Sirt1 Involvement in rd10 Mouse Retinal Degeneration

Abstract

purpose. Sirtuin1 (Sirt1) is an NAD⁺-dependent deacetylase involved in development, cell survival, stress resistance, energy metabolism, and aging. It is expressed in the mammalian central nervous system (CNS) and is activated during processes associated with neuroprotection. The retinal degeneration 10 (rd10) mouse model of retinitis pigmentosa (RP) was used to investigate the possible role of Sirt1 in this type of retinal degeneration.

methods. Eyes from control and rd10 mice were used. Sirt1 mRNA was detected by in situ hybridization, and its abundance was estimated by semiquantitative RT-PCR. The presence of Sirt1 protein was investigated by immunohistofluorescence and Western blot analysis. The apoptosis of photoreceptor cells was analyzed by terminal dUTP transferase nick-end labeling (TUNEL). Immunolabeling for Sirt1, apoptosis-inducing factor (Aif), and caspase-12 (Casp-12) was performed on retinal tissue sections.

results. Sirt1 mRNA and immunoreactivity were observed in normal adult mouse eyes. In the control retina, Sirt1 was immunolocalized mostly to the nucleus. In rd10 mice with retinal degeneration, changes in Sirt1 immunolabeling were observed only in the retinal outer nuclear layer (ONL). The pathologic pattern of Sirt1 immunoreactivity correlated with the start of retinal degeneration in rd10 mice.

conclusions. The results suggest a link between Sirt1 production and retinal degeneration in rd10 mice. The anti-apoptotic, neuroprotective role of Sirt1 in the mouse retina is based on the involvement of Sirt1 in double DNA strand-break repair mechanisms and in maintaining energy homeostasis in photoreceptor cells. The results suggest that the neuroprotective properties of Sirt1 may gradually weaken in rd10 mouse photoreceptor cells.

Sirtuins, mammalian silent information regulator 2 (Sir2) homologues, comprise a family of seven proteins involved in the cell cycle, chromatin structure regulation and transcription, apoptosis regulation, DNA repair, development, inflammation, regulation of energy metabolism, neuronal survival, and processes impaired during aging. ¹ One member of this family, Sirt1, is an NAD⁺-dependent class III histone deacetylase known to be involved in the response to molecular damage and metabolic imbalance, such as that triggered by fasting, exercise, nutritional deficits, caloric restriction, and reactive oxygen species (ROS). Sirt1 deacetylates proteins involved in stress resistance, such as p53, the 70-kDa Ku autoantigen (Ku70), NF-κβ, and FOXOs. ² Sirt1 also deacetylates and modulates the activity of proteins involved in chromatin remodeling, transcription silencing, genome integrity, cell division, energy metabolism, and apoptosis. ³ Most Sirt1-deficient mice die soon after birth, but those that survive display eye abnormalities during development, and surviving adult mice have abnormal retinas, ^{4 5} demonstrating the importance of Sirt1 in eye development. 

Sirt1 protein and mRNA are present in the adult mammalian brain. ^{6 7 8 9} Sirt1 activity has been studied in various animal models of neurodegenerative diseases, including Huntington’s disease, ¹⁰ Alzheimer’s disease and Tauopathies, ⁸ and amyotrophic lateral sclerosis, ¹¹ and in stressful situations, such as caloric restriction and strong oxidative stress. ⁹ A decrease in Sirt1 levels in response to a high-fat diet has been observed in the rat hippocampus, but the mechanisms by which oxidative stress affects local levels of Sirt1 production remain poorly understood. ⁹ In the adult squirrel monkey, amyloidosis in an Alzheimer’s-like disease is attenuated by caloric restriction through SIRT1 activation in the temporal cerebral cortex. ⁸ Although the characterization of Sirt1 production and activation in normal and diseased mammalian brains is currently underway, nothing is known about Sirt1 expression in normal and degenerative mouse retina. 

Retinitis pigmentosa (RP) is a heterogeneous collection of inherited forms of retinal degeneration that is currently untreatable and generally leads to blindness. These neurodegenerative retinal disorders are characterized by a progressive loss of retinal photoreceptor cells. The many molecular mechanisms underlying photoreceptor cell death are far from completely understood, but apoptosis has been identified as the final common pathway of photoreceptor death in RP. ¹² The retinal degeneration 1 (rd1) mouse strain is a genuine model of human autosomal recessive RP. This model is caused by a nonsense mutation that truncates the β-subunit of the rod photoreceptor cGMP phosphodiesterase type 6 (PDE6-β) protein ¹³ and is characterized by progressive degeneration of the retinal ONL, resulting in total blindness. ¹⁴ In this model, apoptosis involves the translocation to the nucleus of both apoptosis-inducing factor (Aif), an apoptogenic factor released after mitochondrial depolarization, and caspase-12 (Casp-12), an endoplasmic reticulum (ER) protein that mediates apoptosis in response to ER stress. Thus, cross-talk occurs between the two apoptotic pathways—that involving the mitochondria and that involving the ER—in the degenerating photoreceptors of rd1 mice. ¹⁵ Another spontaneous mutant has been isolated—the rd10 mouse—and has been shown to carry a homozygous missense mutation, also in the PDE6-β gene. Photoreceptor degeneration is slower in this mutant than in rd1 mice, making it easier to distinguish between the apoptosis occurring during postnatal retinal development and retinal degeneration per se. ¹⁶ The retinal degeneration observed in the rd10 mouse has a later onset and is milder than that observed in the rd1 mouse and therefore provides a better model for studying the mechanisms underlying this form of RP. ¹⁶  

