UCP2–SIRT3 Signaling Relieved Hyperglycemia-Induced Oxidative Stress and Senescence in Diabetic Retinopathy

Abstract

Purpose: Diabetic retinopathy (DR) is one of the most common reasons for blindness. uncoupling protein 2 (UCP2), an uncoupling protein located in mitochondria, has been reported to be related to metabolic and vascular diseases. This research aimed to illustrate the function and mechanism of UCP2 in the pathogenesis of DR.

Methods: Human epiretinal membranes were collected to investigate the expression of UCP2 by quantitative real-time polymerase chain reaction (qRT-PCR) and immunofluorescence. Primary human retinal microvascular endothelial cells (HRECs) were cultured in high glucose (HG) to establish an in vitro cell model for DR. Flow cytometry analysis was used to measure intracellular reactive oxygen species (ROS). Senescence levels were evaluated by the senescence-associated beta-galactosidase (SA-β-gal) assay, the expression of senescence marker P21, and cell-cycle analysis. Adenovirus-mediated UCP2 overexpression or knockdown and specific inhibitors were administered to investigate the underlying regulatory mechanism.

Results: Proliferative fibrovascular membranes from patients with DR illustrated the downregulation of UCP2 and sirtuin 3 (SIRT3) by qRT-PCR and immunofluorescence. Persistent hyperglycemia-induced UCP2 downregulation in the progress of DR and adenovirus-mediated UCP2 overexpression protected endothelial cells from hyperglycemia-induced oxidative stress and senescence. Under hyperglycemic conditions, UCP2 overexpression attenuated NAD⁺ downregulation; hence, it promoted the expression and activity of SIRT3, an NAD⁺-dependent deacetylase regulating mitochondrial function. 3-TYP, a selective SIRT3 inhibitor, abolished the UCP2-mediated protective effect against oxidative stress and senescence.

Conclusions: UCP2 overexpression relieved oxidative stress and senescence based on a novel mechanism whereby UCP2 can regulate the NAD⁺–SIRT3 axis. Targeting oxidative stress and senescence amelioration, UCP2–SIRT3 signaling may serve as a method for the prevention and treatment of DR and other diabetic vascular diseases.

Diabetic retinopathy (DR) is one of the most common reasons for blindness or severe vision impairment. DR is considered a microvascular complication of diabetes and is receiving more attention due to the increasing prevalence of diabetes.¹ This disease impairs the quality of life among patients with diabetes and carries a heavy economic burden.¹ Because a series of complicated risk factors underlies the onset and progression of DR, efficient methods are needed for DR treatment. 

Chronic hyperglycemia increases oxidative stress in endothelial cells. The classic unifying mechanism underlies the common pathogenesis of diabetic vascular complications: Mitochondrial dysfunction drives superoxide formation and stimulates downstream harmful pathways, leading to oxidative stress and cellular damage.^2–4 Blocking production of or neutralizing reactive oxygen species (ROS) can be effective ways to defend against hyperglycemia-induced cellular damage. 

Uncoupling proteins (UCPs), located in the inner membrane of mitochondria, transport protons from the intermembrane space to the matrix of the mitochondria.⁵ Of the UCP family, uncoupling protein 2 (UCP2) is the most ubiquitously distributed and is involved in several metabolic processes.⁵ UCP2 has been implicated in glucose and lipid metabolism in physiological and pathological processes.^6,7 Genetic variants in the UCP2 gene relate to the onset of diabetes and diabetic complications.^8–10 With regard to DR, several researchers have reported an association between UCP2 polymorphisms and the DR phenotype.^11,12 Although previous research has reported that UCP2 alleviates ROS in endothelial cells,¹³ the effect of UCP2 on antioxidant enzymes and metabolic substrates in endothelial mitochondria has not yet been investigated. 

This study demonstrated that UCP2 was downregulated in the progress of DR, and adenovirus-mediated UCP2 overexpression protected endothelial cells from hyperglycemia-induced oxidative stress and senescence based on a new mechanism whereby UCP2 can regulate the nicotinamide adenine dinucleotide (NAD⁺)–sirtuin 3 (SIRT3) axis. UCP2 upregulation increased the critical redox co-enzyme NAD⁺ content and NAD⁺-dependent SIRT3 expression under hyperglycemic conditions. As SIRT3 is a significant regulator of mitochondria metabolism and homeostasis, SIRT3 inhibition by 3-TYP abolished UCP2-mediated anti-oxidative stress and the senescence effect. Therefore, UCP2–SIRT3 signaling upregulation can serve as a protective method for DR and other diabetic vascular diseases. 

