig. 3.5 D), suggesting that the laser capture was effective in isolating material

specific to each of the INL and ONL. Consistent with the in situ hybridization results, Lif was found

to be 4.1-fold up-regulated in Rds+/- compared to wild-type INL samples (p<0.04), while it was not

differentially expressed in Rds+/- compared to wild-type ONL samples (Fig. 3.5 E).

Immunofluorescence analysis was then employed to identify the retinal cell types that expressed

increased amounts of Lif in the Rds+/- retinas. Increased Lif protein staining was observed in Müller

glia projections and PRs of Rds+/- compared to wild-type retinas (Fig. 3.5 F,G); Müller cell

projections were identified by Gfap, which co-localized with Lif (Fig. 3.5 H,I). In summary, my

expression analyses suggest that many cell types produce Lif in wild-type retinas, but that the

greatest increase in Lif expression in Rds+/- retinas occurred in Müller glia.

To identify the cell type(s) that express C/EBPδ transcript, I used laser capture followed by

qPCR analysis. A significant increase in the C/EBPδ transcript was present in the ONL of Rds+/-

retinas, with a 6.2-fold increase in transcript levels (p<0.04), while it was not significantly

differentially expressed in the INL of Rds+/- compared to wild-type retinas (Fig. 3.5 F). Interestingly,

the basal levels of C/EBPδ transcript in the wild-type INL was ~3 fold greater compared to the

97

Figure 3.4: Lif and C/EBPδ are up-regulated in several cell types in the retinas of 7-week

old Rds+/- mice

A, Lif mRNA is present in the outer nuclear layer (ONL), inner nuclear layer (INL) and ganglion cell layer (GCL) of Rds+/+ retinas and is up-regulated in the ONL and INL of Rds+/- retinas (B). C, No staining was observed when a sense negative control was used for either Lif. F, Antibodies recognizing Lif display staining in Müller cell projections of Rds+/+ retinas; increased staining was observed in the ONL and Müller cell projections (arrowhead) in Rds+/- retinas (G), which were identified using Gfap, a Müller cell marker (H). I, Co-localization of Gfap and Lif. J, No staining was observed in the secondary-only negative control. Scale bars indicate 10µm (A-J)

98

Figure 3.5: Lif up-regulation occurs predominantly in the inner nuclear layer (INL), while C/EBPδ up-regulation is observed mainly in the outer nuclear layer (ONL) Laser capture microdissection followed by qPCR was used to quantify Lif and C/EBPδ transcripts in Rds+/- and Rds+/+ retinas. A, Hematoxylin and Eosin stained frozen section prior to laser capture microdissection with intact ONL and INL. B, Following laser capture, cDNA from the ONL and INL were isolated. C, Chx10, a marker of INL cells was present at 37.5 units+/-2.4 (n=3) in wild-type INL cells and 29.0 units+/-3.2 (n=3) in Rds+/- INL cells, while only present at 1.2 units+/-0.42 (n=3) in wild-type ONL cells and 2.0 units+/-0.36 (n=3) in Rds+/- ONL cells. D, Rhodopsin, a marker of ONL cells was present at 138.3 units +/-49.5 (n=3) in wild-type ONL cells and 105.6 units+/-21.1 (n=3) in Rds+/- ONL cells, while only present at 4.9 units+/-2.5 (n=3) in wild-type INL cells and 2.0 units+/-1.3 (n=3) in Rds+/- INL cells. E, Lif was up-regulated 4.1 fold (p<0.04) in the INL of Rds+/- retinas (6.6 units+/-1.5 (n=3) in Rds+/+ INL cells compared to 26.7 units+/-8.4 (n=3) in Rds+/- INL cells , whereas Lif was not differentially expressed between wild-type and Rds+/- ONL cells. F, C/EBPδ was up-regulated 6.2 fold (p<0.04) in the ONL of Rds+/- retinas (1.1 units+/-0.11 in Rds+/+ ONL cells compared to 6.9 units+/-2.5 in Rds+/- ONL cells; n=3), whereas C/EBPδ was not differentially expressed between wild-type and Rds+/- retinas. All units values were followed by +/- standard error of the mean and all p values were calculated using a t test.

99

ONL (Fig. 3.5 F), suggesting that C/EBPδ is normally expressed at low levels in wild-type ONL

cells.

Treatment with IL-6 cytokines have been shown to slow PR cell death in rodent IPD and

light-mediated damage models, concomitant with an increase in pStat-3 levels (LaVail et al., 1992;

Song et al., 2003; Ueki et al., 2008). To determine if endogenous increases mediate a survival response

in mutant retinas, I generated mice carrying both an IPD mutation and loss-of-function mutations

of Osmr, Lif or C/EBPδ. Osmr loss-of-function decreased PR cell survival in two models of IPD: 4-

month old Rds+/-;Osmr-/- retinas had 12.5% fewer PRs compared to Rds+/-;Osmr+/+ (n=9, p<0.05)

(Fig. 3.6 A), whereas 31-day old P347S;Osmr-/- retinas had 13.5% fewer PRs (n=6, p<0.01) compared

to P347S;Osmr+/+ retinas (Fig. 3.6 B). No difference in PR number was observed between 4-month

old Osmr+/+ and Osmr-/- retinas (n=3) (Fig. 3.6 C). These results suggest that the endogenous up-

regulation of the Osmr that occurs in IPD retinas is protective, supporting my hypothesis that the

endogenous activation of IL-6 cytokine pathways in mutant retinas is a survival response. These

results also demonstrate that the loss of the Osmr, in a wild-type background, has no effect on PR

survival.

It has been suggested that the rescue of mutant PRs by cytokine treatment is mediated by

the increased expression of pStat-3 in Müller glia and PRs (LaVail et al., 1992; Song et al., 2003; Ueki

et al., 2008). To determine if pStat-3 levels were affected by the loss of the Osmr, immunoblot

analysis was performed. Surprisingly, there was no significant change in the levels of either Stat-3 or

pStat-3 in Rds+/-;Osmr-/- compared to Rds+/-;Osmr+/+ retinas (Fig. 3.6 D). Similarly, there was no

difference in pERK-1,2 or ERK-1,2 levels in Rds+/-;Osmr-/- compared to Rds+/-;Osmr+/+ retinas (Fig.

3.6 D), suggesting that Osmr-mediated protection is occurring independently of Stat-3 or ERK-1,2.

In contrast to the protective effects of Osmr in two models of IPD, we were surprised to

observe that Lif , a cytokine known to activate Stat-3, was pathogenic in the Rd1-/- model of IPD.

There was a 14% increase in PR cell number in PN13 Rd1-/-;Lif -/- compared to Rd1-/-;Lif +/+ retinas

(n=6, p<0.003) (Fig. 3.7 A); there was no difference in PR cell number between PN13 Lif- /-

compared to Lif +/+ retinas (Fig. 3.7 B). To determine if pStat-3 and total Stat-3 levels were altered as

a result of the loss of Lif, immunoblot analysis was performed. We identified a 1.7 fold decrease in

pStat-3 in PN13 Rd1-/-;Lif -/- compared to Rd1-/-;Lif +/+ retinas, while no significant difference in Stat-

3 levels were observed (Fig. 3.7 C). Similarly, C/EBPδ, a gene down-stream of Stat-3 (Hutt et al.,

2000), was also found to be pathogenic. C/EBPδ loss-of-function increased mutant PR survival:

100

Figure 3.6: In two models of PR degeneration, Osmr is protective in a pStat-3 independent mechanism Representative toluidine blue stained retinal sections (A-C) and immunoblot analysis comparing the total and phosphorylated forms of Stat-3 and ERK-1,2 in Rds+/-;Osmr+/+ and Rds+/-;Osmr -/- retinas (D). A, The outer nuclear layer (ONL) of Rds+/-;Osmr -/- is 12.5% thinner compared to Rds+/-;Osmr+/+ retinas harvested at four months-of-age (Rds+/-;Osmr -/- ONL is 23.8µm +/-.88µm (n=9) versus Rds+/-;Osmr+/+ ONL which is 28.0µm +/-1.6µm (n=9); p<.05). B, The ONL of P347S;Osmr -/- is 13.5% thinner compared to P347S;Osmr+/+ retinas harvested at 31 days of age (P347S;Osmr -/- ONL is 13.9µm +/-0.25µm (n=6) compared to P327S;Osmr+/+ ONL, which is 16.0µm +/-0.67µm (n=6); p<0.01). C, There was no significant difference in ONL thickness between Osmr+/+ compared to Osmr -/- retinas harvested at 4 months of age. Scale bars indicate 25µm. All values were mean ONL thickness+/- standard error of the mean; p values were computed using a two-tailed t test. D, Representative immunoblot showing that the levels of pStat-3, Stat-3, pERK-1,2, and ERK-1,2 are not significantly different in retinal lysates harvested from 7-week old Rds+/-;Osmr+/+ and Rds+/-;Osmr -

/- mice (n=4).

101

Figure 3.7: Lif is pathogenic in the Rd1 -/- model of PR degeneration Representative toluidine blue stained retinal sections (A,B) and immunoblot analysis comparing the total and phosphorylated forms of Stat-3 in Rd1-/-;Lif +/+ and Rd1-/-;Lif -/- retinas (C). A, The outer nuclear layer (ONL) of Rd1-/-;Lif -/- is 14.0% thicker compared to Rd1-/-;Lif +/+ retinas harvested at PN13 (Rd1-/-;Lif -/- ONL is 33.2µm +/-.0.60µm (n=7) versus Rd1-/-;Lif +/+ ONL which is 28.5µm +/-0.59µm (n=6); p<0.003). B, There was no significant difference in ONL thickness between Lif +/+ compared to Lif -/- retinas harvested at PN13. Scale bars indicate 25µm. All values were mean ONL thickness+/- standard error of the mean; p values were computed using an ANOVA. C, Immunoblot showing that the levels of pStat-3 decreases in Rd1-/-;Lif -/- compared to Rd1-/-;Lif +/+ retinas. Densitometry analysis revealed: a 1.7-fold down-regulation of pStat-3 in Rd1-/-;Lif -/- (n=4, 18730 pixels +/- 2227) compared to Rd1-/-;Lif +/+ (n=4, 33318 pixels +/- 2744; p<0.05). There was no significant decrease in Stat-3 levels. p values were computed using a paired two-tailed t test.

102

8-month old Rds+/-;C/EBPδ -/- retinas had 18% more PRs compared to Rds+/-;C/EBPδ+/+ (n=5,

p<0.005). The loss of C/EBPδ in a wild-type background had no effect on the ONL thickness (Fig.

3.8 B).

Discussion

We have identified several up-regulated IL-6 cytokines, receptors and downstream pathway

members in three models of IPD. For most of these molecules, the magnitude of the up-regulation

correlated with the rate of cell death exhibited by each model. For example, the Osmr transcript was

up-regulated 5.6-fold in the faster Rd1-/- model compared to a 3.4-fold increase in the intermediate

P347S model, and a 2.6-fold up-regulation in the Rds+/- model, the slowest degeneration of these

three models. These findings are consistent with the IL-6 pathway functioning in response to cell

stress and may represent a global response to IPD. One notable exception to this trend was the

oncostatin ligand, which was up-regulated in the Rds+/- and P347S but not the Rd1-/- model of IPD.

This result was unexpected because this cytokine can only bind Osmr (Heinrich et al., 2003), which

was up-regulated in all three models examined. If the abundance of the Osmr is rate-limiting in the

mutant retina, then increased signalling could occur due to the increase in receptor abundance,

without an increase in the cytokine. However, other IL-6 receptors do not appear to be rate-limiting,

as illustrated by the increase in Stat-3 phosphorylation upon cytokine addition (Peterson et al., 2000;

Song et al., 2003). Alternatively, Osmr signalling might only occur in a subset of IPD models.

The expression of all IL-6 cytokines and receptors was assayed in Rds+/- and wild-type

retinas. Of the 8 cytokines and six receptors in the IL-6 family, only Osm, Lif, Osmr, IL-6r and gp130

transcripts were identified as up-regulated, suggesting a specific IL-6 response rather than a general

up-regulation of the whole family of IL-6 cytokines and receptors in the mutant retina. The

observation that the Osm transcript was up-regulated in the Rds+/-and P347S but not the Rd1-/- model

of IPD, is consistent with another report which concluded that different models of IPD can

differentially activate specific members of the IL-6 family of cytokines/receptors; for example, Lif

and Clc mRNA were up-regulated in the Rd1-/- and Vpp-/- models of IPD, while Cntf was found to

be up-regulated only in the Rd1-/- model (Samardzija et al., 2006). Taken together with our findings,

Lif and Osmr up-regulation may be a general response to IPD, while other cytokines such as Osm,

Clc, and Cntf may be up-regulated in a subset of IPD models.

