Elevated Retina-Specific Expression of the Small Heat Shock Protein, αA-crystallin, Is Associated with Photoreceptor Protection in Experimental Uveitis

Abstract

purpose. During the early phase of experimental autoimmune uveitis (EAU), before macrophages infiltrate the retina and uvea, photoreceptor mitochondrial oxidative stress, nitration of photoreceptor mitochondrial proteins, and release of cytochrome c have been observed. However, no apoptosis has been detected during this phase. In this study, αA-crystallin upregulation in the retina and its antiapoptotic protective role were evaluated in early EAU.

methods. Gene microarrays were first used to identify upregulated genes in retinas with early EAU. Among highly upregulated crystallins, αA was confirmed by real-time polymerase chain reaction and Western blot, and the site of upregulation was localized by immunohistochemistry. The association of αA-crystallin to nitrated cytochrome c and interaction with a procaspase-3 subunit was assayed. Photoreceptor apoptosis in αA knockout mice was compared with that in wild-type animals with EAU, by using the terminal transferase dUTP nick-end labeling assay and polymerase chain reaction.

results. In early EAU, αA-crystallin was increased 33-fold, and the site of increase was localized to the photoreceptor inner segments. This crystallin suppressed apoptosis by associating with the nitrated cytochrome c and p24. The association with nitrated cytochrome c, in particular, appeared to be restricted to nitrated cytochrome c, and thus, no association of non-nitrated cytochrome c was detected. The knockout mice showed signs of EAU development early and showed apoptosis in the retina; no such changes were seen in the wild-type control animals.

conclusions. αA-Crystallin is highly upregulated in the retina during early EAU. This upregulation is localized primarily in the photoreceptor inner segments, the site of mitochondrial oxidative stress. Further, in early EAU, the photoreceptors preferentially use αA-crystallin to suppress mitochondrial oxidative stress-mediated apoptosis.

Intraocular inflammation, commonly referred to as uveitis, is a leading cause of blindness from retinal photoreceptor cell degeneration. ¹ The rodent model of experimental autoimmune uveitis (EAU) closely resembles the human condition and has been used for the study of T-cell-mediated inflammatory processes in the retina. ² Blood-borne, activated macrophages are major effectors of retinal damage in EAU. In this model, infiltration of the retina by macrophages usually takes place on days 11 to 14 after the injection of a uveitogenic antigen into susceptible animals. ^{3 4} We have shown that in photoreceptor cells, nitration of mitochondrial proteins, including cytochrome c (cyto c) as an indicator of mitochondrial oxidative stress precedes macrophage infiltration. ^{5 6} It is known that oxidative stress in general is accompanied by upregulation of a host of heat shock proteins (Hsp), including Hsp 27, -70 and -90 and the α-crystallins. ⁷ Several of these have been shown to be antiapoptotic through the inhibition of one or more components of the apoptotic machinery. ^{8 9 10}  

Crystallins, the major structural proteins of the eye lens, are primarily categorized into three distinct families: α, β and γ. The two α-crystallins, αA and αB, are the principal members of the small Hsp family of molecular chaperones. ¹¹ They are expressed in multiple tissues, including retina, brain, heart, kidney, spinal cord, and lungs. ^{12 13 14 15} Although αA- and αB-crystallin have related amino acid sequences with similar structural properties, they vary significantly in their tissue specificity and phosphorylation sites. ^{16 17} αA- and αB-crystallins have different functions: They protect different proteins and are active under different conditions. ^{18 19 20} αB-crystallin is upregulated in various neurodegenerative and autoimmune diseases. In contrast, neither αA-crystallin upregulation nor any protective effects of αA-crystallin have been reported under such conditions. 

The present study was designed to investigate the expression of genes that are modulated in this EAU model, particularly at the site of mitochondrial oxidative stress at an early time point that precedes the onset of visible pathogenesis involving macrophage infiltration into the retina and uvea and the apoptosis attendant to photoreceptor cell death. The study revealed a selective upregulation of αA-crystallin, but not of αB or other Hsps, in the photoreceptor inner segments, the site of mitochondrial oxidative stress. Moreover, αA-crystallin effectively intercepted the mitochondrial oxidative stress-related apoptotic processes. 

Materials and Methods

Induction of EAU and Detection of Apoptosis in the EAU Retina

Animal care and use was in compliance with institutional guidelines and with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. EAU was induced in 8-week-old B10.RIII mice (Jackson Laboratory, Bar Harbor, ME) as described earlier. ²¹ The control group was injected with complete Freund’s adjuvant (CFA) alone. Another group of noninjected mice was used as an additional control. 

Six EAU, six CFA control, and six noninjected control mice were euthanatized on day 7 after immunization (early EAU). An additional six EAU, six CFA, and six noninjected animals were euthanatized on day 14. The eyes were fixed in 4% formalin, embedded in paraffin, and sectioned for hematoxylin-eosin (H&E) and terminal transferase dUTP nick-end labeling (TUNEL) staining. The TUNEL procedure was performed with an apoptosis detection kit (ApopTag Plus Peroxidase In Situ; Chemicon, Temecula, CA) according to manufacturer’s instructions. The sections were stained with peroxidase substrate and diaminobenzidine and then counterstained with H&E and viewed by light microscopy. Phosphate-buffered saline (PBS) was used in place of TdT enzyme as the negative control, and positive slides provided with the kit were used as the positive control. The staining was performed in triplicate. 

Detection of Upregulated Genes by Microarray in Early EAU

Twenty early EAU, 20 CFA control, and 20 noninjected animals were euthanatized on day 7 after immunization. The retinas were dissected without lens material contamination and snap frozen. The total RNA was extracted (TriZol method; Invitrogen, Carlsbad, CA) from control and experimental retinas. Each group was further divided into two groups. RNA was quantitated, checked for purity, and analyzed for integrity by microanalysis (Agilent Bioanalyzer; Agilent Technologies, Santa Clara, CA). Probes for array analysis were prepared according to the manufacturer’s protocol (Affymetrix, Inc., Santa Clara, CA). ²² Expression analysis was performed by hybridization on U74Av2 murine whole-genome chips (Affymetrix). The hybridized array was washed, labeled with phycoerythrin-conjugated streptavidin (Invitrogen-Molecular Probes, Eugene, OR), scanned (Affymetrix scanner), and analyzed with Wing Wong DCHIP software. ²² Two sets of each control and two sets of EAU retina samples were run in parallel and compared for differential expression. 

