Protection of Retina by αB Crystallin in Sodium Iodate Induced Retinal Degeneration

Age-related macular degeneration (AMD) is a leading cause of blindness in the developed world. The retinal pigment epithelium (RPE) is a critical site of pathology in AMD and αB crystallin expression is increased in RPE and associated drusen in AMD. The purpose of this study was to investigate the role of αB crystallin in sodium iodate (NaIO3)-induced retinal degeneration, a model of AMD in which the primary site of pathology is the RPE. Dose dependent effects of intravenous NaIO3 (20-70 mg/kg) on development of retinal degeneration (fundus photography) and RPE and retinal neuronal loss (histology) were determined in wild type and αB crystallin knockout mice. Absence of αB crystallin augmented retinal degeneration in low dose (20 mg/kg) NaIO3-treated mice and increased retinal cell apoptosis which was mainly localized to the RPE layer. Generation of reactive oxygen species (ROS) was observed with NaIO3 in mouse and human RPE which increased further after αB crystallin knockout or siRNA knockdown, respectively. NaIO3 upregulated AKT phosphorylation and peroxisome proliferator–activator receptor–γ (PPARγ) which was suppressed after αB crystallin siRNA knockdown. Further, PPARγ ligand inhibited NaIO3-induced ROS generation. Our data suggest that αB crystallin plays a critical role in protection of NaIO3-induced oxidative stress and retinal degeneration in part through upregulation of AKT phosphorylation and PPARγ expression.


Introduction
Age-related macular degeneration (AMD) is characterized by progressive degeneration of the macular region of the retina resulting in loss of central vision. AMD is the leading cause of irreversible blindness in the developed world [1]. Clinically, AMD manifests in two forms; a non-exudative dry form and an exudative, neovascular wet form [1,2]. Geographic atrophy (GA) is an advanced form of dry AMD with extensive atrophy and loss of the retinal pigment epithelium (RPE) and overlying photoreceptors and is responsible for 10-20% of cases of legal blindness from AMD [3,4]. At present, there is no available effective treatment for GA.
A number of murine models have been generated that simulate features of dry AMD including RPE degeneration, lipofuscin accumulation, subretinal deposits, and loss of photoreceptors [5][6][7][8][9][10][11][12]. Our laboratory recently showed that bone morphogenetic protein-4 (BMP-4) is highly expressed in dry AMD and mediates oxidative stress-induced senescence in RPE in in vitro dry AMD thus serving as a molecular switch between atrophic and neovascular AMD [13,14]. Localized RPE debridement or genetic ablation of RPE can lead to a profound reduction in RPE cells and consequent loss of photoreceptors [15]. Retinotoxicity can also be induced by endogenous and exogenous agents in laboratory animals. Mice receiving polyinosine-polycytidylic acid (Poly I: C) had morphological changes similar to that of humans with dry ARMD exhibiting soft and/or hard Drusen, GA [16]. Recently, the conditional ablation of the microRNA processing enzyme DICER1 was shown to induce RPE degeneration in mice [17]. Genetic or pharmacological inhibition of inflammasome components (NLRP3, MYD88) was reported to prevent RPE degeneration induced by DICER1 loss or AluRNA exposure [18]. While most of the animal models for GA mentioned above are long-term involving prolonged treatment regimens, the NaIO 3 -induced retinal degeneration model has proven to be a convenient and widely used model, because it is rapid, reproducible and has a primary site of pathology in the RPE [6,[19][20][21][22]. Thus, in the present study, we have utilized the NaIO 3 model in 129S6/SvEvTac mice to study mechanisms of retinal degeneration.
Crystallins are members of the small heat shock protein (sHSP) family, and aB crystallin has been found to have high chaperone efficiency, and bind misfolded proteins with high affinity and stoichiometry [23]. An increased expression of aB crystallin was found in RPE and associated drusen in dry AMD [24,25]. Both aA and aB crystallin are expressed in the mouse retina [26][27][28]. Mice lacking aB crystallin have provided considerable insights into the functional roles of this protein [29]. Our laboratory has shown that RPE cells from mice lacking aB crystallin are more susceptible to oxidative and endoplasmic reticulum stress as compared to wild type RPE [28,30,31]. Further we found that RPE cells overexpressing aB crystallin showed resistance to apoptosis, suggesting that aB crystallin may prevent stress-induced cell death [32]. Recently, evidence for the secretion of aB crystallin by RPE exosomes and protection of neighboring photoreceptors and RPE by exogenous aB crystallin was presented by our laboratory [33] suggesting that aB crystallin has significant potential in retinal therapy.
This study was undertaken to investigate the role of aB crystallin in a model of NaIO 3 induced retinal degeneration in 129S6/ SvEvTac mice. Further, using cultured mouse and human RPE cells, we also investigated the mechanism of regulation of cell death from NaIO 3 -induced oxidative stress by aB crystallin. Our major finding is that absence of aB crystallin in aB crystallin knockout mice causes more severe degeneration of the retina in NaIO 3 -treated mice as compared to wild type mice treated with NaIO 3 . Further, our studies also suggest that aB crystallin plays a critical role in protection of NaIO 3 -induced oxidative stress and retinal degeneration in part through upregulation of AKT phosphorylation and PPAR c expression.