103

Figure 3.8: C/EBPδ is pathogenic in the Rds+/- model of PR degeneration

Representative toluidine blue stained retinal sections from 8-month old animals. A, The ONL is 18.2% thicker in Rds+/-;C/EBPδ -/- compared to Rds+/-;C/EBPδ+/+ mice (Rds+/-;C/EBPδ -/- ONL is 28µm+/-1.7 compared to Rds+/-;C/EBPδ+/+ ONL which is 22.9µm +/-0.38; (n=5, p<0.006). B, The ONL is the same thickness in C/EBPδ+/+ and C/EBPδ -/- animals (n=3). Scale bars indicate 10µm. All values were mean ONL thickness+/- standard error of the mean; p values were computed using a two-tailed t test.

104

The up-regulation of Osm, Lif and the Osmr transcripts in Rds+/- retinas suggested that the

signalling pathways downstream of the IL-6 cytokines may be more active in IPDs. To investigate

this possibility, I examined the activity of the ERK-1,2 and STAT pathways, the two major pathways

downstream of the IL-6 family of cytokines (Kamimura et al., 2003). On examining the

phosphorylation state of Stat-1, Stat-3 and ERK-1,2, I found that only pStat-3 was significantly up-

regulated. Interestingly, ERK-1,2 has been shown to be up-activated in Cntf-treated retinas

(Peterson et al., 2000) and in light-induced PR degeneration but not in IPD models (Samardzija et

al., 2006). Similarly, pStat-1, a pro-apoptotic STAT family member, has been identified as up-

activated in retinas after light-mediated damage, but not in IPD models (Samardzija et al., 2006). My

results are consistent with these other reports and suggest that Stat-3 may have a key role in

inherited PR degeneration, while Stat-1 and ERK-1,2 may respond to light-mediated damage but not

to genetically mediated damage.

Phosphorylated Stat-3 has been shown to be up-regulated in Müller glia of retinas treated

with IL-6 cytokines. Wild-type animals treated with Cntf (Wang et al., 2002), axokine (a Cntf analog)

(Peterson et al., 2000), or Lif (Ueki et al., 2008) and transgenic rats carrying the rhodopsin mutation

S334ter treated with cardiotrophin-1 (Song et al., 2003) all displayed increased Stat-3 and pStat-3 in

Müller glia. Only Lif treatment resulted in increased Stat-3 and pStat-3 staining in wild-type PRs

(Ueki et al., 2008). While a previous report has shown increased Stat-3 and pStat-3 protein in the

retinas of the Rd1-/- and Vpp-/- mouse models of IPD (Samardzija et al., 2006), the localization of

Stat-3 or pStat-3 has not been reported in any IPD model. We found that the endogenous protein

levels of Stat-3 and pStat-3 were up-regulated in the Müller glia of Rds+/- retinas. Interestingly, in

Rds+/- retinas, the increase in pStat-3 was detected predominantly in a subset of Müller glia, creating

foci or “hotspots” of pStat-3 that were situated at fairly regular intervals throughout the retina (C.

Jiang, McInnes Lab, unpublished). In contrast, Stat-3 was increased uniformly throughout all Müller

glia of the Rds+/- retinas. The origin of these pStat-3 “hotspots” is not currently known. One

hypothesis is that glia in close proximity to a dying neuron might respond through Stat-3

phosphorylation. Immunofluorescence analysis suggests Stat-3 might also be present in PRs. The

presence of the Osmr in Müller glia suggests that signalling from the Osmr can activate Stat-3 in

these cells; the lack of Osmr staining in PRs suggests that any PR specific activation of Stat-3 would

have to occur through an Osmr-independent mechanism. Since mutant PRs are protected by

cytokine treatment, Wahlin et al. suggested that the up-regulation of cytokine receptors in PRs as a

possible mechanism accounting for the PR-specific protective effects of neurotrophin treatment

105

(Wahlin et al., 2001). Our results do not support this hypothesis for the Osmr, which was not

detected in PRs. Instead, our results support the notion of the Müller cell hypothesis, which

proposes that Muller glia activation represents a survival response for mutant PRs (Harada et al.,

2000; Zack, 2000).

Lif protein was identified as up-regulated in Müller glia and PRs; Lif mRNA was also present

in all inner nuclear layer cells and ganglion cells. One interpretation of these results is that although

multiple cell types are capable of producing Lif, the receptor is present predominantly on Müller glia

and PRs. Treatment with Lif increases pStat-3 in both Müller glia and PRs, supporting the

hypothesis that Lif can bind to receptors on those cell types (LaVail et al., 1998; Ueki et al., 2008);

however, there is no published data on the localization of the Lifr in the retina. C/EBPδ is a

transcription factor known to be activated by pStat-3 (Yamada et al., 1997b; Hutt et al., 2000).

Consequently, I initially interpreted the microarray and qPCR results which indicated that C/EBPδ

was up-regulated in the Rds+/-, P347S and Rd1-/- retinas as being a result of Osmr activation. I

therefore extrapolated that the increase in C/EBPδ mRNA expression would be in Müller glia.

Although I did identify C/EBPδ transcript expression in the inner nuclear layer (which includes

Müller glia nuclei) of Rds+/- and wild-type retinas, I observed a significant increase in

C/EBPδ mRNA expression in the PRs of Rds+/- mice using laser capture microdissection and

qPCR. These results suggest that the up-regulation of C/EBPδ transcript in PRs must be mediated

by another cytokine receptor, since I was unable to detect the Osmr in PRs (Fig. 3 A). Thus the

putative IL-6 pathway in the retina would appear to be more complex than the simple linear

pathway proposed in Chapter 2, because 1) the loss of the Osmr does not affect the level of pStat-3

expression; 2) C/EBPδ transcript was up-regulated only in the ONL of Rds+/- retinas, whereas

Osmr up-regulation was only detected in Muller cells and astrocytes of Rds+/- retinas. Rather, it

would appear that the level of pStat-3 is maintained by IL-6 receptors other than Osmr, and/or that

pStat-3 regulation varies in different retinal cell types.

The absence of the Osmr in the Rds+/- and P347S models of IPD led to increased PR cell

death, suggesting that the Osmr is required for an optimal survival response in retinas with PR

mutations. Many reports have shown that IL-6 cytokines are protective in both light-induced and

IPD retinas; pStat-3 and/or pERK-1,2 have been hypothesized as possible mediators of PR

protection (Liu et al., 1998; Song et al., 2003; Samardzija et al., 2006; Ueki et al., 2008). Interestingly,

we did not observe any changes in either pStat-3 or pERK-1,2 levels in Rds+/-;Osmr-/- compared to

106

Rds+/-;Osmr+/+ retinas, suggesting that the Osmr-mediated protection may not be mediated solely or

predominantly by pStat-3 and pERK-1,2. Further work is required to determine whether Stat-

3/pStat-3 are essential to the endogenous retinal responses to the presence of a PR mutation. If they

are, then the identification of downstream targets of pStat-3 will provide valuable insight into the

molecular mechanisms that underlie the protective effect.

The failure of pStat-3 levels to decrease upon loss of the Osmr could be interpreted in

several ways. First, if signalling through the Osmr is not the primary pathway resulting in Stat-3

phosphorylation, then the loss of the Osmr would have a negligible affect on pStat-3 levels. Second,

although the overall retinal abundance of pStat-3 did not change in the absence of the Osmr, its

levels in specific retinal cells may have decreased, but was not detectable by immunoblotting. Third,

a compensatory system might exist such that other factors are up-regulated in the absence of the

Osmr. The transcript levels of Lif, Lifr, Cntf, and Ct-1 were examined and were not up-regulated in

response to the loss of the Osmr in Rds+/- retinas, suggesting that these IL-6 cytokines are not

participating in any compensatory response that sustains the level of pStat-3.

In contrast to the protective effects of the Osmr in Rds+/- and P347S retinas, we identified

Lif and C/EBPδ as essential components of the pathogenic consequences of a PR carrying an IPD

mutation. A pro-apoptotic role for Lif and C/EBPδ has been reported during mammary

development; for example, during mammary glad involution, Lif signalling results in Stat-3

phosphorylation followed by C/EBPδ up-regulation, which then leads to the activation of pro-

apoptotic genes and the suppression of anti-apoptotic genes (Thangaraju et al., 2005; Clarkson et al.,

2006). C/EBPδ null allele mice also exhibit mammary gland hyperplasia, consistent with reduced

developmental apoptosis (Gigliotti et al., 2003). The fact that the Osmr was up-regulated in Müller

glia while C/EBPδ was predominantly up-regulated in PRs may account for the apparently

contradictory roles of these two members of the same pathway. We hypothesize that C/EBPδ may

be required for the pathogenic effect of PR mutations. In contrast, the up-regulation of the Osmr in

Müller glia mediates a protective response.

The finding that Lif is pathogenic in the Rd1-/- model of IPD was unexpected given the

protective affects of Lif treatment in other IPD models and in light-mediated PR cell death (LaVail

et al., 1998; Ueki et al., 2008). Furthermore, pStat-3 levels were reduced 1.7-fold in Rd1-/-;Lif-/-

compared to Rd1-/-;Lif+/+ retinas, suggesting that maintenance of pStat-3 is not essential for

protection, and/or that, paradoxically in this model, the reduction of pStat-3 was actually protective.

107

One or more alternative mechanisms could account for these results. First, Lif has been established

as an important cytokine involved in PR differentiation during development (Elliott et al., 2006).

Although the total number and histological appearance of PRs was not different between Lif-/- and

Lif+/+ retinas, the lack of Lif during development may have somehow modified the response to the

Rd1 mutation in Rd1-/-;Lif-/- retinas, changing the rate of cell death. Second, Lif may actually be

pathogenic in some models of IPD. Lif treatment is protective in the RCS rat, and two rhodopsin

transgenic rat models (Matthew LaVail, personal communication), whereas Rd1-/- retinas treated with

Lif did not show any PR protection (LaVail et al., 1998). Since the treatment involved only a single

injection of Lif, it is possible that a more constant albeit endogenous exposure to Lif may be

pathogenic in the Rd1-/- model. Since pStat-3 has been shown to be activated in PRs in response to

Lif injection (Ueki et al., 2008), Lif may be capable of activating C/EBPδ in PRs, resulting in PR cell

death. Conversely, the lack of Lif may decrease C/EBPδ levels, leading to increased PR survival.

The finding that Lif and C/EBPδ are pathogenic in Rd1-/- and Rds+/- retinas, respectively,

may have important therapeutic implications. Animal studies indicating Cntf treatment is effective in

delaying PR cell death in animal IPD models led to further investigations of Cntf as a possible

therapy in humans (Sieving et al., 2006). While the efficacy of Cntf treatment in humans remains to

be determined, the possible activation of C/EBPδ in PRs- a potentially pathogenic response- should

also be investigated. The presence of the Cntf receptor on PRs has been identified by one study

(Valter et al., 2003), consistent with the possibility that Cntf treatment could increase C/EBPδ levels

in mutant PRs. The long-term activation of C/EBPδ may have harmful effects and partially mitigate

the protective effects of Cntf mediated through other cells such as Müller glia. If further studies

support a pathogenic role for C/EBPδ in mutant PRs, then the use of reagents such as PR-specific

siRNAs to C/EBPδ, may prolong the survival of mutant PRs. Similarly, the finding that long-term

exposure to increased levels of Lif may be pathogenic has obvious implications for the use of Lif as

a treatment. Further studies are required to confirm the pathogenic role for Lif and determine

whether its pathogenicity is unique to the Rd1-/- model.

108

4 General discussion and future directions

4.1 Significance of findings

Inherited PR degenerations result from mutations in many different functional categories of

genes, but it is unlikely that each mutation triggers a unique pathway that results in cell death.

Historically, research on IPDs has focused on understanding the normal function of the proteins

encoded by IPD genes; however, with few exceptions, this approach has not led to significant

insights into why mutant PRs die. In this thesis I have described an unbiased, whole-genome

approach to identify transcripts with altered expression in the retinas of animals with IPDs. I

focused my attention on a subset of the differentially expressed genes comprising a putative IL-6

signalling pathway because i) they were among the most significantly altered in their expression; ii)

studies by others had shown that some of these genes displayed increased expression that correlated

with- but had not been mechanistically linked to- a protective response initiated by IL-6 cytokines,

suggesting that the responses of these genes I identified might also mediate an endogenous survival

response to the presence of a PR mutation; and iii) many of the differentially expressed genes I

found had also been identified in studies by other groups examining altered gene expression in a

wide variety of IPD and light-damaged models. That these transcripts are up-regulated in a range of

genetic and non-genetic insults to the retina is consistent with my hypothesis that they represent a

general response that are not related to the specific biochemical abnormalities conferred by the

disruption of any particular IPD protein. In none of these other studies, however, was the

pathogenic significance of these changes in gene expression investigated. In my work, in contrast, I

took advantage of the availability of mouse strains carrying loss-of-function alleles of several of the

genes I found to be differentially expressed in IPD models, to examine the role of these genes in the

pathogenesis of PR degeneration.