The subcellular distribution of Sirt1 has been studied in various cultured cells and tissues. ^{5 7 17} A nuclear location for Sirt1 has been reported in early mouse embryonic cells, large oocytes, proliferating granulosa cells, and spermatogenic cells. Sirt1 appears to be associated with euchromatin rather than heterochromatin in the nuclei of these cells. ⁵ In the pancreas, β cells display immunolabeling for Sirt1 in both the nucleus and the cytoplasm. ¹⁷ By contrast, Sirt1, which is mostly found in the nucleus in cells of the embryonic mouse brain, has been localized to the cytoplasm of pyknotic cell nuclei after apoptotic induction. This finding provided the first evidence linking Sirt1 nucleocytoplasmic shuttling with cell death. ¹⁸ The Sirt1 protein has two nuclear localization signals and two cytoplasmic localization signals, which may be involved in the activation or inhibition of apoptosis. ^{19 20} Despite the known involvement of Sirt1 in apoptosis regulation and stress resistance, little is known about its nucleocytoplasmic localization and potential involvement in retinal degeneration. 

In this study, we undertook a detailed analysis of Sirt1 gene expression in normal adult retinas and during photoreceptor degeneration in rd10 mice. We detected a particularly strong immunolabeling signal for Sirt1 in the central ONL at the onset of retinal degeneration in rd10 mice. This signal gradually weakened with the progression of ONL degeneration. 

Materials and Methods

Animals and Tissues

C57BL6/J (Charles River L’Arbresle, France) and rd10 mice (The Jackson Laboratory, Bar Harbor, ME, kindly provided by Bo Chang) were kept at 21°C, under a 12-hour light/12-hour dark cycle, with food and water supplied ad libitum. Adult (3 months old) and postnatal (P) mice (P14 to 5 months), were handled in compliance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mice were killed (between 10 AM and noon) by cervical dislocation. The eyes were rapidly removed and processed as previously described. ²¹  

Semiquantitative RT-PCR

Total RNA was isolated from adult mouse neuroretina, retinal pigmented epithelium (RPE), ciliary body, and lens. Tissue dissection, RNA extraction, and reverse transcription were performed as previously described. ²¹ RT-PCR amplification was performed over 32 cycles of 30 seconds at 94°C, 30 seconds at 55°C, and 1 minute at 72°C. Sirt1 primers S1F (5′-CAG AAC CAC CAA AGC GGA AA-3′, position 709-728) and S1R (5′-GGC ACT TCA TGG GGT ATA GA-3′, position 1402-1383) amplified a 693-bp product. The cyclophilin A (Cypa) primers CF (5′-TGG TCA ACC CCA CCG TGT TCT TCG-3′) and CR (5′-TCC AGC ATT TGC CAT GGA CAA GA-3′) amplified a 311-bp product. This fragment was coamplified as an internal control, as Cypa gene expression levels are similar in normal ocular tissues and during retinal degeneration processes. ²²  

DNA Probes for In Situ Hybridization

Probes were selected on the basis of the mouse Sirt1 mRNA sequence (NM_019812.1), using commercial sequencing (OLIGO 4; Molecular Biology Insights, Inc., Cascade, CO; 55% GC content, ΔG>-4 kcal/mol for hairpins, and self-pairing). BLAST procedures were used to compare primer sequences with sequences in the GenBank and EMBL nucleotide sequence databases (GenBank; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the National Center for Biotechnology Information, Bethesda, MD, and EMBL, http://www.embl-heidelberg.de/; provided in the public domain by the European Molecular Biology Laboratory, Heidelberg, Germany) and to ensure their specificity. The antisense sequences were Sirt1 to -5: 5′-AAC CTC TTG ATC CCC TCC ATC AGC TCC AAA TCC AGA TCC TCC AGC ACA TTC GGG CCT CTC-3′ (position 2136-2195) and Sirt1 to -3: 5′-GGG AGG TCT GGG AAG TCC ACC GCA AGG CGA GCA TAG ATA CCG TCT CTT GAT CTG AAG TCA-3′ (position 836-895). The corresponding sense Sirt1 and scrambled probes (5′-ATC GTC AGC TGA GAT CAA TAA TGG CCC CGG TTA GAG CTC TAC TGC GAT AAT GGC TTG CCA-3′) were used as negative controls. The scrambled probe contained nucleotides randomly selected by an appropriate algorithm. Comparison of this scrambled probe with the entire GenBank and EMBL nucleotide databases retrieved no matching sequence. All 60-mer oligonucleotide probes (synthesized and HPLC-purified by Invitrogen, Cergy Pontoise, France) were 3-′ end-labeled with [³⁵S]dATP (PerkinElmer, Courtabeuf, France) and terminal deoxyribonucleotidyl transferase (15 U/mL; Invitrogen), to a specific activity of approximately 7 × 10⁸ cpm/mg, as previously described. ²³ The probes were purified (Biospin columns; Bio-Rad, Ivry-sur-Seine, France) before use. 

In Situ Sirt1 mRNA Hybridization

In situ hybridization was performed, as previously described. ^{21 24} As a positive control, a mouse tyrosine hydroxylase oligoprobe (5′-CTG AAG CTC TCT GAC ACG AAGTAC ACC GGC TGG TAG GTT TGA TCT TGG TAG GGC TGC ACG-3′) was radioactively labeled and hybridized with substantia nigra tissue sections, according to the same protocol. The classic butterfly-wing pattern of in situ hybridization was obtained (data not shown). Reproducible patterns of Sirt1 labeling were obtained by incubating brain coronal tissue sections from the hippocampus with the same antisense Sirt1 oligoprobe. 

