HUVECs were obtained from ScienCell Research Laboratories (San Diego, CA, USA). The cells were cultured in endothelial cell medium (ECM) (ScienCell Research Laboratories) with 5% (v/v) fetal bovine serum (FBS) at 37°C in an atmosphere with 5% (v/v) CO₂ and 95% humidity. When they reached confluence, the cells were maintained in 1% (v/v) fetal calf serum and exposed to normal amounts of glucose (5.5 mmol/l) or high amounts of glucose (30 mmol/l) for 3–7 days, during which time the medium was changed every 2 days. When HUVECs reached approximately 50% confluence in fresh serum-free medium, they were transiently transduced with control adenovirus β-galactosidase (Ad-β-gal) or with adenovirus overexpressing UCP2 (Genechem, Shanghai, China) at multiplicities of infection (MOI) of 10. The cells were further cultured in ECM with 5% (v/v) FBS after infection for 4 h and then selected using 200 µm/ml puromycin (Thermo Fisher Scientific, Waltham, MA, USA). The stable overexpression lines were established when more than 95% of the transduced cells were found to strongly express green fluorescent protein (GFP) under a fluorescence microscope (BX51; Olympus, Tokyo, Japan).

All experiments in the present study comply with the requirements of the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research. Eighteen-week-old male C57 mice weighing ~20 g and db/db diabetic mice weighing ~20 g were obtained from the Shanghai Laboratory Animal Center of Chinese Academy of Sciences (Jiuting, Shanghai, China). UCP2-deficient mice weighing ~20 g were obtained from the Shanghai Biomodel Organism Research Center (Jiuting, Shanghai, China). This study was approved by the Ethics Committee of Shanghai Jiao Tong University School of Medicine.

Target sequences were aligned to the human genome database in a BLAST search to ensure that the chosen sequences were not highly homologous with those of other genes. Cells were seeded in 6-well plates and cultured in drug-free medium (ScienCell Research Laboratories, Carlsbad, CA, USA). When cells reached 90–95% confluence they were washed twice with phosphate-buffered saline (PBS) and grown in 2 ml ECM without antibiotics. Using Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA, USA), the indicated concentration (50 nM) of UCP2 siRNA oligo (Genechem) was transfected into HUVECs according to the manufacturer's instructions. Cells transfected with control siRNA served as a negative control. Forty-eight hours later, the expression levels of UCP2 were evaluated by western blot analysis. The cells transfected with siRNA were used for experiments 48 h after transfection.

Western blot analysis was performed, as previously described (14). Briefly, the primary antibodies used to probe the membranes included anti-UCP2 (1:500, cat. no. ab67241; Abcam, Cambridge, UK), anti-cytochrome c (1:200, cat. no. 1896-1; Epitomics, Burlingame, CA, USA), anti-caspase-3 (1:5,000, cat. no. ab32351; Abcam) and anti-β-actin (1:1,000, cat. no. A1978; Sigma-Aldrich, St. Louis, MO, USA). Experiments were repeated in triplicate.

For the analysis of the release of cytochrome c from mitochondria into the cytosol, proteins in the cytosol and mitochondrial fractions were separated by SDS/PAGE. Bands of proteins were transferred onto a PVDF membrane (Millipore, Bedford, MA, USA). Cytochrome c was detected with an enhanced chemiluminescence western blot analysis system (Amersham, Buckinghamshire, UK) with monoclonal antibodies specific to cytochrome c (1:200, cat. no. 1896-1; Epitomics). The same method for western blot analyses was used on cytosolic and mitochondrial preparations using antibody against human UCP2 (1:500, cat. no. ab67241; Abcam) and subunit IV of cytochrome oxidase (1:500, cat. no. ab110272; Abcam).

For quantification of cellular viability, cells were double-stained with Annexin V and propidium iodide (PI) according to the manufacturer's instructions [Annexin V-fluorescein isothiocyanate (FITC) apoptosis detection kit (BD Biosciences, Franklin Lakes, NJ, USA)]. The proportion of apoptotic and necrotic cells was determined using FACSCalibur (Becton-Dickinson, San Jose, CA, USA). Annexin-V is a marker of apoptosis, and PI reflects the integrity of the cell membrane, thereby serving as a marker of necrosis.