109

In the second chapter of this thesis, I described the microarray and quantitative PCR (qPCR)

experiments in which I examined the transcriptomes of the Rom1-/- and Rds+/- murine models of

IPD. I sampled the mRNA population of those models at times when >90% of PRs were present in

each model. My studies are one of the first to have identified differentially expressed transcripts in

IPD retinas at time-points before significant PR degeneration has occurred. The differentially

expressed genes represent early responses to a pathogenic mutation; these responses may be

pathogenic, protective or neutral in their relationship to PR death.

I identified up-regulated transcripts in the retinas of Rds+/- mice that I hypothesized to be

participating in a common IL-6 signalling pathway, including the mRNAs of Lif, Osm, Osmr, Stat-3,

C/EBPδ, Socs3, and α2macroglobulin. Since treatment with IL-6 cytokines is neuroprotective in

light-damaged and IPD models, as indicated above, my working hypothesis was that IL-6 signalling

in the IPD retina was also a protective response (LaVail et al., 1992; LaVail et al., 1998).

In chapter three, I further examined the role of members of the putative IL-6 pathway in

IPDs. After examining the expression levels of all eight IL-6 cytokines and six IL-6 receptors in the

retinas of Rds+/- mice, I determined that Lif, Osmr and gp130 were also differentially expressed in

the Rd1-/- and P347S models of IPD, suggesting that an IL-6 cytokine response may occur generally

in IPDs. The observation that phosphorylated Stat-3 (pStat-3) was increased in Rds+/- retinas

suggests that an IL-6 pathway is up-regulated in mutant retinas. Although increased pStat-3 had

been identified in light-damaged retinas and wild-type retinas treated with IL-6 cytokines, my work is

one of the first studies to identify increased pStat-3 in IPD retinas, and the first to establish the

identity of the cell types expressing Lif, Osmr, Stat-3, pStat-3, and C/EBPδ in IPD retinas. While

Osmr was localized exclusively to Müller glia and astrocytes, Stat-3, pStat-3, Lif and C/EBPδ were

expressed in multiple cell types, including PRs, in the Rds+/- retina. The observation that Stat-3 and

its downstream target, C/EBPδ (Yamada et al., 1997a), are both expressed in PRs, suggests that an

IL-6 receptor other than Osmr must mediate the activation of Stat-3 in PRs. My work provides the

first indication that the IL-6 signalling responses in IPD retinas may be more complicated than a

simple linear model in which all components are functioning in a single cell type.

My use of genetic analyses to determine if the loss of Lif, Osmr or C/EBPδ expression

would modify the rate of PR death in IPD models is also novel. As predicted by the protective

effect of IL-6 cytokines on IPDs, the Osmr loss-of-function allele in the Rds+/- and P347S models of

IPD increased PR death, demonstrating that the presence of the Osmr is required for a protective

110

response. Unexpectedly, pStat-3 levels did not decrease in Rds+/-;Osmr-/- retinas (i.e. in the absence of

Osmr), suggesting either that pStat-3 is not solely regulated by Osmr in the mutant retina, or that the

protective role of Osmr does not require a change in the steady-state level of pStat-3 in the whole

IPD retina. Given that C/EBPδ is a downstream target of IL-6 signalling (Hutt et al., 2000), I

originally predicted that C/EBPδ would be a protective molecule. However, my genetic analyses

suggest that C/EBPδ plays a previously unrecognized pathogenic role in the IPD retina, which is

consistent with its pro-apoptotic role during mammary gland development (Thangaraju et al., 2005).

Thus, we propose that C/EBPδ signalling is pathogenic when expressed in mutant PRs and may be

a novel therapeutic target. My finding that the Lif loss-of-function allele in Rd1-/- retinas decreases

the expression of pStat-3 and increases the number of surviving mutant PRs was also unexpected

given that treatment with Lif is protective in light-damage and several IPD models (Mathew Lavail,

personal communication and (Ueki et al., 2008)). Thus, Lif may be pathogenic in the Rd1-/- model.

Alternatively, a secondary affect of Lif loss-of-function in IPD retinas may account for the increased

PR cell number. These results add new insight into IL-6 signalling in the retina. They may also have

therapeutic implications, since an IL-6 cytokine (Cntf) is being tested as a therapy in human retinitis

pigmentosa patients (Sieving et al., 2006).

4.2 Identifying the mechanism(s) of IL-6 mediated photoreceptor protection

The mechanism by which IL-6 cytokines protect PRs is not known, although the correlation

between activation of Stat-3 expression following IL-6 cytokine administration, and the protective

effect of IL-6 cytokines, suggests that pStat-3 mediated signaling may be critical. (Kamimura et al.,

2003; Samardzija et al., 2006). The work presented here and by Samardzija et al. show that Stat-3

activation occurs in many models of IPD, which may be a protective response to a pathogenic

mutation (Peterson et al., 2000; Ueki et al., 2008). I demonstrated that Osmr loss-of-function

increased the rate of PR death in the Rds+/- model, although the levels of both pStat-3 and pERK-

1,2, the two major mediators of IL-6 signalling, remained unchanged in the absence of Osmr. This

result does not directly refute the putative protective role of pStat-3 or pERK-1,2, but it does raise

the possibility that the Osmr may activate other pathways involved in protecting PRs. One strategy

to identify pathways downstream of Osmr would be to compare the retinal transcriptomes of Rds+/-

111

;Osmr+/+ mice and Rds+/-;Osmr-/- mice. Microarray analysis might reveal differentially expressed

transcripts that may be involved in a non-Stat-3 and non-ERK-1,2 pathway downstream of Osmr.

Other molecular components of the IL-6 pathway could also be identified by examining the

retinal transcriptome after IL-6 cytokine treatment of the mutant retina, assuming that cytokine-

mediated PR rescue requires de novo transcription. Treatment with Cntf has been shown to protect

PRs in IPD models; however at high doses, PR morphology and function is abnormal (Liang et al.,

2001). A recent examination of the transcriptional effects of sustained high doses of Cntf in the

murine retina was used to gain insight into Cntf-induced toxicity (Rhee et al., 2007). Rhee et al.

compared the retinal transcriptome of Rds+/- mice treated with adeno-associated virus (rAAV)

expressing Cntf to the retinal mRNA population of untreated Rds+/- mice (Rhee et al., 2007). Among

the 50 significantly differentially expressed genes identified, several members of the

phototransduction cascade were down-regulated, which may account for the reduced ERG activity

in rAAV-Cntf treated retinas (Rhee et al., 2007). Since Cntf slows mutant PR death, the microarray

analysis comparing rAAV-Cntf treated Rds+/- retinas and untreated Rds+/- controls may have also

identified candidate neuroprotective genes. Stat-3 and ERK-1,2, were both up-regulated in treated

retinas, while two pro-apoptotic genes, programmed cell death protein 4 (pdcd4) and programmed

cell death protein 7 (pdcd7), were down-regulated ~2-fold (Rhee et al., 2007). These findings are

consistent with the deceased PR death observed in Cntf-treated retinas (Rhee et al., 2007). While

pdcd7 is a tumour suppressor gene involved in the repression of cell cycle proteins (Goke et al.,

2004), pdcd4 is a pro-apoptotic gene involved in inducing cell death in T-cells (Park et al., 1999).

Further characterization will be required before the role of these genes in the Cntf protection of

mutant PRs can be understood. One drawback to the design of this experiment for the identification

of genes involved in PR protection is the toxic affect of high doses of Cntf on PRs, as indicated by

the reduced ERGs of treated PRs. A better experimental design to identify protective genes would

be to treat Rds+/- retinas with a lower dose of Cntf, using, for example, encapsulated cell technology

to provide prolonged Cntf exposure, and to repeat the microarray experiment on retinas from

treated and untreated Rds+/- animals. The identification of differentially expressed genes may identify

the molecules that mediate neuroprotection following endogenous increases in Cntf expression or

exogenous Cntf treatment.

112

4.3 The mode of IL-6 signalling: direct or indirect action on photoreceptors

The identification of the cell types that respond to IL-6 cytokines is important in order to

understand the mechanism of IL-6 cytokine mediated PR protection. The Müller cell hypothesis

suggests that Müller glia respond to exogenous or endogenous signals to protect PRs from cell death

(Zack, 2000). Data presented in this thesis and by Ueki et al. suggest that IL-6 cytokines may directly

activate downstream responses in PRs (Ueki et al., 2008). There are at least four possible types of

retinal cell responses following IL-6 cytokine administration in IPDs: 1) PRs may respond directly to

IL-6 cytokines; 2) Müller glia may respond to IL-6 signalling and in turn protect PRs; 3) both cell

types may respond to IL-6 signalling; 4) other retinal cell types may respond, although to date there

is no evidence bearing on this possibility.

Consistent with the Müller cell hypothesis, treatment with Cntf or Lif results in the up-

regulation of pStat-3 predominantly in Müller glia concomitant with PR cell protection in light-

damaged or IPD models (Peterson et al., 2000; Ueki et al., 2008). Harada et al. identified several

molecules involved in Müller glia-PR signalling in the light-damage model (Harada et al., 2000).

Müller cells respond to nerve growth factor (NGF) or neurotrophin-3 (NT-3) by modulating the

release of basic fibroblast growth factor (bFGF); increased NGF correlated with decreased bFGF

release by Müller cells and increased PR cell death, while increased NT-3 correlated with increased

bFGF release by Müller glia and PR rescue (Harada et al., 2000). Some of the work presented in this

thesis is also consistent with a Müller glia response influencing PR survival. For example, I found

that the Osmr was exclusively expressed in Müller glia projections, and the absence of the Osmr

increased the rate of PR death in P347S and Rds+/- retinas. However, I also obtained evidence that an

IL-6 response also occurs in mutant PRs: Lif, Stat-3, pStat-3 and C/EBPδ were all up-regulated in

mutant PRs, suggesting that the direct reponses of mutant PRs may also decrease the rate of PR

death. These findings are consistent with work done by Ueki et al. that showed that Lif treatment of

light-damaged retinas resulted in the up-regulation of pStat-3 in both PRs and Müller cells, and in

PR protection (Ueki et al., 2008).

Several approaches could be taken to examine the importance of direct IL-6 signalling in

protecting PRs. Imaging studies to determine whether IL-6 cytokine receptors are present in PRs

would establish a pre-requisite for IL-6 signalling to occur. The observation that Lif treatment

results in pStat-3 activation in PRs is consistent with the presence of Lif receptors on PRs (Ueki et

al., 2008), but this data should be independently confirmed. I have presented results which suggests

113

that the Osmr is not expressed in PRs, while work by others suggests that the Cntf receptor and

gp130 are expressed in both the inner and outer segments of PRs (Beltran et al., 2003; Rhee and

Yang, 2003). However, the expression of the appropriate receptors is only a minimum requirement

for IL-6 signalling to occur in PRs. Several studies have shown that retinas treated with Cntf or

axokine (a Cntf analog) do not exhibit pStat-3 expression in PRs (Peterson et al., 2000; Wang et al.,

2002), despite the presence of all the receptors necessary for Stat-3 signalling to occur.

Consequently, a functional approach is also required to determine if IL-6 signalling occurs in PRs.

A functional approach that directly examines the ability of IL-6 cytokines to activate

downstream signalling molecules in PRs would be strong evidence that IL-6 cytokines act directly on

PRs. Traverso et al. have developed a technique to isolate PRs from a porcine retina and grow them

in culture for two weeks (Traverso et al., 2003). They used this cell culture system to test and

confirm that epidermal growth factor (Egf) and bFGF act directly to protect PRs. Cells grown in the

presence of either factor survived for longer and exhibited increased phosphotyrosine abundance, an

indication of signalling downstream of bFGF or Egf (Traverso et al., 2003). An analogous

experiment could be performed to determine if any of the IL-6 cytokines can affect PR survival in

this assay system, an indication that the cytokine can act directly on PRs. Furthermore, downstream

IL-6 signalling molecules could be directly examined in PRs to determine if cytokine treatment

increased pStat-3 or pERK-1,2 in PRs. If IL-6 cytokines increase the survival of mutant PRs in this

context, and activate Stat-3 or ERK-1,2 signaling molecules, the transcriptomes of isolated wild-type

and mutant PRs could be compared before and after IL-6 treatment, as a first step in dissecting

direct PR-specific responses to cytokine administration.

An in vivo approach could also be taken to determine if IL-6 signalling directly acts on PRs.

Since gp130 is the common co-receptor of all IL-6 cytokine receptors, its presence is necessary to

activate downstream signalling pathways (Heinrich et al., 1998). Transgenic mice expressing a

dominant negative gp130 transgene have been used to assess the role of IL-6 signalling during

antibody production (Kumanogoh et al., 1997). The dominant negative gp130 protein is a truncated

form of the wild-type protein, lacking part of the cytoplasmic domain, which is required for Stat

and Erk binding; expression of the truncated gp130 significantly reduced Stat-3 phosphorylation in

the spleen (Kumanogoh et al., 1997). A dominant negative gp130 transgene could be expressed in

several different retinal cell types of IPD mice. Since IL-6 signalling is required for retinal

development, an inducible truncated gp130 transgene could be created that can be activated in adult

mice (Kistner et al., 1996). If PRs primarily respond to IL-6 signalling, then transgenic mice

114

expressing the dominant negative gp130 transgene in PRs alone should be resistant to the protective

effects of IL-6 cytokine treatment after damaging light treatment or in models of IPD. The

analogous experiment could also be performed in transgenic mice expressing the dominant negative

gp130 transgene in Müller glia to determine if IL-6 signalling in these glia contributes to PR

protection. These experiments may determine whether Müller glia, or PRs, or both, underlie the

protection of mutant or light-damaged PRs mediated by IL-6 cytokines.