Western Blot Analysis

Mouse tissues were homogenized in lysis buffer (150 mM NaCl, 10 mM Tris-HCl, 1 mM EDTA, 1 mM EGTA, 2 mM Na₃VO₄, 180 mM NaF, 2 mM PMSF, 1% Triton X-100, 0.5% NP40); protease inhibitor cocktail [pH 7.4]; Sigma-Aldrich, Saint-Quentin Fallavier, France). Homogenates were centrifuged, and protein concentration was determined by the Lowry protein assay (DC protein assay; Bio-Rad). Each sample (100 μg of protein) was mixed with loading buffer (pH 6.8, 60 mM Tris, 10% glycerol, 2% sodium dodecyl sulfate, 5% 2-betamercaptoethanol, 0.01% bromophenol blue) and boiled for 5 minutes. The proteins were separated by electrophoresis in a 10% polyacrylamide gel and were transferred onto nitrocellulose membranes (for 1 hour at 90 V). We used two different primary polyclonal anti-Sirt1 antibodies (a rabbit antibody from Upstate Biotechnology, Lake Placid, NY) and a goat antibody (Abcam, Cambridge, UK); both incubated with membranes overnight, at a dilution of 1:500 and at 4°C. We used horseradish peroxidase-conjugated secondary antibodies: a goat anti-rabbit IgG antibody or a rabbit anti-goat IgG antibody (1:6000, 2 hours at room temperature; both from Sigma-Aldrich), depending on the primary antibody used. An anti-β-actin antibody (1:2000; Santa Cruz Biotechnology, Santa Cruz, CA) was used as an internal control. Bands corresponding to Sirt1 or β-actin were visualized by enhanced chemiluminescence (ECL; Perkin Elmer), on light-sensitive film (Biomax Light Kodak; Sigma-Aldrich). 

Immunohistochemistry and TUNEL

For immunoperoxidase staining, the paraffin was removed from paraffin-embedded eye sections by incubation in xylene, and these sections were then rehydrated by passage through a graded series of alcohol solutions. The sections were labeled with the detection kit (ChemMate; Dako, Trappes, France), according to the manufacturer’s instructions. For both immunoperoxidase staining and immunofluorescence studies, Sirt1 was detected with the two antibodies previously used for Western blot analysis (diluted 1:300 in Dako antibody diluent), at 4°C overnight). The specificity of the staining achieved with each antibody was confirmed by incubating adjacent tissue sections in the absence of primary antibody. For double immunohistofluorescence labeling, sections were incubated overnight at 4°C with a goat polyclonal anti-recoverin antibody, a goat polyclonal anti-Aif antibody (both from Santa Cruz Biotechnology; 1:50), and a rat monoclonal anti-caspase-12 antibody (1:200; Sigma-Aldrich). The following day, sections were incubated with Alexa Fluor 488-conjugated donkey anti-rabbit, Alexa Fluor 546-conjugated donkey anti goat, Alexa Fluor 546-conjugated donkey anti-rat, Alexa Fluor 488-conjugated donkey anti-goat, Alexa Fluor 647-conjugated goat anti-hamster (all Alexa Fluor antibodies were purchased from Molecular Probes-Invitrogen), and Texas red-conjugated donkey anti-rabbit (Rockland Immunochemicals, Gilbertsville, PA) antibodies, as appropriate (1:300, 2 hours at room temperature). The nuclei were counterstained with DAPI, propidium iodide (PI, 1:1000; both from Sigma-Aldrich), or TO-PRO-3 (1:1000; Invitrogen), depending on the secondary antibodies used. The sections were mounted and stored at 4°C. Negative control experiments included incubation in the absence of primary antibody. 

Apoptotic nuclei were detected in eye sections by TUNEL (Promega Corp. Madison, WI). For negative controls, the enzyme was omitted and, for positive controls, samples were first incubated with DNase, according to the manufacturer’s protocol. 

Confocal fluorescence micrographs were captured with a confocal microscope (TCS-NT; Leica, Wetzlar, Germany) and analyzed with the system software (Microsystems LAS AF software ver. 1.8.2; Leica). We quantified three adjacent tissue sections for each of the six eyes analyzed. The number of positive cells was determined by eye, for each staining, by two independent investigators blind to the type of tissue section. 

Results

Sirt1 in Normal Ocular Tissues

RT-PCR analysis showed the bands for Sirt1 and Cypa to be of the expected sizes in the neuroretina, RPE, ciliary body, and lens (Fig. 1A) . The Sirt1 cDNA product amplified by RT-PCR was sequenced, and the sequence obtained was found to be identical with that of Sirt1 (NM_019812.1; data not shown). The largest amounts of Sirt1 mRNA were found in the neuroretina, and the lowest amounts were found in the RPE and lens (Fig. 1B) . We localized Sirt1 mRNA in adult mouse eye tissue sections, by performing in situ hybridization with radioactive probes. Sirt1 mRNA was detected in the ONL, inner nuclear layer (INL), and ganglion cell layer (GCL) of the retina (Figs. 1C 1D 1E 1F 1G 1H 1I 1J) . A comparison of dark- and bright-field labeling images and the high hybridization signal-to-noise ratio confirmed the specificity of radioactive Sirt1 probe labeling. An Sirt1 sense probe and a scrambled oligonucleotide probe were used for negative control hybridization experiments on adjacent ocular tissue sections, in which no specific labeling was observed (data not shown). 