Selection of optimal in vivo dose and duration of NaIO 3 treatment
Preliminary experiments were performed to select an optimal dosage of NaIO 3 that was used in all subsequent in vivo experiments in mice. We tested the effect of a single intravenous injection of 20, 35, 50 and 70 mg/kg NaIO 3 for 1, 2 and 3 weeks on retinal morphology. The histologic data from varying doses of NaIO 3 treatment are presented in Figure 1. NaIO 3 -induced retinal degeneration increased with the dose. The extent of degeneration was absent to mild with 20 mg/kg, moderate with 35 and 50 mg/kg and was severe with 70 mg/kg dose 3 week after NaIO 3 treatment ( Figure 1). We chose the 20 mg/kg NaIO 3 dose that produced no more than mild degeneration in all our subsequent experiments to enable studying the exacerbating effects of aB crystallin knockout on retinal damage (see below). While high doses of NaIO 3 resulted in retinal degeneration as early as 1 week post-injection, the low dose (20 mg/kg) NaIO 3 showed damage localized to the RPE at the 3 week time-point ( Fig  S1).

Fundus photography shows accelerated NaIO 3 -induced retinal degeneration in aB crystallin knockout mice
To determine the extent of NaIO 3 -induced retinal degeneration, we compared the fundus photographs of mice from PBStreated WT, NaIO 3 -treated WT, PBS-treated aB crystallin knockout, and NaIO 3 -treated aB crystallin knockout groups at the end of 3 weeks. The dose of NaIO 3 in these studies was 20 mg/kg. The retinal degeneration induced by NaIO 3 in mice appeared as patchy white retinal lesions when observed by fundus photography (Fig. 2).
The fundus photographs of thirteen out of fourteen NaIO 3treated aB crystallin knockout mice showed patchy retinal degeneration three weeks after injection (Fig.2D, E). Only three out of fourteen eyes of NaIO 3 -treated wild type mice showed retinal degeneration (Fig 2B, E). Thus, the difference in the number of mice with retinal degeneration between NaIO 3 -treated aB crystallin knockout mice and NaIO 3 -treated wild type mice was highly significant (P,0.001). No apparent degeneration could be seen in control, untreated wild type or aB crystallin knockout retina.
Histopathology shows accelerated NaIO 3 -induced degeneration in aB crystallin knockout mice The primary site of pathology after NaIO 3 injection (20 mg/kg) was the RPE layer; we observed that the RPE layer was discontiguous and damaged in all aB crystallin knockout mice, while only two out of seven wild type mice showed these changes in the RPE (Fig. 3). Using TUNEL staining we confirmed that with NaIO 3 (20 mg/kg; 3 week time point) cell death was localized to the RPE layer in the aB crystallin knockout mice ( Figure S2). Significant differences were found between NaIO 3 -treated wild type mice and NaIO 3 -treated aB crystallin knockout mice in the extent of RPE degeneration (P,0.01). Retinas from aB crystallin knockout mice (Fig. 3D) revealed more severe degeneration from NaIO 3 injection as compared to wild-type retinas (Fig. 3B). Total retinal thickness was significantly decreased (P,0.01) in aB crystallin knockout mice with NaIO 3 treatment as compared to untreated aB crystallin knockout group (P,0.01). In contrast, in wild type mice, no significant difference in retinal thickness after treatment with NaIO 3 was found vs. untreated controls. (Fig. 3F). An assessment of the localization of retinal damage by NaIO 3 was made by counting the number of nuclei in the inner nuclear layer (INL), outer nuclear layer (ONL) and ganglion cell layer (GCL) of wild type and aB crystallin knockout retina ( Fig. 3G-I). This analysis revealed that the loss of nuclei was more prominent at 3 weeks post-NaIO 3 injection in aB crystallin knockout retina vs. that of wild type. The number of nuclei per unit area showed a significant decrease with NaIO 3 injection in the ONL of aB crystallin knockout mice which was statistically significant (P, 0.01; Fig. 3I). No significant differences in the number of nuclei in any of the other nuclear layers (GCL, INL) were found between the NaIO 3 -injected and PBS-injected groups of wild type mice ( Fig. 3-G,H).