4.4 Is Stat-3 signalling a protective response?

Many studies have suggested that Stat-3 signalling is protective to light-damaged and IPD

retinas, based on its up-regulation in Müller cells following IL-6 cytokine treatment (Song et al.,

2003; Ueki et al., 2008). Few studies, however, have addressed the fundamental question of whether

the increase in Stat-3 signalling is mechanistically linked to the protective effect of IL-6 cytokines, or

whether it is merely correlative. Samardzija et al. employed the AG490 Jak kinase inhibitor in an

attempt to prevent the phosphorylation of Stat-3 in IPD retinas (Samardzija et al., 2006). Although

AG490 treated IPD retinas did not exhibit any change in the rate of PR death, the result was not

informative because AG490 treated IPD retinas did not exhibit decreased Stat-3 signalling

(Samardzija et al., 2006). However, light-damaged retinas treated with AG490 did exhibit a

significant reduction in both pStat-1 and pStat-3 and increased PR cell survival (Samardzija et al.,

2006). Since Stat-1 is a pro-apoptotic member of the Stat family, whereas Stat-3 is regarded as an

anti-apoptotic protein (Stephanou and Latchman, 2005) and both proteins exhibited decreased

phosphorylation with AG490 treatment, it was not possible to determine if the decrease in pStat-1

or pStat-3 was responsible for the increased PR cell survival. In the experiments presented in this

thesis, I employed a genetic approach to determine if the Osmr or Lif loss-of-function mutants

affected Stat-3 phosphorylation in several IPD models, but these experiments are also correlative in

nature. Consequently, other approaches must be used to ascertain if Stat-3 signalling is protective,

pathogenic or a bystander to cell death.

A genetic approach could be used to determine if Stat-3 directly mediates PR protection. A

Stat-3 floxed allele has already been created to study the effect of Stat-3 in mammary glad

development and could be employed in the retina (Chapman et al., 2000). Since Stat-3 is known to

be expressed in Müller glia and PRs (Ueki et al., 2008), this approach could also be used to identify

not only whether Stat-3 signalling is protective, but in which cell type its expression is required for

115

Stat-3 signalling to be protective. An IPD model lacking Stat-3 in PRs could be created. After IL-6

treatment, the rate of cell death could be compared between an IPD model lacking Stat-3 in PRs and

the same IPD model with normal Stat-3 PR expression. If the rate of cell death was affected by the

presence of Stat-3 in PRs, then a direct relationship between PR survival and Stat-3 expression could

be made. A similar methodology could be employed to prevent Stat-3 expression in Müller glia to

determine if Stat-3 expression in Muller glia is responsible for neuroprotection upon IL-6 treatment.

If Stat-3 expression in Müller cells is protective, then another question is to determine what

signal(s) Müller cells are releasing that is mediating PR protection. If both PRs and Müller glia are

capable of responding to IL-6 signalling, resulting in a decrease in PR death, then the role of the

Müller glia signalling may be to amplify the IL-6 cytokine signal. IL-6 cytokines are known to

function as part of a positive feedback loop (Davey et al., 2007). Thus, upon binding an IL-6

cytokine, Müller glia may respond by secreting more of the cytokine, which would increase the

amount of cytokine available to bind to PRs, thereby increasing the neurotrophic effect.

Alternatively, Müller glia may respond to IL-6 cytokines by secreting another protein, which then

protects PRs from cell death. Identifying protective factors produced by Müller glia could be

performed by exposing a Müller cell culture to an IL-6 cytokine and then examining changes in

transcription using microarray technology. A mass spectrometry based method to identify any

proteins that are up-regulated in Müller cell media in response to cytokine treatment might also be

informative, although the sensitivity of this approach may be limited.

4.5 Determining whether C/EBPδ expression in photoreceptors is pathogenic

The finding that C/EBPδ is a pathogenic molecule was unexpected given that its

expression is activated by pStat-3 (Hutt et al., 2000), a putative protective molecule. At least two

possible explanations may account for this data: 1) C/EBPδ expression in the retina may be

regulated by factors in addition to Stat-3 and this alternative regulatory pathway(s) may be

pathogenic in IPDs; 2) in some physiological contexts, perhaps increased pStat-3 expression is,

paradoxically, pathogenic. Quantitative PCR performed on isolated PRs and inner nuclear layer

(INL) cells revealed that C/EBPδ expression is significantly increased in the PRs of Rds+/- vs. wild-

type mice. C/EBPδ has been identified as a pro-apoptotic factor during mammary gland involution

116

(Thangaraju et al., 2005), providing a precedent for the pro-apoptotic role I observed in the IPD

retina.

To support the evidence that C/EBPδ is a pathogenic factor in the IPD retina, a transgenic

approach could be used to over-express C/EBPδ moderately in wild-type and IPD PRs. A

significant increase in PR death with increased C/EBPδ PR expression in the IPD and wild-type

retinas would be supportive of the pathogenic role for C/EBPδ in PRs. Alternatively, increased PR

death may only occur in the context of increased C/EBPδ expression in the IPD retina, which

would suggest a pathogenic role for C/EBPδ in the mutant PR. A similar strategy could be

employed to determine if C/EBPδ expression in Müller glia is pathogenic to PRs. My findings

suggest that C/EBPδ is a pathogenic factor in IPDs; however, since Rds+/-;C/EBPδ-/- mice lack

C/EBPδ expression in all cell types, it is not possible to conclude that the PR expression of

C/EBPδ is pathogenic. For example, C/EBPδ expression in Müller glia alone may cause the release

of a pathogenic factor resulting in PR cell death. To distinguish between the roles of C/EBPδ in

PRs vs. Müller cells, one could delete C/EBPδ in either PRs or Müller cells alone, as I described

above with reference to Stat-3. If the loss of C/EBPδ only in PRs of an IPD model resulted in a

slower rate of PR death, then C/EBPδ would be confirmed as a direct mediator of pathogenesis in

the mutant PR.

The downstream targets of C/EBPδ represent candidate genes involved in IPD PR death. A

microarray experiment to identify genes regulated by C/EBPδ might lead to the identification of

genes and pathways that are pathogenic to PRs.

The pathogenic role of C/EBPδ in PRs could have important therapeutic implications. If

IL-6 cytokines act directly on PRs, then pathogenic C/EBPδ expression may be an unavoidable

consequence of IL-6 treatment, which is protective overall. However, if Müller cells are the primary

mediators of PR protection as a result of IL-6 cytokine treatment, then IL-6 cytokines should be

screened for their ability to up-regulate C/EBPδ in PRs and not used as a potential therapy for

patients suffering from PR degeneration. Different IL-6 cytokines may have different capacities to

up-regulate C/EBPδ in PRs. PR cultures could be treated with each IL-6 cytokine and the degree of

C/EBPδ up-regulation could be assayed and compared. All other things being equal, the IL-6

cytokine that activates C/EBPδ in PRs the least may be the best candidate for human therapy. Some

117

studies have indirectly addressed this question for Lif and Cntf. While Cntf treatment did not lead to

pStat-3 expression in PRs, Lif expression did, suggesting that Cntf might not be able to activate

C/EBPδ in PRs (Peterson et al., 2000; Ueki et al., 2008).

4.6 Reconciling the pathogenic and protective affects of IL-6 signalling

My finding in Chapter 3 that Lif loss-of-function decreased the rate of cell death in Rd1-/-

retinas was unexpected, since Lif is neuroprotective in both light-damaged and IPD models (LaVail

et al., 1998; Ueki et al., 2008). How can Lif be both protective and pathogenic in these different

contexts? While the lack of Lif during development had no effect on PR number in the work I

reported, it may have changed the responsiveness of PRs to the Rd1 mutation. Lif expression has

been shown to influence blood vessel formation in the retina (Ash et al., 2005; Kubota et al., 2008).

Transgenic mice expressing Lif exhibited a decrease in normal vasculature, resulting in retinal

ischemia and pathogenic neovascularization (Ash et al., 2005). In the Lif -/- murine retina, increased

microvessel density and a resistance to hyperoxic insults was documented (Kubota et al., 2008).

Since oxidative stress is observed in Rd1-/- retinas (Hackam et al., 2004b; Ahuja-Jensen et al., 2007;

Sharma and Rohrer, 2007), the loss of Lif in that model might also protect mutant PRs, accounting

for increased PR number in retinas of Rd1-/-;Lif-/- mice compared to Rd1-/-;Lif+/+ mice. Thus, an

alteration in the retinal vasculature may be a secondary effect of Lif loss-of-function, accounting for

the decreased rate of cell death. Mutant retinas treated with Lif, on the other hand, would not

exhibit altered vasculature and Lif could exert its protective affects.

4.7 Stat-3 may be involved in retinal regeneration in lower vertebrates

The regenerative ability of the adult mammalian retina is limited (Fischer and Reh, 2003). In

contrast, lower vertebrates can undergo retinal regeneration in response to injury (Gilbert, 2006).

For example, the retina of the teleost fish regenerates after surgical lesions to produce a functional

retina (Fischer and Reh, 2003). Similarly, the PR death elicited by light-damage is followed by

regeneration in zebrafish (Bernardos et al., 2007). Using a transgenic cell marking strategy, Müller

cells were observed to re-enter the cell cycle and produce progenitors that could differentiate into

rods and cones (Bernardos et al., 2007).

118

To identify some of the molecular events occurring during the regeneration process,

microarray analysis was performed on retinal samples taken at several time-points after zebrafish

were exposed to damaging light (Kassen et al., 2007). Stat-3 was identified as significantly up-

regulated at all time-points, including during the periods of cell death, progenitor proliferation,

migration, and PR differentiation. Immunoblot analysis confirmed that pStat-3 was also up-

regulated; significantly, although pStat-3 was up-regulated greater than 3-fold at all time-points

examined, the peak of pStat-3 expression (~7 fold up-regulated) occurred at 68 hours post light

damage, a time-point characterized by progenitor proliferation and migration (Kassen et al., 2007).

Co-localization experiments demonstrated that the Stat-3 positive cells were Müller glia, a subset of

which also expressed PCNA, a marker of cell proliferation (Kassen et al., 2007). In this context, Stat-

3 may represent one of the first genes to be up-regulated during regeneration (Kassen et al., 2007).

The co-localization of PCDNA in a subset of Stat-3 labeled Müller glia may indicate that Müller glia

proliferation and differentiation occurs in an asynchronous fashion (Kassen et al., 2007).

Retinal regeneration has also been observed in the avian retina after cell death in the inner

retina was induced using N-methyl-D-aspartate (NMDA) (Fischer and Reh, 2003). Two days after

NMDA injection, a large number of mitotically active Müller glia were observed (Fischer and Reh,

2003). Fate mapping experiments suggested that Müller cells de-differentiate, undergo one round of

mitosis and then become one of three possible cell types: neurons, Müller glia, or undifferentiated

progenitor cells (Fischer and Reh, 2003). Subsequent studies showed that Müller cells are capable of

de-differentiation, cell division, and differentiation into ganglion cells (Fischer and Reh, 2003).

Similar experiments in rats completely lacking PRs after N-methyl-N-nitrosourea injection suggest

that some Müller derived cells expressed rhodopsin, possibly representing an attempt at regeneration

in the mammalian retina (Wan et al., 2008). Rodent Müller cells, grown under conditions normally

used for neural stem cell cultures, proliferate into sphere colonies and express genes normally found

in neuronal progenitor cells, confirming the ability of mammalian Müller glia to de-differentiate and

proliferate in vitro.

I initially hypothesized that the up-regulation of Stat-3 I observed in Rds+/- retinas was a

protective response. Given the aforementioned data, an alternative explanation is the increased Stat-

3 might represent a vestigial mammalian attempt to regenerate a damaged retina from Müller cells.

119

References

Abd-El-Barr MM, Sykoudis K, Andrabi S, Eichers ER, Pennesi ME, Tan PL, Wilson JH, Katsanis N, Lupski JR, Wu SM (2007) Impaired photoreceptor protein transport and synaptic transmission in a mouse model of Bardet-Biedl syndrome. Vision Res 47:3394-3407.

ABI (2001) User Bulletin #2: ABI Prism 770 sequence detection system. In: (Biosystems A, ed).

Abramoff MD, Magelhaes PJ, Ram SJ (2004) Image Processing with ImageJ. Biophotonics International 11:36-42.

Acland GM, Aguirre GD, Ray J, Zhang Q, Aleman TS, Cideciyan AV, Pearce-Kelling SE, Anand V, Zeng Y, Maguire AM, Jacobson SG, Hauswirth WW, Bennett J (2001) Gene therapy restores vision in a canine model of childhood blindness. Nat Genet 28:92-95.