Real-Time PCR Confirmation of Crystallin mRNA Upregulation

Real-time polymerase chain reaction (PCR) was performed to confirm the microarray analysis, using the same retinal samples from EAU and control animals (iCycler Optical System; Bio-Rad Laboratories, Hercules, CA). Similarly, the brains, hearts, and livers of these animals were used as separate samples. Total RNA was prepared (TriZol reagent; Invitrogen), and the cDNA template was generated (Omniscript RT kit; Qiagen, Valencia, CA). The tissues from the CFA control group were similarly processed. Each 25-μL PCR reaction mixture contained a master mix (SYBR Green I; Bio-Rad Laboratories); 0.5 μM of gene-specific primers for αA-, βA1-, βB2-, and γS-crystallin; and the cDNA template. In the quantification analysis, all crystallin mRNAs were normalized with glyceraldehyde-3-phosphate dehydrogenase (GAPDH), as it remained constant in our experimental conditions. The sequences of primers used are shown in Table 1 . 

The PCR reactions for each gene in each experiment were performed in triplicate on each cDNA template, along with triplicate reactions of the housekeeping gene GAPDH. The specificity of PCR amplification products was checked by performing dissociation melting curve analysis. The threshold cycle (Ct) difference between the experimental and control groups, for each crystallin gene in each tissue, was calculated and normalized to GAPDH, and the increase (x-fold) in mRNA expression was determined by the 2^−ΔΔCt method. ²³ Statistical analysis of ΔΔCt was performed with a Student’s t-test for three independent samples, with significance set as P < 0.05, and compared between the EAU and control CFA groups. Animals injected with CFA were also compared with noninjected control animals. The experiments were performed in triplicate. Similarly, gene expression of αB-crystallin was quantified in the retina, brain, heart, and liver of early-EAU animals, by using gene-specific primers (Table 1) , and compared with the control groups. 

Western Blot Analysis of αA- and αB-Crystallins

Retinas of five early-EAU mice and five CFA control animals were dissected and homogenized in PBS containing protease inhibitors. The homogenates were sonicated for 30 seconds and centrifuged at 13,000 rpm for 30 minutes at 4°C. The protein concentration of the supernatant was determined (Bio-Rad Laboratories) using bovine serum albumin as the standard. Equal amounts of protein samples were loaded and run on SDS-PAGE (15% Tris-HCl polyacrylamide ready gels; Bio-Rad). After electrophoresis, the proteins were transferred onto nitrocellulose membranes using a transblot semidry system (Bio-Rad). The membranes were blocked in 5% skim milk and then probed with a polyclonal anti-αA-crystallin (provided by author SB) at a 1:1000 dilution overnight at 4°C. After incubation for 30 minutes with the secondary antibody tagged with horseradish peroxidase, signals were detected by the chemiluminescence system (GE Healthcare, Piscataway, NJ). Similarly, αB-crystallin was detected by probing the membranes with monoclonal anti-αB-crystallin (Stressgen Bioreagents, Victoria, BC, Canada) at 1:1000 dilution. Equal protein loading of retinal lysates from control and experimental animals was confirmed by reprobing blots with a monoclonal antibody to GAPDH. The anti-αA crystallin used was made commercially (Sigma Genosys, Woodlands, TX) against the N-terminal peptide (residues 2-13): CDVTIQHPWFKRA of rat αA crystallin using the peptide-KLH conjugate for immunization. (C is not in the primary sequence of αA crystallin). The anti-serum was then purified on a protein G column and tested for its specificity and cross-reactivity with αB by Western blot analysis. The result indicated that the anti-αA crystallin was selective, with no cross-reactivity against αB crystallin. 

Localization of αA- and αB-Crystallins in the Early-EAU Retina

Ten-micrometer cryosections of the retina were obtained from six mice each from the early-EAU and the CFA control groups. To detect αA-crystallin, sections were probed with a polyclonal anti-αA-crystallin at 1:50 dilution and then with Alexa-Fluor⁴⁸⁸-conjugated goat anti-rabbit IgG (Invitrogen-Molecular Probes) at 1:200 dilution. For αB-crystallin, an anti-αB-crystallin at 1:100 dilution was used as the primary antibody and Alexa-Fluor⁴⁸⁸-conjugated goat anti-mouse IgG (Invitrogen-Molecular Probes) at 1:200 dilution was used as the secondary antibody. The slides were viewed by confocal microscope (Carl Zeiss, Oberkochen, Germany). Isotype controls and primary antibody replaced by PBS were used as the negative control. The experiments were performed in triplicate. 

Heat Shock Treatment and Detection of Crystallins

Six mice were anesthetized with ketamine-xylazine and placed in a heating blanket until their colonic core temperatures reached 42°C for 10 minutes. ²⁴ The animals were euthanatized 24 hours later. Retina, brain, heart, and liver were subjected to RNA extraction. cDNA was synthesized and subjected to quantitative real-time PCR analysis, as described earlier. In addition to mRNA expression of αA-, βA1-, βB2-, and γS-crystallins, mRNA expression of Hsp 27 and 70 was analyzed by using gene-specific primers (Superarray Bioscience Corp., Frederick, MD). The increases in expression levels of different crystallins, as well as that of Hsp 27 and -70, in different tissues from the experimental animals were calculated and compared with levels in control animal tissues, as described earlier. The experiments were performed in triplicate. 

Detection of Cyto c Release in EAU

Nine early-EAU and nine CFA control animals were euthanatized, six retinas were combined for each determination, and assays were performed in triplicate. A previously described method was used to obtain the cytosolic and mitochondrial fractions. ^{25 26} As a positive control, mitochondria were briefly sonicated (Branson Ultrasonic Corp., Danbury, CT) before centrifugation. The released cyto c in the cytosol was determined by Western blot (15% gel) probed with monoclonal anti-rat cyto c (BD PharMingen, San Diego, CA) as the primary antibody and biotinylated goat anti-mouse IgG (Dako, Carpinteria, CA) as the secondary antibody. No detergent was used to solubilize membranes in these procedures. For each run om these Western blot analyses, the same number of retinas (six per preparation) was used to obtain cytosolic fractions. At this point, total cytosolic proteins were determined by a commercial method (Bio-Rad), with bovine serum albumin used as the standard. Exactly the same amount of cytosolic proteins (9 μg) was loaded in lane 1, control; lane 3, day 7; and lane 4, day 10. The positive control was obtained by sonicating mitochondria to attain artificial release most of the nitrated cyto c. This preparation was initially used to determine the accuracy of pipetting by running the positive control samples at 4, 6, 8, 9, and 10 μg and by repeating the 9 μg three times. The densitometric quantitation of these results indicated that the pipetting error was less than 3%. We then decided that this pipetting accuracy was sufficient for qualitative determination of the release of cyto c during the early EAU. The determination was performed in triplicate, using six retinas for each determination, and the results are consistent. 