Reduced ERG amplitudes in NaIO 3 -treated aB crystallin knockout mice
To determine whether the absence of aB crystallin had an effect on the retinal function of NaIO 3 -treated mice, we compared mesopic (mixed rod and cone) ERG responses. These studies to assess the functional response of neural retina were performed in four groups of mice (PBS-treated WT, NaIO 3treated WT, PBS-treated aB crystallin knockout, and NaIO 3treated aB crystallin knockout) that received a dose of 20 mg/kg NaIO 3 at the end of 3 weeks. Significant differences were observed in the ERGs of NaIO 3 -treated aB crystallin knockout mice compared with the PBS-treated aB crystallin knockout mice ( Fig 4A). The amplitude of the a wave of the ERG, that originates from the photoreceptors, of NaIO 3 -treated aB crystallin knockout mice decreased by 68.3% compared with that of PBS-treated aB crystallin knockout mice ( Fig 4B). The amplitude of the b wave of the ERG, (that originates from the bipolar cells), of NaIO 3treated aB crystallin knockout mice decreased by 55.3% compared with that of PBS-treated aB crystallin knockout mice ( Fig 4C). No significant differences were found between the ERGs of the NaIO 3 -treated and control wild type mice at this low dose (Fig 4B,C).
Increased production of reactive oxygen species (ROS) in aB crystallin knockout RPE and cultured human RPE cells transfected with aB crystallin siRNA after NaIO 3 treatment These experiments were performed in both mouse and human RPE cultured in 0.5% FBS-containing DMEM; cells were treated with 200 mg/ml NaIO 3 for 24 h. Treatment with NaIO 3 induced ROS production in RPE from WT mice which was not found in untreated controls (Fig. 5A-B). ROS partially co-localized with mitochondria. The ROS production was even higher in aB crystallin knockout RPE after NaIO 3 treatment (arrows, Fig. 5D). Negligible ROS was produced in aB crystallin knockout RPE without NaIO 3 (Fig.5C). To further evaluate the effect of NaIO 3 on ROS production, we studied primary human RPE cells after aB crystallin knockdown. The percentage of knockdown of aB crystallin in human RPE by siRNA transfection was about 80% as determined by Western blot analysis ( Fig 5E). Treatment with NaIO 3 resulted in a pronounced intracellular generation of ROS that was predominantly localized to the mitochondria in aB crystallin siRNA-transfected RPE cells ( Fig. 5I-K). However, the staining for ROS was much less prominent in NaIO 3 -treated RPE cells with scrambled siRNA (Fig. 5F-H). Thus, these results show that knockout or siRNA knockdown of aB crystallin results in increased generation of ROS in RPE cells treated with NaIO 3 .
Mode of cell death in RPE exposed to low dose of NaIO 3 is not by necrosis Propidium iodide (PI) staining was performed to assess necrotic features in RPE cells incubated with different doses of NaIO 3 . The NaIO 3 treatment of RPE was performed for 24 h in 0.5% FBScontaining DMEM at doses of 200, 500, or 1000 mg/ml, respectively. Confocal microscopy images are presented in Fig. 6 (A-H) that show PI staining in control (scrambled siRNA) RPE nuclei ( Fig. 6A-D) and aB crystallin siRNA-transfected RPE nuclei ( Fig. 6E-H) with or without NaIO 3 treatment. No significant differences were found in the number of PI positive cells between control group and the group treated with 200 mg/ml of NaIO 3 in both scrambled siRNA and aB crystallin siRNAtransfected human RPE cells. Treatment with 500 and 1000 mg/ ml of NaIO 3 resulted in increased PI positive cells both in scrambled siRNA RPE and aB crystallin siRNA pretreated RPE (P,0.01). However, no significant differences were found between the number of PI positive cells in control RPE and aB crystallin siRNA-transfected groups (Fig. 6I). Therefore, it can be concluded that high dose NaIO 3 (500 mg/ml and 1000 mg/ml) induced predominantly RPE cell necrosis, while induction of necrosis was negligible or insignificant with low dose NaIO 3 (200 mg/ml).
Increased apoptosis in aB crystallin siRNA-transfected RPE with low dose NaIO 3 TUNEL staining was performed to assess the extent of apoptosis with NaIO 3 in RPE cells. However, no significant difference was found between control and NaIO 3 -treated RPE cells without aB crystallin siRNAtransfection. Thus NaIO 3 induces apoptosis in aB crystallin siRNA-transfected RPE cells that were exposed to low doses.