Affymetrix (2004) Array design and performance of the GeneChip(r) mouse expression set 430. In.

Ahuja-Jensen P, Johnsen-Soriano S, Ahuja S, Bosch-Morell F, Sancho-Tello M, Romero FJ, Abrahamson M, van Veen T (2007) Low glutathione peroxidase in rd1 mouse retina increases oxidative stress and proteases. Neuroreport 18:797-801.

Allison DB, Cui X, Page GP, Sabripour M (2006) Microarray data analysis: from disarray to consolidation and consensus. Nat Rev Genet 7:55-65.

Anderson DH, Stern WH, Fisher SK, Erickson PA, Borgula GA (1983) Retinal detachment in the cat: the pigment epithelial-photoreceptor interface. Invest Ophthalmol Vis Sci 24:906-926.

Ansley SJ, Badano JL, Blacque OE, Hill J, Hoskins BE, Leitch CC, Kim JC, Ross AJ, Eichers ER, Teslovich TM, Mah AK, Johnsen RC, Cavender JC, Lewis RA, Leroux MR, Beales PL, Katsanis N (2003) Basal body dysfunction is a likely cause of pleiotropic Bardet-Biedl syndrome. Nature 425:628-633.

Anthony JG, Weidenhammer EM, Woolford JL, Jr. (1997) The yeast Prp3 protein is a U4/U6 snRNP protein necessary for integrity of the U4/U6 snRNP and the U4/U6.U5 tri-snRNP. Rna 3:1143-1152.

Ash J, McLeod DS, Lutty GA (2005) Transgenic expression of leukemia inhibitory factor (LIF) blocks normal vascular development but not pathological neovascularization in the eye. Mol Vis 11:298-308.

Badano JL, Mitsuma N, Beales PL, Katsanis N (2006) The ciliopathies: an emerging class of human genetic disorders. Annu Rev Genomics Hum Genet 7:125-148.

Barbeito LH, Pehar M, Cassina P, Vargas MR, Peluffo H, Viera L, Estevez AG, Beckman JS (2004) A role for astrocytes in motor neuron loss in amyotrophic lateral sclerosis. Brain Res Brain Res Rev 47:263-274.

120

Barbin G, Manthorpe M, Varon S (1984) Purification of the chick eye ciliary neuronotrophic factor. J Neurochem 43:1468-1478.

Barnett NL, Pow DV (2000) Antisense knockdown of GLAST, a glial glutamate transporter, compromises retinal function. Invest Ophthalmol Vis Sci 41:585-591.

Bascom RA, Manara S, Collins L, Molday RS, Kalnins VI, McInnes RR (1992) Cloning of the cDNA for a novel photoreceptor membrane protein (rom-1) identifies a disk rim protein family implicated in human retinopathies. Neuron 8:1171-1184.

Beere HM (2005) Death versus survival: functional interaction between the apoptotic and stress-inducible heat shock protein pathways. J Clin Invest 115:2633-2639.

Beltran WA, Zhang Q, Kijas JW, Gu D, Rohrer H, Jordan JA, Aguirre GD (2003) Cloning, mapping, and retinal expression of the canine ciliary neurotrophic factor receptor alpha (CNTFRalpha). Invest Ophthalmol Vis Sci 44:3642-3649.

Berglin L, Algvere PV, Seregard S (1997) Photoreceptor decay over time and apoptosis in experimental retinal detachment. Graefes Arch Clin Exp Ophthalmol 235:306-312.

Bernardos RL, Barthel LK, Meyers JR, Raymond PA (2007) Late-stage neuronal progenitors in the retina are radial Muller glia that function as retinal stem cells. J Neurosci 27:7028-7040.

Biswas SK, Zhao Y, Nagalingam A, Gardner TW, Sandirasegarane L (2008) PDGF- and insulin/IGF-1-specific distinct modes of class IA PI 3-kinase activation in normal rat retinas and RGC-5 retinal ganglion cells. Invest Ophthalmol Vis Sci 49:3687-3698.

Blackshaw S, Fraioli RE, Furukawa T, Cepko CL (2001) Comprehensive analysis of photoreceptor gene expression and the identification of candidate retinal disease genes. Cell 107:579-589.

Bok D (2005a) The role of RPE65 in inherited retinal diseases. Retina 25:S61-S62.

Bok D (2005b) Cellular mechanisms of retinal degenerations: RPE65, ABCA4, RDS, and bicarbonate transporter genes as examples. Retina 25:S18-S20.

Bok D, Yasumura D, Matthes MT, Ruiz A, Duncan JL, Chappelow AV, Zolutukhin S, Hauswirth W, LaVail MM (2002) Effects of adeno-associated virus-vectored ciliary neurotrophic factor on retinal structure and function in mice with a P216L rds/peripherin mutation. Exp Eye Res 74:719-735.

Boon CJ, den Hollander AI, Hoyng CB, Cremers FP, Klevering BJ, Keunen JE (2008) The spectrum of retinal dystrophies caused by mutations in the peripherin/RDS gene. Prog Retin Eye Res 27:213-235.

Bowne SJ, Sullivan LS, Blanton SH, Cepko CL, Blackshaw S, Birch DG, Hughbanks-Wheaton D, Heckenlively JR, Daiger SP (2002) Mutations in the inosine monophosphate dehydrogenase 1 gene (IMPDH1) cause the RP10 form of autosomal dominant retinitis pigmentosa. Hum Mol Genet 11:559-568.

121

Bringmann A, Reichenbach A (2001) Role of Muller cells in retinal degenerations. Front Biosci 6:E72-92.

Bringmann A, Pannicke T, Grosche J, Francke M, Wiedemann P, Skatchkov SN, Osborne NN, Reichenbach A (2006) Muller cells in the healthy and diseased retina. Prog Retin Eye Res 25:397-424.

Burns ME, Arshavsky VY (2005) Beyond counting photons: trials and trends in vertebrate visual transduction. Neuron 48:387-401.

Bush RA, Lei B, Tao W, Raz D, Chan CC, Cox TA, Santos-Muffley M, Sieving PA (2004) Encapsulated cell-based intraocular delivery of ciliary neurotrophic factor in normal rabbit: dose-dependent effects on ERG and retinal histology. Invest Ophthalmol Vis Sci 45:2420-2430.

Bustin SA (2002) Quantification of mRNA using real-time reverse transcription PCR (RT-PCR): trends and problems. J Mol Endocrinol 29:23-39.

Campochiaro PA, Chang M, Ohsato M, Vinores SA, Nie Z, Hjelmeland L, Mansukhani A, Basilico C, Zack DJ (1996) Retinal degeneration in transgenic mice with photoreceptor-specific expression of a dominant-negative fibroblast growth factor receptor. J Neurosci 16:1679-1688.

Cao W, Wen R, Li F, Lavail MM, Steinberg RH (1997) Mechanical injury increases bFGF and CNTF mRNA expression in the mouse retina. Exp Eye Res 65:241-248.

Carter-Dawson LD, LaVail MM (1979) Rods and cones in the mouse retina. I. Structural analysis using light and electron microscopy. J Comp Neurol 188:245-262.

Cayouette M, Behn D, Sendtner M, Lachapelle P, Gravel C (1998) Intraocular gene transfer of ciliary neurotrophic factor prevents death and increases responsiveness of rod photoreceptors in the retinal degeneration slow mouse. J Neurosci 18:9282-9293.

Chang B, Hawes NL, Hurd RE, Davisson MT, Nusinowitz S, Heckenlively JR (2002) Retinal degeneration mutants in the mouse. Vision Res 42:517-525.

Chang GQ, Hao Y, Wong F (1993) Apoptosis: final common pathway of photoreceptor death in rd, rds, and rhodopsin mutant mice. Neuron 11:595-605.

Chapman RS, Lourenco P, Tonner E, Flint D, Selbert S, Takeda K, Akira S, Clarke AR, Watson CJ (2000) The role of Stat3 in apoptosis and mammary gland involution. Conditional deletion of Stat3. Adv Exp Med Biol 480:129-138.

Chaum E (2003) Retinal neuroprotection by growth factors: a mechanistic perspective. J Cell Biochem 88:57-75.

Chen H, Weber AJ (2002) Expression of glial fibrillary acidic protein and glutamine synthetase by Muller cells after optic nerve damage and intravitreal application of brain-derived neurotrophic factor. Glia 38:115-125.

122

Chen L, Wu W, Dentchev T, Zeng Y, Wang J, Tsui I, Tobias JW, Bennett J, Baldwin D, Dunaief JL (2004) Light damage induced changes in mouse retinal gene expression. Exp Eye Res 79:239-247.

Chen S, Guttridge DC, You Z, Zhang Z, Fribley A, Mayo MW, Kitajewski J, Wang CY (2001) Wnt-1 signaling inhibits apoptosis by activating beta-catenin/T cell factor-mediated transcription. J Cell Biol 152:87-96.

Chen S, Wang QL, Nie Z, Sun H, Lennon G, Copeland NG, Gilbert DJ, Jenkins NA, Zack DJ (1997) Crx, a novel Otx-like paired-homeodomain protein, binds to and transactivates photoreceptor cell-specific genes. Neuron 19:1017-1030.

Cheng JG, Patterson PH (1997) LIF is an autocrine factor for sympathetic neurons. Mol Cell Neurosci 9:372-380.

Chong NH, Alexander RA, Waters L, Barnett KC, Bird AC, Luthert PJ (1999) Repeated injections of a ciliary neurotrophic factor analogue leading to long-term photoreceptor survival in hereditary retinal degeneration. Invest Ophthalmol Vis Sci 40:1298-1305.

Chowers I, Gunatilaka TL, Farkas RH, Qian J, Hackam AS, Duh E, Kageyama M, Wang C, Vora A, Campochiaro PA, Zack DJ (2003) Identification of novel genes preferentially expressed in the retina using a custom human retina cDNA microarray. Invest Ophthalmol Vis Sci 44:3732-3741.

Clarke G, Lumsden CJ (2005) Scale-free neurodegeneration: cellular heterogeneity and the stretched exponential kinetics of cell death. J Theor Biol 233:515-525.

Clarke G, Lumsden CJ, McInnes RR (2001) Inherited neurodegenerative diseases: the one-hit model of neurodegeneration. Hum Mol Genet 10:2269-2275.

Clarke G, Collins RA, Leavitt BR, Andrews DF, Hayden MR, Lumsden CJ, McInnes RR (2000a) A one-hit model of cell death in inherited neuronal degenerations. Nature 406:195-199.

Clarke G, Goldberg AF, Vidgen D, Collins L, Ploder L, Schwarz L, Molday LL, Rossant J, Szel A, Molday RS, Birch DG, McInnes RR (2000b) Rom-1 is required for rod photoreceptor viability and the regulation of disk morphogenesis. Nat Genet 25:67-73.

Clarkson RW, Boland MP, Kritikou EA, Lee JM, Freeman TC, Tiffen PG, Watson CJ (2006) The genes induced by signal transducer and activators of transcription (STAT)3 and STAT5 in mammary epithelial cells define the roles of these STATs in mammary development. Mol Endocrinol 20:675-685.

Collins CA, Guthrie C (2000) The question remains: is the spliceosome a ribozyme? Nat Struct Biol 7:850-854.

Comitato A, Spampanato C, Chakarova C, Sanges D, Bhattacharya SS, Marigo V (2007) Mutations in splicing factor PRPF3, causing retinal degeneration, form detrimental aggregates in photoreceptor cells. Hum Mol Genet 16:1699-1707.

123

Connell G, Bascom R, Molday L, Reid D, McInnes RR, Molday RS (1991) Photoreceptor peripherin is the normal product of the gene responsible for retinal degeneration in the rds mouse. Proc Natl Acad Sci U S A 88:723-726.

Curcio CA, Sloan KR, Kalina RE, Hendrickson AE (1990) Human photoreceptor topography. J Comp Neurol 292:497-523.

Daiger SP, Rossiter BJF, Greenberg J, Christoffels A, Hide W (2008) RetNet. In.

Davey RE, Onishi K, Mahdavi A, Zandstra PW (2007) LIF-mediated control of embryonic stem cell self-renewal emerges due to an autoregulatory loop. Faseb J 21:2020-2032.

Dejneka NS, Surace EM, Aleman TS, Cideciyan AV, Lyubarsky A, Savchenko A, Redmond TM, Tang W, Wei Z, Rex TS, Glover E, Maguire AM, Pugh EN, Jr., Jacobson SG, Bennett J (2004) In utero gene therapy rescues vision in a murine model of congenital blindness. Mol Ther 9:182-188.

Derouet D, Rousseau F, Alfonsi F, Froger J, Hermann J, Barbier F, Perret D, Diveu C, Guillet C, Preisser L, Dumont A, Barbado M, Morel A, deLapeyriere O, Gascan H, Chevalier S (2004) Neuropoietin, a new IL-6-related cytokine signaling through the ciliary neurotrophic factor receptor. Proc Natl Acad Sci U S A 101:4827-4832.