Binding of αA-Crystallin to Nitrated Cyto c In Vivo

Nine early-EAU mice and nine CFA control mice were euthanatized, and six eyes were combined for each determination. The retinas were dissected, homogenized, and centrifuged. Aliquots of supernatant (each containing 50 μg protein) were incubated with polyclonal αA-crystallin antibody (Stressgen Bioreagents) overnight at 4°C. Protein A agarose slurry (50 μL; Sigma-Aldrich, St. Louis, MO) was added to each sample and incubated at 4°C for another 3 hours. The protein mixture was then washed by repeated centrifugation with Tris-buffered saline containing 0.1% Triton X-100 (four times) and PBS (once). The resultant pellet was resuspended in 2× SDS sample buffer. After boiling, the samples were analyzed by 15% SDS-PAGE and transferred to polyvinylidene difluoride membranes (Millipore, Billerica, MA). Western blot analysis was performed with monoclonal anti-cyto c as the primary antibody and biotinylated goat anti-mouse IgG as the secondary antibody. Visualization was performed with 3,3′diaminobenzidine/NiCl₂ reagent. 

Binding of αA-Crystallin to Nitrated Cyto c In Vitro

For detecting the in vitro binding of αA-crystallin and nitrated cyto c, bovine αA-crystallin (Assay Design, Ann Arbor, MI), 0.5 mg/mL in phosphate buffer, was immobilized on the sensor chip using the amine coupling method (Amine Coupling Kit; Biacore, Piscataway, NJ) by injecting for 5 minutes at a flow of rate of 10 μL/min. Nitrated cyto c was passed over the chip surface. Surface plasmon resonance measurements were performed (T100 instrument; Biacore) using a series S sensor chip (CM5; Biacore AB, Uppsala, Sweden). The kinetics of the association phase were recorded. The results were analyzed with the system’s evaluation software (T100; Biacore). 

Binding of αA-Crystallin to p24 and Procaspase-3

Eighteen non-EAU eyes were divided into three groups, with each group containing six eyes as one sample. Retinas were suspended in 1.0 mL of hypotonic extraction buffer (50 mM PIPES [pH 7.4], 50 mM KCl, 5 mM EGTA, 2 mM MgCl₂, 1 mM dithiothreitol, and 0.1 mM phenylmethylsulfonyl fluoride), ^{27 28} Homogenized (Dounce; Bellco Glass Co., Vineland, NJ), and centrifuged for 30 minutes at 14,000g. The supernatant was either used immediately or stored in aliquots at −70°C. Cytosolic procaspase-9 and -3 were activated by adding 10 μM nitrated cyto c, 10 mM dATP (Sigma-Aldrich) and incubated for 2 hours at 37°C. For the controls, no nitrated cyto c or dATP was added. To establish the protective effect of crystallins, two concentrations, 20 and 35 μM, of αA-crystallin (Sigma-Aldrich) were used. The reaction products were resolved by 15% SDS-PAGE, and the Western blot was probed with anti-caspase-3 (Stressgen). 

EAU Induction and Apoptosis Detection in αA-Knockout Mice

A group of 24, 8-week-old αA knockout mice ²⁹ (provided by author EW) and 24 corresponding wild-type, 129SvEv mice (Taconic Laboratory, Germantown, NY) were injected with IRBP, as described earlier. ²¹  

Six animals from each group were euthanatized on day 14. Their enucleated eyes were fixed in 4% formalin and embedded in paraffin, and sections were stained with H&E and processed for TUNEL staining, as described earlier. As the negative control, PBS was added in place of the TdT enzyme, and the positive control slides provided by the manufacturer were used. Experiments were performed in triplicate. Another group of six knockout and six wild-type animals were euthanatized on day 18 after immunization with IRBP, and the enucleated eyes were processed for H&E staining. 

To detect genes related to apoptosis and its signaling pathway, retinas from a group of 12 αA-crystallin-knockout mice and 12 129SvEv mice immunized with IRBP and euthanatized on day 14 were subjected to PCR array (Superarray). The increases in the gene expression levels in αA-crystallin knockout mice were compared with those in the wild-type control mice. 

Results

Induction of EAU and Detection of Apoptosis

There was no inflammatory cell infiltration in the retinas of animals euthanatized on day 7 (Fig. 1B) , whereas the animals immunized with the antigen and euthanatized on day 14 showed inflammatory cell infiltration with disruption of photoreceptors (Fig. 1E) . None of the control animals euthanatized on day 7 revealed inflammation (Fig. 1A)or TUNEL-positive staining in the retina (Fig. 1C) . On day 7, the early-EAU experimental animals also showed no apoptosis (Fig. 1D) . In contrast, numerous TUNEL-positive cells were seen in the retinas of EAU animals on day 14 (Fig. 1F) . The TUNEL-positive nuclei were seen mostly in the photoreceptor layer. In these positive eyes, TUNEL-staining was totally abolished when TdT enzyme was replaced with PBS (not shown). 

Detection of Upregulated Genes by Microarray in Early-EAU Retinas

Of the 12,489 genes tested, 48 were upregulated more than 2-fold and 22 were upregulated between 1.5- and 2-fold in the experimental group, compared with the control retinas. Further, in the EAU group, 34 genes were downregulated more than 2-fold and 17 genes between 1.5- and 2-fold, compared with the control retinas. Among the highly upregulated genes, there was profound upregulation of a few crystallin genes, erythropoietin, and cytokines (Table 2) . There was no upregulation of αB-crystallin or of Hsp 27 and -70. 