We performed Western blot analysis to test for the presence of Sirt1 protein in brain and retina. Proteins extracted from adult mouse testis were used as a positive control. The two purified polyclonal antibodies used specifically recognized the same band, corresponding to a protein of ∼120 kDa in the testis, brain, and retinal protein extracts, consistent with the molecular mass previously reported for the mouse Sirt1 protein and with the results obtained by other teams with the same antibodies ⁴ (Fig. 2A) . Immunohistochemical staining was used to study Sirt1 protein distribution in normal mouse eyes. Identical results were obtained with the two antibodies previously used for Western blot analysis. All ocular structures examined displayed strong, specific Sirt1 immunoreactivity in cell nuclei, with Sirt1 immunostaining of various intensities also observed in the cytoplasm. Significant Sirt1 immunolabeling was observed in the nuclei of corneal epithelial cells, and Sirt1 was also detected in the cytoplasm of these cells. Sirt1 immunolabeling was observed in the nuclei of flattened corneal endothelial cells and in keratocytes. No Sirt1 immunoreactivity was detected in the acellular part of the corneal stroma (Fig. 2B) . In the ciliary body, Sirt1 immunoreactivity was found mostly in the nuclei of ciliary process cells, in the pigmented ciliary epithelial cell layer and in the nonpigmented ciliary epithelial cell layer (Fig. 2C) . The adult mouse lens displayed principally nuclear immunolabeling for Sirt1 in the epithelial and fiber cells. No Sirt1 immunoreactivity was detected in the lens capsule. Sirt1 immunolabeling was detected in the RPE and melanocyte nuclei. Endothelial cells of choroidal vessels displayed Sirt1 immunolabeling, mostly in the cytoplasm (Fig. 2D) . No specific Sirt1 immunoreactivity was apparent in negative control sections (Supplementary Fig. S1). 

The retinal and subcellular distributions of Sirt1 were evaluated by confocal imaging on normal adult eye tissue sections. Specific Sirt1 immunolabeling was observed in the ONL, INL, and GCL. It was granular and mostly nuclear (Fig. 3A) . In the mammalian ONL, rod photoreceptors have small nuclei, containing a central mass of dense heterochromatin surrounded by euchromatin. ²⁵ Photoreceptor nuclear heterochromatin was strongly stained with PI, whereas euchromatin, which is generally peripherally located in nuclei, displayed only weak PI staining (Fig. 3B ; PI). Sirt1 immunostaining in the ONL was confined to the periphery of photoreceptor nuclei (Fig. 3B ; Sirt1). Sirt1 immunoreactivity was systematically colocalized with peripheral nuclear DNA staining (Fig. 3B ; merged images). We investigated the subcellular distribution of Sirt1 in the ONL further, using an anti-recoverin antibody for double retinal immunolabeling. Recoverin is a cytoplasmic protein involved in the visual transduction cascade in both rods and cones. Immunolabeling for recoverin surrounded areas of TO-PRO-3 nuclear staining (Fig. 3C ; TO-PRO-3+ recoverin). Sirt1-immunolabeling was colocalized with weak peripheral nuclear TO-PRO-3 staining (Fig. 3C ; TO-PRO-3+ Sirt1). Recoverin immunolabeling consistently surrounded the nuclear Sirt1 immunolabeling (Fig. 3C ; Sirt1+ recoverin). No merging of the immunoreactivity signals for recoverin and Sirt1 was observed in the ONL. In the INL and GCL, granular Sirt1 immunolabeling was also observed in nuclear regions weakly stained with the DNA marker TO-PRO-3 (Figs. 3D 3E) . Moreover, we identified some cells in the ONL localized close to the outer limiting membrane, in which Sirt1-immunoreactivity was similar to that showed by cells of the INL and GCL. The location and frequency at which these cells were observed (Fig. 3B ; Supplementary Fig. S2) in the ONL and their nuclear structures resembled those of the cones of the rodent retina. 

Sirt1 Expression during rd10 Mouse Retinal Degeneration

There is evidence to suggest that Sirt1 is involved in neuroprotective processes during neurodegenerative diseases. ^{7 8 9 10 11} We therefore studied Sirt1 protein expression during retinal degeneration in rd10 mice, from P14 to the age of 5 months (M). On P15, Sirt1 immunolabeling was observed in the ONL, INL, and GCL of normal retinas. In rd10 ONL on P15, strong Sirt1 immunoreactivity was observed in scattered cells in a central part of the retina delimited by two sharply delineated outer boundaries (Fig. 4) . The INL and GCL displayed Sirt1 immunostaining similar to that observed in the normal retina. From P20 to 3M, the number of cellular structures in the degenerative peripheral ONL displaying strong Sirt1 immunolabeling gradually decreased, whereas levels of Sirt1 immunostaining remained similar to those in the control for the other retinal layers. The strong and abnormal Sirt1 immunolabeling observed in rd10 mouse ONL, from P14 onward, was present in only the central part of this layer and the density of stained structures was highest on P15. Beyond P16, this staining expanded more slowly outside the central part of the ONL, following an apparent center-to-periphery gradient. This staining is visible on a representative tissue section from an rd10 mouse taken on P24 (Fig. 4) . By 5M, the ONL had completely degenerated and the Sirt1 immunoreactivity of the INL and GCL was similar to that in the control. The distribution of Sirt1 immunostaining in the other eye tissues did not change during rd10 retinal degeneration and was similar to that observed in the normal adult mouse eye. 

We studied retinas from different animals and at different stages of development, to characterize this particularly strong pattern of Sirt1 immunostaining. Sirt1 immunolabeling in the retinas of rd10 mice was similar to that observed in control mice from birth to P13 (data not shown). The most significant differences during rd10 retinal degeneration in mouse retinas were observed between P14 and P16, when significant changes in the pattern of Sirt1 immunolabeling were observed in the central region of the ONL in rd10 mouse retinas (Fig. 5) . Abnormally widespread, strongly clustered, heterogeneous Sirt1 immunolabeling was observed in the central ONL of rd10 mice from P14. Sirt1 immunolabeling in the central part of the ONL was most intense on P15. By P16, the density of the Sirt1-immunolabeled cellular structures had already begun to decrease and strong Sirt1-immunolabeling was observed in the area of the retina located between the periphery and the optic nerve. This pattern gradually faded, leading eventually to widespread abnormal Sirt1 immunolabeling across the whole ONL (Figs. 4 5) . We performed Western blot analysis to assess Sirt1 protein levels throughout whole neural retinas from normal and rd10 mice on P15. We observed no significant quantitative difference between the groups (Fig. 5C) . 