Increased caspase 3 activation with NaIO 3 and aB crystallin siRNA
Cleaved caspase 3 staining was performed to confirm that the mechanism of cell death was by apoptosis. Fig. 8 shows immunostaining of cleaved caspase 3 in control human RPE cells  Signaling molecules associated with NaIO 3 -induced RPE cell death We investigated the mechanism of the exacerbation of NaIO 3induced RPE cell apoptosis in the absence of aB crystallin by analyzing several apoptotic signaling proteins (Fig. 9). Expression of phosphorylated AKT at serine 473 increased after treatment with low dose (200 mg/ml) NaIO 3 (Fig. 9B). However, the increase of phospho-AKT was much lower in aB crystallin siRNAtransfected RPE cells treated with NaIO 3 compared to controls. A similar trend was observed for phospho-GSK 3b and phosphorylated -c-Raf, signaling proteins downstream of AKT. Low dose NaIO 3 markedly increased PPAR c expression in scrambled transfected RPE cells, while this increase could not be seen in aB crystallin siRNA-transfected RPE cells (Fig. 9C). However, no significant changes were evident with NaIO 3 treatment in phospho-PDK1, a kinase upstream of AKT (Fig. 9A). This suggests that AKT may be the point of interaction with NaIO 3 in the AKT signaling pathway (Fig. 9A).
The treatment with low dose (200 mg/ml) NaIO 3 increased PPAR c expression in RPE. The increased expression of PPAR c possibly exerts a reactive protective role, as we observed A-PAF, a PPAR c ligand, significantly decreased the generation of ROS (Fig. 10A). On the other hand, GW9662, a PPAR c antagonist, caused a significant increase in NaIO 3 -induced production of ROS (P,0.01; Fig. 10B). The increase in PPAR c by NaIO 3 was attenuated in NaIO 3 -treated aB crystallin siRNA-transfected RPE cells thereby compromising the protective defense by PPAR c under these conditions (Fig. 10A, B).

