Human Retinal Capillary Endothelial Cells (HRCECs) were acquired from the China Center for Type Culture Collection (Wuhan, China) and were cultured in high glucose Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco, 11965092, Grand Island, New York, USA) supplied with 10% Fetal Bovine Serum (FBS) (Sijiqing, 70220-8611, Hangzhou, Zhejiang, China) and 1% penicillin-streptomycin solution (Sangon-Biotech, E607011, Shanghai, China). HRCECs were incubated at 37°C in 5% CO₂.

HRCECs were implanted into the 6 well plates and then processed with 10 μg/mL BAE for 24 h. Then, the HRCECs were processed with 30 mM high glucose or 5.5 mM normal glucose medium. Cells were processed with 30 mM glucose for 48 h to simulate the DR microenvironment [20]. The blueberry anthocyanins extract (BAE) was purchased from the Daxinganling Lingonberry Organic Foodstuffs Co., Ltd. (Daxinganling, China).

For cell experiments, the cells were randomly divided into 4, 5, or 6 groups. 4 groups: Control group, high glucose (HG) group, BAE group, HG + BAE group. 5 groups: Control group, HG group, HG + BAE group, HG + BAE + mimic NC group, HG + BAE + miR-33 group. 6 groups: Control group, HG group, HG + BAE group, HG + BAE + mimic NC + pcDNA-NC group, HG + BAE + miR-33 + pcDNA-NC group, HG + BAE + miR-33 + pcDNA-GLCCI1 group.

The miR-33 mimic and miR-33 inhibitor were purchased from Gene Pharm (Shanghai, China), and their sequences were as follows: hsa-miR-33a mimic: GUGCAUUGUAGUUGCAUUGCA, hsa-miR-33a inhibitor: UGCAAUGCAACUACAAUGC; The mimic NC and inhibitor NC were served as control. Cells were planted on 6-well plates for 24 h. Then the miR-33 mimic (or miR-33 inhibitor) and Lipofectamine™ 3000 (Invitrogen, L3000015, Carlsbad, CA, USA) were respectively incubated in OPTI-MEM reduced serum medium (Invitrogen, 31985070, Carlsbad, CA, USA) for 20 min. The mixture of miR-33 mimic (or miR-33 inhibitor) and Lipofectamine™ 3000 was transfected to the cells in the plate. The cells were replaced with the normal glucose medium 5 h later. 48 h later, the cells on the plates were harvested for further study. For GLCCI1 plasmid transfection, cells were transfected with a mixture of GLCCI1 plasmid and Lipofectamine™ 3000 in OPTI-MEM reduced serum medium.

Cell viability was detected using the CCK-8 kit (Yeasen, 40203ES60, Shanghai, China). 5 × 10³ HRCECs were seeded into one cell of the 96 well plates and cultured for 24 h. 24 h later, the cells in each well were added 10 μL CCK-8 reagent for 4 h incubation in 5% CO₂. After 4 h incubation, the optical densities (OD) at 450 nm were measured by a microplate reader (Beckman coulter, DTX 880, Brea, CA, USA).

FITC Annexin V Apoptosis Detection Kit I (BD, No. 556547, Franklin Lakes, New York, USA) was used to detect apoptosis in HRCECs. HRCECs were collected when they reached the logarithmic growth phase. 1 × 10⁵ HRCECs were suspensed with 100 μl Binding Buffer. Next, 5 μL of Annexin-V and 5 μL of Propidium Iodide (PI) were added to the buffer and incubated at room temperature for 15 min in the darkness for Accuri C6 flow cytometer (BD Biosciences, USA) analysis.

RNA was isolated from HRCECs with TRIzol™ reagent (Invitrogen, 15596026CN, Carlsbad, CA, USA). 1 μg isolated RNA was prepared using Prime Script™ RT reagent kit (Takara, RR037A, Shiga, Japan) or Prime Script™ miRNA RT-PCR Kit (Takara, RR716, Shiga, Japan). TB Green® Premix Ex Taq™ II (Takara, RR820Q, Shiga, Japan) was used for real-time PCR, which was run on Step One Plus™ Real-Time PCR System (Applied Biosystems, 4376600, Foster City, CA, USA). Relative gene expression was normalized with GAPDH or U6 and calculated with the 2^−ΔΔCT method. The sequences of the primers used in this study are listed in Table 1.