Real-Time PCR Confirmation of the Upregulated Crystallins

αA-Crystallin increased 33.13-fold; βA1, 30.06-fold; βB2, 44.63-fold; and γS, 34.78-fold in the early EAU retinas, when compared with the CFA control retinas (P < 0.0001). However, these increases were not detected in the brain, heart, or liver of these animals (Fig. 2) . βB2-crystallin was not detected in the heart or liver, and βA1- and γS-crystallin were not detected in the heart. αB-crystallin was seen in all tissues tested; however, it was elevated only 1.52-fold in the EAU retina, 0.66-fold in the brain, 1.07-fold in the heart, and 1.87-fold in the liver. Within the two control groups, the CFA-injected and the nonimmunized controls, there was no significant difference in the levels of expression of any of these genes in the retina, brain, heart, or liver (data not shown). 

Western Blot Analyses of αA- and αB-Crystallin

In the retinal homogenates of EAU animals and CFA-injected control animals, a prominent band of αA-crystallin was detected as expected at 20 kDa (Fig. 3A) . A second, less intense band was seen at 22 kDa, corresponding to an αA insert. In the early-EAU retinas, there was a 10-fold increase in the αA-crystallin protein level in the retina when compared with the control groups, as revealed by Western blot and densitometry measurements (*P < 0.001; Fig. 3B ). A similar increase was seen in the αA insert levels in early EAU. For the upregulation of αA in early EAU, the real-time PCR showed a nearly 30-fold increase, whereas in Western blot analysis, the increase in protein was only 10-fold. The reason for this discrepancy could be multifactorial. The upregulation of gene and protein in vivo does not always have a direct correlation for several reasons: (1) the low stability of transcripts generated; (2) low translational efficiency; and (3) for the low-molecular-weight proteins, such as αA, the run-off from SDS-PAGE gel could become appreciable. Under the same Western blot conditions, αB-crystallin was detected at 21 kDa. This coincides with the published results from other laboratories. ¹⁴ No increase in αB-crystallin was seen in the day 7 EAU retina, compared with control retina (Fig. 3A) . 

Localization of αA and αB-Crystallins in Early EAU Retina

In control retina, αA-crystallin was localized mostly in the ganglion cell and inner nuclear layers (Fig. 4B) . There was minimal staining in the inner segments of the photoreceptors. However, an intense immunostaining was observed in the inner segments of the photoreceptors in early EAU retinas (Fig. 4A) . Less intense levels of staining were also seen in the ganglion and inner nuclear layers. αB-crystallin was also localized to the inner and the outer nuclear layers in EAU, as well as in the control retinas (Figs. 4C 4D , respectively). 

Heat Shock Treatment and Detection of Crystallins

The Hsps 27 and -70 were increased 6.06- and 4.92-fold in the retina, 5.28- and 5.66-fold in the brain, and 9.85-, 6.96-fold and 7.46-, 4.92-fold in the heart and liver, respectively (*P < 0.001). In contrast, there was only a slight change in the expression of αA-, βA-, βB2-, and γS-crystallins during heat shock treatment (Fig. 5) . 

Cyto c Release in the Early Phase of EAU

In the early EAU retina, the cytosolic fraction showed the presence of cyto c (Fig. 6 ; lanes 3, 4). However, in the control animals, no cyto c was detected in the cytosol (Fig. 6 ; lane 1), even after mild sonication (Fig. 6 ; lane 2). Substantially more cyto c release was seen when the outer membranes of mitochondria were sonicated to mimic pathologic opening of the permeability transition pores in the outer membrane. 

Association of αA-Crystallin to Nitrated Cyto c In Vivo

After coimmunoprecipitation, Western blot detection with monoclonal cyto c antibody revealed association of cyto c with αA-crystallin in day 7 EAU retina, but not in the control (Fig. 7A) . Western blot analysis with a nitrotyrosine antibody revealed that the cyto c which associates with αA-crystallin is also nitrated. Retinal supernatant immunoprecipitated with monoclonal β-crystallin antibody and probed with polyclonal nitrotyrosine antibody revealed that nitrated cyto c also associated with β-crystallin in EAU day 7 animals, but not in the control mice (Fig. 7A) . The extent of in vivo binding of nitrated cyto c by αA-crystallin in early-EAU retina was determined by the densitometric reading of nitrated cyto c bands in the Western blot analysis before and after coimmunoprecipitation. The association efficiency αA to nitrated cyto c was in the range of 70% to 80% between runs. 

Association of αA-Crystallin to Nitrated Cyto c In Vitro

In the in vivo association experiment, both αA-crystallin and nitrated cyto c were generated in early EAU. This in vivo experiment was further substantiated by an in vitro association assay in which SIN-1 nitrated cyto c and commercial αA-crystallin were used as two binding components. Real-time association kinetics of nitrated cyto c with αA-crystallin were measured through concentration range of 7.8 nM to 1 μM. The sensorgrams for the surface plasmon resonance analyses indicated that the nitrated cyto c bound to αA-crystallin with high affinity (76%). Under the same conditions, the binding of the non-nitrated cyto c to αA-crystallin was not detected (Fig. 7B) . 

Association of αA-Crystallin to p24, a Procaspase-3 Intermediate Subunit

The addition of nitrated cyto c/dATP to the retinal cell-free system triggered the cleavage of procaspase-3, initially into the intermediate subunit p24. The p24 was then further cleaved to the final executioner subunits p20 and p17. ³⁰ Specifically, the addition of αA-crystallin in the assay led to the dose-dependent accumulation of the p24 partially processed caspase-3 (prodomain and the large subunit) with substantial reduction in the amount of fully processed active caspase-3, p20, and p17 subunits (Fig. 8) . 

EAU Induction and Apoptosis Detection in αA-Knockout Mice

The wild-type animals used as controls for the αA^−/− mice was strain 129SvEv. We had successfully induced EAU in these animals. Inflammation started at a later time point (day 18), which is different from the B10.RIII mice (day 12) we used in our earlier experiments. During early EAU, on day 14 after immunization, there was no inflammatory cell infiltration in the wild-type control (129SvEv) animals (Fig. 9A) , whereas 50% of αA-crystallin-knockout mice showed inflammation in the choroid (Fig. 9B) . On day 18 after immunization, the wild-type animals showed trace inflammation in the uvea (Fig. 9E) , whereas the knockout mice revealed severe inflammation in the uvea (Fig. 9F) . TUNEL-positive cells were seen in the retinas of knockout mice (Fig. 9C) , but such cells were absent in wild-type mice (Fig. 9C) . TUNEL staining was absent in the eyes treated with PBS instead of the TdT enzyme. 