Caspase-3 activity was determined using a caspase-3 colorimetric protease assay kit (Beyotime Institute of Biotechnology, Shanghai, China) following the manufacturer's instructions. Briefly, the kit uses spectrophotometry to detect the chromophore p-nitroaniline (pNA) after cleavage from the labeled substrate DEVD-pNA (DEVD is the sequence recognized by caspases). The free pNA was quantified using a spectrophotometer or a microtiter plate reader at 405 nm. The ratio of absorbance from a sample to that from a control allows for the determination of the fold increase of caspase-3 activity.

To determine the effect of UCP2 overexpression on mPTP opening, a previously validated cell model of mPTP opening was used (15). A microplate spectrofluorometer (SPECTRAmax GEM-INI-XS; Molecular Devices, Sunnyvale, CA, USA) was used to study stimulation of the fluorophore, tetramethylrhodamine methyl ester (TMRM), which accumulates in mitochondria, and generation of ROS within mitochondria. Culture medium was removed and replaced with Krebs imaging buffer. Cells were then loaded with 3 mM TMRM for 15 min at 37°C and washed with Krebs imaging buffer (Guidechem, Shanghai, China). The time taken to induce mitochondrial membrane depolarization is recorded as a measurement of mPTP opening. This was defined as the time taken to reach half the maximum TMRM fluorescence intensity. Twenty transfected cells were randomly selected for the induction and detection of mPTP opening from each treatment group, and this was repeated in four independent experiments, providing a total of 80 cells/treatment group. As a positive control and in order to confirm that mitochondrial membrane depolarization was indicative of mPTP opening, following TMRM loading, a group of cells were pretreated for 10 min with the mPTP inhibitor, cyclosporin A (CsA; 0.2 mM; Selleck Chemicals, Houston, TX, USA), as previously described (16–18). The time taken to induce mPTP opening was recorded.

In the present study, HUVECs were incubated with 10 µmol/ml carboxydichlorodihydrofluorescein diacetate (DCFH2-DA; Molecular Probes, Inc., Eugene, USA) at 37°C. After 15 min incubation, the increase in DCFH2 oxidation was measured using a FACSCalibur (Becton-Dickinson) (19,20). The NO level in HUVECs was measured in situ using DAF-FM diacetate (Sigma-Aldrich), as previously described (21).

Mice were sacrificed by cervical dislocation. Eyes were removed from four 8 week-old db/db mice, four age-matched UCP2-KO mice and four age-matched normal mice were used as control retinas. The eyes were immediately enucleated and the retina was separated. The retina from one of the eyes was frozen in liquid nitrogen and stored at −80°C. TUNEL was performed on frozen sections using the DeadEnd™ TUNEL assay kit (Promega, Madison, WI, USA) and were counterstained with PI, according to the manufacturer's suggestions. Briefly, sections were hydrated with alcohol (100, 95 and 70%), and then fixed in 3.7% paraformaldehyde. After washing, the slides were incubated in a mixture of TdT, Mn²⁺, and TdT dNTP (Sangon Biotech, Shanghai, China) for 1 h at 37°C. The reaction was stopped with TdT Stop Buffer (Trevigen, Gaithersburg, MD, USA) for 5 min. After washing with deionized water, the slides were incubated with streptavidin-FITC (diluted 1:200) solution for 20 min at room temperature. Slides were counterstained, mounted, covered with coverslips, and visualized by confocal microscopy (LSM 510; Carl Zeiss, Inc., Oberkochen, Germany). Apoptotic cells were identified as doubly labeled with TdT fluorescein and PI, and only the nuclei which were clearly labeled yellow were scored.

Tissue processing, electron microscopy, morphometric measurements [retinal capillary basement membrane thickness (BMT)], and statistical analysis were performed as detailed in our previous study (12). Eyes were removed from four 8 week-old db/db mice, four age-matched UCP2-KO mice and four age-matched normal mice were used as control. Briefly, enucleated eyes were fixed in 2.5% (w/v) glutaraldehyde in 0.1 M cacodylate buffer (pH 7.4) containing 0.2% (v/v) tannic acid, washed in the same buffer, and post-fixed in 0.5% (v/v) osmium tetroxide. Tissue sections were block stained with uranyl acetate, lead stained, dehydrated through a graded series of ethanol, and embedded in Epon. One-micrometer-thick sections were examined with a JEM-1200EX transmission electron microscope (JEOL, Ltd., Akishima, Japan). Computer-assisted morphometric measurements (The Image Center of Beijing University of Aeronautics and Astronautics, Beijing, China) were performed on electron micrographs taken from 12 randomly selected capillaries of the outer plexiform layer from four different tissue blocks of the same retina. Only cross-sectioned capillaries were considered. A total of 96 capillaries were evaluated in each experimental group.

