Abstract
Background
Diabetic retinopathy (DR), a complication of diabetes, damages microvascular of retina through various molecular pathways. Emerging evidence points to miRNAs as key players in progression of DR. This study aims to investigate whether miR-103a-3p contributes to pathological processes of DR through MFN2.
Methods
Serum samples were collected from type 2 diabetes mellitus patients and divided into NDR, NPDR, and PDR groups based on fundus lesions. miR-103a-3p levels in each group were quantified by qRT-PCR. ARPE-19 cells were cultured under high-glucose (HG). Cell viability and apoptosis rates were evaluated by CCK-8 assay and flow cytometry. Activities of MDA and GSH-Px were detected by specific kits. Luciferase reporter gene confirmed MFN2 as a direct target of miR-103a-3p.
Results
Clinical sample detection revealed that miR-103a-3p levels were higher in NPDR and PDR patients compared to control and NDR groups, and it was an independent risk factor for DR. In vitro experiments confirmed that HG treatment greatly increased miR-103a-3p levels, decreased cell viability, accelerated apoptosis, elevated MDA content, and reduced GSH-Px activity, while transfection of miR-103a-3p inhibitor reversed these effects. Using luciferase reporter assay, we identified MFN2 as a direct target of miR-103a-3p. Moreover, rescue experiments demonstrated that silencing MFN2 effectively reversed the cell functions induced by miR-103a-3p inhibitor.
Conclusion
miR-103a-3p is upregulated in DR and accelerates its progression by directly targeting and inhibiting MFN2 expression. This study suggested the molecular mechanism of miR-103a-3p/MFN2 axis in DR, identifying a novel potential target for early diagnosis and therapy of DR.
Supplementary Information
The online version contains supplementary material available at 10.1186/s13098-026-02165-5.
Keywords: Diabetic retinopathy, miR-103a-3p, MFN2, Cell viability, Apoptosis
Introduction
Diabetic retinopathy (DR) is a major microvascular complication of diabetes, which has become main cause of blindness in working-age adults worldwide [1, 2]. The pathological process is complex, and prolonged chronic hyperglycemia triggers a series of molecular cascades. This may cause retinal vascular endothelial cell dysfunction, disruption of the blood-retinal barrier, and ultimately trigger a series of pathological changes such as retinal ischemia, neovascularization, and fibroproliferation [3]. Hyperglycemia state induces oxidative stress conditions, which activate multiple signaling pathways and generate various metabolites, thereby exacerbating complications [4]. Although current treatments for DR have achieved some efficacy, a portion of patients still suffer from poor treatment, and early prevention and intervention remain limited [5]. Therefore, elucidating the molecular mechanisms driving DR progression is crucial for developing novel early diagnostic biomarkers and effective therapeutic targets.
MicroRNAs (miRNAs) regulate target genes, thereby participating in various disease-related processes such as cell proliferation, apoptosis, oxidative stress, and inflammatory responses [6]. In DR research, multiple miRNAs have been identified as being abnormally expressed in serum or aqueous humor of patients and participated in DR progression by regulating specific target genes [7]. For example, miR-15b is involved in regulating angiogenesis in DR by targeting VEGF [8]. In DR patients, miR-181 levels are elevated, and it raises proliferation and migration potential of retinal endothelial cells by modulating KLF6 [9]. These results point to miRNAs as key molecules linking hyperglycemia with retinal dysfunction. Through preliminary literature review and initial screening, we highlighted miR-103a-3p as our primary focus. Research indicates that miR-103a-3p markedly elevated in vitreous exosomes from DR patients [10]. Additionally, miR-103a-3p mediates hyperglycemia-induced cardiomyocyte senescence by regulating Rnd3 [11]. However, its specific mechanism in DR remains unclear.
Mitochondrial dynamics abnormalities represent an emerging focus in diabetic complications [12]. Mitofusin 2 (MFN2) is a key protein regulating mitochondrial outer membrane fusion, which is essential for maintaining mitochondrial functional homeostasis [13]. Research has demonstrated that decreased expression of MFN2 is associated with endothelial cell apoptosis and oxidative stress [14]. Furthermore, oxidative stress significantly reduces MFN2 levels by inducing miR-195 expression, thereby promoting diabetic retinal endothelial cell injury and increased vascular permeability [15]. Notably, G protein-coupled receptors also play a significant role in the pathological processes of diabetes and its complications, such as diabetic retinopathy, suggesting the existence of diverse molecular regulatory networks in this field [16]. Bioinformatics analysis identified MFN2 to be a putative target of miR-103a-3p. However, no studies have directly investigated whether the regulatory relationship between miR-103a-3p and MFN2 exists in retinal cells, nor have any reports describing its role in DR context been published. Although it is known that miR-103a-3p is abnormally expressed in DR and that MFN2 is involved in the pathological process of DR, the scientific question of whether these two constitute an independent regulatory axis remains unexplained.
