Diabetic retinopathy is a multifactorial disease, and the exact mechanism of its pathogenesis remains obscure. Sirtuin 1 (Sirt1), a multifunctional deacetylase, is implicated in the regulation of many cellular functions and in gene transcription, and retinal Sirt1 is inhibited in diabetes. Our aim was to determine the role of Sirt1 in the development of diabetic retinopathy and to elucidate the molecular mechanism of its downregulation. Using Sirt1-overexpressing mice that were diabetic for 8 months, structural, functional, and metabolic abnormalities were investigated in vascular and neuronal retina. The role of epigenetics in Sirt1 transcriptional suppression was investigated in retinal microvessels. Compared with diabetic wild-type mice, retinal vasculature from diabetic Sirt1 mice did not present any increase in the number of apoptotic cells or degenerative capillaries or decrease in vascular density. Diabetic Sirt1 mice were also protected from mitochondrial damage and had normal electroretinography responses and ganglion cell layer thickness. Diabetic wild-type mice had hypermethylated Sirt1 promoter DNA, which was alleviated in diabetic Sirt1 mice, suggesting a role for epigenetics in its transcriptional suppression. Thus strategies targeted to ameliorate Sirt1 inhibition have the potential to maintain retinal vascular and neuronal homeostasis, providing opportunities to retard the development of diabetic retinopathy in its early stages.

Diabetic retinopathy remains the major cause of acquired blindness in working-age adults, and high circulating glucose is considered to be the major instigator of deleterious functional, structural, and metabolic changes (13). Chronic hyperglycemia increases oxidative stress, activates protein kinases and polyol pathways, and results in neuronal and vascular damage including loss of ganglion cells and formation of degenerative capillaries (1,2,4,5), but the exact molecular mechanism of the development of diabetic retinopathy remains to be established.

Sirtuin 1 (Sirt1), a member of the silent information regulator 2 family, is a class III histone deacetylase that interacts with target proteins and regulates many cellular functions including cell proliferation, apoptosis, and inflammatory responses (68). Sirt1 is mainly a nuclear protein, and its activity depends on cellular NAD availability (9). It is expressed throughout the retina, and upregulation of Sirt1 protects against various ocular diseases including retinal degeneration, cataract, and optic neuritis (10). Our previous work has shown that Sirt1 expression and activity are decreased in the retina and its capillary cells in diabetes (11). However, the direct role of Sirt1 in the development of diabetic retinopathy remains elusive.

Sirt1 also regulates gene transcription, and this is mediated either by altering the acetylation status of the transcription factor or by regulating epigenetic modifications at the transcriptional factor binding site of a gene (12). In the pathogenesis of diabetic retinopathy, inhibition of Sirt1 is implicated in the hyperacetylation and activation of nuclear transcription factor-κB (NF-κB), and NF-κB plays a major role in the transcriptional activation of mitochondria-damaging matrix metalloproteinase (MMP)-9 (11,13,14). Sirt1 is also a redox-sensitive enzyme (15), and oxidative stress, in addition to regulating NAD levels, affects Sirt1 activity by regulating posttranslational modifications and protein-protein interactions (16); in diabetes, regulation of oxidative stress prevents decreases in Sirt1 activity in the retinal vasculature (11). How diabetes regulates Sirt1 is, however, not clear.

The expression of a gene, along with its DNA sequence, is also regulated by epigenetic modifications (17,18). Diabetes alters the epigenetic machinery (activates/inhibits) in the retina, and many genes considered to play important roles in mitochondrial homeostasis are epigenetically modified (2,3,11,1921). We have shown that dynamic activation of DNA methylating–hydroxymethylating enzymes, DNA methyltransferases (Dnmts) and ten-eleven translocation enzymes, maintain the DNA methylation status of the retinal MMP-9 promoter to keep it transcriptionally active (19). The role of epigenetics in the regulation of Sirt1 in the pathogenesis of diabetic retinopathy, however, remains to be explored.

Sirt1 is a multifunctional protein implicated in a wide range of molecular and epigenetic pathways (68). The goal of this study was to determine whether regulation of Sirt1 ameliorates the development of diabetic retinopathy and to elucidate the mechanism responsible for Sirt1 regulation. Using mice overexpressing Sirt1, we investigated its role in structural, functional, and metabolic abnormalities critically associated with the development of diabetic retinopathy. The possible mechanism of Sirt1 transcriptional suppression is elucidated through investigation of the DNA methylation status of its promoter.