Methods and Materials

Cell Culture and Adenovirus Infection

Primary human retinal microvascular endothelial cells (HRECs, three to four passages; Cell Systems, Kirkland, WA, USA) were cultured in Endothelial Cell Medium (ScienCell, Carlsbad, CA, USA) containing 5% fetal bovine serum, 1% endothelial cell growth supplement, 100 U/mL penicillin, and 100 µg/mL streptomycin at a 5.5-mM d-glucose concentration for normal glucose (NG) and 30-mM for high glucose (HG). HRECs cultured in 5.5-mM d-glucose and 24.5-mM mannitol were used as an osmotic control. 

Adenoviruses (adenovirus serotype 5) Ad-shUCP2 (expressing short hairpin RNA [shRNA] targeting human UCP2; GenBank NM_003355.3), Ad-UCP2 (overexpressing human UCP2 mRNA), vehicle control Ad-sh-ctrl (Ad-U6-CMV-MCS), and Ad-ctrl (Ad-CMV-MCS-HA) were purchased from OBiO Technology (Shanghai, China). The knockdown target sequences are available in Supplementary Table S1. Endothelial cells were infected with Ad-UCP2 or Ad-shUCP2 at a multiplicity of infection of 10 or 150, respectively, or with corresponding vehicle control at approximately 70% confluence. Inhibitors for SIRT3 (3-TYP, 16 µM; MedChemExpress, Monmouth Junction, NJ, USA) and UCP2 (genipin, 20 µM; MedChemExpress) were used individually or in combination with the adenoviruses. 

Human Epiretinal Membrane Tissues

Human epiretinal membrane (ERM) tissues harvested from nondiabetic patients (idiopathic epiretinal membrane and retinal detachment–induced proliferative vitreoretinopathy) and pre-retinal proliferative fibrovascular membranes (PFVMs) from patients with proliferative DR (PDR) were collected during surgery. This study followed the tenets of the Declaration of Helsinki and was approved by the ethics committee of Shanghai General Hospital, Shanghai Jiao Tong University School of Medicine. All of the involved patients had signed informed consent forms, and their personal details are provided in the Table. 

Table.

 

Demographic Characteristics of the Patients Involved in the Analysis of UCP2 and SIRT3 mRNA Expression

 

For quantitative polymerase chain reaction (qPCR) analysis, epiretinal membranes were collected and quickly preserved in liquid nitrogen until RNAs were isolated. For the immunofluorescence assay, the epiretinal membranes were immersed and fixed in 4% paraformaldehyde overnight. Then, membranes were embedded in paraffin and sliced into 4-µm sections. 

Western Blot Analysis

HRECs were lysed in cell lysis buffer (Beyotime Biotechnology, Shanghai, China) containing 1-mM phenylmethylsulfonyl fluoride and a protease/phosphatase inhibitor cocktail (Roche, Indianapolis, IN, USA). Total protein concentrations were measured using a bicinchoninic acid assay (Beyotime Biotechnology). Protein samples (30–50 µg) were used for western blot using 4% to 20% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) gels and transferred to a 0.22-µm polyvinylidene fluoride membrane (Roche). After being blocked in nonfat milk, the membranes were incubated in the primary antibody at 4°C overnight and in the secondary antibody (The Jackson Laboratory, West Grove, PA, USA) at room temperature for 1 hour. The ECL Plus HRP substrate (EMD Millipore, Burlington, MA, USA) and Amersham Imager 600 (General Electric, Boston, MA, USA) were used to visualize the immunoreactive bands. The primary antibodies were as follows: anti-UCP2 (#89326), anti-SIRT3 (#2627), anti-SOD2 (#13141), anti-P21 (#2947), and anti-Actin (#5125), all of which were purchased from Cell Signaling Technology (Danver, MA, USA), and Anti-beta Tubulin antibody (#ab6046), which was purchased from Abcam (Cambridge, UK). 