The total protein content was extracted from HRCECs with RIPA lysis buffer (Beyotime, P0013B, Shanghai, China) and quantified. 30 μg was loaded in the concentration gel and resolved by separation gel. The objective protein was transferred to the polyvinylidene difluoride membrane (Millipore, ISEQ10100, Boston, MA, USA). Next, the membrane was blocked with 5% skim milk for 2 h. The membranes were incubated with primary antibodies anti-GLCCI1 (Abcam, ab107491, Cambridge, UK), and GAPDH (Abcam, ab181602, Cambridge, UK) at 4°C overnight. Subsequently, the membranes were washed with Tris-buffered saline tween (TBST) three times and then incubated with Goat Anti-Rabbit IgG (H + L) HRP (Abways, AB0101, Shanghai, China) at room temperature for 2 h. GAPDH served as an internal control. The protein band was visualized with an enhanced chemiluminescent detection kit (NCM Biotech, P2300, Suzhou, China) and analyzed by ImageJ software (Bio-Rad, Image Lab 6.1, Hercules, CA, USA).

Reactive oxygen species (ROS) generation in HRCECs was examined with the Reactive Oxygen Species Assay Kit (Beyotime, S0033s, Shanghai, China). 1 × 10⁵ HECRCs were seeded into one well of 6 well plates and then stained with 10 μM 2,7-Dichlorodi-hydro fluorescein diacetate (DCFH-DA) probes at 37°C in darkness for 30 min. The fluorescence intensity of ROS was measured at Ex/Em = 488/525 nm wavelength by fluorescence microscope (Nikon, Ts2R-FL, Tokyo, Japan).

The MDA level and SOD activity were detected with a Malondialdehyde (MDA) assay kit (Nanjing Jiancheng, A003-1, Nanjing, China) and Superoxide Dismutase Activity Assay kit (amyjet, STA-340, Wuhan, China) according to the manufacturer’s instructions.

We added 250 μL Matrigel (Corning, 354234, Corning, New York, USA) to pre-cooled 24-well plates for polymerization at 37°C for 60 min. A total of 1 × 10⁵ treated HRCEC cells were plated onto the Matrigel matrix. After incubation at 37°C 48 h, the formation of tubes was counted in 6 random microscopic fields with a computer-assisted microscope (Olympus, IXplore Pro, Tokyo, Japan) and quantified using ImageJ software (Bio-Rad, Image Lab 6.1, Hercules, CA, USA).

Dual-luciferase reporter assay was carried out with the method in this article [22]. Briefly, the wild-type (wt) GLCCI1 3′UTR mRNA and mutated (mut) GLCCI1 3′UTR mRNA was inserted into the luciferase pmir-GLO reporter vector. Cells were co-transfected with either wt GLCCI1 or mut GLCCI1 luciferase pmir-GLO reporter vector plus NC mimic (miR-33a-5p mimic) or NC mimic (miR-33b-5p mimic) by Lipofectamin™ 3000 (Invitrogen, L3000015, Carlsbad, CA, USA). After 48 h, cells were harvested and the luciferase activities were measured using Beyotime’s Dual-Lumi™ Luciferase Assay System (Beyotime, RG088S, Shanghai, China).

The two groups’ differences were analyzed using an unpaired two-tailed Student’s t-test. If there were more than two groups, one-way ANOVA or two-way ANOVA with Tukey’s post hoc analysis was used. GraphPad Prism 9.0 software (GraphPad Software, Inc., La Jolla, CA, USA) was used for statistical analysis. All values are presented as mean ± standard deviation (SD). p values < 0.05 was considered statistically significant. Each assay was performed at least 3 times.

To test the potential of BAE in alleviating DR, HRCECs were first treated with HG to simulate the DR microenvironment. Based on the findings of Huang et al., 10 mg/L BAE was selected as the optimal concentration for our subsequent experiments [20]. Proanthocyanidins inhibit the miR-33 expression in obese rats [23]. miR-33 has emerged as a drug target for a variety of metabolic diseases [24]. BAE partially inhibited the miR-33 expression increased by high-glucose treatment (Fig. 1A). The GLCCI1 level in high-glucose treated HRCECs cells was lower than the GLCCI1 in the control group, but the result was reversed after adding BAE (Fig. 1B,C). Moreover, the oxidative stress indexes in HRCECs were evaluated. Results revealed that ROS level was elevated in HG-treated HRCECs, which was abolished by BAE treatment (Fig. 1D,E). Similarly, BAE treatment reversed HG effects on the SOD activity and the MDA levels in HRCECs (Fig. 1F,G). Furthermore, the influence of BAE on cell apoptosis, cell viability, and angiogenesis of HRCECs was evaluated. As shown in Fig. 1H,I, HG treatment down-regulated the apoptosis rate of HRCECs, while BAE restored that. CCK-8 and tube formation assay results showed cell viability and angiogenesis of HRCECs were elevated by HG treatment. BAE treatment significantly reduced the proliferation and angiogenesis of HG-treated HRCECs (Fig. 1J–L). According to these data, BAE relieved oxidative stress and angiogenesis in HG-treated HRCECs.