Based on the above background, this study aims to clarify the expression, clinical impact, and mechanism of miR-103a-3p in DR. miR-103a-3p levels were validated in samples to assess its correlation with DR severity and its diagnostic capability. Effect of miR-103a-3p on cell viability, apoptosis, and oxidative stress were investigated by an in vitro high-glucose (HG) model. The targeted regulatory association of miR-103a-3p and MFN2 was confirmed by rescue experiments. This aims to reveal novel mechanisms of DR pathogenesis and provides new theoretical basis and potential targets for prevention and treatment of this disease.
Materials and methods
Clinical samples collection
A total of 202 type 2 diabetic mellitus (T2DM) patients admitted to Daping Hospital, Army Medical University from September 2022 to March 2025 were included in the study. Meanwhile, 62 healthy controls (HC) matched for age and gender were included. Inclusion criteria for T2DM patients were as follows: age ranged from 40 to 75 years, patients met American Diabetes Association diagnostic criteria for diabetes [17] and received daily subcutaneous insulin injections. Exclusion criteria were as follows: patients had type 1 diabetes or gestational diabetes. Patients with other ocular diseases such as cataracts or glaucoma. Patients had a recent history of ocular surgery. Patients had severe cardiac, hepatic, or renal insufficiency, tumors, or autoimmune diseases. All patients underwent fundus examinations and were categorized into no diabetic retinopathy (NDR), non-proliferative diabetic retinopathy (NPDR), and proliferative diabetic retinopathy (PDR) groups based on fluorescein angiography results. No retinal microangiopathy caused by diabetes was found in NDR group. The NPDR group exhibited microangiopathy confined within the retinal layer without the presence of neovascularization. The PDR group showed retinal neovascularization or vitreous and pre-retinal hemorrhage. Fasting venous blood samples were obtained from study subjects to detect lipid and blood glucose related indicators.
This study was approved by the Ethics Committee of Daping Hospital, Army Medical University, and all study subjects signed informed consent.
Cell culture and treatment
Human retinal pigment epithelial cells ARPE-19 were acquired from ATCC and maintained in DMEM medium (Gibco, USA) added with 10% FBS (Gibco, USA) and 1% penicillin-streptomycin (Sigma, USA) at 37℃ in a 5% CO2 incubator. Cells in logarithmic growth phase were exposed to either 5 mM (control) or 25 mM glucose (HG group) for 24 h. The rationale for selecting a 25 mM glucose concentration is as follows: This concentration is the standard HG concentration widely used in DR in vitro studies to simulate the hyperglycemic state in diabetic patients [18, 19]. Meanwhile, this study established an isotonic mannitol (MNT) group (5 mM glucose + 20 mM mannitol) and glucose concentration gradients (5 mM, 15 mM, 25 mM, 30 mM) in the pre-experiment. Results demonstrated that 24 h treatment with 25 mM glucose stably induced changes in miR-103a-3p expression and cell viability (Supplementary Fig.1).
Cell transfection
All miRNA reagents (mimic, inhibitor, and NC) were from Ribobio (Guangzhou, China), while si-MFN2 and si-NC were from GenePharma (Shanghai, China). At 80% confluency, cells were transfected of respective plasmids using Lipofectamine 3000 (Invitrogen, USA). The medium was replaced after 6 h transfection, and the incubation continued for 24 h before high-glucose treatment.
Quantitative real RT-PCR assay
RNA isolation was performed on serum and cell using Trizol reagent (Invitrogen, USA). Reverse transcription was carried according to the PrimeScript RT Master Mix kit instructions (Takara, Japan). The reaction system was prepared following the SYBR Green PCR Master Mix kit (Thermo Fisher, USA) instructions for thermal cycling. U6 and GAPDH were as reference genes, and relative expression of genes were determined via 2−ΔΔCt method. The primer sequences used in this study were as follows: miR-103a-3p forward primer: 5’- GGGAGCAGCATTGTACAG-3’, reverse primer: 5’-CTCAACTGGTGTCGTGGA-3’; MFN2 forward primer: 5’-CTTGAAGACACCCACAGGAACA-3’, reverse primer: 5’-GGCCAGCACTTCGCTGATAC-3’; U6 forward primer: 5’-CTCGCTTCGGCAGCACA-3’, reverse primer: 5’-AACGCTTCACGAATTTGCGT-3’; GAPDH forward primer: 5’-CATCAACGGGAAGCCCATC-3’, reverse primer: 5’-CTCGTGGTTCACACCCATC-3’.