RNA Extraction and Quantitative Real-Time PCR

Total RNAs from epiretinal membranes were isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) and reversed to cDNAs using a PrimeScript RT Reagent Kit (Takara, Tokyo, Japan). Analysis of mRNA levels was performed using TB Green Premix Ex Taq (Takara) by ViiA 7 (Applied Biosystems, Waltham, MA, USA). Relative mRNA levels were calculated using the ∆∆C_t method by normalizing with β-actin (Sangon Biotech, Shanghai, China). Specific primers for the target gene were as follows: UCP2: forward 5′-CCCCGAAGCCTCTACAATGG-3′, reverse 5′-CTGAGCTTGGAATCGGACCTT-3′; SIRT3: forward 5′- CAGCGGCTCCCCAAAGAACAC-3′, reverse 5′-CGGCTCTACACGCAGAACATC-3′. 

Immunofluorescence

Epiretinal membranes were cut in 4-µm sections and fixed on slides. To remove paraffin, the slides were immersed in xylene twice for 15 minutes and then in graded ethanol (5 minutes in 100% ethanol, 5 minutes in 85% ethanol, and 5 minutes in 75% ethanol). After being washed with distilled water, the slides were immersed in Tris-EDTA buffer (pH = 8.0) in a microwave for antigen retrieval. The slides were washed three times with PBS after cooling. A 3% BSA solution was used to block the sections for 30 minutes at room temperature. The diluted primary antibody was incubated with sections in a wet box at 4°C overnight. The corresponding secondary antibodies and 4′,6-diamidino-2-phenylindole (DAPI) for nucleus staining were incubated the next day. Images were captured under confocal microscopy (Leica, Wetzlar, Germany). The antibodies involved were as follows: anti-UCP2 (ab97931, 1:400; Abcam), anti-SIRT3 (#2627, 1:400; Cell Signaling Technology), anti-CD31 (GB13063, 1:100; Servicebio, Wuhan, China), anti-goat FITC (GB22404, 1:100; Servicebio), and anti-rabbit Cyanine3 (GB21403, 1:100; Servicebio). 

Animals

Eight-week-old male C57BL/6J mice (20–25 g body weight) were purchased from Shanghai Laboratory Animal Center (Shanghai, China). Diabetes was induced by intraperitoneal injection of streptozotocin (Sigma-Aldrich, St. Louis, MO, USA). Mice were injected with 55 mg/kg streptozotocin in sodium citrate buffer (pH = 4.5) for 6 consecutive days, and the control animals were injected with the sodium citrate buffer only. The animals were fasted for 12 hours before injection, and they were categorized as diabetic when their blood glucose level exceeded 16.7 mM. Mice were sacrificed 5 months after diabetes induction. Eyes were carefully enucleated, and retinas were carefully removed and sonicated in lysis buffer on ice for western blot analysis. 

The in vivo animal experiments complied with the requirements of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Animal research was approved by the Laboratory Animal Ethics Committee of Shanghai General Hospital. 

Measurement of Intracellular ROS

Total ROS in HRECs were measured using DCFH-DA probes (Molecular Probes, Eugene, OR, USA). HRECs were incubated with DCFH-DA probes (10 µM) diluted in serum-free medium in a 37°C incubator for 30 minutes, protecting them from light. After being washed with cold PBS to remove uncombined probes, the cells were collected for flow cytometry analysis in the dark (Beckman Coulter, Brea, CA, USA). 

Measurement of Intracellular NAD⁺ Level

Intracellular NAD⁺ was measured using an NAD⁺/NADH quantification kit (Abcam). HRECs were scraped, washed in cold PBS, and centrifuged for cell pellets. After the supernatant was discarded, the cell pellets were lysed using an extraction buffer for two freeze/thaw cycles. Based on the instructions of the manufacturer's manual, extractions and diluted standard solutions were incubated with a reaction mix or background reaction mix. Intracellular NAD⁺ levels were calculated by a standard curve using the colorimetric method at 450-nm absorbance at 25°C. The quantification of intracellular NAD⁺ was normalized to the total protein concentrations. 

Measurement of Intracellular Adenosine Triphosphate

Intracellular adenosine triphosphate (ATP) level was measured using an ATP determination kit (Molecular Probes) according to the manufacturer's instructions. The level of intracellular ATP was measured using a luminescence assay with a microplate reader (Synergy 2; BioTek, Winooski, VT, USA). 