In retinas from knockout mice, the proapoptotic genes, caspases-1, -12, -11, -7, and -8 were upregulated 2.41-, 1.85-, 3.17-, 1.68-, and 1.55-fold, respectively. The TNF and its receptors were elevated 3.55- and 2.05-fold, and the TNF-associated factor 1 was also increased 2.34-fold. Fas was elevated 1.87-fold, and TNF receptor family member CD40 and its ligand CD40l were upregulated 3.02- and 1.60-fold. In contrast, the antiapoptotic genes Bcl10, transformation related protein 73 (Trp73), Birc 2 (a member of the apoptosis-inhibitor proteins), and NFkB1 were downregulated 2.72-, 1.62-, 1.58- and 1.56-fold, respectively (Table 3) . 

Discussion

In the current report, gene array studies, confirmed by real-time PCR and protein expression studies, revealed a selective upregulation of a limited number of crystallins, including the antiapoptotic small Hsp αA-crystallin in the photoreceptor inner segments, the site of mitochondrial oxidative stress during early EAU. Moreover, αA-crystallin inhibited the apoptotic process effectively, in both the in vitro and in vivo assays, as shown in αA-crystallin-knockout mice during early EAU. These observations suggest that upregulation of αA affords a protective mechanism, directed against mitochondrial oxidative stress. Furthermore, no significant change was seen in gene and protein expression profiles of αB-crystallin or other Hsps, including Hsps 27 and -70, which are known to be upregulated during oxidative and other stress conditions. ^{31 32 33 34} Such findings indicate that, in EAU, photoreceptors selectively use αA-crystallin upregulation to protect themselves against mitochondrial oxidative stress-induced apoptosis. 

Mitochondria are the primary intracellular source of reactive oxygen species. When under stress, they generate massive amounts of nitric oxide metabolites, including peroxynitrite. ³⁵ We have reported nitration of cyto c in the photoreceptors during early EAU. ⁵ Nitration of this protein results in its dissociation from the electron transport chain and subsequent release into the cytosol, which activates caspase-3 and initiates apoptotic cell death. ³⁶ In the present study, we detected the release of cyto c into the cytosol. However, we did not detect apoptotic cells in the retina, indicating that activation of antiapoptotic factors such as Hsps in the cytosol may play a role in inhibiting the activation of caspase-3. 

To assess the molecular make-up of the early EAU response, we harvested retinas on day 7 after immunization, when there was no evidence of inflammatory cell infiltration (Fig. 1B)or apoptosis in the photoreceptor cell layer (Fig. 1D)and subjected to a systematic study of the profiles of mRNA expression changes in different retinal genes using DNA microarray analysis. This method allows a thorough analysis of all the up- and downregulated genes by which the potential cellular pathways and molecular complexes that are activated during oxidative stress can be identified. ^{37 38} Microarray analyses of the retinal RNA transcripts revealed differential expression of several genes, among these the crystallins, cytokines, and erythropoietin were highly expressed. Crystallins αA, βA1, βB2, and γS were upregulated 10- to 15-fold in the retina of EAU animals. Of interest, no change was seen in the gene expression of αB-crystallin and Hsps 27 and -70, which are known to be upregulated during various stress conditions affecting neuronal tissue, liver, and others. The limited number of crystallins that were upregulated in EAU appeared to be antigen driven rather than a nonspecific response or epiphenomenon, since no such crystallin upregulation was observed in CFA-injected animals or in animals exposed to heat shock (Fig. 5) . 

α-Crystallins are principal members of the small Hsp family. They interact directly with various components of the programmed cell death machinery. ³⁹ Even though αA and αB share a 58% sequence homology, ⁴⁰ in the early-EAU retina, only αA-crystallin protein expression was elevated. Such an observation suggests that in the retina, during immune-mediated mitochondrial oxidative stress (such as EAU), αA-crystallin is selectively used to prevent apoptosis of photoreceptors. 

We validated the gene array results by quantitating mRNA expression of four highly upregulated crystallins in the early EAU retina, by real-time PCR analysis. Changes in mRNA expression of these crystallins were also studied in the brain, heart, and liver of the same animals to determine whether this upregulation was specific to the retina or was a global change caused by the immune stress. The crystallins were increased by 30- to 40-fold in retinas of the EAU animals, whereas no significant change was observed in the brain, heart, and liver tissues of these same animals, suggesting that the increase in crystallin gene expression occurring in EAU is retina specific. We also quantitated the mRNA levels of αB-crystallin in the four tissues studied. There was no significant change in its expression level. A similar finding was reported by Kumar et al., ⁴¹ who detected a fourfold increase in gene expression of αA-crystallin in the retinas of diabetic rats while αB-crystallin expression remained unaltered. However, αB-crystallin increased fivefold in skeletal muscles of these diabetic rats. Such studies, and our observations in EAU, indicate that photoreceptors may selectively use αA-crystallin for protection against the stress. Among the crystallins that were induced in the EAU retina, we focused on α-crystallins (αA and αB), the two principal members of the small Hsp family known for their antiapoptotic activities in cultured cells ¹⁰ and for their antiaggregation properties. ⁴² It is noteworthy that elevated expression of αB-crystallin is associated with various neurodegenerative and autoimmune diseases such as Alzheimer’s disease and multiple sclerosis. ¹² The αB-crystallin gene that has a canonical heat shock promoter responds to heat, hyperglycemia, light damage, ischemia-reperfusion, hypoxia/reoxygenation and retinal injury, ^{37 43 44 45 46} and has been shown to interfere with the processing of procaspase-3. ⁹ In contrast, the αA-crystallin gene promoter lacks a heat shock promoter and is not induced in response to heat stress. In the retina, these two proteins have been shown to be present in the same ratio as in the ocular lens (1-αB: 3-αA). ⁴⁷  

It is interesting to note that the microarray analyses of the retinal expression profiles during early EAU reveal very little or no induction of αB-crystallin or of Hsp 27 and -70 (Hsps whose induction in different tissues is commonly associated with exposure to varied physical and chemical stresses ⁷ ) indicating operational differences in the pathophysiology of EAU and stresses like heat shock. These data thus point to a specific role for αA and other crystallins (βA1, βB2, and γS) in early EAU. 