Derouiche A, Rauen T (1995) Coincidence of L-glutamate/L-aspartate transporter (GLAST) and glutamine synthetase (GS) immunoreactions in retinal glia: evidence for coupling of GLAST and GS in transmitter clearance. J Neurosci Res 42:131-143.

Donovan M, Cotter TG (2002) Caspase-independent photoreceptor apoptosis in vivo and differential expression of apoptotic protease activating factor-1 and caspase-3 during retinal development. Cell Death Differ 9:1220-1231.

Doonan F, Donovan M, Cotter TG (2003) Caspase-independent photoreceptor apoptosis in mouse models of retinal degeneration. J Neurosci 23:5723-5731.

Doonan F, Donovan M, Cotter TG (2005) Activation of multiple pathways during photoreceptor apoptosis in the rd mouse. Invest Ophthalmol Vis Sci 46:3530-3538.

Dowling JE, Sidman RL (1962) Inherited retinal dystrophy in the rat. J Cell Biol 14:73-109.

Dryja TP, Hahn LB, Kajiwara K, Berson EL (1997) Dominant and digenic mutations in the peripherin/RDS and ROM1 genes in retinitis pigmentosa. Invest Ophthalmol Vis Sci 38:1972-1982.

Dryja TP, McGee TL, Reichel E, Hahn LB, Cowley GS, Yandell DW, Sandberg MA, Berson EL (1990) A point mutation of the rhodopsin gene in one form of retinitis pigmentosa. Nature 343:364-366.

Duenker N (2005) Transforming growth factor-beta (TGF-beta) and programmed cell death in the vertebrate retina. Int Rev Cytol 245:17-43.

124

Elliott J, Cayouette M, Gravel C (2006) The CNTF/LIF signaling pathway regulates developmental programmed cell death and differentiation of rod precursor cells in the mouse retina in vivo. Dev Biol 300:583-598.

Fain GL (2006) Why photoreceptors die (and why they don't). Bioessays 28:344-354.

Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVail MM (1990) Photoreceptor degeneration in inherited retinal dystrophy delayed by basic fibroblast growth factor. Nature 347:83-86.

Faktorovich EG, Steinberg RH, Yasumura D, Matthes MT, LaVail MM (1992) Basic fibroblast growth factor and local injury protect photoreceptors from light damage in the rat. J Neurosci 12:3554-3567.

Fan J, Woodruff ML, Cilluffo MC, Crouch RK, Fain GL (2005) Opsin activation of transduction in the rods of dark-reared Rpe65 knockout mice. J Physiol 568:83-95.

Farber DB, Lolley RN (1974) Cyclic guanosine monophosphate: elevation in degenerating photoreceptor cells of the C3H mouse retina. Science 186:449-451.

Fayein NA, Courtois Y, Jeanny JC (1990) Ontogeny of basic fibroblast growth factor binding sites in mouse ocular tissues. Exp Cell Res 188:75-88.

Feng W, Yasumura D, Matthes MT, LaVail MM, Vollrath D (2002) Mertk triggers uptake of photoreceptor outer segments during phagocytosis by cultured retinal pigment epithelial cells. J Biol Chem 277:17016-17022.

Fernald RD (2006) Casting a genetic light on the evolution of eyes. Science 313:1914-1918.

Fiore R, Puschel AW (2003) The function of semaphorins during nervous system development. Front Biosci 8:s484-499.

Fischer AJ, Reh TA (2003) Potential of Muller glia to become neurogenic retinal progenitor cells. Glia 43:70-76.

Fischer P, Lehmann U, Sobota RM, Schmitz J, Niemand C, Linnemann S, Haan S, Behrmann I, Yoshimura A, Johnston JA, Muller-Newen G, Heinrich PC, Schaper F (2004) The role of the inhibitors of interleukin-6 signal transduction SHP2 and SOCS3 for desensitization of interleukin-6 signalling. Biochem J 378:449-460.

Fontaine V, Kinkl N, Sahel J, Dreyfus H, Hicks D (1998) Survival of purified rat photoreceptors in vitro is stimulated directly by fibroblast growth factor-2. J Neurosci 18:9662-9672.

Francke M, Makarov F, Kacza J, Seeger J, Wendt S, Gartner U, Faude F, Wiedemann P, Reichenbach A (2001) Retinal pigment epithelium melanin granules are phagocytozed by Muller glial cells in experimental retinal detachment. J Neurocytol 30:131-136.

Frasson M, Sahel JA, Fabre M, Simonutti M, Dreyfus H, Picaud S (1999) Retinitis pigmentosa: rod photoreceptor rescue by a calcium-channel blocker in the rd mouse. Nat Med 5:1183-1187.

125

Furukawa T, Morrow EM, Cepko CL (1997) Crx, a novel otx-like homeobox gene, shows photoreceptor-specific expression and regulates photoreceptor differentiation. Cell 91:531-541.

Garcia M, Vecino E (2003) Role of Muller glia in neuroprotection and regeneration in the retina. Histol Histopathol 18:1205-1218.

Gaur VP, Liu Y, Turner JE (1992) RPE conditioned medium stimulates photoreceptor cell survival, neurite outgrowth and differentiation in vitro. Exp Eye Res 54:645-659.

Gauthier R, Joly S, Pernet V, Lachapelle P, Di Polo A (2005) Brain-derived neurotrophic factor gene delivery to muller glia preserves structure and function of light-damaged photoreceptors. Invest Ophthalmol Vis Sci 46:3383-3392.

Gigliotti AP, Johnson PF, Sterneck E, DeWille JW (2003) Nulliparous CCAAT/enhancer binding proteindelta (C/EBPdelta) knockout mice exhibit mammary gland ductal hyperlasia. Exp Biol Med (Maywood) 228:278-285.

Gilbert SF (2006) Developmental biology, 8th Edition. Sunderland, Mass.: Sinauer Associates, Inc. Publishers.

Goke R, Barth P, Schmidt A, Samans B, Lankat-Buttgereit B (2004) Programmed cell death protein 4 suppresses CDK1/cdc2 via induction of p21(Waf1/Cip1). Am J Physiol Cell Physiol 287:C1541-1546.

Hackam AS, Qian J, Liu D, Gunatilaka T, Farkas RH, Chowers I, Kageyama M, Parmigiani G, Zack DJ (2004a) Comparative gene expression analysis of murine retina and brain. Mol Vis 10:637-649.

Hackam AS, Strom R, Liu D, Qian J, Wang C, Otteson D, Gunatilaka T, Farkas RH, Chowers I, Kageyama M, Leveillard T, Sahel JA, Campochiaro PA, Parmigiani G, Zack DJ (2004b) Identification of gene expression changes associated with the progression of retinal degeneration in the rd1 mouse. Invest Ophthalmol Vis Sci 45:2929-2942.

Harada T, Harada C, Nakayama N, Okuyama S, Yoshida K, Kohsaka S, Matsuda H, Wada K (2000) Modification of glial-neuronal cell interactions prevents photoreceptor apoptosis during light-induced retinal degeneration. Neuron 26:533-541.

Harada T, Harada C, Kohsaka S, Wada E, Yoshida K, Ohno S, Mamada H, Tanaka K, Parada LF, Wada K (2002) Microglia-Muller glia cell interactions control neurotrophic factor production during light-induced retinal degeneration. J Neurosci 22:9228-9236.

Hargrave PA (2001) Rhodopsin structure, function, and topography the Friedenwald lecture. Invest Ophthalmol Vis Sci 42:3-9.

Hartmann J, Konnerth A (2005) Determinants of postsynaptic Ca2+ signaling in Purkinje neurons. Cell Calcium 37:459-466.

Hartong DT, Berson EL, Dryja TP (2006) Retinitis pigmentosa. Lancet 368:1795-1809.

126

Hastings ML, Krainer AR (2001) Pre-mRNA splicing in the new millennium. Curr Opin Cell Biol 13:302-309.

Hauswirth W, Aleman TS, Kaushal S, Cideciyan AV, Schwartz SB, Wang L, Conlon T, Boye SL, Flotte TR, Byrne B, Jacobson SG (2008) Phase I Trial of Leber Congenital Amaurosis due to RPE65 Mutations by Ocular Subretinal Injection of Adeno-Associated Virus Gene Vector: Short-Term Results. Hum Gene Ther.

Hawkins RK, Jansen HG, Sanyal S (1985) Development and degeneration of retina in rds mutant mice: photoreceptor abnormalities in the heterozygotes. Exp Eye Res 41:701-720.

Heinrich PC, Behrmann I, Muller-Newen G, Schaper F, Graeve L (1998) Interleukin-6-type cytokine signalling through the gp130/Jak/STAT pathway. Biochem J 334 ( Pt 2):297-314.

Heinrich PC, Behrmann I, Haan S, Hermanns HM, Muller-Newen G, Schaper F (2003) Principles of interleukin (IL)-6-type cytokine signalling and its regulation. Biochem J 374:1-20.

Hojo M, Abe T, Sugano E, Yoshioka Y, Saigo Y, Tomita H, Wakusawa R, Tamai M (2004) Photoreceptor protection by iris pigment epithelial transplantation transduced with AAV-mediated brain-derived neurotrophic factor gene. Invest Ophthalmol Vis Sci 45:3721-3726.

Horsford DJ, Nguyen MT, Sellar GC, Kothary R, Arnheiter H, McInnes RR (2005) Chx10 repression of Mitf is required for the maintenance of mammalian neuroretinal identity. Development 132:177-187.

Hsiau TH, Diaconu C, Myers CA, Lee J, Cepko CL, Corbo JC (2007) The cis-regulatory logic of the mammalian photoreceptor transcriptional network. PLoS ONE 2:e643.

Hubel DH (1995) Eye, brain, and vision, [New Edition. New York: Scientific American Library : Distributed by W.H. Freeman.

Humphries MM, Rancourt D, Farrar GJ, Kenna P, Hazel M, Bush RA, Sieving PA, Sheils DM, McNally N, Creighton P, Erven A, Boros A, Gulya K, Capecchi MR, Humphries P (1997) Retinopathy induced in mice by targeted disruption of the rhodopsin gene. Nat Genet 15:216-219.

Hutt JA, O'Rourke JP, DeWille J (2000) Signal transducer and activator of transcription 3 activates CCAAT enhancer-binding protein delta gene transcription in G0 growth-arrested mouse mammary epithelial cells and in involuting mouse mammary gland. J Biol Chem 275:29123-29131.

Irizarry RA, Hobbs B, Collin F, Beazer-Barclay YD, Antonellis KJ, Scherf U, Speed TP (2003) Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 4:249-264.

jackson-laboratories (2008) In.

Jager RD, Mieler WF, Miller JW (2008) Age-related macular degeneration. N Engl J Med 358:2606-2617.

127

Jomary C, Cullen J, Jones SE (2006) Inactivation of the Akt survival pathway during photoreceptor apoptosis in the retinal degeneration mouse. Invest Ophthalmol Vis Sci 47:1620-1629.

Jones SE, Jomary C, Grist J, Stewart HJ, Neal MJ (2000) Altered expression of secreted frizzled-related protein-2 in retinitis pigmentosa retinas. Invest Ophthalmol Vis Sci 41:1297-1301.

Kamimura D, Ishihara K, Hirano T (2003) IL-6 signal transduction and its physiological roles: the signal orchestration model. Rev Physiol Biochem Pharmacol 149:1-38.

Kandel ER (2008) Principles of neural science, 5th, rev. Edition. New York: McGraw Hill.

Kassen SC, Ramanan V, Montgomery JE, C TB, Liu CG, Vihtelic TS, Hyde DR (2007) Time course analysis of gene expression during light-induced photoreceptor cell death and regeneration in albino zebrafish. Dev Neurobiol 67:1009-1031.

Katsanis N (2004) The oligogenic properties of Bardet-Biedl syndrome. Hum Mol Genet 13 Spec No 1:R65-71.

Kennan A, Aherne A, Palfi A, Humphries M, McKee A, Stitt A, Simpson DA, Demtroder K, Orntoft T, Ayuso C, Kenna PF, Farrar GJ, Humphries P (2002) Identification of an IMPDH1 mutation in autosomal dominant retinitis pigmentosa (RP10) revealed following comparative microarray analysis of transcripts derived from retinas of wild-type and Rho(-/-) mice. Hum Mol Genet 11:547-557.

Kistner A, Gossen M, Zimmermann F, Jerecic J, Ullmer C, Lubbert H, Bujard H (1996) Doxycycline-mediated quantitative and tissue-specific control of gene expression in transgenic mice. Proc Natl Acad Sci U S A 93:10933-10938.

Klionsky DJ, Emr SD (2000) Autophagy as a regulated pathway of cellular degradation. Science 290:1717-1721.

Kolb H, Linberg KA, Fisher SK (1992) Neurons of the human retina: a Golgi study. J Comp Neurol 318:147-187.

Kubota Y, Hirashima M, Kishi K, Stewart CL, Suda T (2008) Leukemia inhibitory factor regulates microvessel density by modulating oxygen-dependent VEGF expression in mice. J Clin Invest 118:2393-2403.