Next, we analyzed, as accurately as possible, where the strong Sirt1 immunoreactivity is localized in the central ONL of rd10 mice at P15 and tried to determine what subcellular compartment (nucleus versus cytoplasm) displays this immunoreactivity. We first confirmed that the recoverin immunoreactivity was always detected exclusively in the cytoplasm of all photoreceptor cells located in the P15 ONL of both normal and rd10 mice retinas (Fig. 6) . Then, we determined that Sirt1 immunolabeling in normal P15 ONL retinas was mostly, if not exclusively, restricted to the photoreceptor cell nuclei. The strong Sirt1 immunolabeling observed in the ONL of rd10 P15 central retina did not appear to be granular but rather corresponded to dispersed distinct immunostained clusters (Figs. 6A 6B ; Sirt1). Of importance, we found out after repeated cell counting, that 77% ± 3% (mean ± SEM; n = 6) of the positive Sirt1-immunolabeled cells of the central rd10 ONL displayed a colocalization of their Sirt1 immunoreactivity with their recoverin immunoreactivity, whereas 23% ± 5% (n = 6) of these cells exhibited a colocalization of their Sirt1 immunoreactivity with nuclear staining markers. Thus, in the P15 rd10 central ONL, most Sirt1 immunostaining was unexpectedly colocalized with recoverin immunolabeling, in striking contrast with our observation in the P15 central ONL of normal mice (Fig. 6C) . Afterward, we investigated whether Sirt1 was linked to the apoptosis observed during rd10 retinal degeneration. Only rare TUNEL⁺ cells were detected at P15 in normal retinas from control mice ²⁶ At this stage in the rd10 retinas, numerous cells were TUNEL⁺ (Figs. 7A 8A) . Moreover, 85% ± 3% (n = 6) of the Sirt1-immunoreactive cell structures were found to be TUNEL⁺. By contrast, 60% ± 2% (n = 6) of the TUNEL⁺ cells were Sirt1-immunostained. In addition we analyzed the rate of TUNEL⁺ nuclei that also showed Sirt1 labeling in the nucleus or close to it. Of the TUNEL⁺ nuclei, 33% ± 5% showed Sirt1 labeling in the nucleus or surrounding it. 

Finally, we analyzed the distribution of Casp-12 and Aif in the retinas of rd10 mice. In the GCL and INL of rd10 mouse retinas on P15, Casp-12 immunostaining was confined to the cytoplasm, and no staining was observed in the nuclei. In the ONL of rd10 retinas on P15, Casp-12 immunostaining was more diffuse in both the nuclear and cytoplasmic compartments than that in the other retinal layers (Figs. 7B 8B) , and 82% ± 2% (n = 6) of the strongly Sirt1-immunolabeled cell structures were also positively immunostained for Casp-12. In the central ONL of rd10 retinas, 20% ± 5% (n = 6) of the cellular structures immunolabeled for both Sirt1 and Casp-12 displayed nuclear labeling. Aif immunoreactivity seemed at a first glance to be mostly cytoplasmic in normal and rd10 retinas. However, a major difference of Aif immunodetection could actually be observed in the central ONL of rd10 mice compared with P15 normal control mice: Aif immunoreactivity was also detected in the photoreceptor cell nuclei of P15 central ONL of rd10 mice but not in any part of the P15 ONL of normal mice. In the central rd10 ONL, 71% ± 5% (n = 6) of the Sirt1-immunolabeled cellular structures were also positively immunostained for Aif (Figs. 7C 8C) . However, 25% ± 6% (n = 6) of these cellular structures immunolabeled for both Sirt1 and Aif also displayed nuclear immunolabeling. 

Discussion

We provide a report of the first detailed analysis of Sirt1 mRNA and protein levels in normal adult mouse eyes and during retinal degeneration in the rd10 mouse model of RP. 

Sirt1 mRNA and protein were ubiquitous in normal adult mouse eyes, in which Sirt1 was mostly restricted to cell nuclei in the retina. The detection of Sirt1 mRNA and protein in ocular compartments, such as the cornea, ciliary body, lens, RPE, choroid and neuroretina, in normal mice is consistent with previous reports that this protein is ubiquitous in a wide variety of tissues. ⁷ Sirt1 mRNA levels were highest in the neuroretina. This finding is consistent with previous studies showing Sirt1 mRNA levels to be highest in the normal brain. ⁷ Sirt1 was originally described as a nuclear protein. ^{5 7} However, recent reports have suggested that Sirt1 may be present in both the nuclei and cytoplasm of non-neural tissues and cell lines. ^{19 20} We found that Sirt1 was produced in non-neuronal ocular tissues, such as the cornea, lens, ciliary body, and RPE. Sirt1 immunolabeling was detected in the nuclei and cytoplasm of corneal epithelial cells. Sirt1 immunolabeling was also found in the nucleus of RPE cells from adult mouse eyes, consistent with the results obtained for the ARPE-19 human RPE cell line. ²⁷  

Sirt1 production has been demonstrated in the mammalian CNS. ^{7 8} In this study, it was also detected in the GCL, INL, and ONL of the normal adult mouse retina. In all the neuroretinal cells of the normal mice analyzed in this study, Sirt1 seemed to be exclusively present in the nucleus and was never detected in the cytoplasm. It has been suggested that Sirt1 is a neuronal enzyme conferring resistance to stress in the normal adult CNS and helping to maintain energy homeostasis and anti-apoptotic mechanisms essential for optimal normal brain function. ⁹ One of the major findings of this study is the demonstration of a predominantly, if not exclusively, nuclear location for the Sirt1 protein in all adult normal retinal neurons, including photoreceptors. Sirt1 immunolabeling was observed in regions of retinal cell nuclei only weakly stained with DNA-specific markers, consistent with previous reports. ^{5 7} At first glance, the normal ONL seemed to have a pattern of Sirt1 immunolabeling different from that observed in the INL and GCL. However, the particular arrangement of the euchromatin/heterochromatin and nucleoplasm in the photoreceptor cells may provide an explanation for this apparent discrepancy. Higher magnification observations of immunolabeling for Sirt1 and recoverin in the ONL showed Sirt1 to be predominantly, if not exclusively, localized in the nucleus of the photoreceptors and surrounding the central mass of heterochromatin. 