Measurement of SIRT3 Activity

SIRT3 activity was measured using a fluorometric SIRT3 assay kit (Abcam). Mitochondria were isolated for SIRT3 activity detection using a mitochondria isolation kit (Beyotime Biotechnology). The fluorescent intensity was measured using the BioTek microplate reader at 360/460 nm every 4 minutes and normalized to mitochondria protein concentration. The deacetylation activity was calculated when the reaction velocity was constant. 

Senescence-Associated β-Galactosidase Assay

The senescence-associated beta-galactosidase (SA-β-gal) assay was conducted to detect senescent cells using an assay kit (Beyotime Biotechnology). HRECs were seeded in plates and pretreated as mentioned previously. After washing with PBS, the cells were fixed for 15 minutes at room temperature and incubated in X-Gal solution at 37°C in a CO₂-free incubator overnight. After removing the staining buffer, the cells were washed in PBS. Images were captured randomly under a microscope (Olympus, Tokyo, Japan), and SA-β-gal–positive cells were counted. 

Cell-Cycle Analysis

HRECs were collected and fixed in 70% ethanol overnight at 4°C. After removing ethanol, the cells were washed with PBS and stained with FxCycle PI/RNase Staining Solution (Invitrogen) for 30 minutes at room temperature in the dark. Flow cytometry was conducted to analyze the cell cycle. 

Statistical Analysis

All experimental data are expressed as the mean ± standard deviation (SD) of at least three independent experiments. Statistical analyses were performed using a two-tailed Student's t-test or one-way ANOVA followed by Tukey's test in Prism 8 (GraphPad, Boston, MA, USA) or SPSS Statistics 22.0 (IBM, Chicago, IL, USA). P < 0.05 was considered significant. 

Results

Endogenous UCP2 Was Downregulated in Hyperglycemia-Induced Vascular Injury

To determine the change in UCP2 expression in the pathogenesis of DR, epiretinal membranes were collected and a qPCR assay was further conducted. Information about the patients involved in this study is summarized in the Table. UCP2 mRNA was significantly downregulated in PFVMs from patients with PDR compared to ERMs from nondiabetic patients (Fig. 1A). Immunofluorescence staining was conducted in membrane sections, demonstrating that more endothelial cells marked by CD31 were observed and the expression of UCP2 was downregulated in PFVMs from patients with PDR when compared to that in ERMs from nondiabetic patients (Fig. 1B). Meanwhile, UCP2 expression was significantly downregulated in diabetic mouse retinas when compared to that in age-matched nondiabetic controls (Fig. 1C). A short period of HG exposure (2 days) increased UCP2 expression, but chronic hyperglycemia (5 days) significantly downregulated UCP2 expression (Fig. 1D) in endothelial cells, indicating a dynamic change in UCP2 expression in diabetic vascular disease. To imitate chronic hyperglycemia-induced endothelial dysfunction, HRECs were cultured in HG for 5 days. 

Figure 1.

 

Endogenous UCP2 was downregulated in hyperglycemia-induced vascular injury. (A) qPCR analysis of UCP2 mRNA levels in ERMs from nondiabetic control patients and PFVMs from patients with PDR. (B) Immunofluorescence illustrated the localization and expression of endothelial cell marker CD31 (green) and UCP2 (red) in ERMs from nondiabetic control patients and PFVMs from patients with PDR. DAPI (blue) indicated staining of a nucleus. Scale bar: 50 µm. (C) Western blot analysis of UCP2 expression in control and diabetic mouse retinas 5 months after diabetes induction. (D) Western blot analysis of UCP2 expression in HRECs exposed to normal glucose (NG), brief high glucose (HG 2 days), and chronic high glucose (HG 5 days). All the results were analyzed by two-tailed unpaired t-test or ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

UCP2 Alleviated HG-Induced Oxidative Stress and Senescence in Endothelial Cells

To further determine the effect of UCP2 in endothelial cells incubated in hyperglycemic states, adenoviruses overexpressing or knocking down UCP2 were employed, and the levels of overexpression and knockdown were confirmed by western blot (Fig. 2A). The accumulation of ROS in endothelial cells induces oxidative stress, leading to inflammation and endothelial cell death.¹⁴ Endothelial dysfunction breaks the balance of vascular permeability, which contributes to vascular leakage and vision loss. Therefore, total intracellular ROS was measured using DCF probes. HG incubation increased intracellular ROS, whereas UCP2 overexpression decreased ROS in hyperglycemic conditions (Fig. 2B). Conversely, knocking down endogenous UCP2 further increased intracellular ROS accumulation (Fig. 2C). 