Mice

Diabetes was induced in wild-type C57BL/6J (WT) and Sirt1-overexpressing (St, C57BL/6-Actbtm3.1 [Sirt1] Npa/J; Sirt1) mice (The Jackson Laboratory, Bar Harbor, ME), with a body weight (BW) of ∼20 g (either sex), by streptozotocin injection (55 mg/kg BW for four consecutive days). Mice presenting blood glucose >250 mg/dL 2 days after the last injection were considered to have diabetes (19,22). Age-matched normal WT and Sirt1 mice were used as their respective controls. Compared with normal WT mice, although Sirt1 expression was significantly increased in the retina of Sirt1 mice, we found no increase in kidneys from the same animals (Supplementary Fig. 1). Mice were sacrificed ∼8 months after diabetes was induced; one eye was fixed in 10% buffered formalin, and the retina from the other eye was removed immediately to obtain biochemical measurements. Glycated hemoglobin was measured after the mice had diabetes for ∼6 months using a kit from Helena Laboratories (Beaumont, TX), and serum HDL was quantified as described previously (23). Normal Sirt1 and WT mice had similar glucose and HDL levels, and the severity of hyperglycemia (blood glucose and glycated hemoglobin) was also similar in WT and Sirt1 mice with diabetes (Supplementary Table 1). The treatment of animals conformed to the Association for Research in Vision and Ophthalmology Resolution on the Use of Animals in Research and to Wayne State University’s guidelines.

Retinal microvessels were prepared with a hypotonic shock method. The retina was incubated in 5–6 mL deionized water in a shaking water bath at 37°C for 60 min. The nonvascular tissue was gently removed under a microscope (22).

Gene Expression

cDNA was synthesized from the retinal microvessels, and gene expression was quantified with SYBR green–based quantitative PCR using gene-specific primers (Table 1). Ribosomal 18S RNA was used as a housekeeping gene, and the fold change was calculated using the ΔΔCt method (19,22).

Histopathology and Apoptosis in Retinal Microvessels

The whole retina was isolated from the formalin-fixed eyes and rinsed overnight in running water and then incubated in 3% crude trypsin (Thermo Fisher Scientific, Waltham, MA) containing 200 mol/L sodium fluoride at 37°C for 45–70 min. After gently brushing away the neuroretinal tissue under a microscope, the vasculature was stained with terminal deoxyribonucleotide TUNEL (In Situ Cell Death Kit; Roche Molecular Biochemicals, Indianapolis, IN). As a control, retinal vasculature treated with DNAse was also stained with TUNEL. The TUNEL-positive capillary cells were counted under a microscope, and the slides then were stained with periodic acid Schiff–hematoxylin to count acellular capillaries and pericyte ghosts by light microscopy (23).

Immunofluorescence Staining

Using specific primary antibodies, 8-μm-thick retinal cryosections were stained for MMP-9, Sirt1 (Abcam, Cambridge, MA), and Dnmt1 (Sigma-Aldrich, St. Louis, MO) following the method reported previously (24). The slides then were incubated with fluorescence-labeled secondary antibodies: DyLight 488 (green) for Sirt1 and Texas Red for Dnmt1 and MMP-9. The sections were mounted with medium containing DAPI (blue) and photographed under a ZEISS ApoTome fluorescence microscope (Carl Zeiss, Chicago, IL) at ×40 magnification.

Vascular Permeability

Approximately 2 weeks before the experiment was terminated, vascular permeability was quantified by fluorescein angiography using a Micron IV Retinal Imaging Microscope (Phoenix Research Labs, Pleasanton, CA). Animals were anesthetized with a mixture of ketamine (67 mg/kg) and xylazine (10 mg/kg) (intraperitoneal injection), the pupil was dilated with 0.1% tropicamide ophthalmic solution, and the cornea was lubricated with Goniovisc (hypromellose 2.5%). The fundus was photographed using a fundus camera for small animals. AK-FLUOR (0.5% solution, 0.01 mL/g BW) was then injected intraperitoneally, and photographs were taken 10 min after fluorescein injection using a barrier filter for fluorescein angiography (25).

Vascular permeability was confirmed by quantifying albumin leakage into the retina using the Evans blue technique (26). Evans blue (30 mg/kg) was injected into the tail vein, and after mice were kept on a heating pad for 2 h, paraformaldehyde was perfused and the eye globes were enucleated immediately. The dye was extracted from the retina using formamide, and absorbance was measured at 620 nm.

Vascular Density

Vascular density was measured on fluorescein angiograms after they were converted to grayscale (to eliminate the fluorescein background) using AngioTool software from the National Cancer Institute (27) and by staining of the retinal flatmount with isolectin (28) using fluorescein isothiocyanate–conjugated Isolectin B4 (Alexa Fluor 488, 1:100 dilution; Life Technologies, Carlsbad, CA). The flatmounts were visualizing under a Leica SP5 confocal microscope at ×10 magnification (Leica Microsystems, Wetzlar, Germany), and the images were converted to grayscale by AngioTool for analysis.

Mitochondrial DNA

Mitochondrial DNA (mtDNA) damage was determined by analyzing the sequence variants in the regulatory region of the mtDNA (displacement loop [D-Loop])—the region that experiences more damage in diabetes than other regions of the mtDNA (29)—using a Surveyor Mutation Detection Kit (IDT Inc., Coralville, IA). Sequence variants were determined by digesting the amplicons with a surveyor nuclease, a mismatch-specific endonuclease with high specificity for the sites of base substitution sequence variants. The digested products were electrophoresed on a 2% agarose gel and analyzed upon visualization under an ultraviolet transilluminator, as reported previously (29).

mtDNA copy numbers were quantified in retinal microvessels using primers for cytochrome B (CytB) as a marker for mtDNA and β-Actin for nuclear DNA (Table 1). SYBR green–based quantitative PCR was carried out by amplification at 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 60 s. The ratio of mtDNA to nuclear DNA (CytB:β-Actin) within each sample was used to calculate mtDNA copy numbers (29).