Kumanogoh A, Marukawa S, Kumanogoh T, Hirota H, Yoshida K, Lee IS, Yasui T, Yoshida K, Taga T, Kishimoto T (1997) Impairment of antigen-specific antibody production in transgenic mice expressing a dominant-negative form of gp130. Proc Natl Acad Sci U S A 94:2478-2482.

Kumanogoh A, Marukawa S, Suzuki K, Takegahara N, Watanabe C, Ch'ng E, Ishida I, Fujimura H, Sakoda S, Yoshida K, Kikutani H (2002) Class IV semaphorin Sema4A enhances T-cell activation and interacts with Tim-2. Nature 419:629-633.

Kunchithapautham K, Rohrer B (2007) Autophagy is one of the multiple mechanisms active in photoreceptor degeneration. Autophagy 3:65-66.

128

Lau D, McGee LH, Zhou S, Rendahl KG, Manning WC, Escobedo JA, Flannery JG (2000) Retinal degeneration is slowed in transgenic rats by AAV-mediated delivery of FGF-2. Invest Ophthalmol Vis Sci 41:3622-3633.

LaVail MM, Unoki K, Yasumura D, Matthes MT, Yancopoulos GD, Steinberg RH (1992) Multiple growth factors, cytokines, and neurotrophins rescue photoreceptors from the damaging effects of constant light. Proc Natl Acad Sci U S A 89:11249-11253.

LaVail MM, Yasumura D, Matthes MT, Lau-Villacorta C, Unoki K, Sung CH, Steinberg RH (1998) Protection of mouse photoreceptors by survival factors in retinal degenerations. Invest Ophthalmol Vis Sci 39:592-602.

Lenzi L, Coassin M, Lambiase A, Bonini S, Amendola T, Aloe L (2005) Effect of exogenous administration of nerve growth factor in the retina of rats with inherited retinitis pigmentosa. Vision Res 45:1491-1500.

Levy DE, Lee CK (2002) What does Stat3 do? J Clin Invest 109:1143-1148.

Li T, Snyder WK, Olsson JE, Dryja TP (1996) Transgenic mice carrying the dominant rhodopsin mutation P347S: evidence for defective vectorial transport of rhodopsin to the outer segments. Proc Natl Acad Sci U S A 93:14176-14181.

Liang FQ, Aleman TS, Dejneka NS, Dudus L, Fisher KJ, Maguire AM, Jacobson SG, Bennett J (2001) Long-term protection of retinal structure but not function using RAAV.CNTF in animal models of retinitis pigmentosa. Mol Ther 4:461-472.

Liu C, Peng M, Laties AM, Wen R (1998) Preconditioning with bright light evokes a protective response against light damage in the rat retina. J Neurosci 18:1337-1344.

Liu C, Li Y, Peng M, Laties AM, Wen R (1999) Activation of caspase-3 in the retina of transgenic rats with the rhodopsin mutation s334ter during photoreceptor degeneration. J Neurosci 19:4778-4785.

Liu IS, Chen JD, Ploder L, Vidgen D, van der Kooy D, Kalnins VI, McInnes RR (1994) Developmental expression of a novel murine homeobox gene (Chx10): evidence for roles in determination of the neuroretina and inner nuclear layer. Neuron 13:377-393.

Livesey FJ, Cepko CL (2001) Vertebrate neural cell-fate determination: lessons from the retina. Nat Rev Neurosci 2:109-118.

Livesey FJ, Furukawa T, Steffen MA, Church GM, Cepko CL (2000) Microarray analysis of the transcriptional network controlled by the photoreceptor homeobox gene Crx. Curr Biol 10:301-310.

Loewen CJ, Molday RS (2000) Disulfide-mediated oligomerization of Peripherin/Rds and Rom-1 in photoreceptor disk membranes. Implications for photoreceptor outer segment morphogenesis and degeneration. J Biol Chem 275:5370-5378.

129

Ma J, Norton JC, Allen AC, Burns JB, Hasel KW, Burns JL, Sutcliffe JG, Travis GH (1995) Retinal degeneration slow (rds) in mouse results from simple insertion of a t haplotype-specific element into protein-coding exon II. Genomics 28:212-219.

MacDonald IM, Sauve Y, Sieving PA (2007) Preventing blindness in retinal disease: ciliary neurotrophic factor intraocular implants. Can J Ophthalmol 42:399-402.

Makarova OV, Makarov EM, Liu S, Vornlocher HP, Luhrmann R (2002) Protein 61K, encoded by a gene (PRPF31) linked to autosomal dominant retinitis pigmentosa, is required for U4/U6*U5 tri-snRNP formation and pre-mRNA splicing. Embo J 21:1148-1157.

Maniatis T, Fritsch EF, Sambrook J (1987) Molecular cloning : a laboratory manual, 2nd Edition. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory.

Marlhens F, Bareil C, Griffoin JM, Zrenner E, Amalric P, Eliaou C, Liu SY, Harris E, Redmond TM, Arnaud B, Claustres M, Hamel CP (1997) Mutations in RPE65 cause Leber's congenital amaurosis. Nat Genet 17:139-141.

Masland RH (2001) The fundamental plan of the retina. Nat Neurosci 4:877-886.

Mata NL, Weng J, Travis GH (2000) Biosynthesis of a major lipofuscin fluorophore in mice and humans with ABCR-mediated retinal and macular degeneration. Proc Natl Acad Sci U S A 97:7154-7159.

McGill TJ, Prusky GT, Douglas RM, Yasumura D, Matthes MT, Nune G, Donohue-Rolfe K, Yang H, Niculescu D, Hauswirth WW, Girman SV, Lund RD, Duncan JL, LaVail MM (2007) Intraocular CNTF reduces vision in normal rats in a dose-dependent manner. Invest Ophthalmol Vis Sci 48:5756-5766.

Miron M, Woody OZ, Marcil A, Murie C, Sladek R, Nadon R (2006) A methodology for global validation of microarray experiments. BMC Bioinformatics 7:333.

Moiseyev G, Chen Y, Takahashi Y, Wu BX, Ma JX (2005) RPE65 is the isomerohydrolase in the retinoid visual cycle. Proc Natl Acad Sci U S A 102:12413-12418.

Molday RS (1998) Photoreceptor membrane proteins, phototransduction, and retinal degenerative diseases. The Friedenwald Lecture. Invest Ophthalmol Vis Sci 39:2491-2513.

Molday RS, Molday LL (1979) Identification and characterization of multiple forms of rhodopsin and minor proteins in frog and bovine rod outer segment disc membranes. Electrophoresis, lectin labeling, and proteolysis studies. J Biol Chem 254:4653-4660.

Moritz OL, Molday RS (1996) Molecular cloning, membrane topology, and localization of bovine rom-1 in rod and cone photoreceptor cells. Invest Ophthalmol Vis Sci 37:352-362.

Morrow EM, Furukawa T, Lee JE, Cepko CL (1999) NeuroD regulates multiple functions in the developing neural retina in rodent. Development 126:23-36.

130

Mullen RJ, LaVail MM (1976) Inherited retinal dystrophy: primary defect in pigment epithelium determined with experimental rat chimeras. Science 192:799-801.

Mykytyn K, Sheffield VC (2004) Establishing a connection between cilia and Bardet-Biedl Syndrome. Trends Mol Med 10:106-109.

Nakanishi M, Niidome T, Matsuda S, Akaike A, Kihara T, Sugimoto H (2007) Microglia-derived interleukin-6 and leukaemia inhibitory factor promote astrocytic differentiation of neural stem/progenitor cells. Eur J Neurosci 25:649-658.

Newman EA (1993) Inward-rectifying potassium channels in retinal glial (Muller) cells. J Neurosci 13:3333-3345.

Nir I, Kedzierski W, Chen J, Travis GH (2000) Expression of Bcl-2 protects against photoreceptor degeneration in retinal degeneration slow (rds) mice. J Neurosci 20:2150-2154.

Nishimura DY, Fath M, Mullins RF, Searby C, Andrews M, Davis R, Andorf JL, Mykytyn K, Swiderski RE, Yang B, Carmi R, Stone EM, Sheffield VC (2004) Bbs2-null mice have neurosensory deficits, a defect in social dominance, and retinopathy associated with mislocalization of rhodopsin. Proc Natl Acad Sci U S A 101:16588-16593.

O'Driscoll C, O'Connor J, O'Brien CJ, Cotter TG (2008) Basic fibroblast growth factor-induced protection from light damage in the mouse retina in vivo. J Neurochem 105:524-536.

Okoye G, Zimmer J, Sung J, Gehlbach P, Deering T, Nambu H, Hackett S, Melia M, Esumi N, Zack DJ, Campochiaro PA (2003) Increased expression of brain-derived neurotrophic factor preserves retinal function and slows cell death from rhodopsin mutation or oxidative damage. J Neurosci 23:4164-4172.

Pacione LR, Szego MJ, Ikeda S, Nishina PM, McInnes RR (2003) Progress toward understanding the genetic and biochemical mechanisms of inherited photoreceptor degenerations. Annu Rev Neurosci 26:657-700.

Paquet-Durand F, Johnson L, Ekstrom P (2007) Calpain activity in retinal degeneration. J Neurosci Res 85:693-702.

Paquet-Durand F, Azadi S, Hauck SM, Ueffing M, van Veen T, Ekstrom P (2006) Calpain is activated in degenerating photoreceptors in the rd1 mouse. J Neurochem 96:802-814.

Park EJ, Kim JH, Seong RH, Kim CG, Park SD, Hong SH (1999) Characterization of a novel mouse cDNA, ES18, involved in apoptotic cell death of T-cells. Nucleic Acids Res 27:1524-1530.

Pasterkamp RJ, Kolodkin AL (2003) Semaphorin junction: making tracks toward neural connectivity. Curr Opin Neurobiol 13:79-89.

Pennacchio LA, Ahituv N, Moses AM, Prabhakar S, Nobrega MA, Shoukry M, Minovitsky S, Dubchak I, Holt A, Lewis KD, Plajzer-Frick I, Akiyama J, De Val S, Afzal V, Black BL,

131

Couronne O, Eisen MB, Visel A, Rubin EM (2006) In vivo enhancer analysis of human conserved non-coding sequences. Nature 444:499-502.

Pennesi ME, Cho JH, Yang Z, Wu SH, Zhang J, Wu SM, Tsai MJ (2003) BETA2/NeuroD1 null mice: a new model for transcription factor-dependent photoreceptor degeneration. J Neurosci 23:453-461.

Peterson WM, Wang Q, Tzekova R, Wiegand SJ (2000) Ciliary neurotrophic factor and stress stimuli activate the Jak-STAT pathway in retinal neurons and glia. J Neurosci 20:4081-4090.

Phelan JK, Bok D (2000) A brief review of retinitis pigmentosa and the identified retinitis pigmentosa genes. Mol Vis 6:116-124.

Philpott KL, McCarthy MJ, Klippel A, Rubin LL (1997) Activated phosphatidylinositol 3-kinase and Akt kinase promote survival of superior cervical neurons. J Cell Biol 139:809-815.

Punzo C, Cepko C (2007) Cellular responses to photoreceptor death in the rd1 mouse model of retinal degeneration. Invest Ophthalmol Vis Sci 48:849-857.

Qian J, Esumi N, Chen Y, Wang Q, Chowers I, Zack DJ (2005) Identification of regulatory targets of tissue-specific transcription factors: application to retina-specific gene regulation. Nucleic Acids Res 33:3479-3491.

Ramón y Cajal S, Bresler J (1900) Studien über die Hirnrinde des Menschen. Leipzig: J. A. Barth.

Rattner A, Nathans J (2005) The genomic response to retinal disease and injury: evidence for endothelin signaling from photoreceptors to glia. J Neurosci 25:4540-4549.

Rattner A, Nathans J (2006) Macular degeneration: recent advances and therapeutic opportunities. Nat Rev Neurosci 7:860-872.

Redmond TM, Yu S, Lee E, Bok D, Hamasaki D, Chen N, Goletz P, Ma JX, Crouch RK, Pfeifer K (1998) Rpe65 is necessary for production of 11-cis-vitamin A in the retinal visual cycle. Nat Genet 20:344-351.

Rhee KD, Yang XJ (2003) Expression of cytokine signal transduction components in the postnatal mouse retina. Mol Vis 9:715-722.

Rhee KD, Ruiz A, Duncan JL, Hauswirth WW, Lavail MM, Bok D, Yang XJ (2007) Molecular and cellular alterations induced by sustained expression of ciliary neurotrophic factor in a mouse model of retinitis pigmentosa. Invest Ophthalmol Vis Sci 48:1389-1400.

Rice DS, Huang W, Jones HA, Hansen G, Ye GL, Xu N, Wilson EA, Troughton K, Vaddi K, Newton RC, Zambrowicz BP, Sands AT (2004) Severe retinal degeneration associated with disruption of semaphorin 4A. Invest Ophthalmol Vis Sci 45:2767-2777.