Figure 2.

 

UCP2 alleviated HG-induced oxidative stress and senescence in endothelial cells. (A) Western blot analysis confirmed the efficiency of adenovirus-mediated UCP2 overexpression and knockdown. (B) Intracellular ROS was measured in HRECs cultured in NG, HG, and HG with UCP2 overexpression (HG+Ad-UCP2). (C) Intracellular ROS was measured in HRECs cultured in NG, HG, and HG with UCP2 knockdown (HG+Ad-shUCP2). (D) Representative images and quantification of SA-β-gal–positive cells in HRECs cultured in NG, HG, and HG+Ad-UCP2. Scale bar: 50 µm. (E) Western blot analysis of P21 in HRECs cultured in NG, HG, and HG+Ad-UCP2. (F) Cell-cycle analysis of HRECs cultured in NG, HG, and HG+Ad-UCP2. (G) Representative images and quantification of SA-β-gal–positive cells in HRECs exposed to NG, hyperosmolarity (NG+mannitol), and HG. Scale bar: 50 µm. (H) Western blot analysis of P21 in HRECs exposed to NG, NG+mannitol, and HG. (I) Cell-cycle analysis of HRECs exposed to NG, NG+mannitol, and HG. (J) Intracellular ROS was measured in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+genipin. (K) Representative images and quantification of SA-β-gal–positive cells in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+genipin. (L) Western blot analysis of P21 in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+genipin. (M) Cell-cycle analysis of HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+genipin. All of the results were analyzed by ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Intracellular oxidative stress induces DNA damage and shortens telomeres, contributing to premature senescence.¹⁵ The presence of senescent cells in retinas from the postmortem globes of patients with PDR was found to have significantly increased when compared to age-matched nondiabetic patients,¹⁶ indicating the association between DR and senescence. Senescence is considered to be a permanent cell-cycle arrest, and senescent cells have increased lysosomal content and activity, as indicated by SA-β-gal assay.¹⁷ Aging phenotypes were observed in HRECs cultured in HG, verified by biomarkers of cellular senescence, including SA-β-gal assay, cell-cycle inhibitor P21, and cell-cycle analysis. UCP2 overexpression alleviated endothelial cell senescence in HG, confirmed by a decrease of SA-β-gal–positive cells (Fig. 2D), downregulation of P21 expression (Fig. 2E), and the relief of G0/G1cell-cycle arrest (Fig. 2F). 

Because chronic hyperglycemia may induce osmotic pressure and contribute to cell senescence, mannitol was used to observe the effect of hyperosmolarity in HRECs. No obvious increase of SA-β-gal–positive cells (Fig. 2G), upregulation of P21 (Fig. 2H), or cell-cycle arrest in the G0/G1 phase (Fig. 2I) was observed in HRECs exposed to mannitol when compared to those incubated in NG. The effect on relieving ROS accumulation and cell senescence mediated by UCP2 overexpression can be reversed by the UCP2 inhibitor genipin (Figs. 2J–2M). 

Knocking Down or Inhibiting Endogenous UCP2 Exacerbated Senescence

To further confirm the effect of UCP2 on senescence in HRECs, adenovirus-mediated UCP2 knockdown or UCP2-specific inhibitor genipin was administrated to observe the senescence phenotype in HRECs incubated in NG and HG. Knocking down UCP2 increased SA-β-gal–positive cells (Fig. 3A), upregulated P21 expression (Fig. 3B), and arrested the cell cycle (Fig. 3C) in both NG and HG. A similar effect was observed when UCP2 activity was inhibited by genipin (Figs. 3D–3F). In brief, UCP2 knockdown or inhibition aggravated HREC senescence in both NG and HG. 

Figure 3.