Epigenetic Modifications

DNA methylation of the Sirt1 promoter was determined by quantifying 5-methylcytosine (5mC) using a methylated DNA immunoprecipitation kit (Epigentek, Farmingdale, NY) (19).

Acetylation of Dnmt1 promoter was determined by quantifying acetylated H3K9 using a chromatin immunoprecipitation technique, as reported previously (21). Dnmt1 protein acetylation was analyzed by immunoprecipitating acetyl lysine (Abcam), followed by Western blotting for Dnmt1 (22).

Electroretinography

Electroretinography (ERG) was performed in dark-adapted mice anesthetized with ketamine and xylazine. After dilating the pupil with tropicamide ophthalmic solution and lubricating the cornea with Goniovisc, ERG responses were measured using an OcuScience HMsERG Lab System by placing a thread electrode embedded in silver over the cornea, above the lubricant solution (hypromellose 2.5%). A contact lens was used to keep the electrode in place. ERG responses were recorded using a series of Ganzfeld flashes with intensities ranging from 100 to 25,000 mcds/m2. The amplitudes and the implicit times of both a-waves and b-waves were measured using ERGview software (23).

Retinal Thickness

Retinal thickness was quantified by optical coherence tomography (OCT) using an OCT module customized to image the retinas of small animals on the Micron IV microscope. The anesthetized animals were positioned in front of the camera, and a high-resolution B-scan of the retinal cross sections was obtained from both eyes by averaging and spatially aligning 50 individual B-scans along the same horizontal axis through the optic disc (30). Thickness of the ganglion cell layer (GCL) + inner plexiform layer (IPL), and of the inner nuclear layer (INL), was measured at a 200- to 400-µm distance on either side of the optic disc using the caliper tool available in InSight software.

To confirm changes in the retinal layers, retinal cryosections were stained with hematoxylin-eosin (31), and the images were analyzed at three random places using ImageJ software.

Statistical Analysis

Statistical analysis was performed using SigmaStat software (Systat Software Inc., San Jose, CA). Comparison among groups was analyzed using one-way ANOVA followed by the Student-Newman-Keuls test for data with a normal distribution. For data for which a normality test failed, one-way ANOVA was performed, followed by the Dunn test. A P value <0.05 was considered statistically significant.

Retinal vasculature from diabetic WT mice had a significantly higher number of acellular capillaries compared with age-matched normal WT mice, as expected (13): ∼8 acellular capillaries in normal mice compared with ∼20 in diabetic mice (Fig. 1A). Similarly, the number of pericytes and TUNEL-positive cells was increased from 5–8 in normal to 16–25 in diabetic mice. However, both diabetic and normal Sirt1 mice had approximately nine acellular capillaries and six to eight TUNEL-positive capillary cells. Figure 1B–D shows a significant decrease in Sirt1 expression (protein and mRNA) in retinal vasculature in diabetic WT mice, and its prevention in diabetic Sirt1 mice.

Because vascular leakage is one of the hallmarks of diabetic retinopathy (1), the effect of Sirt1 overexpression on vascular health was determined by fluorescein angiography. Although some vascular leakage was observed in diabetic WT mice, consistent with normal WT and Sirt1 mice, no leakage was observed in diabetic Sirt1 mice (Fig. 2A). Similarly, protection by Sirt1 against diabetes-induced retinal vascular leakage was also confirmed using the Evans blue method (Fig. 2B). The effect of Sirt1 on retinal vascular health was evaluated further by quantifying capillary density. Both fluorescein angiography and isolectin staining showed a significant decrease in the retinal vascular density in diabetic WT mice compared with normal WT mice (Fig. 2C and D); however, diabetes had no effect on retinal vascular density in Sirt1 mice, and the density was similar to that seen in normal WT mice.

To investigate the effect of Sirt1 on mitochondrial damage, mtDNA damage was analyzed in the retinal microvasculature. Consistent with our previous results (29), diabetic WT mice had a higher number of sequence variants in the D-Loop and a lower parent amplicon band intensity (Fig. 3A and B), which was prevented in diabetic Sirt1 mice. Sirt1 also plays a significant role in mtDNA biogenesis (32), and in diabetes, mtDNA copy numbers are decreased (33). Compared to normal WT mice, mtDNA copy numbers were significantly lower in the retinal microvessels of diabetic WT mice (Fig. 3C), but diabetic Sirt1 mice had an amount of copy numbers similar to those obtained from normal mice (WT or Sirt1).

Because upregulation of MMP-9 in diabetes is implicated in the mitochondrial damage seen in the retina and its vasculature (11,13), we evaluated the effect of Sirt1 on MMP-9 expression. The retinal microvasculature of diabetic Sirt1 mice was also protected from an increase in MMP-9 expression, and the values obtained from normal WT and Sirt1 mice and diabetic Sirt1 mice were not different from each other (Fig. 4A). Figure 4B further confirms the decrease in immunostaining of MMP-9 in the retinal cryosections from diabetic Sirt1 mice compared with that in diabetic WT mice.