Discussion
In an attempt to understand the protective role of aB crystallin in stress-induced RPE degeneration, we have investigated the effect of suppression of aB crystallin on apoptosis and have studied the signaling mechanisms associated with this phenomenon. For this purpose, we used a murine model of NaIO 3 -induced retinal degeneration in vivo and cell death in human RPE in vitro. NaIO 3 has been previously shown to induce selective degeneration of the RPE and consequent retinal degeneration [6,19,20]. A low dose of NaIO 3 was used to induce retinal degeneration in this study. No significant retinal degeneration was found on fundus photography in wild type mice three weeks after tail vein injection of low dose NaIO 3 , while 93.7% (15/16) eyes of aB crystallin knockout mice exhibited retinal degeneration after the same treatment. A much higher dose and duration of treatment (100 mg/kg NaIO 3 ; 6 weeks) was required in ICR strain of mice to induce changes in morphology in the retina [21] indicating that differences among mouse strains can also play a role [19][20][21]. Furthermore, previous ultrastructural and TUNEL labeling studies showed that RPE cell death induced by 100 mg/kg NaIO 3 was from necrosis and that of the photoreceptors was from apoptosis [21]. In the present study, increased RPE apoptosis was found in low dose NaIO 3 -treated RPE cells from aB crystallin knockout mice and in human RPE after aB crystallin knockdown; however, necrosis was minimal. Necrotic cells increased in a dose dependent manner with high dose of NaIO 3 . Therefore, we may conclude that low dose of NaIO 3 induces RPE cell apoptosis, while high dose of NaIO 3 results in RPE cell necrosis.
The role of aB crystallin in cellular protection is becoming increasingly important because aB crystallin acts on a variety of cellular processes [23,27]. Newer studies have taken advantage of aB crystallin's antiapoptotic and anti-inflammatory properties in devising therapy [34,35]. For example, intravenous administration of aB crystallin in mice was found to reduce inflammation and thus play a protective role in experimental autoimmune demyelination [34,36]. aB crystallin may play an important role in protection of retinal neurons from damage by metabolic and environmental stress as seen by evidence of elevated crystallin expression in light damaged photoreceptors and in models of retinal degeneration [37,38]. aB crystallin could be important in the development of, or in response to, AMD since aB crystallin was found to be accumulated in RPE, drusen and Bruch membrane tissues from AMD patients [23,24]. In a recent study, we found that aB crystallin is secreted via exosomes by RPE cells and presented evidence for its extracellular function in protecting neighboring RPE cells and photoreceptors from oxidative injury [33].
In our present studies, we found that lack of aB crystallin accelerated and augmented the retinal degeneration in NaIO 3treated mice in vivo and was associated with increased RPE cell apoptosis in vitro. In previous studies from our laboratories, we had reported that lack of aB crystallin renders RPE cells more susceptible to apoptosis from oxidative stress induced by H 2 O 2 [28]. Similarly, when RPE cells were exposed to ER stress, Thirteen NaIO 3 -treated aB-/mice showed patchy retinal degeneration three weeks after NaIO 3 injection (E). Only three out of fourteen eyes of NaIO 3 -treated WT mice showed retinal degeneration (E). A statistically significant difference was found between NaIO 3 -treated aB-/and NaIO 3 -treated WT mice (P,0.001). doi:10.1371/journal.pone.0098275.g002 apoptosis ensues and aB crystallin regulated ER-stress induced cell death [31]. Silencing of aB crystallin by siRNA knockdown exacerbated apoptosis while overexpression attenuated apoptotic cell death in RPE cells [30,31]. Our present data show that induction of apoptosis by NaIO 3 occurs through generation of ROS consistent with the known oxidative properties of iodate ions involving mitochondria [6,19,20].
In the aB crystallin knockout mouse, knockout of the aB crystallin gene also disrupted the closely related gene HSPB2 [29]. However, while HSPB2 is expressed in muscle tissues, our previous work established that HSPB2 is not expressed in normal or pathologic murine posterior eye cups, including the retina and RPE [29,39]. Therefore, loss of HSPB2 has no effect on the evaluation of aB crystallin knockout in studies of retinal degeneration.
It was reported that there was no apparent phenotype in the retina of aB crystallin knockout mice [29]; however, our recent studies found that while the histology of the neural retina was unaffected, there was a mild decrease in retinal vessel density in the inner plexiform layer in aB crystallin knockout mice compared to wild type [39]. We found that absence of aB crystallin accelerated and augmented the degeneration of the retina in NaIO 3 treated mice in this study. Further, apoptosis was exacerbated in RPE after aB crystallin siRNA knockdown. For example, we observed increased production of ROS in RPE cells from aB crystallin knockout mice and human RPE cells transfected with aB crystallin siRNA upon NaIO 3 treatment. These results suggest that while an apparent retinal phenotype could not be found in vivo in normal conditions, suppression of aB crystallin does indeed cause injury and death at a cellular level in RPE after oxidative stress. Furthermore, the mode of cell death was via apoptosis and necrosis was not seen in RPE cells treated with low NaIO 3 doses.
It is of interest that very recently it was shown that knockout of aA crystallin also exacerbates retinal degeneration in the NaIO 3 model [40]. We have previously shown that aB crystallin is expressed at much higher levels in RPE than aA crystallin but that  knockout of either aA or aB crystallin in RPE renders them more susceptible to oxidative stress [28]. Thus, it might be interesting to study the effects of double knockout of aA and aB crystallin on the extent of retinal degeneration in this model.
We used human RPE cultures in vitro to elucidate the mechanism of NaIO 3 induced apoptosis under conditions of aB crystallin deficiency. PPARc , a member of a nuclear receptor superfamily, plays a key role in numerous cellular functions and is 24 h. In panels A-D, DAPI is shown in blue, ROS staining in green and mitotracker in red. ROS staining was observed in WT RPE cells treated with NaIO 3 (A-B). RPE cells from aB-/mice (C-D, see white arrows) showed increased accumulation of ROS that partially colocalized with mitochondria. For studies with human RPE cells, aB crystallin was knocked down (,80%) by siRNA transfection (Fig. 5E). Increased ROS production was observed in aB crystallin si-RNA transfected human RPE cells with NaIO 3 as compared to scrambled siRNA transfected cells which showed negligible ROS staining (I-K). (F-H). Scale bars represent 20 mm for A-D and 10 mm for F-K, respectively. doi:10.1371/journal.pone.0098275.g005 a key regulator of mitochondrial biogenesis and of ROS metabolism [41]. We found that the expression of PPAR c protein in RPE cells increased after treatment with low dose NaIO 3 . The increase in PPAR c expression was significantly lower in aB crystallin siRNA-transfected RPE cells treated with NaIO 3 . Furthermore, the PPAR c ligand A-PAF inhibited ROS production in aB crystallin knockdown RPE cells treated with NaIO 3 . This finding of ROS inhibition by PPAR c ligand in RPE cells is consistent with some recent reports in other cell types. In mesangial cells, PPAR c ligand rosiglitazone abolished ROS generation during exposure to high glucose, while inhibition of PPAR c by GW9662 caused ROS generation in normal glucose [41]. Further, aB crystallin was shown to effectively inhibit both ROS formation and apoptosis in cultured vascular endothelial cells [42]. The ROS-inhibitory function of PPAR c could arise from the antioxidative properties reported for PPAR c . For example, it was shown recently that thiazolidinediones, synthetic ligands of PPAR c , effectively protected pancreatic beta-cells from oxidative stress by an increase in the expression of the antioxidative enzyme catalase [43]. Similarly, antioxidative, neuroprotective function for PPAR c was reported in a model of Parkinson's disease [44]. It will be of interest to investigate whether the observed antiapoptotic function of PPAR c in RPE is linked to any changes in endogenous antioxidant enzymes. In this context, our laboratory has shown that overexpression of aB crystallin protects human RPE from oxidative and ER stress and upregulation of GSH and its biosynthetic enzymes are involved in this process [31,32,45].
The phosphoinositide 3-kinase (PI3K)-Akt pathway serves to coordinate the cellular response and ultimately determine cell fate  [46]. Akt activation enhances RPE cell survival. It was reported that H 2 O 2 induced PI3K and thereby activated Akt in human RPE cells [47]. AKT activation occurs through direct oxidation of phosphatase tensin homologue (PTEN) in acute oxidative stress [48]. We found in the present study that phosphorylated Akt and the signaling proteins downstream of AKT increased in RPE cells after treatment with NaIO 3 , Further, knockdown of aB crystallin by siRNA suppressed the activation of Akt. Together, these data suggest aB crystallin mediated protection of RPE cells from NaIO 3 induced oxidative stress involves AKT. Working with HeLa cells, Pasupuleti et al. found evidence for activation of the PI3K/Akt cell survival pathway by alphaA crystallin by promoting phosphorylation of PDK1, AKT and PTEN [49]. It is of interest that Zhao et al reported that RPE dedifferentiation and hypertrophy in a model of oxidative phosphorylation (OXPHOS) deficiency or NaIO 3 administration to B6 mice resulted in the stimulation of AKT/ mammalian target of rapamycin (AKT/mTOR pathway [7]. Further, evidence for RPE oxidative damage and a rapid reduction of RPE65 and several other RPE-characteristic proteins was found [7]. This led the authors to suggest that mTOR pathway inhibition could be an effective therapeutic strategy for retinal degenerative diseases involving RPE stress [50].
In conclusion, our data show that aB crystallin plays a critical role in protection of NaIO 3 induced oxidative and retinal degeneration in part through upregulation of AKT phosphorylation and PPAR c expression.