To correlate the mRNA expression of αA-crystallin to its gene product, its expression was evaluated at the protein level by immunoblot analysis. We observed two protein bands, one with a molecular mass of 20 kDa, and the second with a band at 22 kDa, corresponding to an alternate gene product, αA insert crystallin, which has been described in rodent lens. ⁴⁸ This protein product arises from alternate splicing of mRNA, producing a protein with a 23-amino acid insert between residues 63 and 64 of αA-crystallin. A similar observation was reported by Kapphahn et al. ⁴⁹ in aged retinas and by Xi et al. ¹⁴ in normal mouse retina. Densitometric analysis of the two immunoreactive bands in the EAU retinas showed a remarkable 10-fold increase in αA-crystallin compared with the control retinas, thus confirming the gene array and real-time PCR analysis. 

The immunohistologic localization of αA-crystallin protein in early-EAU retinas induced an intense immunostaining in the outer retina, particularly in the inner segments of the photoreceptors. However, in control retinas, αA-crystallin was distributed mostly in the ganglion cell and inner nuclear layers, with minimal or no staining in the inner and outer segments of the photoreceptors (Fig. 4B) . This finding is remarkable because (1) in agreement with previous studies, ¹⁰ under normal conditions αA-crystallin is mostly seen in the ganglion cell and inner nuclear layers, with no staining in the inner and outer segments of the photoreceptors, and (2) we have established that in EAU retinas, photoreceptor cells are the site of oxidative damage. ^{5 6} Together, these data suggest involvement of αA in photoreceptor cell degeneration in EAU. This involvement could be either as part of a causative cascading that leads to apoptosis and degeneration or as part of a protective physiological response. 

Previous studies have reported induction of crystallins and Hsps during various physiological and environmental stresses, including thermal shock. Under normal physiological conditions, the level of Hsps is very low, whereas under stress situations such as heat shock treatment, a very strong synthesis of these proteins, especially Hsps 27 and 70, has been observed. ^{50 51} Similarly, the present study revealed an increased expression of Hsp 27 and -70 in the brain, heart, liver, and retina, after whole-body hyperthermia. Such results indicate that retina is capable of overexpressing Hsp 27 and -70 during heat shock, but not in uveitis. Such findings also suggest a selective use of αA-crystallin in early EAU, which may be related to photoreceptor mitochondrial oxidative stress. However, further studies are needed to address the mechanism of crystallin upregulation during such mitochondrial stress. Using microarray analysis (Affymetrix) and real-time PCR, in the early phase of EAU, we found that four crystallins, αA, βB1, βB2, and γS, were upregulated. No significant upregulation of αB was seen in these analyses. The functions of β- and γ-crystallins are still largely unknown, and therefore the protective effects of β- and γ-crystallins are not immediately apparent in intraocular inflammation. Thus, we elected to focus on αA-crystallin first in this study. Our system may allow us to understand the function of at least one of the two α-crystallins, the αA. 

Thus far, all evidence that the small Hsps αA- and αB-crystallin inhibit apoptosis is based on in vitro or tissue culture experiments. ⁵² No animal models have yet been analyzed to elucidate the roles of these proteins in protecting cells from programmed cell death. However, a recent study using αA- and αB-crystallin double-knockout mice suggested that the absence of α-crystallin causes elevates caspase activity in lens secondary fiber cells. ⁵³ In experiments using αA-crystallin-knockout mice, αA-crystallin expression in vivo was found to protect against cell death during mitosis in the lens epithelium. ⁵⁴ However, no studies have yet been done in the retina to elucidate the roles of αA-crystallin during oxidative stress, particularly in photoreceptor mitochondrial oxidative stress. Our preliminary study also indicated that oxidative or nitrosative stress induced by the iNOS is responsible for the upregulation of αA-crystallin seen in the present study, since in iNOS knockout mice, similar upregulation was not observed (Saraswathy S, unpublished observations, 2007). 

Our in vivo study of knockout mice revealed the protective function of αA-crystallin, since these animals developed EAU early on and experienced extensive photoreceptor damage compared with the wild-type (129SvEv). Moreover, EAU induction caused apoptosis of photoreceptors in knockout mice, but such changes were not observed in the wild-type mice (129SvEv). These findings were further validated by the PCR array analysis of retinal tissue, which showed upregulation of proapoptotic genes and downregulation of antiapoptotic genes in the αA knockout mice, compared with the wild-type animals (Table 3) . The apoptosis-inducing caspase-1, -7, -8, -11, and -12 were elevated in the knockout mice. Caspase-1 and -12 are known to play critical roles in inflammation by activating interleukin-1β and -18. ⁵⁵ Caspase 11 is also an important mediator in activating caspase-1 and -3 under the pathologic conditions that induce apoptosis. ⁵⁶ TNF ligand and its receptors, which were increased in the knockout mice, are known to play an important role in inducing apoptosis. ⁵⁷ TNF receptor associated factor 1 (TRAF-1), which is involved in the apoptotic signal transduction pathway, was also markedly upregulated, indicating that, in knockout mice, activation signals from TNF receptors bind to TRAF-1 to induce apoptosis. ⁵⁸ Other proapoptotic genes, such as Fas and CD40 and its ligand CD40l, were also elevated. Fas is known to participate with TNF receptors in inducing apoptosis. ⁵⁹ Moreover CD40 induces the expression of Fas and TNF ligand. ⁶⁰  

The antiapoptotic genes, primarily BCL10, NFκB, transformation-related protein 73, and Birc 2, were downregulated in the knockout mice. BCL10 induces NFκB activation and contributes to antiapoptotic action through the NFκB-mediated upregulation of apoptotic inhibitor genes. Birc 2 is a member of the inhibitor of apoptosis protein family which also activates NFκB and inhibits apoptosis. ⁶¹ These findings and the above upregulation of proapoptotic genes in the knockout mice indicate that αA-crystallin may protect the photoreceptors from mitochondrial oxidative stress-induced apoptosis. 