Cytosine methylation results in transcriptional repression (17,18,34), and DNA methylation machinery is activated in diabetes (19,20). To understand the mechanism responsible for Sirt1 suppression, its promoter DNA methylation was investigated. Compared with normal WT mice, retinal microvasculature from diabetic WT mice had over 3-fold more 5mC at the Sirt1 promoter, and ∼2.5-fold higher Dnmt1 expression (Fig. 5A). However, in diabetic Sirt1 mice, although 5mC levels were higher than those in normal mice, these values were significantly lower than those obtained from diabetic WT mice, suggesting a role for DNA methylation in diabetes-induced Sirt1 suppression. In a similar way, a diabetes-induced increase in Dnmt1 expression was ameliorated in diabetic Sirt1 mice (Fig. 5B). Immunohistochemical data confirmed the protective effect of Sirt1 against an increase in Dnmt1 staining (Fig. 5C and D). Sirt1 can deacetylate both histones and proteins, and deacetylation of Dnmt1 by Sirt1 deactivates it (35). To investigate the role of Sirt1 in a diabetes-induced increase in Dnmt1, we investigated the acetylation of histone at the Dnmt1 promoter. As shown in Fig. 5E, although diabetes increased acetylated H3K9 levels, which were ameliorated in diabetic Sirt1 mice, it had no effect on Dnmt1 protein acetylation (Supplementary Fig. 2).

Diabetes also affects nonvascular cells of the retina, resulting in abnormal electrical responses (5). ERG was performed to investigate the effect of Sirt1 on functional changes in nonvascular cells. Compared with normal WT mice, diabetic WT mice showed an ∼20% decrease in a-wave and b-wave amplitudes; these changes were prevented in diabetic Sirt1 mice, and the values were similar to those obtained from age-matched normal WT mice (Fig. 6A–C). Similarly, the implicit times of a- and b-waves were also increased by 15–20%. Consistent changes in a- and b-waves were also observed at 3,000 and 10,000 mcds/m2.

Because of the accelerated loss of retinal cells in diabetes, the retinal layers commonly become thinner (36); we determined the effect of Sirt1 overexpression on diabetes-induced changes in retinal layer thickness. Diabetic WT mice showed a significant decrease in the thickness of the GCL + IPL 200 µm from the optic disc (40.43 ± 1.10 µm) compared with normal WT mice (44.00 ± 1.26 µm; P < 0.0001). Similar decreases were seen in the INL (22.83 ± 1.72 vs. 20.29 ± 1.3 µm; P < 0.013). However, in diabetic Sirt1 mice, thicknesses of both the GCL + IPL and the INL were significantly different from those observed in diabetic WT mice (43.83 ± 0.82 and 23.00 ± 0.60 µm, respectively; P < 0.0001). Similar differences in the thickness in diabetic WT and Sirt1 animals were also observed 300 µm from the optic disc (Fig. 7A). Retinal cryosections stained with hematoxylin-eosin also presented similar protection against retinal layer thinning in diabetic Sirt1 mice (Fig. 7B).

Diabetic retinopathy is a multifactorial disease, and many structural, biochemical, molecular, and functional abnormalities are implicated in its development (1,2). Sirt1 is a cellular energy sensor that plays a critical role in linking metabolic stress with cellular response (6,7); in diabetes, Sirt1 is downregulated in the retina and its vasculature (11). Downregulation of Sirt1 hyperacetylates many regulatory proteins, including transcription factors associated with oxidative stress and apoptosis (6,8,11). Here we show that Sirt1 is also intimately associated with the development of diabetic retinopathy and related retinal vascular and neuronal damage. Our novel data clearly demonstrate that diabetic mice overexpressing Sirt1 are protected from the development of degenerative capillaries, a histopathological characteristic of retinopathy, and their retinas have normal vascular density and do not leak. Sirt1 also protects mitochondria from diabetes-induced increases in mtDNA damage and decreases in copy numbers, and it obliterates the activation of mitochondria-damaging MMP-9. Diabetic Sirt1 mice are spared from retinal neuronal damage; ERG responses are normal and GCL-IPL and INL thicknesses are not different than those seen in normal WT or Sirt1 mice, suggesting a major role of Sirt1 in preventing diabetes-induced damage to the overall retinal health. Furthermore, we elucidated the role of DNA methylation in diabetes-induced Sirt1 transcriptional suppression and showed that, to the contrary, Sirt1 regulates its transcription by modulating histone acetylation of the Dnmt1 promoter. These results together imply a significant role for Sirt1 in the development of diabetic retinopathy, and the mechanism of Sirt1 suppression seems to be the epigenetic modification of its promoter.