 

Knocking down or inhibiting endogenous UCP2 exacerbated senescence. (A) Representative images and quantification of SA-β-gal–positive cells in HRECs cultured in NG and HG with or without UCP2 knockdown (shUCP2). Scale bar: 50 µm. (B) Western blot analysis of P21 in HRECs cultured in NG and HG with or without shUCP2. (C) Cell-cycle analysis of HRECs cultured in NG and HG with or without shUCP2. (D) Representative images and quantification of SA-β-gal–positive cells in HRECs cultured in NG and HG with or without UCP2 inhibition (genipin). Scale bar: 50 µm. (E) Western blot analysis of P21 in HRECs cultured in NG and HG with or without genipin. (F) Cell-cycle analysis of HRECs cultured in NG and HG with or without genipin. All of the results were analyzed by ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

UCP2 Overexpression Upregulated Intracellular NAD⁺ Levels in Hyperglycemic States

As an uncoupling protein, UCP2 translocates protons from the intermembrane space to the matrix of the mitochondria, thereby dissipating the driving force of ATP production.⁵ In hyperglycemic states, UCP2 overexpression decreases intracellular ATP content (Fig. 4A). With regard to mitochondrial ATP synthesis, the redox state matters. During oxidative phosphorylation, electron flow was driven from the reduced substrate (mainly NADH) to oxygen, leading to the increase of oxidized substrate (NAD⁺).¹⁸ UCP2 overexpression increased intracellular NAD⁺ levels when compared to cells incubated in HG (Fig. 4B), which may be interpreted as the facilitation of NADH re-oxidation.¹⁸ 

Figure 4.

 

UCP2 overexpression upregulated intracellular NAD⁺ levels in hyperglycemic states. (A) Intracellular ATP was measured in HRECs cultured in NG, HG, and HG+Ad-UCP2. (B) Intracellular NAD⁺ was measured in HRECs cultured in NG, HG, and HG+Ad-UCP2. All of the results were analyzed by ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

UCP2 Overexpression Upregulated NAD⁺-Dependent Deacetylase SIRT3

In addition to being a critical redox co-enzyme, NAD⁺ is also an important signaling molecule that controls cell function and survival.¹⁹ Sirtuins, one of the main NAD⁺-responsive signaling protein families, regulate various processes, including mitochondrial metabolism, inflammation, meiosis, autophagy, circadian rhythms, and apoptosis.¹⁹ SIRT3, a deacetylase mainly located in mitochondria, is involved in mitochondrial metabolism and homeostasis.²⁰ To investigate the role of SIRT3 in the pathogenesis of DR, qPCR analysis was conducted to measure SIRT3 mRNA in membrane samples. SIRT3 mRNA decreased in the PFVMs from patients with DR when compared to that in ERMs from nondiabetic patients (Fig. 5A). Apart from downregulation of the SIRT3 mRNA level, immunofluorescence demonstrated less SIRT3 expression in PFVMs (Fig. 5B). Hyperglycemia downregulated the expression of SIRT3 in HRECs, whereas UCP2 overexpression reversed it (Fig. 5C). Knocking down endogenous UCP2 aggravated the downregulation of SIRT3 (Fig. 5D). HG incubation decreased SIRT3 activity in HRECs when compared to NG. Apart from the upregulation of SIRT3 expression levels, UCP2 overexpression increased SIRT3 activity in HRECs under hyperglycemia conditions (Fig. 5E). Superoxide dismutase 2 (SOD2), which efficiently scavenges superoxide, was detected in HRECs incubated in HG. UCP2 overexpression increased SOD2 expression in hyperglycemic conditions (Fig. 5F). 

Figure 5.

 

UCP2 overexpression upregulated NAD⁺-dependent deacetylase SIRT3. (A) qPCR analysis of SIRT3 mRNA levels in ERMs from nondiabetic control patients and PFVMs from patients with PDR (two-tailed unpaired t-test). (B) Immunofluorescence indicated the localization and expression of endothelial cell marker CD31 (green) and SIRT3 (red) in ERMs from nondiabetic control patients and PFVMs from patients with PDR. DAPI (blue) indicated staining of a nucleus. Scale bar: 50 µm. (C) Western blot analysis of SIRT3 in HRECs cultured in NG, HG, and HG+Ad-UCP2. (D) Western blot analysis of SIRT3 in HRECs cultured in NG, HG, and HG+Ad-shUCP2. (E) Measurement of SIRT3 activity in HRECs cultured in NG, HG, and HG+Ad-UCP2. (F) Western blot analysis of SOD2 in HRECs cultured in NG, HG, and HG+Ad-UCP2. All of the results were analyzed by ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

UCP2 Relieved Hyperglycemia-Induced Oxidative Stress and Senescence Via SIRT3 Activity

As UCP2 overexpression upregulated SIRT3, a selective SIRT3 inhibitor, 3-TYP, was administrated to explore the mechanism. SIRT3 inhibition reversed UCP2 overexpression-mediated ROS alleviation (Fig. 6A). Consistently, the decline of SA-β-gal–positive cells and downregulation of P21 mediated by UCP2 overexpression were abolished after SIRT3 inhibition (Figs. 6B, 6C). Collectively, these results indicate that UCP2 relieves hyperglycemia-induced oxidative stress and senescence via SIRT3 activity. 