Ethics statement
This study conforms to applicable regulatory guidelines at the University of Southern California, principles of human subject The 129S6/SvEvTac wild type mice were purchased from Taconic Farms (Germantown, NY), and the aB crystallin knockout mice in 129S6/SvEvTac background were obtained from the National Eye Institute [28,29]. Mice aged between 6 and 8 weeks maintained on a standard laboratory chow in an airconditioned room equipped with a 12-hour light/12-hour dark cycle were used in all studies.

Experimental groups and NaIO 3 treatment
The mice were divided into four groups of seven mice per group: control wild type (PBS-treated WT), NaIO 3 -treated wild type (NaIO 3 -treated WT), control (PBS) aB crystallin knockout mice, and NaIO 3 -treated aB crystallin knockout mice.
Experiments to determine the dose and time-dependent effect of NaIO 3 were performed using doses of 20 mg/kg, 35 mg/kg, 50 mg/kg and 70 mg/kg body weight and duration of the study was one week to three weeks post NaIO 3 administration. Briefly, varying doses of sodium iodate (NaIO 3 ; Sigma, St. Louis, MO) diluted with Phosphate buffered saline (PBS) were injected through the tail vein to restrained mice. Animals injected with equivalent volumes of PBS served as controls. Electroretinography and fundus photograph (see below) were assessed 21 days postinjection. After the tests were performed, mice were euthanized with CO 2 and their eyes processed for histology.