Rohrer B, Pinto FR, Hulse KE, Lohr HR, Zhang L, Almeida JS (2004) Multidestructive pathways triggered in photoreceptor cell death of the rd mouse as determined through gene expression profiling. J Biol Chem 279:41903-41910.

132

Rosenbaum JL, Witman GB (2002) Intraflagellar transport. Nat Rev Mol Cell Biol 3:813-825.

Ross AJ, May-Simera H, Eichers ER, Kai M, Hill J, Jagger DJ, Leitch CC, Chapple JP, Munro PM, Fisher S, Tan PL, Phillips HM, Leroux MR, Henderson DJ, Murdoch JN, Copp AJ, Eliot MM, Lupski JR, Kemp DT, Dollfus H, Tada M, Katsanis N, Forge A, Beales PL (2005) Disruption of Bardet-Biedl syndrome ciliary proteins perturbs planar cell polarity in vertebrates. Nat Genet 37:1135-1140.

Samardzija M, Wenzel A, Aufenberg S, Thiersch M, Reme C, Grimm C (2006) Differential role of Jak-STAT signaling in retinal degenerations. Faseb J 20:2411-2413.

Sanges D, Comitato A, Tammaro R, Marigo V (2006) Apoptosis in retinal degeneration involves cross-talk between apoptosis-inducing factor (AIF) and caspase-12 and is blocked by calpain inhibitors. Proc Natl Acad Sci U S A 103:17366-17371.

Sanyal S, Jansen HG (1981) Absence of receptor outer segments in the retina of rds mutant mice. Neurosci Lett 21:23-26.

Sanyal S, De Ruiter A, Hawkins RK (1980) Development and degeneration of retina in rds mutant mice: light microscopy. J Comp Neurol 194:193-207.

Schaeren-Wiemers N, Gerfin-Moser A (1993) A single protocol to detect transcripts of various types and expression levels in neural tissue and cultured cells: in situ hybridization using digoxigenin-labelled cRNA probes. Histochemistry 100:431-440.

Schalken JJ, Janssen JJ, Sanyal S, Hawkins RK, de Grip WJ (1990) Development and degeneration of retina in rds mutant mice: immunoassay of the rod visual pigment rhodopsin. Biochim Biophys Acta 1033:103-109.

Schulz HL, Goetz T, Kaschkoetoe J, Weber BH (2004) The Retinome - defining a reference transcriptome of the adult mammalian retina/retinal pigment epithelium. BMC Genomics 5:50.

Schutte M, Werner P (1998) Redistribution of glutathione in the ischemic rat retina. Neurosci Lett 246:53-56.

Sharma AK, Rohrer B (2004) Calcium-induced calpain mediates apoptosis via caspase-3 in a mouse photoreceptor cell line. J Biol Chem 279:35564-35572.

Sharma AK, Rohrer B (2007) Sustained elevation of intracellular cGMP causes oxidative stress triggering calpain-mediated apoptosis in photoreceptor degeneration. Curr Eye Res 32:259-269.

Sieving PA, Caruso RC, Tao W, Coleman HR, Thompson DJ, Fullmer KR, Bush RA (2006) Ciliary neurotrophic factor (CNTF) for human retinal degeneration: phase I trial of CNTF delivered by encapsulated cell intraocular implants. Proc Natl Acad Sci U S A 103:3896-3901.

133

Song Y, Zhao L, Tao W, Laties AM, Luo Z, Wen R (2003) Photoreceptor protection by cardiotrophin-1 in transgenic rats with the rhodopsin mutation s334ter. Invest Ophthalmol Vis Sci 44:4069-4075.

SP Daiger BR, J Greenberg, A Christoffels, W Hide (2008) RetNet. In.

Staley JP, Guthrie C (1998) Mechanical devices of the spliceosome: motors, clocks, springs, and things. Cell 92:315-326.

Steinberg RH, Fisher SK, Anderson DH (1980) Disc morphogenesis in vertebrate photoreceptors. J Comp Neurol 190:501-508.

Stephanou A, Latchman DS (2005) Opposing actions of STAT-1 and STAT-3. Growth Factors 23:177-182.

Stryer L (1996) Vision: from photon to perception. Proc Natl Acad Sci U S A 93:557-559.

Swaroop A, Zack DJ (2002) Transcriptome analysis of the retina. Genome Biol 3:REVIEWS1022.

Takano Y, Ohguro H, Dezawa M, Ishikawa H, Yamazaki H, Ohguro I, Mamiya K, Metoki T, Ishikawa F, Nakazawa M (2004) Study of drug effects of calcium channel blockers on retinal degeneration of rd mouse. Biochem Biophys Res Commun 313:1015-1022.

Takeuchi K, Nakazawa M, Mizukoshi S (2008) Systemic administration of nilvadipine delays photoreceptor degeneration of heterozygous retinal degeneration slow (rds) mouse. Exp Eye Res 86:60-69.

Tao W, Wen R, Goddard MB, Sherman SD, O'Rourke PJ, Stabila PF, Bell WJ, Dean BJ, Kauper KA, Budz VA, Tsiaras WG, Acland GM, Pearce-Kelling S, Laties AM, Aguirre GD (2002) Encapsulated cell-based delivery of CNTF reduces photoreceptor degeneration in animal models of retinitis pigmentosa. Invest Ophthalmol Vis Sci 43:3292-3298.

Tcheng M, Fuhrmann G, Hartmann MP, Courtois Y, Jeanny JC (1994) Spatial and temporal expression patterns of FGF receptor genes type 1 and type 2 in the developing chick retina. Exp Eye Res 58:351-358.

Thangaraju M, Rudelius M, Bierie B, Raffeld M, Sharan S, Hennighausen L, Huang AM, Sterneck E (2005) C/EBPdelta is a crucial regulator of pro-apoptotic gene expression during mammary gland involution. Development 132:4675-4685.

Torp R, Su JH, Deng G, Cotman CW (1998) GADD45 is induced in Alzheimer's disease, and protects against apoptosis in vitro. Neurobiol Dis 5:245-252.

Traverso V, Kinkl N, Grimm L, Sahel J, Hicks D (2003) Basic fibroblast and epidermal growth factors stimulate survival in adult porcine photoreceptor cell cultures. Invest Ophthalmol Vis Sci 44:4550-4558.

Travis GH (1998) Mechanisms of cell death in the inherited retinal degenerations. Am J Hum Genet 62:503-508.

134

Travis GH, Brennan MB, Danielson PE, Kozak CA, Sutcliffe JG (1989) Identification of a photoreceptor-specific mRNA encoded by the gene responsible for retinal degeneration slow (rds). Nature 338:70-73.

Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 98:5116-5121.

Ueki Y, Wang J, Chollangi S, Ash JD (2008) STAT3 activation in photoreceptors by leukemia inhibitory factor is associated with protection from light damage. J Neurochem 105:784-796.

Valter K, Bisti S, Stone J (2003) Location of CNTFRalpha on outer segments: evidence of the site of action of CNTF in rat retina. Brain Res 985:169-175.

Vandesompele J, De Preter K, Pattyn F, Poppe B, Van Roy N, De Paepe A, Speleman F (2002) Accurate normalization of real-time quantitative RT-PCR data by geometric averaging of multiple internal control genes. Genome Biol 3:RESEARCH0034.

Vazquez-Chona F, Song BK, Geisert EE, Jr. (2004) Temporal changes in gene expression after injury in the rat retina. Invest Ophthalmol Vis Sci 45:2737-2746.

Venters HD, Dantzer R, Kelley KW (2000) A new concept in neurodegeneration: TNFalpha is a silencer of survival signals. Trends Neurosci 23:175-180.

Vollrath D, Feng W, Duncan JL, Yasumura D, D'Cruz PM, Chappelow A, Matthes MT, Kay MA, LaVail MM (2001) Correction of the retinal dystrophy phenotype of the RCS rat by viral gene transfer of Mertk. Proc Natl Acad Sci U S A 98:12584-12589.

Wahlin KJ, Campochiaro PA, Zack DJ, Adler R (2000) Neurotrophic factors cause activation of intracellular signaling pathways in Muller cells and other cells of the inner retina, but not photoreceptors. Invest Ophthalmol Vis Sci 41:927-936.

Wahlin KJ, Adler R, Zack DJ, Campochiaro PA (2001) Neurotrophic signaling in normal and degenerating rodent retinas. Exp Eye Res 73:693-701.

Walsh N, Valter K, Stone J (2001) Cellular and subcellular patterns of expression of bFGF and CNTF in the normal and light stressed adult rat retina. Exp Eye Res 72:495-501.

Wan J, Zheng H, Chen ZL, Xiao HL, Shen ZJ, Zhou GM (2008) Preferential regeneration of photoreceptor from Muller glia after retinal degeneration in adult rat. Vision Res 48:223-234.

Wang Y, Smith SB, Ogilvie JM, McCool DJ, Sarthy V (2002) Ciliary neurotrophic factor induces glial fibrillary acidic protein in retinal Muller cells through the JAK/STAT signal transduction pathway. Curr Eye Res 24:305-312.

Wen R, Song Y, Cheng T, Matthes MT, Yasumura D, LaVail MM, Steinberg RH (1995) Injury-induced upregulation of bFGF and CNTF mRNAS in the rat retina. J Neurosci 15:7377-7385.

135

Wen R, Cheng T, Song Y, Matthes MT, Yasumura D, LaVail MM, Steinberg RH (1998) Continuous exposure to bright light upregulates bFGF and CNTF expression in the rat retina. Curr Eye Res 17:494-500.

Weng J, Mata NL, Azarian SM, Tzekov RT, Birch DG, Travis GH (1999) Insights into the function of Rim protein in photoreceptors and etiology of Stargardt's disease from the phenotype in abcr knockout mice. Cell 98:13-23.

Wenzel A, Grimm C, Samardzija M, Reme CE (2005) Molecular mechanisms of light-induced photoreceptor apoptosis and neuroprotection for retinal degeneration. Prog Retin Eye Res 24:275-306.

Westerfeld C, Mukai S (2008) Stargardt's disease and the ABCR gene. Semin Ophthalmol 23:59-65.

Woodruff ML, Wang Z, Chung HY, Redmond TM, Fain GL, Lem J (2003) Spontaneous activity of opsin apoprotein is a cause of Leber congenital amaurosis. Nat Genet 35:158-164.

Yamada T, Tobita K, Osada S, Nishihara T, Imagawa M (1997a) CCAAT/enhancer-binding protein delta gene expression is mediated by APRF/STAT3. J Biochem (Tokyo) 121:731-738.

Yamada T, Tobita K, Osada S, Nishihara T, Imagawa M (1997b) CCAAT/enhancer-binding protein delta gene expression is mediated by APRF/STAT3. J Biochem 121:731-738.

Yen HJ, Tayeh MK, Mullins RF, Stone EM, Sheffield VC, Slusarski DC (2006) Bardet-Biedl syndrome genes are important in retrograde intracellular trafficking and Kupffer's vesicle cilia function. Hum Mol Genet 15:667-677.

Young RW (1971) Shedding of discs from rod outer segments in the rhesus monkey. J Ultrastruct Res 34:190-203.

Young RW (1976) Visual cells and the concept of renewal. Invest Ophthalmol Vis Sci 15:700-725.

Young RW (1985) Cell proliferation during postnatal development of the retina in the mouse. Brain Res 353:229-239.

Yuan L, Kawada M, Havlioglu N, Tang H, Wu JY (2005) Mutations in PRPF31 inhibit pre-mRNA splicing of rhodopsin gene and cause apoptosis of retinal cells. J Neurosci 25:748-757.

Yuen T, Wurmbach E, Pfeffer RL, Ebersole BJ, Sealfon SC (2002) Accuracy and calibration of commercial oligonucleotide and custom cDNA microarrays. Nucleic Acids Res 30:e48.

Zack DJ (2000) Neurotrophic rescue of photoreceptors: are Muller cells the mediators of survival? Neuron 26:285-286.

Zacks DN, Han Y, Zeng Y, Swaroop A (2006) Activation of signaling pathways and stress-response genes in an experimental model of retinal detachment. Invest Ophthalmol Vis Sci 47:1691-1695.

136

Zeiss CJ, Neal J, Johnson EA (2004) Caspase-3 in postnatal retinal development and degeneration. Invest Ophthalmol Vis Sci 45:964-970.

Zhang SS, Xu X, Li J, Liu MG, Zhao H, Soares MB, Barnstable CJ, Fu XY (2005) Comprehensive in silico functional specification of mouse retina transcripts. BMC Genomics 6:40.

Zhang X, Darnell JE, Jr. (2001) Functional importance of Stat3 tetramerization in activation of the alpha 2-macroglobulin gene. J Biol Chem 276:33576-33581.

Zhou X, Li F, Ge J, Sarkisian SR, Jr., Tomita H, Zaharia A, Chodosh J, Cao W (2007) Retinal ganglion cell protection by 17-beta-estradiol in a mouse model of inherited glaucoma. Dev Neurobiol 67:603-616.