Figure 6.

 

UCP2 relieved hyperglycemia-induced oxidative stress and senescence via SIRT3 activity. (A) Intracellular ROS was measured in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+3-TYP. (B) Representative images and quantification of SA-β-gal–positive cells in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+3-TYP. Scale bar: 50 µm. (C) Western blot analysis of P21 in HRECs cultured in NG, HG, HG+Ad-UCP2, and HG+Ad-UCP2+3-TYP. All of the results were analyzed by ANOVA with Tukey's multiple comparison test and are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001.

Discussion

HG induces oxidative stress and damages vascular endothelial cells, leading to the onset of diabetic vascular diseases. Several clinical studies have illustrated the association between UCP2 polymorphisms and diabetes mellitus (DM) or DR.^8–12 However, the change in UCP2 expression in the progression of DM or DR may be complicated. Zheng et al.²¹ reported that UCP2 was upregulated in diabetic rat retinas and bovine retinal endothelium cells (BRECs) incubated in HG, a finding that was consistent with the research result that rat retinal mitochondria undergo mild uncoupling in the early stage of diabetes.²² The upregulation of UCP2 in the early stage of DR was considered an adaptation to prevent ROS production, as the ROS scavenger N-acetyl-l-cysteine inhibited the upregulation of UCP2 in BRECs under hyperglycemic conditions.²¹ In the brief hyperglycemia exposure, the upregulation of UCP2 in HRECs was observed. However, the expression of UCP2 decreased in PFVMs and HRECs exposed to chronic hyperglycemia, which may be interpreted as a failure to adapt due to long-term hyperglycemia damage. The significance of UCP2 has gradually been realized in various diseases, and the regulation of UCP2 expression at multiple levels was reported.²³ As UCP2 is an unstable protein with a short half-life, the accommodation of UCP2 requires subtle regulation.²³ Thus, this adaptation may be disturbed due to chronic hyperglycemia exposure. 

The effect of UCP2 may be controversial because UCP2 can dissipate proton gradients and influence intracellular energy and substrate metabolism. With regard to diabetes, UCP2 protected islet beta cells against oxidative stress and glucotoxicity, but UCP2 inhibition reversed obesity and HG-induced beta cell dysfunction and stimulated insulin secretion.^24–26 Besides its controversial role in metabolic diseases, UCP2 plays an important role in the development of vascular diseases due to its effect on ROS accumulation, proinflammatory and proatherogenic signals, and mitochondrial-induced cell death.²⁷ UCP2 was supposed to be a potential therapeutic target in vascular diseases, including atherosclerosis, diabetic vascular diseases, hypertension, stroke, and peripheral vascular diseases.²⁷ Endothelial-specific UCP2 deletion promoted atherogenesis and collagen production, whereas endothelial-specific UCP2 overexpression inhibited carotid atherosclerotic plaque formation in mice.²⁸ In addition, it has been reported that UCP2 deficiency triggers pseudohypoxic pulmonary vascular remodeling and pulmonary hypertension.²⁹ In this study, it was found that UCP2 efficiently relieved hyperglycemia-induced endothelial ROS accumulation and senescence. Therefore, in addition to improving DR, targeting UCP2 can be a potential therapy for other vascular diseases. 

Endothelial cells remain quiescent for years unless they are stimulated for angiogenesis.³⁰ They prefer not to maximize energy (ATP) production and rely on glycolysis.³⁰ Due to the characteristic of endothelial cells, UCP2-mediated energy dissipation may have few impacts on physiological functions as a vascular endothelial barrier. However, UCP2 overexpression maintained intracellular NAD⁺. Importantly, NAD⁺ declines in age-related diseases and diabetes, and NAD⁺ supplementation enhances the life span of mice.^31–33 Nicotinamide mononucleotide, a key NAD⁺ intermediate, ameliorated glucose intolerance in HFD-induced type 2 diabetes (T2D) mice and improved glucose intolerance and lipid profiles in age-induced T2D mice by restoring NAD⁺ levels.³² Due to the critical role of NAD⁺, UCP2 overexpression-mediated NAD⁺ maintenance may improve the outcomes of age-related diseases and diabetes. 