Electroretinography (ERG)
Mice were dark-adapted overnight and anesthetized by intraperitoneal injection of ketamine (100 mg/kg body weight) and xylazine (10 mg/kg body weight). Pupils were dilated with topical administration of 2.5% phenylephrine containing 0.5% tropicamide, and the cornea was anesthetized with 0.5% proparacaine. Mesopic ERGs were measured using a nonattenuated light stimulus. To measure cone responses, a 6 lux white background light was delivered through the other arm of the coaxial cable to suppress rod responses, and a non-attenuated light stimulus was applied. a-Wave amplitude was measured from the baseline to the trough of the a-wave, while b-wave amplitude was measured from the trough of the a-wave to the peak of the b-wave [51].

Fundus photography
Mice were anesthetized by administration of ketamine and xylazine as described above. Pupils were dilated and the cornea was anesthetized with 0.5% proparacaine. Images were captured

Histopathologic analysis
Eyes were enucleated and the anterior segments were removed. The remaining posterior eye cups were snap-frozen in tissue freezing medium (Triangle Biomedical Sciences, Durham, NC), Optimal cutting temperature (OCT). Cryostat Sections (8 mm) were stained with hematoxylin and eosin (H&E), to assess the histopathologic changes.
Mouse retinal sections were scanned and retinal thickness was measured (Aperio ScanScope; Leica Biosystems) using Aperio software. Cell numbers in RPE layer, GCL layer, INL layer and ONL layer were determined by counting the nuclei in a 50 mm wide region of retinal section located at equal distance from the ora serrata and the optic disc. For each group, three eyes were dissected. For each, three different regions were counted by Image J 4.3.2 (NIH Image). Average cell numbers and standard deviation were calculated using Statlab (SPSS Inc, Chicago, Illinois, USA).

Human and mouse RPE cell cultures
All procedures conformed to the Declaration of Helsinki for research involving human subjects and were performed with the approval of the institutional review board (IRB) of the University of Southern California. Human RPE cells were isolated from fetal human eyes of 16-18 wks gestation (Advanced Bioscience Resources, Inc., Alameda, CA) as previously described [52,53]. Cells were cultured in DMEM (Fisher Scientific, Pittsburgh, PA) with 2 mM L-glutamine, 100 U/ml penicillin, 100 mg/ml streptomycin (Sigma, St. Louis, MO), and 10% heat-inactivated fetal bovine serum (FBS, Irvine Scientific, Santa Ana, CA). The preparations contained .95% RPE cells (cytokeratin-positive). Cells used were from passages 2 to 4. Primary mouse RPE cells were isolated as previously described [54]. Primary mouse RPE cells were isolated from 4 to 6 week old WT (129S6/SvEvTac) and aB crystallin knockout mice. RPE cells were cultured in DMEM containing 20% FBS and antibiotics until confluent and P3 cells were used for experiments. aB crystallin small interfering RNA (siRNA) transfection Human RPE cells were switched to DMEM containing 0.5% FBS shortly before transfection. siRNA targeting aB crystallin was diluted in DMEM without serum. HiPerFect Transfection Reagent (Qiagen, Valencia, CA) was added to the diluted siRNA and mixed by vortexing. After incubation for 10 min at room temperature, the complexes were added dropwise to RPE cells. The final siRNA concentration was 5 nM. The cells were harvested or fixed for further assay 24 hours later. The sequence for siRNA targeting aB crystallin was: sense: r(CCA GGG AGU UCC ACA GGA A)dTdT; antisense:r(UUC CUG UGG AAC UCC CUG G) dTdT; nonsilencing control siRNA (scrambled siRNA): sense r(UUC UCC GAA CGU GUC ACG U) dTdT; antisense:r(ACG UGA CAC GUU CGG AGA A) dTdT. Fortyeight hours after transfection, in vitro effects of NaIO 3 were studied either with a fixed final concentration 200 mg/ml added to the culture medium or at different doses as specified.