Cell viability assay.
Cells in well-grown condition were harvested and resuspended at a density of 1 × 104 cells/mL. Afterwards, 200 µL of suspension was seeded into 96-well plate. After incubation for 0, 24, 48, and 72 h, 10 µL of CCK-8 reagent (AbMole, USA) was added to every well. The absorbance was recorded at 450 nm by a microplate reader.
Apoptosis assay
Cells from each group were digested with trypsin, washed with PBS, and resuspended in 1× Binding Buffer before being seeded in 6-well plates. Subsequently, the cell suspension was incubated with 5 µL Annexin V-FITC (BD, USA) and 10 µL PI for 15 min in the dark. The proportion of apoptotic cells was examined by flow cytometry.
Oxidative stress assay
The concentrations of MDA and GSH-Px in each group of ARPE-19 cells were detected using the cellular glutathione peroxidase assay kit with NADPH and the lipid peroxidation MDA assay kit from Beyotime (Shanghai, China).
Luciferase assay
The 3’-UTR fragment of MFN2 containing miR-103a-3p binding sites was cloned into the pGL3 vector to generate MFN2 wild-type (MFN2-WT) vector. Binding sites were changed and inserted into the pGL3 vector to construct MFN2 mutant (MFN2-MUT) vector. MFN2-WT or MFN2-MUT was co-transfected with miR-103a-3p mimic, inhibitor, or negative control into ARPE-19 cells using Lipofectamine 3000. After 48 h of transfection, luciferase activity was detected by dual luciferase reporter assay kit (Promega, USA).
Statistical analysis
Data in this study are expressed as mean ± SD, with all experiments independently replicated three times. Statistical analysis and figure plotting were conducted by SPSS 27.0 and GraphPad Prism 9.0. Differences between two groups were assessed by t-test, while multiple group comparisons were employed by one-way ANOVA, followed by Tukey’s post hoc test for pairwise comparisons between groups. Diagnostic capability of miR-103a-3p was evaluated using ROC curves. Multivariate logistic analysis was utilized to explore related risk factors of DR. To assess multicollinearity among variables, the variance inflation factor (VIF) was calculated, with VIF < 5 used as the criterion for absence of multicollinearity. Statistical significance set at P < 0.05.
Results
Comparison of basic clinical information on study subjects
The study cohort comprised 202 T2DM patients, including 82 NDR patients, 75 NPDR patients, and 45 PDR patients. Their basic information was shown in Table 1. Compared with HC group, all T2DM patients showed significant differences in TG, HDL-C, FPG, and HbA1c (P < 0.001), while no significant differences in age, gender, BMI, TC, and LDL-C (P > 0.05). In addition, comparison between DR and NDR patients revealed significant differences in TG, disease duration, FPG, and HbA1c (P < 0.05), while other test indicators showed insignificant differences (P > 0.05).
Table 1.
Basic information of research subjects
| HC (n = 68) |
NDR (n = 82) |
NPDR (n = 75) |
PDR (n = 45) |
P valuea | P valueb | |
|---|---|---|---|---|---|---|
| Age (range, years) | 57.09 ± 10.13 (40–75) | 57.54 ± 10.95 (40–75) | 57.64 ± 10.73 (40–75) | 57.91 ± 11.07 (40–75) | 0.703 | 0.895 |
| Gender, male, n (%) | 36 (52.9%) | 49 (59.8%) | 47 (62.7%) | 28 (62.2%) | 0.220 | 0.694 |
| BMI (kg/m2) | 24.39 ± 2.94 | 24.58 ± 2.96 | 25.16 ± 2.84 | 25.73 ± 3.27 | 0.116 | 0.065 |
| TC (mmol/L) | 4.61 ± 0.79 | 4.72 ± 1.03 | 5.01 ± 1.23 | 4.97 ± 1.14 | 0.065 | 0.092 |
| TG (mmol/L) | 1.44 ± 0.31 | 1.81 ± 0.56 | 1.94 ± 0.59 | 2.08 ± 0.54 | < 0.001 | 0.025 |
| LDL-C (mmol/L) | 2.76 ± 0.64 | 2.80 ± 0.82 | 2.97 ± 0.93 | 3.04 ± 1.02 | 0.171 | 0.140 |