In addition, the maintenance of intracellular NAD⁺ promotes the function of SIRT3, one of the most important deacetylases in mitochondria. SIRT3, located in the mitochondrial matrix, regulates the acetylation levels of metabolic enzymes in mitochondria and responds to a variety of stresses, including calorie restriction and metabolic stress.³⁴ SIRT3 depletion increased SOD2 acetylation but reduced SOD2 activity in SIRT3 knock-out mice, accompanied by increased vascular oxidative stress, reduced endothelial nitric oxide, and exaggerated hypertension.³⁵ Several researchers have reported that diabetes downregulated SIRT3 expression and increased SOD2 acetylation.^36–38 As SOD2 efficiently scavenges superoxide, the effect of ROS scavenging mediated by UCP2 may partially be attributed to the promotion of SIRT3 expression and the recovery of SOD2 activity, further confirmed by the fact that 3-TYP abolished the downregulation of ROS. 

The free-radical theory of aging illustrates the connection between oxidative stress and senescene.^39,40 Accumulating evidence suggests the role of mitochondria in aging.^41,42 Loss of mitochondrial SIRT function, especially loss of SIRT3, has been related to many age-related diseases.⁴³ Depletion of SIRT3 accelerates the pace of aging in several diseases, including cancer, metabolic syndrome, cardiovascular disease, and neurodegenerative diseases.⁴⁴ In this study, UCP2 overexpression decreased intracellular ROS and upregulated NAD⁺-dependent SIRT3, protecting endothelial cells from hyperglycemia-induced senescence. The same effect was reported elsewhere, that loss of SIRT3 induced senescence in IMR-90 human fibroblast cells.⁴⁵ SIRT3 inhibition blocked the protective effect against senescence mediated by UCP2 overexpression in hyperglycemic states. Shimasaki et al.⁴⁶ reported that the UCP2-null endothelium exhibited impaired G1–S cell-cycle transition and upregulated cyclin-dependent kinase inhibitors p16 and p21, showing a premature senescence phenotype, which was consistent with our results. In their study, overexpressing SOD2 in UCP2-null endothelium rescued cell senescence when compared to the UCP2-null endothelium, illustrating that the UCP2-null endothelium exhibited a dysfunctional phenotype, such as premature senescence via excess mitochondrial superoxide.⁴⁶ In addition to the positive effect on mitochondrial superoxide, this research verified that UCP2 relieved senescence via NAD⁺ and SIRT3 upregulation. 

During the advanced stages of DR, pathological neovascularization in retinopathy contributes to vascular leakage and hemorrhage, leading to severe vision loss. The development of fibrous bands may lead to retinal traction, retinal detachment, and ultimately vision loss.⁴⁷ Cell senescence has been observed in diseases with pathological neovascularization, including DR and retinopathy of prematurity.¹⁶ Crespo-Garcia et al.¹⁶ reported that senescent cells accumulated in the retinas of patients with PDR and in a mouse model of oxygen-induced retinopathy. In contrast to normal vessels, pathological vessels engage pathways of cellular senescence.¹⁶ Worse yet, senescent endothelial cells induce a deleterious tissue microenvironment and age-related organ dysfunction through senescence-associated secretory phenotype (SASP).⁴⁸ Due to the fact that premature senescence of retinal cells and subsequent SASP exacerbate DR, senolytics, selective drugs for the elimination of senescent cells, have been studied in clinical trials for DR treatment.⁴⁹ UCP2 shows promise for DR and other pathological neovascularization diseases through attenuating senescence. 

In summary, this research demonstrated the endogenous change of UCP2 in DR progression. In HG-incubated endothelial cells, UCP2 relieved hyperglycemia-induced oxidative stress and senescence via SIRT3 upregulation. The beneficial effect of UCP2 in endothelial cells may provide a potential target for DR and a series of diseases, including cardiovascular diseases and other pathological neovascularization diseases. 

Acknowledgments

Supported by the Science and Technology Research Project of Songjiang District (No. 2020SJ300). 

Disclosure: S. Li, None; D. Sun, None; S. Chen, None; S. Zhang, None; Q. Gu, None; Y. Shen, None; L. Xu, None; X. Xu, None; F. Wei, None; N. Wang, None