The rd10 mouse strain displays a characteristic photoreceptor degeneration. ¹⁶ Given the known role of Sirt1 in neurodegenerative diseases, we investigated the possible involvement of Sirt1 in the retinal degeneration of rd10 mice, during which we observed an abnormal heterogeneous distribution of Sirt1 protein, which was present in large amounts, in the ONL. This distribution was most marked on P15. The Sirt1 immunolabeling found in many photoreceptor cell bodies in the ONL of rd10 mouse retinas on P15 was strongly colocalized with recoverin, a cytoplasmic photoreceptor protein. We therefore decided to investigate the relationship between Sirt1 and apoptosis in rd10 photoreceptor cells, as a cytoplasmic distribution of Sirt1 has been shown to be associated with apoptosis. ¹⁹ TUNEL staining was first detected on P14 in rd10 mice (data not shown) and peaked on P15. These results suggest that: (1) apoptosis may begin earlier than previously suggested ^{16 28} and (2) the patterns of Sirt1-immunolabeled cells and TUNEL-positive cells were similar during retinal degeneration in rd10 mice. Almost all the cellular structures strongly immunolabeled for Sirt1 were colocalized with TUNEL staining, suggesting a specific role for Sirt1 in the apoptosis of photoreceptors during retinal degeneration in rd10 mice. Our observations are consistent with previous reports demonstrating a strong link between the presence of Sirt1 in the cytoplasm and apoptosis in other experimental settings and cell types. ^{18 19 20} It has also been demonstrated that cell survival and stress responses are determined by a functional interplay between Sirt1 and poly(ADP-ribose) polymerase 1 (Parp-1). ²⁹ Parp-1 is a multifunctional protein with DNA-repair activity. It is as sensitive as Sirt1 to the ratio of NAD⁺/NADH and contributes to both neuronal death and survival under stress conditions. The strong Sirt1 immunolabeling observed in the ONL of rd10 mice on P15 was similar to that in the in situ Parp-1 activity pattern reported for the ONL of rd1 mice on P11. ³⁰ The pattern of Parp-1 activity also correlated with TUNEL assay findings, both types of staining observed at the onset of retinal degeneration in rd1 mice, suggesting a role for Parp-1 in rod cell death in rd1 mice. ³⁰ These results suggest that Sirt1 may be involved, probably through interaction with Parp-1, in the apoptosis of photoreceptors during retinal degeneration. Further investigations are needed to test this hypothesis in rd10 mice displaying retinal degeneration. 

We report the delocalization of Aif and Casp-12 immunostaining in the ONL of rd10 mice on P15. These results are consistent with those previously reported for retinal degeneration in rd1 mice. ^{15 30 31} The cleaved active form of Casp-12 is involved in photoreceptor apoptosis through its translocation to the nucleus and reinforcement of the principal proapoptotic activity of Aif in retinal degeneration in rd1 mice. ¹⁵ The colocalization of Sirt1 with Aif and Casp-12 in the central region of the ONL in the rd10 mouse model suggests that Sirt1 may be another key player in photoreceptor apoptosis during retinal degeneration in rd10 mice. A small percentage of photoreceptor cells displayed both nuclear Sirt1 immunoreactivity and significant Aif or Casp-12 nuclear immunostaining. The pattern of intense Sirt1 immunolabeling of cellular structures in the ONL of rd10 mice began in the central part of this layer on P14. In this area, immunostaining intensity and density peaked on P15. Immunostaining spread beyond the ONL only after P16, albeit at a slower rate and lower density and intensity, following a center-to-periphery gradient. The abnormal intense and regionally concentrated Sirt1 immunolabeling of cell structures in the ONL was not observed in any other retinal layer at any other stage of retinal degeneration in the rd10 mice. These data, together with the high level of nucleocytoplasmic shuttling of Sirt1 observed in the central ONL of the rd10 mice on P15, strongly suggest that Sirt1 is involved in triggering the onset of retinal degeneration in rd10 mice. A plausible explanation for the Sirt1 involvement in triggering the rd10 photoreceptor cell degeneration is its exclusion from photoreceptor cell nuclei leading to its inactivation or a blockade of its survival-enhancing effects. This explanation necessitates experimental validation. 

The unusual spatial and temporal pattern of Sirt1 expression during rd10 retinal degeneration may be linked to one or several well-established center-to-periphery gradients occurring during postnatal retinal development in normal or pathologic conditions. ^{28 32 33 34 35} Several intriguing features common to postnatal choroidal and retinal vascular development and postnatal neuroretinal development may account for the pathologic Sirt1 immunolabeling of the central part of the ONL firstly and the peripheral part of ONL secondarily in rd10 mice. Further experiments are needed to test this hypothesis. The spatial and temporal pattern of Sirt1 immunolabeling in rd10 mice reported herein provides the first demonstration of the existence of a relationship between Sirt1 production and retinal degeneration in rd10 mice. This finding suggests that Sirt1 is involved in the rd10 mouse model of retinal degeneration and, potentially, in other types of inherited retinal degeneration. However, further experiments on other mouse models of retinal degeneration caused by mutations in other genes are needed to demonstrate a possible general role of Sirt1 in the pathophysiology of neuroretinal degenerative processes. In general, neurodegenerative diseases have been associated with decreases in Sirt1 levels, whereas increases in Sirt1 levels have been associated with neuronal protection. ³⁷ However, the amount of Sirt1 protein in the retinas of rd10 mice on P15 was similar to that in normal retinas. As we used a homogenate of the entire retina for Western blots and changes in Sirt1 levels are restricted to the ONL, this technique may not have been sensitive enough to detect differences in Sirt1 protein levels. 