Several studies have indicated that the small Hsps and αB-crystallin confer an antiapoptotic effect by specifically inhibiting one or more components of the apoptotic machinery, ^{8 61} and αB-crystallin, in particular, has been shown to bind to cyto c, thus negatively regulating the subsequent proteolytic generation of active caspase-3 subunits. ^{62 63} Similar antiapoptotic effects, however, have not been reported for αA-crystallin. The present study shows that αA-crystallin intercepts the apoptotic processes by associating upstream to nitrated cyto c and downstream to the processed procaspase-3 subunit p24, thereby eliminating the subsequent formation of executioner p20/p17 subunits (Fig. 8) . Further, the extent of the association of αA-crystallin to nitrated cyto c appears to be appreciable, since the assays revealed the association of these two components to be on the order of 70% to 80% in early-EAU retina and close to 76% in the combination of in vitro nitrated cyto c and authentic αA-crystallin. Most important, in both in vivo and in vitro assays, αA-crystallin associates with only nitrated cyto c, not non-nitrated cyto c. Therefore, αA appears to be an efficient inhibitor of oxidative stress-mediated apoptosis of the photoreceptors. 

The mechanism of action of the protective function of αA remains to be elucidated. Large Hsps (Hsp60, -70, and -90) have been shown to be involved in several inflammatory diseases and have been discussed as immunoregulatory modulators with potential anti-inflammatory roles. ⁶⁴ Among the small Hsps αB-crystallin has been directly implicated in autoimmune disease in multiple sclerosis. ¹² It is possible that in addition to its involvement in inhibition of apoptosis brought about by oxidative stress, αA may also be involved in yet to be elucidated anti-inflammatory immunoregulatory functions in the retina. 

Finally, it is obvious from the data presented herein that there are several gene products involved in the response to the immune challenge in early EAU; an intriguing finding was that, in addition to αA, these include such crystallins as βA1, βB2, and γS. At this time, the functions of βA1, βB2, and γS are unknown; thus, the significance of their elevated expression in the diseased retina at best remains conjectural and must therefore await further investigations. In the case of αA, however, the data presented in this investigation are instructive: It inhibits procaspase processing in vitro, its elevated expression is localized to the photoreceptor cell layer, the site of oxidative stress that attends EAU leading to apoptosis and degeneration. These data and the observation that absence of αA (in αA-null mice) predisposes the retina to earlier onset of apoptosis and retinal degeneration point to the involvement of αA in photoreceptor cell protection in EAU. 

Addressing the functional significance of crystallins in EAU will not only enhance understanding of the importance of crystallin upregulation in neuronal inflammations, it will also enhance their role in preventing retinal degeneration mediated by oxidative stress. 

 Supported in part by National Eye Institute Grants EY015714 and EY03040 and by a grant from Research to Prevent Blindness, Inc.

 Submitted for publication September 26, 2007; revised October 23, and November 6, 2007; accepted January 28, 2008.

 Disclosure: N.A. Rao, None; S. Saraswathy, None; G.S. Wu, None; G.S. Katselis, None; E.F. Wawrousek, None; S. Bhat, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Sequence of Primers Specific for Crystallins Used for Real-Time PCR Analyses

Gene	Forward Primer	Reverse Primer
αA	CAACGAGAGGCAGGATGAC	AGGGGACAACCAAGGTGAG
βA1	AACTTCCAGGGCAAGAGGAT	AGATGGGTCGGAAGGACAT
βB2	GGCTACGAGCAGGCTAATTG	CCTTGTAATCCCCCTTCTCC
γS	TGCGGGAATCAACCTTTGC	GCCTTCTACCACCTTACAGGAATG
αB	GCGGTGAGCTGGGATAATAA	GCTTCACGTCCAGATTC
GAPDH	TCACCACCATGGAGAAGGC	GCTAAGCAGTTGGTGGTGCA

 Quantitative (real-time) PCR was performed on the tissues obtained from B10.RIII mice induced with EAU, using the primer pairs listed in the table. GAPDH served as the housekeeping gene for normalization. The specificity of each primer set was verified by melting-curve analysis.

Figure 1.

 

Absence of inflammatory cells and apoptosis in early EAU retinas of B10.RIII mice. Histologically, the enucleated globes of controls on days 7 (A) and of early EAU on day 7 (B) revealed no inflammatory cell infiltration of the retina. The EAU retina on day 14 showed marked inflammatory cell infiltration in the retina and destruction of the outer retina (E). Paraffin-embedded sections were also subjected to TUNEL staining with an apoptosis detection kit. Diaminobenzidine was used as the chromogen. The sections were counterstained with hematoxylin-eosin. Note the absence of TUNEL-positive cells in the control (C) and early EAU retina (D) and the presence of such cells in the retina of day 14 EAU (F). OS, outer segments; IN, inner segments; ONL, outer nuclear layer; INL, inner nuclear layer; GCL, ganglion cell layer.

Table 2.

 

Differential Gene Expression Profiles of Elevated Genes in the Retina during Early EAU by Microarray Analyses

Genes	Gene Bank Accession No.	Change (x-Fold)
αA-crystallin	J00376	12.55
βA1-crystallin	AJ239052	15.22
γS-crystallin	AF032995	15.16
βB2-crystallin	M60559	10.66
γB-crystalline	Z22573	7.95
γC-crystallin	Z22574	9.77
Erythropoeitin	AW049031	17.61
Hsp 70	L27086	1.41
TNF	X87128	1.55
TNf receptor	L35302	2.4

 There was marked upregulation of a few crystallin mRNAs in retinas of early-EAU B10.RIII mice, compared with animals injected with CFA. The profiles of mRNA expression changes of different genes in the retina during early EAU were studied by microarray analysis.

Figure 2.

 

Marked upregulation of αA-, βA1-, βB2-, and γS-crystallin genes were found in early EAU retina; such upregulation was absent in the brain, heart, and liver. Total RNA was extracted from the tissues and reverse transcribed. Real-time PCR was performed using gene specific primers and normalized with GAPDH. The relative multiple of change in mRNA expression was determined using the method 2^−ΔΔCT. Data represent the mean ± SD of results in triplicate experiments. *P < 0.0001.

Figure 3.