The Sirtuin family of proteins comprises highly conserved NAD-dependent deacetylases; humans encode seven different sirtuins (Sirt1–Sirt7); among them Sirt1 is the most extensively studied member (6,8). Sirt1 is linked to cellular energy metabolism and the redox state through multiple signaling and survival pathways, and it is also implicated in the pathophysiology of many chronic diseases, including diabetes, neurodegenerative disorders, and cardiovascular disease (37). Sirt1 controls the redox environment and counteracts oxidative damage by converting NAD to its reduced form, NADH (8); in coronary heart and cerebrovascular diseases, activation of Sirt1 protects against oxidative stress at the cellular level and increases survival at the systemic level to further limit these diseases (38). Sirt1-transgenic mice are protected from diabetes induced by obesity (through both diet and genetics) (39). However, although Sirt1 is a core systemic regulator of cellular metabolism, mice overexpressing Sirt1 have a normal life span, and the protection is restricted to specific tissues (40). Here we provide convincing data showing a direct role for Sirt1 in diabetic retinopathy: the retinal vasculature of diabetic mice overexpressing Sirt1 is protected from the development of histopathology and from accelerated apoptosis, a phenomenon that precedes the development of histopathology (41), implying that Sirt1 activators could potentially impede the development of retinopathy in patients with diabetes. Consistent with this, decreased Sirt1 mRNA and activity are also seen in retinal microvasculature and peripheral blood mononuclear cells from human donors with diabetic retinopathy (11,42). A decrease in Sirt1, via regulating acetylation of the liver X receptor, has been shown to increase both hyperglycemia and dyslipidemia in type 2 diabetes, contributing to the progression of diabetic retinopathy (43). Furthermore, administration of the Sirt1 activator resveratrol ameliorates age-related increases in retinal degeneration and cell apoptosis in rats (44).

Clinical and experimental studies have documented many alterations in the retinal vasculature in diabetic retinopathy, including increased vascular permeability, leukostasis, and capillary degeneration (1,2), and the vascular density is decreased (45). Damage to the blood-retina barrier is considered to be an early clinical sign of the development of diabetic retinopathy (1). We show that, whereas the retinas of diabetic WT mice have less vascular density and more vascular leakage than normal WT mice, diabetic Sirt1 mice have normal retinal vascular density with no vascular leakage, suggesting that, in addition to preventing increased capillary cell apoptosis and the formation of degenerative capillaries, Sirt1 overexpression also helps to maintain the overall health of the retinal vasculature.

Mitochondrial damage plays a critical role in the pathogenesis of diabetic retinopathy; in diabetic rodents, increased functional and structural damage to mitochondria is observed before capillary cell apoptosis and the number of degenerative capillaries increase within the retinal vasculature (13,29). Compromised mtDNA biogenesis and fewer mtDNA copy numbers are observed in the retinal microvasculature, and damaged mtDNA continues to fuel the vicious cycle of free radicals by compromising the electron transport chain system (2,3). Here we show that the retinal microvasculature of these diabetic Sirt1 mice are also protected from mtDNA damage, and their mtDNA copy numbers are normal. In support of this, Sirt1 is critically involved in maintaining mitochondrial quality; it regulates mtDNA biogenesis and turnover of damaged mitochondria (32), and manipulation of Sirt1 has been shown to regulate hyperglycemia-induced mitochondrial dysfunction and apoptosis in human umbilical cord vascular cells (46). Furthermore, retinal poly(ADP-ribose) polymerase (PARP) is also activated in diabetes (22), and chronic activation of PARP-1 inactivates Sirt1 by depleting NAD and results in mitochondrial dysfunction (47), further fueling mitochondrial damage.

Activation of proteinases, especially MMP-2 and MMP-9, is implicated in the retinal mitochondrial damage seen in diabetes (13,24). The MMP-9 promoter has multiple transcriptional factor binding sites, and many transcriptional factors are involved in its transcription (48). We have shown that, because of hyperacetylation of transcriptional factors NF-κB and activator protein-1, their binding at the MMP-9 promoter is increased, resulting in MMP-9 transcriptional activation (11,14). Furthermore, Sirt1 also regulates the binding of PARP-1 at the MMP-9 promoter, which regulates MMP-9 expression by manipulating the binding of NF-κB/activator protein-1 (22). The results presented here clearly show that the retinal vasculature of the same diabetic Sirt1 mice protected from the development of diabetic retinopathy also have normal MMP-9 expression, further confirming the role of Sirt1 in the mitochondrial damage seen in diabetes.

Diabetes downregulates Sirt1 in many tissues, including retina and kidney (11,49); one of the mechanisms implicated in its inactivation is increased oxidative stress (8). In a disease state, epigenetic modifications also regulate gene transcription, and diabetes favors many of these epigenetic modifications, including DNA methylation and histone modifications (2,19). DNA methylation is considered to be a gene suppressor (34), and the machinery responsible for maintaining DNA methylation is activated in the retina in diabetes (19). Our results demonstrating hypermethylation of the Sirt1 promoter in retinal microvessels from diabetic mice are supported by other reports showing increased Sirt1 promoter methylation in peripheral blood from patients with Alzheimer disease (50). Consistent with the regulation of DNA methylation of the Sirt1 promoter, in the same diabetic Sirt1 mice, the Dnmt1 promoter (the only member of the Dnmt family that is upregulated in the retina in diabetes [19]) is protected from increased acetylation of H3K9, which results in the regulation of Dnmt1 transcription. These results suggest that Sirt1 ameliorates Dnmt1 activation by inhibiting H3K9 acetylation at the promoter of Dnmt1. Inhibition of Dnmt1, in turn, impedes Sirt1 promoter DNA methylation, and regulates transcription of Sirt1.