SPSS 17.0 was used to analyze the experimental data. All experimental data are represented as the means ± standard deviation (SD). One-way analysis of variance (ANOVA) followed by Student-Newman-Keuls test was used to compare the effect of treatment on the various parameters. Non-parametric data was analyzed using the Chi-square test or Fisher's exact method. A P-value <0.05 was considered to indicate a statistically significant difference.

Loss of cells in the ganglion cell layer was due to apoptosis, and this was reflected as an increase in TUNEL in the ganglion cell layer of both the UCP2-knockout mice and db/db mice. We performed TUNEL on retinal sections to measure cell death in the UCP2-knockout mice and db/db mice. Notably, we found that UCP2-knockout mice exhibited a DR-like phenotype, similar to db/db mice. Fig. 1A and C illustrate increased TUNEL in the ganglion cell layer of UCP2-knockout mice and db/db mice compared to their normal counterparts. Furthermore, compared to the non-diabetic mice, the BMT was significantly increased in the UCP2-knockout mice and db/db mice (P<0.01) (Fig. 1B and D). However, there was no significant difference in BMT between UCP2-knockout mice and db/db mice. Therefore, these results demonstrate that knockdown of UCP2 is a critical factor for retinal cell death and damage.

UCP2 siRNA effectively reduced endogenous levels of UCP2 protein compared to control siRNA (Fig. 2A). UCP2 siRNA increased caspase-3 expression to a significant extent, whereas control siRNA did not (Fig. 2B). UCP2 siRNA exaggerated HG-induced apoptosis compared to control-siRNA infected cells (Fig. 2C and D). These data demonstrate that endogenous UCP2 expression is important for preventing apoptosis in HUVECs.

In the present study, HUVECs were infected with adenovirus carrying UCP2 cDNA in sense orientation at an MOI of 1, 5 and 10 pfu/cell for 24 h. Immunofluorescence microscopy was then used in order to observe UCP2 eGFP expression in infected cells. At an MOI of 10 pfu/cell, UCP2 eGFP expression was detected in >95% of cultured cells (Fig. 3A). UCP2 was overexpressed mainly within the mitochondria (Fig. 3B and C).

HG increased HUVEC apoptosis by 4.1-fold; UCP2 overexpression reduced HG-induced apoptosis by 65% in HUVECs (Fig. 4A). FACScan analysis revealed that UCP2 overexpression inhibited both HG-induced apoptosis and necrosis of HUVECs (Fig. 4A and B). HG induced a significant increase in caspase-3 activity in HUVECs (Fig. 4C and D). The HG-induced increase in caspase-3 activity was inhibited by transfection with Ad-UCP2. The generation of active caspase-3 from the proenzyme was enhanced in the presence of HG (Fig. 4C and D). HG-induced cleavage of caspase-3 into active subunits was thus prevented in the presence of UCP2.

UCP2 overexpression in HUVECs decreased the susceptibility of the cells to mPTP opening. UCP2 overexpression in HUVECs delayed the time of mPTP opening by 1.5±0.2-fold when compared to control values (P<0.01) (Fig. 5A). In addition, treatment with HG resulted in the liberation of cytochrome c from mitochondria to the cytosol (Fig. 5B). The liberation of cytochrome c by HG was blocked in cells overexpressing UCP2 (Fig. 5B).

Previous research has indicated that UCP2 is activated by ROS and attenuates excessive ROS production in the form of a negative feedback mechanism (22–24). Thus, in the present study, the levels of ROS were determined in HUVECs. Incubation with HG significantly increased intracellular ROS generation. However, upon UCP2 overexpression, we noted that the increase in ROS production effected by HG was suppressed (Fig. 6A).

It has previously been noted that NO produced by eNOS exerts anti-inflammatory effects on the vascular wall, and inhibits the migration and proliferation of vascular smooth muscle cells (25). Thus, in the present study we detected the intracellular NO level in HUVECs. We noted that intracellular NO levels were decreased by HG, whereas UCP2 overexpression increased intracellular NO levels (Fig. 6B).