 

αA-Crystallin was dramatically upregulated during early EAU (A). Equal amounts of total retinal proteins from day 7 control and EAU mice were separated on a 15% SDS-polyacrylamide gel. Protein bands were transferred to nitrocellulose membrane and probed with polyclonal anti-αA-crystallin and monoclonal anti αB-crystallin as the primary antibodies and with a secondary antibody tagged with horseradish peroxidase. αA-crystallin was detected at the molecular mass indicated (∼20 kDa). Equal protein loading was confirmed by reprobing blots with monoclonal antibody to GAPDH. αB-Crystallin was detected in the retina as ∼21 kDa protein. There was no significant increase in the αB in the day 7 EAU retina compared with the control retina. (B) Densitometry measurements showing 10-fold increase in αA-crystallin protein in early EAU retina compared with the control. *P < 0.001.

Figure 4.

 

Immunofluorescence localization of αA and αB-crystallin in the retina. Tissues were labeled using polyclonal antibody against αA and monoclonal antibody against αB-crystallin (primary antibodies), and Alexa-fluor⁴⁸⁸ goat anti-rabbit IgG and Alexa-fluor⁴⁸⁸ goat anti-mouse IgG, respectively (secondary antibodies). αA-Crystallin immunostaining: EAU day 7 postimmunization (A) and control day 7 postinjection (B) with adjuvant only. αB staining: EAU day 7 postimmunization (C) and control day 7 postinjection (D) with adjuvant only. Note the high immunoreactivity for αA-crystallin in the photoreceptor inner segments (arrow) in the retina of day 7 EAU (A) which is devoid of αB-crystallin signal (C). Such staining was not detectable in the controls (B, D). αB-Crystallin was localized in the inner nuclear and the outer nuclear layers in both day 7 EAU and control retinas (C, D).

Figure 5.

 

Hsp27 and -70 genes are upregulated during heat shock treatment. Changes in mRNA expression levels of αA-, βA1-, βB2-, γS-crystallins, Hsp27 and -70 in the retina, brain, heart, and liver of heat shock-treated and normal mice were quantitated by qRT-PCR and normalized to GAPDH. mRNA expression levels of Hsp27 and -70 increased significantly in all the tissues (*P < 0.001), whereas there was no apparent change in crystallin mRNA gene expression.

Figure 6.

 

Release of nitrated cyto c in the early phase of EAU. In EAU day 7 postimmunization, when mitochondria and cytosol were separated, a small amount of cyto c was detected in the cytosol (lane 3); a slightly higher concentration of cyto c release was also seen in the cytosol of day 10 preparation (lane 4). This was not observed in the cytosol from the control (lane 1). Positive control (sonicated mitochondria) for cyto c is shown (lane 2).

Figure 7.

 

(A) Interaction of αA-crystallin with nitrated cyto c. Lane 1: day 7; lane 2: day 0; lane 3: day 7; and lane 4: day 0, the retinal extracts were incubated with polyclonal αA-crystallin antibody, and coimmunoprecipitated proteins were detected with monoclonal cyto c antibody (lanes 1, 2), and polyclonal nitrotyrosine antibody (lanes 3, 4). Immunoprecipitated cyto c seen in lane 1 is also nitrated as shown in lane 3 with nitrotyrosine antibody. Lanes 5 and 6: retina extracts were immunoprecipitated with monoclonal β-crystallin antibody, then probed with polyclonal nitrotyrosine antibody. A band corresponding to nitrated cyto c is seen in EAU (lane 5), but not in the control (lane 6). (B) In vitro interaction between nitrated cyto c and αA-crystallin measured by surface plasmon resonance. Nitrated cyto c (1 μM) was injected with a rate of 30 μL/min to the immobilized αA-crystallin. The measurement are expressed a resonance units against function of time in seconds. No association was detected in un-modified cyto c (a) nitrated cyto c and (b) native cyto C.

Figure 8.

 

Inhibition of autoproteolytic maturation of procaspase-3 by αA-crystallin. Control retinal cytosolic extracts, free of nuclei and mitochondria were used. Apoptosis was induced by the addition of nitrated cyto c and dATP, followed by incubation for 2 hours at 37°C. αA-crystallin was added for inhibitory experiments. Western blot analysis was used to detect the generation of p20/p17 active caspase-3 subunits. Lane 1: no nitrated cyto c/dATP were added; lane 2: apoptosis was induced by nitrated cyto c and dATP. Lane 3: 20 μM αA-crystallin was added to the mixture. Lane 2: active caspase-3 subunits p20/p17 was observed. These active subunits, however, were not seen in lane 3 when αA-crystallin was present, indicating association of αA-crystalin to the intermediate subunit p24 and preventing its further proteolytic processing to p20/p17.

Figure 9.

 

Early inflammation and apoptosis in αA-crystallin-knockout mice. There was mild inflammation in the choroid and presence of TUNEL-positive cells (arrow) in the retinas of αA knockout mice on day 14 postimmunization, (B, D, respectively), whereas these phenomena were not detected in the wild-type control animals (129SvEv) (A, C). On the other hand, histology revealed trace inflammation in the wild-type uvea (E) and severe inflammation in the αA knockout uvea on day 18 after immunization (F).

Table 3.

 

Apoptotic PCR Array Analyses of Apoptotic Genes in αA^−/− Mice Compared to Wild-Type Controls during Early EAU

Genes	Change (x-fold)
Bag3	1.53
Birc1a	3.70
Birc3	2.88
Caspase 1	2.41
Caspase 12	1.85
Caspase 11	3.17
Caspase 7	1.68
Caspase 8	1.55
Fas	1.87
IL-10	1.51
Ltbr	1.52
TNF	3.55
TNF receptor sf11b	2.05
TNF receptor sf1a	1.69
Cd40	3.02
Cd40lg	1.60
Traf1	2.34
Bcl10	−2.72
Birc2	−1.58
NfkB1	−1.56
Trp73	−1.62

 Increased expression of proapoptotic genes and decreased expression of antiapoptotic genes during early EAU in the retina of αA-crystallin-knockout mice. Retinas of αA knockout and wild-type control mice (129SvEv) on postimmunization day 14 were subjected to apoptosis PCR array analysis specific for apoptotic genes and genes related to its signaling pathway. Expression levels of genes related to apoptosis and its signaling pathway were quantitated by real-time PCR, and the changes in expression were calculated by the method of 2^−ΔΔCT in the knockout mice and compared to the wild-type control levels.

The authors thank Terry Lee for advice and discussion and Jignesh Parikh for helping in the preparation of the figure illustrations.