Although the histopathology of diabetic retinopathy is observed in the retinal vasculature, neuronal cells also undergo many structural, functional, and metabolic changes (5,36). Neuronal cell apoptosis and GCL thinning occur before vascular cell apoptosis, and patients (and rodents) with diabetes present functional neurogenic changes with prolonged implicit times and decreased amplitudes before histopathological characteristics of diabetic retinopathy are observed (5,23). Here, our exciting data clearly show that diabetic Sirt1 mice are also protected from neuronal abnormalities, their GCL does not show any thinning, and the ERG signals are not different from those observed in normal mice. Although the mechanism by which Sirt1 overexpression could prevent neuronal damage remains to be investigated, our results clearly suggest that Sirt1 has an important protective role against both retinal neuronal damage and vascular damage associated with the development of diabetic retinopathy.

In summary, our convincing structural, functional, and molecular data demonstrate that Sirt1, a multifunctional protein that deacetylates proteins (including transcriptional regulators) and histones, has a protective role against the development of diabetic retinopathy. Amelioration of its inhibition maintains retinal vascular homeostasis by preventing mitochondrial damage and protects the vasculature from undergoing accelerated apoptosis and becoming leaky. Regulation of Sirt1 inhibition also prevents structural and functional damage to neurons, which is observed before vascular pathology, providing opportunities to ameliorate the development of diabetic retinopathy during its early stages. The mechanistic insight into Sirt1 regulation in diabetes suggests the role of epigenetics in its transcriptional suppression. Thus our study introduces the therapeutic potential of strategies targeting Sirt1 activation in maintaining retinal vascular and neuronal health, and in preventing the continuation of the vicious cycle of mitochondrial damage, ultimately ameliorating the development of this blinding disease.

Funding. This study was supported in part by grants from the National Eye Institute, National Institutes of Health (EY014370, EY017313, and EY022230) and from the Thomas Foundation to R.A.K. and an unrestricted grant from Research to Prevent Blindness to the Department of Ophthalmology, Wayne State University, Detroit, MI.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. M.M. researched data, performed the literature search, and edited the manuscript. A.J.D. researched data and performed the literature search. R.A.K. created the experimental plan, performed the literature search, and wrote and edited the manuscript. R.A.K. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