In conclusion, on the basis of known features of photoreceptor cell metabolism and recent discoveries concerning the role of Sirt1 in DNA repair, we propose a global interpretation of our observations, which obviously must be tested in future experiments. Photoreceptors in normal conditions have the highest cellular rate of oxygen consumption per milligram of tissue of any cell type in mammals. This high level of oxygen consumption is associated with a high rate of ROS production. ROS are highly damaging when not adequately detoxified. The roles of ROS, peroxynitrite, calcium imbalance, intracellular iron metabolism, and oxidative stress in the pathophysiology of various types of retinal degeneration have been demonstrated by several studies. ^{38 39 40 41} Free radicals and ROS in particular constitute a major cause of DNA lesions, such as 8-oxoguanine and double-strand breaks (DSBs). We recently demonstrated extensive production of 8-oxoguanine (Bigot K, et al., manuscript submitted) and DSBs ⁴² in the retinas of normal mice exposed to normal lighting conditions. In degenerating retinas, the production of large amounts of ROS is likely to trigger even more DNA damage. A high load of mitochondrial DNA modifications has repeatedly been associated with aging and neurodegenerative diseases. ^{38 43 44} Sirt1 has been linked to the DNA damage response via the regulation of p53 ^{45 46} and interaction with NBS1, a component of the MRN (MRE11-RAD50-NBS1) DNA damage sensor complex. ⁴⁷ The most recent results strongly suggest that SIRT1 is involved in chromatin remodeling, which is an integral part of the DNA repair response. ^{48 49} Notably, the relocalization of Sirt1 to DSBs was shown to occur during the DNA damage response and to be dependent on Atm and H2ax. ⁴⁸ Our discovery of Sirt1 nucleocytoplasmic shuttling resulting in a cytoplasmic redistribution and exclusion from the nucleus in photoreceptor cells at the onset of their degeneration in rd10 mice may be interpreted as the reaction of photoreceptor cells in the central rd10 retina to overwhelming abnormal redox and metabolic conditions associated with the abnormal nuclear localization of Aif and Casp-12 and the occurrence of irreparable DNA damage. These conditions may force these cells to remove Sirt1 from their nuclei. The expulsion of Sirt1 from the nucleus may prevent its anti-apoptotic action and its major role in the DNA damage-induced remodeling of chromatin. 

We have provided a detailed characterization of Sirt1 immunolabeling in normal mouse eyes and during retinal degeneration in rd10 mice. Our results suggest that the Sirt1 protein may be involved in the multiple molecular mechanisms underlying the onset of photoreceptor degeneration in rd10 mice. It remains unclear why Sirt1 was overproduced between P14 and P16 in the central ONL of rd10 mice first and subsequently in the periphery of the ONL. Further studies are needed to determine the molecular basis underlying the occurrence of this pattern. However, the discovery of the change in the nucleocytoplasmic distribution of Sirt1 coinciding with the start of retinal degeneration in rd10 mice should make it easier to unravel the high level of complexity of the molecular pathways underlying photoreceptor cell death in rd10 mice. Our findings open up new possibilities for designing molecules that could be used to treat inherited retinal degenerative diseases based on the prevention of Sirt1 nucleocytoplasmic shuttling, thereby maintaining survival pathways, balanced energy homeostasis, and physiological DNA repair mechanisms in photoreceptor cells. 

 Supported by Rétina France and Université Paris-Descartes, Ministère de la Recherche et de L’Enseignement Supérieur, AP-HP, INSERM (Institut National de la Santé et de la Recherche Médicale), CNRS (Centre National de la Recherche Scientifique), Fondation pour la Recherche Médicale, Fondation de France, and Fondation de LAvenir pour la Recherche Médicale Appliquée. CJ held postdoctoral fellowships from Fondation pour la Recherche Médicale and Ligue Nationale Contre le Cancer.

 Submitted for publication September 2, 2008; revised March 18 and April 16, 2009; accepted June 4, 2009.

 Disclosure: C. Jaliffa, None; I. Ameqrane, None; A. Dansault, None; J. Leemput, None; V. Vieira, None; E. Lacassagne, None; A. Provost, None; K. Bigot, None; C. Masson, None; M. Menasche, None; M. Abitbol, 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.

 

Figure 1.

 

Sirt1 mRNA distribution in adult mouse eye. (A) Products of semiquantitative RT-PCR for Sirt1 mRNA (693-bp amplification fragment) from neuroretina (NR), RPE, ciliary body (CB) and lens (L). Cypa (311-bp amplification fragment) coamplification was used as an internal control. Sirt1 cDNA amplification was observed for all tissues examined. NR, n = 3 independent samples; RPE, CB, and L, n = 6 samples, repeated three times. MW: molecular weight markers. (B) Densitometric quantification of RT-PCR products. (C–J) In situ hybridization of Sirt1 mRNA was performed with an Sirt1 to 5-′ radioactive antisense probe in (C–D) normal mouse retina. High magnification of the GCL (E, F); INL (G, H); and ONL (I, J). Hybridization with the Sirt1 sense probe was used as a negative control (data not shown). (C, E, G, I) dark-field view; (D, F, H, J) bright-field view. Sirt1 mRNA labeling was repeatedly observed in the ONL, INL, and GCL of the retina. Random labeling was observed with the control probe.