To reveal the mechanism of miR-33 in DR, miR-33 was overexpressed in HRCECs (Fig. A1). Overexpression of miR-33 elevated miR-33 expression in HG + BAE-treated HRCECs (Fig. 2A). Compared with the HG + BAE group, transfection of miR-33 mimic decreased the GLCCI1 mRNA and protein expression (Fig. 2B,C). Significant reduction in ROS and MDA levels can be seen in the group of HG-treated HRCECs after BAE. The above results were rescued by miR-33 overexpression (Fig. 2D–G). Additionally, flow cytometry results demonstrated that HG-induced downregulation of apoptosis rate in HRCECs could be reversed by BAE combined treatment, but the effects of BAE were canceled out when miR-33 overexpression was achieved through transfection with miR-33 miRNA (Fig. 2H,I). MiR-33 overexpression increased the cell viability and angiogenesis of HRCECs compared with the HG + BAE group (Fig. 2J–L). Taken together, these data indicated that BAE alleviated the oxidative stress and angiogenesis of HG-treated HRCECs by regulating miR-33 expression.

The GLCCI1 gene is located on 7p21.3, it contains eight exons and identified functional single nucleotide polymorphism [25,26]. Firstly, we used the online bioinformatics tools Target Scan (www.targetscan.org, accessed on 03/28/2024), miRDB (mirdb.org, accessed on 03/28/2024), and ENCORI (https://starbase.sysu.edu.cn/, accessed on 03/28/2024) to predict the potential target of miR-33. According to these datasets, GLCCI1 was recognized as the latent target of miR-33 and had a potential binding site of miR-33 (Fig. 3A). By dual-luciferase reporter assays, it was found that miR-33 mimic inhibited luciferase activity of the wild-type (WT) GLCCI1 reporter, but this effect was entirely canceled out in the mutant (MUT) GLCCI1 reporter group (Fig. 3B). Subsequently, HRCECs treated with miR-33 mimics exhibited significant downregulation of GLCCI1 mRNA and protein by qRT-PCR and WB. The miR-33 inhibitor led to an up-regulation of GLCCI1 in HRCECs (Fig. 3C–E). Up-regulation of GLCCI1 reversed miR-33 mimic-mediated inhibition of GLCCI1 mRNA and protein expression in HRCECs (Fig. 3F–H). HRCECs were upregulated by miR-33 inhibitors in terms of GLCCI1 mRNA and protein levels, which was abolished by GLCCI1 knockdown (Fig. 3I–K). All these data indicated that miR-33 interacted with GLCCI1 and repressed GLCCI1 expression.

Finally, the mechanism of BAE in ameliorating HG-induced oxidative stress and angiogenesis of HRCECs was investigated. qRT-PCR and WB analysis indicated up-regulation of miR-33 and decreased GLCCI1 mRNA and protein in HRCECs treated with HG. BAE treatment inhibited the miR-33 expression while enhancing GLCCI1 mRNA and protein. In HRCECs treated with HG + BAE, miR-33 overexpression decreased the expression of GLCCI1. This effect was reversed by overexpression of GLCCI1. (Fig. 4A–C). Moreover, the oxidative stress indexes in HRCECs were evaluated. Following BAE treatment, HRCECs treated with HG showed a reduction in ROS and MDA levels and an increase in SOD activity. miR-33 overexpression elevated oxidative stress with increased levels of ROS and MDA and decreased the SOD activity in HG + BAE-treated HRCECs, which was abolished by GLCCI1 overexpression (Fig. 4D–F). Additionally, CCK-8, flow cytometry, and tube formation assay results showed that GLCCI1 overexpression reversed miR-33 mimic-mediated promotion of cell viability, angiogenesis, and inhibition of apoptosis in HG + BAE-treated HRCECs (Fig. 4G–I). Taken together, BAE alleviated the oxidative stress and neovascularization of HG-treated HRCECs by regulating the miR-33/GLCCI1 axis.