1.
Frank
RN
.
Diabetic retinopathy
.
N Engl J Med
2004
;
350
:
48
58
[PubMed]
2.
Kowluru
RA
,
Kowluru
A
,
Mishra
M
,
Kumar
B
.
Oxidative stress and epigenetic modifications in the pathogenesis of diabetic retinopathy
.
Prog Retin Eye Res
2015
;
48
:
40
61
[PubMed]
3.
Kowluru
RA
.
Diabetic retinopathy, metabolic memory and epigenetic modifications
.
Vision Res
2017
;
139
:
30
38
[PubMed]
4.
Mizutani
M
,
Kern
TS
,
Lorenzi
M
.
Accelerated death of retinal microvascular cells in human and experimental diabetic retinopathy
.
J Clin Invest
1996
;
97
:
2883
2890
[PubMed]
5.
Kern
TS
,
Barber
AJ
.
Retinal ganglion cells in diabetes
.
J Physiol
2008
;
586
:
4401
4408
[PubMed]
6.
Brooks
CL
,
Gu
W
.
How does SIRT1 affect metabolism, senescence and cancer?
Nat Rev Cancer
2009
;
9
:
123
128
[PubMed]
7.
Li
X
.
SIRT1 and energy metabolism
.
Acta Biochim Biophys Sin (Shanghai)
2013
;
45
:
51
60
[PubMed]
8.
Hwang
JW
,
Yao
H
,
Caito
S
,
Sundar
IK
,
Rahman
I
.
Redox regulation of SIRT1 in inflammation and cellular senescence
.
Free Radic Biol Med
2013
;
61
:
95
110
[PubMed]
9.
Tanno
M
,
Sakamoto
J
,
Miura
T
,
Shimamoto
K
,
Horio
Y
.
Nucleocytoplasmic shuttling of the NAD+-dependent histone deacetylase SIRT1
.
J Biol Chem
2007
;
282
:
6823
6832
[PubMed]
10.
Mimura
T
,
Kaji
Y
,
Noma
H
,
Funatsu
H
,
Okamoto
S
.
The role of SIRT1 in ocular aging
.
Exp Eye Res
2013
;
116
:
17
26
[PubMed]
11.
Kowluru
RA
,
Santos
JM
,
Zhong
Q
.
Sirt1, a negative regulator of matrix metalloproteinase-9 in diabetic retinopathy
.
Invest Ophthalmol Vis Sci
2014
;
55
:
5653
5660
[PubMed]
12.
Pavlová
S
,
Klucska
K
,
Vašíček
D
, et al
.
The involvement of SIRT1 and transcription factor NF-κB (p50/p65) in regulation of porcine ovarian cell function
.
Anim Reprod Sci
2013
;
140
:
180
188
[PubMed]
13.
Kowluru
RA
,
Mohammad
G
,
dos Santos
JM
,
Zhong
Q
.
Abrogation of MMP-9 gene protects against the development of retinopathy in diabetic mice by preventing mitochondrial damage
.
Diabetes
2011
;
60
:
3023
3033
[PubMed]
14.
Mishra
M
,
Flaga
J
,
Kowluru
RA
.
Molecular mechanism of transcriptional regulation of matrix metalloproteinase-9 in diabetic retinopathy
.
J Cell Physiol
2016
;
231
:
1709
1718
[PubMed]
15.
Caito
S
,
Rajendrasozhan
S
,
Cook
S
, et al
.
SIRT1 is a redox-sensitive deacetylase that is post-translationally modified by oxidants and carbonyl stress
.
FASEB J
2010
;
24
:
3145
3159
[PubMed]
16.
Chung
S
,
Yao
H
,
Caito
S
,
Hwang
JW
,
Arunachalam
G
,
Rahman
I
.
Regulation of SIRT1 in cellular functions: role of polyphenols
.
Arch Biochem Biophys
2010
;
501
:
79
90
[PubMed]
17.
Reddy
MA
,
Zhang
E
,
Natarajan
R
.
Epigenetic mechanisms in diabetic complications and metabolic memory
.
Diabetologia
2015
;
58
:
443
455
[PubMed]
18.
Regha
K
,
Sloane
MA
,
Huang
R
, et al
.
Active and repressive chromatin are interspersed without spreading in an imprinted gene cluster in the mammalian genome
.
Mol Cell
2007
;
27
:
353
366
[PubMed]
19.
Kowluru
RA
,
Shan
Y
,
Mishra
M
.
Dynamic DNA methylation of matrix metalloproteinase-9 in the development of diabetic retinopathy
.
Lab Invest
2016
;
96
:
1040
1049
[PubMed]
20.
Mishra
M
,
Kowluru
RA
.
Epigenetic modification of mitochondrial DNA in the development of diabetic retinopathy
.
Invest Ophthalmol Vis Sci
2015
;
56
:
5133
5142
[PubMed]
21.
Zhong
Q
,
Kowluru
RA
.
Regulation of matrix metalloproteinase-9 by epigenetic modifications and the development of diabetic retinopathy
.
Diabetes
2013
;
62
:
2559
2568
[PubMed]
22.
Mishra
M
,
Kowluru
RA
.
Role of PARP-1 as a novel transcriptional regulator of MMP-9 in diabetic retinopathy
.
Biochim Biophys Acta
2017
;
1863
:
1761
1769
23.
Kowluru
RA
,
Mishra
M
,
Kowluru
A
,
Kumar
B
.
Hyperlipidemia and the development of diabetic retinopathy: comparison between type 1 and type 2 animal models
.
Metabolism
2016
;
65
:
1570
1581
[PubMed]
24.
Mohammad
G
,
Kowluru
RA
.
Matrix metalloproteinase-2 in the development of diabetic retinopathy and mitochondrial dysfunction
.
Lab Invest
2010
;
90
:
1365
1372
[PubMed]
25.
Liu
CH
,
Sun
Y
,
Li
J
, et al
.
Endothelial microRNA-150 is an intrinsic suppressor of pathologic ocular neovascularization
.
Proc Natl Acad Sci U S A
2015
;
112
:
12163
12168
[PubMed]
26.
Gao
G
,
Shao
C
,
Zhang
SX
,
Dudley
A
,
Fant
J
,
Ma
JX
.
Kallikrein-binding protein inhibits retinal neovascularization and decreases vascular leakage
.
Diabetologia
2003
;
46