Determination of ROS
To determine the compartmentalized generation of reactive oxygen species (ROS), mitochondria were labeled by a cellpermeable mitochondria-specific red fluorescent dye (Mito-Tracker, Molecular Probes); stained with carboxy-H2-DCFDA (Molecular Probes; 5 mM for 1 h at 37 uC), and rapidly evaluated by confocal microscopy (LSM510, Zeiss, Thornwood, NY, USA) as previously described [28,31]. A yellow color is observed when ROS (green) are colocalized in the mitochondria (red). In some experiments, the effect of treatment with Azelaoyl PAF (A-PAF) (Sigma, St, MO, USA), a PPAR c ligand at a concentration of 20 mM and GW9662 (Cayman Chemicals, Ann Arbor, Mich, USA), a PPAR c antagonist, at a concentration of 10 mM was determined [55,56].

Determination of necrosis and apoptosis with NaIO 3
Propidium Iodide (PI) stains DNA of necrotic cells [28,31]. Human RPE cells on an eight-well Lab-TekTM chamber were treated with 10 mg/ml PI (Roche Applied Science) for 15 min at 25 uC in the dark. Cells were washed once with ice-cold PBS and observed under a laser scanning confocal microscope (LSM510, Zeiss, Thornwood, NY, USA).
Apoptosis (DNA fragmentation) was detected by the terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end-labeling (TUNEL) method according to the manufacturer's protocol (In situ cell death detection kit-POD; Roche Applied Science). In short, after treatment with several doses of NaIO 3 at room temperature, RPE cells on eight-well Lab-TekTM chambers were fixed in 1% paraformaldehyde solution and rinsed with PBS. Cells were then incubated with the TUNEL reaction mixture containing TdT and fluorescence UTP for 1 hour at 37uC in a humidified chamber. The nucleotides incorporated into DNA breaks were detected by applying anti-fluorescein peroxidase (POD) conjugate and peroxidase substrate.

Immunocytochemistry of Cleaved Caspase-3
Human RPE cells on an eight-well Lab-TekTM chamber were fixed in 4% paraformaldehyde for 30 min, and then permeabilized using 0.2% Triton-X 100 at 37uC for 15 min. Blocking was achieved by addition of 1% goat serum for 20 min. The samples were incubated with primary anti-cleaved caspase-3 antibody (Cell Signaling; 1:200) for 1 hr at room temperature. After washing with PBS, secondary biotinylated conjugated goat anti-rabbit antibody (1:400; Vector, Burlingame, CA, USA) was applied to the slides for 30 min at room temperature. After washing with PBS, streptavidin peroxidase (Invitrogen, Camarillo, CA, USA) was applied to the slides for 30 min. 3-Amino-9-Ethylcarbazole (AEC) was added to the slide (AEC Substrate Kit, Invitrogen, Camarillo) which produced a red colored deposit. Sections were examined and photographed with microscope (Leica, Germany).

Statistics
All experiments were performed at least three times. The data were analyzed using the Student's t-test (amplitudes of ERG; number of nuclei of outer nuclear layer, inner nuclear layer and ganglion layer histopathology; ROS; TUNEL; Cleaved caspase 3; and western blot) or Chi-square (fundus photography, RPE layer histopathology) and P,0.05 was considered as significant. Figure S1 Fundus images showing time-dependent effect of a single dose of NaIO 3 on wild type (WT) and aB crystallin -/-(aB-/-) mice. Representative images from a single mouse from WT and aB-/groups are shown on the left accompanied by data for all experimental animals on the right. Fundus photography was taken one, two and three weeks after tail vein injection of PBS or 20 mg/kg NaIO 3 . PBS-treated WT and PBS-treated aB -/-did not exhibit any degenerative changes (data not shown). NaIO 3treated WT mice did not show retinal degeneration at any time point (A). However, NaIO 3 -treated aB-/mice showed patchy retinal degeneration two and three weeks after NaIO 3 injection (B). Arrow indicates the site of patchy retinal degeneration. aB-/refers to aB crystallin knockout mice. (TIF) Figure S2 NaIO 3 -induced cell death in RPE layer of aB crystallin knockout mouse retina as determined by TUNEL staining. TUNEL staining was performed after WT and aB crystallin knockout mice were injected with 20 mg/kg NaIO 3 . No TUNEL+ cells were observed in the RPE layer of WT retina while TUNEL+ cells could be easily identified in the RPE layer of aB crystallin knockout retina (white arrow). Scale bar = 50 mm. (TIF)