Figure 2.

 

Sirt1 immunolabeling in the adult mouse eye. (A) Western blot showing Sirt1 immunoreactivity (∼120 kDa) in adult normal neuroretina (R), brain (Br) and testis (T; as a positive control) homogenates (repeated three times). Paraffin-embedded sections of adult mouse eye were incubated with anti-Sirt1 antibody. (B–E) Sirt1 localization is shown in the cornea (B); ciliary body (C); lens (D); and RPE and choroids (E). Each panel contains images of nuclear staining (TO-PRO-3 or PI, red), Sirt1 immunolabeling (green), merged images, and merge images plus phase contrast (PhC). Arrows: the corneal epidermal cell nucleus; ( ) the cytoplasm. CEp: corneal epithelium; CEn, corneal endothelial cells; K, keratocytes; PE, pigmented ciliary epithelium; NPE, nonpigmented ciliary epithelium; LEp, lens epithelium; Tz, transitional zone; LFn, lens fiber nuclei; M, melanocytes; V, endothelial cells; Cho, choroids.

Figure 3.

 

Sirt1 distribution in adult mouse neuroretina. Paraffin-embedded sections of adult mouse eye were incubated with anti-Sirt1 antibody. Green: Sirt1 immunolabeling. (A) Normal mouse retina displaying TO-PRO-3 nuclear staining (red), (B) ONL displaying PI nuclear staining (red), (C) ONL magnification showing TO-PRO-3 nuclear staining (light blue), recoverin immunostaining (red; corresponding to the photoreceptor cytoplasm), and Sirt1 immunolabeling (green), merged in different combinations. (D) INL and (E) GCL, showing TO-PRO-3 nuclear staining (red). Sirt1 immunoreactivity was detected in the nuclei of ONL, INL, and GCL cells. Merged images of photoreceptor magnification showing the colocalization of Sirt1 immunoreactivity and nuclear staining but not of Sirt1 immunolabeling and recoverin staining.

Figure 4.

 

Retinal Sirt1 distribution during retinal degeneration in the rd10 mouse model. Paraffin-embedded sections of mouse eyes were incubated with anti-Sirt1 antibody. (A) Sirt1 immunolocalization in P15 control retina and P15, P24, and 5M rd10 mouse retinas. Sirt1 immunostaining was strong in some ONL cells in the central retina of rd10 mice on P15. (B) Detail of Sirt1 levels in the central retina of control mice on P15 control and in the same area of the retina in rd10 mice on P15 and P24 and at 5M.

Figure 5.

 

Sirt1 immunolocalization in rd10 mouse retina on P14, P15, and P16. (A) Sirt1 immunolocalization in the retina of postnatal r10 mice. Strong Sirt1 immunostaining is observed in some ONL cells in the central retina of rd10 mice on P15. (B) Detail of Sirt1 levels in the rd10 mouse retina on P14, P15, and P16. (C) Western blot showing Sirt1 immunoreactivity (∼120 kDa) in retinal homogenates from normal and rd10 mice on P15 (n = 3 retinas per homogenate, repeated three times; testis was used as a positive control, T).

Figure 6.

 

Nucleocytoplasmic distribution of Sirt1 in the ONL of rd10 mouse retina on P15. Confocal sections of rd10 mouse retinal sections were probed with antibodies against Sirt1 and recoverin. Each panel shows nuclear staining (TO-PRO-3 blue), Sirt1 immunolabeling (green), recoverin immunolabeling (red), and merged images from (A) rd10 retina, (B) rd10 retinal ONL, and (C) normal retinal ONL. Merged images for the ONL of rd10 mice, but not of normal mice, showed yellow Sirt1 and recoverin colocalization in the ONL on P15. Scale bar: (A) 50 μm; (B, C) 20 μm.

Figure 7.

 

Colocalization of Sirt1 immunolabeling with apoptotic markers in rd10 mouse retina on P15. Confocal micrographs of entire rd10 mouse retinal sections showing Sirt1 immunolabeling (red) and TUNEL staining (A) or Casp-12 (B) or Aif (C) immunolabeling (green). Merged images with and without nuclear staining show mostly yellow colocalization with apoptotic markers in the rd10 mouse retina on P15. Scale bar, 50 μm.

Figure 8.

 

Magnification of Sirt1 immunolabeling and staining for apoptotic markers in the ONL of rd10 mouse retinas on P15. Confocal micrographs of entire rd10 mouse retinal sections showing Sirt1 immunolabeling (red) and TUNEL stain (A), or Casp-12 (B), or Aif (C) immunolabeling (green). Merged images show mostly yellow colocalization with apoptotic markers in the rd10 mouse retina on P15. Scale bar, 20 μm.

Supplementary Materials

Supplementary Figure S1 - (PDF) 

Supplementary Figure S2 - (PDF) 

The authors thank Axel Kahn (President, Université René Descartes), Patrick Berche (Dean, Faculté de Médecine de Paris-Descartes), Jean-Louis Dufier (President, Scientific Advisory Board of Retina-France), and the members of the board; Jean-Jacques Frayssinet (President, Retina-France) and Paul Kelly (Director, IRNEM [Institut Fédératif de Recherche Necker-Enfants Malades]) for their constant support of EA 2502; Nicholas Staddler and the staff members of the Necker Core Animal Facility; Meriem Garfa-Traoré, Nicolas Goudin, Abdelmounaim Errachid, and Bertrand Bauchat for technical assistance; and Cécile Marsac for critical reading of the article.