:
689
698
[PubMed]
27.
Zudaire
E
,
Gambardella
L
,
Kurcz
C
,
Vermeren
S
.
A computational tool for quantitative analysis of vascular networks
.
PLoS One
2011
;
6
:
e27385
[PubMed]
28.
Hasegawa
E
,
Sweigard
H
,
Husain
D
, et al
.
Characterization of a spontaneous retinal neovascular mouse model
.
PLoS One
2014
;
9
:
e106507
[PubMed]
29.
Mishra
M
,
Kowluru
RA
.
Retinal mitochondrial DNA mismatch repair in the development of diabetic retinopathy, and its continued progression after termination of hyperglycemia
.
Invest Ophthalmol Vis Sci
2014
;
55
:
6960
6967
[PubMed]
30.
Sun
Y
,
Ju
M
,
Lin
Z
, et al
.
SOCS3 in retinal neurons and glial cells suppresses VEGF signaling to prevent pathological neovascular growth
.
Sci Signal
2015
;
8
:
ra94
[PubMed]
31.
Tan
J
,
Liu
Y
,
Li
W
,
Gao
Q
.
Ocular pathogenesis and immune reaction after intravitreal dispase injection in mice
.
Mol Vis
2012
;
18
:
887
900
[PubMed]
32.
Tang
BL
.
Sirt1 and the Mitochondria
.
Mol Cells
2016
;
39
:
87
95
[PubMed]
33.
Santos
JM
,
Tewari
S
,
Goldberg
AFX
,
Kowluru
RA
.
Mitochondrial biogenesis and the development of diabetic retinopathy
.
Free Radic Biol Med
2011
;
51
:
1849
1860
[PubMed]
34.
Deaton
AM
,
Bird
A
.
CpG islands and the regulation of transcription
.
Genes Dev
2011
;
25
:
1010
1022
[PubMed]
35.
Peng
L
,
Yuan
Z
,
Ling
H
, et al
.
SIRT1 deacetylates the DNA methyltransferase 1 (DNMT1) protein and alters its activities
.
Mol Cell Biol
2011
;
31
:
4720
4734
[PubMed]
36.
van Dijk
HW
,
Verbraak
FD
,
Kok
PH
, et al
.
Early neurodegeneration in the retina of type 2 diabetic patients
.
Invest Ophthalmol Vis Sci
2012
;
53
:
2715
2719
[PubMed]
37.
Jęśko
H
,
Wencel
P
,
Strosznajder
RP
,
Strosznajder
JB
.
Sirtuins and their roles in brain aging and neurodegenerative disorders
.
Neurochem Res
2017
;
42
:
876
890
[PubMed]
38.
Chong
ZZ
,
Wang
S
,
Shang
YC
,
Maiese
K
.
Targeting cardiovascular disease with novel SIRT1 pathways
.
Future Cardiol
2012
;
8
:
89
100
[PubMed]
39.
Banks
AS
,
Kon
N
,
Knight
C
, et al
.
SirT1 gain of function increases energy efficiency and prevents diabetes in mice
.
Cell Metab
2008
;
8
:
333
341
[PubMed]
40.
Herranz
D
,
Serrano
M
.
SIRT1: recent lessons from mouse models
.
Nat Rev Cancer
2010
;
10
:
819
823
[PubMed]
41.
Kern
TS
,
Tang
J
,
Mizutani
M
, et al
.
Response of capillary cell death to aminoguanidine predicts the development of retinopathy: comparison of diabetes and galactosemia
.
Invest Ophthalmol Vis Sci
2000
;
41
:
3972
3978
[PubMed]
42.
Liu
S
,
Lin
YU
,
Liu
X
.
Protective effects of SIRT1 in patients with proliferative diabetic retinopathy via the inhibition of IL-17 expression
.
Exp Ther Med
2016
;
11
:
257
262
[PubMed]
43.
Hammer
SS
,
Beli
E
,
Kady
N
, et al
.
The mechanism of diabetic retinopathy pathogenesis unifying key lipid regulators, Sirtuin 1 and liver X receptor
.
EBioMedicine
2017
;
22
:
181
190
[PubMed]
44.
Zeng
Y
,
Yang
K
.
Sirtuin 1 participates in the process of age-related retinal degeneration
.
Biochem Biophys Res Commun
2015
;
468
:
167
172
[PubMed]
45.
Kim
AY
,
Chu
Z
,
Shahidzadeh
A
,
Wang
RK
,
Puliafito
CA
,
Kashani
AH
.
Quantifying microvascular density and morphology in diabetic retinopathy using spectral-domain optical coherence tomography angiography
.
Invest Ophthalmol Vis Sci
2016
;
57
:
OCT362
OCT370
[PubMed]
46.
Wang
S
,
Wang
J
,
Zhao
A
,
Li
J
.
SIRT1 activation inhibits hyperglycemia-induced apoptosis by reducing oxidative stress and mitochondrial dysfunction in human endothelial cells
.
Mol Med Rep
2017
;
16
:
3331
3338
[PubMed]
47.
Guarente
L
.
Linking DNA damage, NAD(+)/SIRT1, and aging
.
Cell Metab
2014
;
20
:
706
707
[PubMed]
48.
Crowe
DL
,
Brown
TN
.
Transcriptional inhibition of matrix metalloproteinase 9 (MMP-9) activity by a c-fos/estrogen receptor fusion protein is mediated by the proximal AP-1 site of the MMP-9 promoter and correlates with reduced tumor cell invasion
.
Neoplasia
1999
;
1
:
368
372
[PubMed]
49.
Yacoub
R
,
Lee
K
,
He
JC
.
The role of SIRT1 in diabetic kidney disease
.
Front Endocrinol (Lausanne)
2014
;
5
:
166
[PubMed]
50.
Hou
Y
,
Chen
H
,
He
Q
, et al
.
Changes in methylation patterns of multiple genes from peripheral blood leucocytes of Alzheimer’s disease patients
.
Acta Neuropsychiatr
2013
;
25
:
66
76
[PubMed]
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at http://www.diabetesjournals.org/content/license.

Supplementary data