| HDL-C (mmol/L) | 1.29 ± 0.28 | 1.18 ± 0.34 | 1.14 ± 0.25 | 1.08 ± 0.25 | < 0.001 | 0.133 |
| Disease duration (years) | / | 6.06 ± 2.65 | 10.13 ± 2.02 | 12.22 ± 1.64 | / | 0.003 |
| FPG (mmol/L) | 4.97 ± 0.47 | 8.66 ± 2.45 | 9.44 ± 3.43 | 10.81 ± 3.02 | < 0.001 | < 0.001 |
| HbA1C (%) | 5.27 ± 0.41 | 7.54 ± 1.66 | 8.79 ± 2.12 | 9.24 ± 2.06 | < 0.001 | < 0.001 |
BMI, body mass index; TC, total cholesterol; TG, triacylglycerol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose; HbA1C, glycosylated hemoglobin. P valuea, comparison between diabetic patients (NDR, NPDR and PDR) versus HC group. P valueb, comparison between diabetic retinopathy (NPDR and PDR) versus NDR patients. Bold, P value < 0.05
Expressions of miR-103a-3p in different groups
We quantified serum miR-103a-3p levels by qRT-PCR, with results shown in Fig. 1A. Levels were significantly elevated in NDR, NPDR, and PDR compared with HC group. Compared with the NDR group, levels were also markedly elevated in NPDR and PDR. ROC curve showed that miR-103a-3p had ability to distinguish T2DM patients from HC, with a sensitivity of 81.19% and specificity of 88.24% (AUC = 0.918, P < 0.001, Fig. 1B). Additionally, miR-103a-3p also significantly distinguish DR patients from NDR patients, with a sensitivity of 85.83% and specificity of 74.39% (AUC = 0.877, P < 0.001, Fig. 1C).
Fig. 1.
Expressions of miR-103a-3p in study subjects. (A) miR-103a-3p levels in the HC, NDR, NPDR, and PDR groups. (B) ROC curve of miR-103a-3p’s ability to distinguish diabetic patients from the HC group. (C) ROC curve miR-103a-3p’s ability to distinguish DR patients from the NDR patients. ***P < 0.001, comparison between diabetic patients (NDR, NPDR and PDR) versus HC group. ###P < 0.001, comparison between diabetic retinopathy patients (NPDR and PDR) versus NDR patients
Assessing risk factors for DR in diabetic patients
Through multivariate logistic regression, we identified miR-103a-3p as an independent risk factor for DR in the T2DM cohort (OR = 7.366, P < 0.001). In addition, disease duration (OR = 2.894, P = 0.007), FPG (OR = 3.001, P = 0.005), and HbA1C (OR = 3.153, P = 0.003) were also independent risk factors (Table 2).
Table 2.
Analysis of risk factors for diabetic retinopathy in patients with diabetes
| OR | 95% CI | P value | VIF | ||
|---|---|---|---|---|---|
| Lower | Upper | ||||
| miR-103a-3p | 7.366 | 3.365 | 16.122 | < 0.001 | 1.264 |
| Age | 1.232 | 0.589 | 2.577 | 0.579 | 1.047 |
| Gender | 1.398 | 0.661 | 2.955 | 0.380 | 1.038 |
| BMI | 1.675 | 0.793 | 3.536 | 0.176 | 1.051 |
| TC | 1.270 | 0.599 | 2.694 | 0.533 | 1.091 |
| TG | 1.996 | 0.959 | 4.158 | 0.065 | 1.031 |
| LDL-C | 1.628 | 0.775 | 3.419 | 0.198 | 1.055 |
| HDL-C | 0.571 | 0.265 | 1.233 | 0.154 | 1.074 |
| Disease duration | 2.894 | 1.342 | 6.242 | 0.007 | 1.334 |
| FPG | 3.001 | 1.405 | 6.413 | 0.005 | 1.079 |
| HbA1C | 3.153 | 1.482 | 6.706 | 0.003 | 1.113 |
BMI, body mass index; TC, total cholesterol; TG, triacylglycerol; LDL-C, low-density lipoprotein cholesterol; HDL-C, high-density lipoprotein cholesterol; FPG, fasting plasma glucose; HbA1C, glycosylated hemoglobin. Bold, P value < 0.05
Effect of inhibition of miR-103a-3p on HG-treated cells
HG treatment resulted in significant rise of miR-103a-3p, while transfection with miR-103a-3p inhibitor decreased its levels (Fig. 2A). In addition, HG treatment inhibited cell viability, whereas inhibiting miR-103a-3p reversed this trend (Fig. 2B). Apoptosis assays indicated that HG-induced accelerated cell apoptosis, whereas inhibiting miR-103a-3p slowed it (Fig. 2C). Furthermore, HG treatment elevated MDA levels and reduced GSH-Px levels, whereas inhibition of miR-103a-3p reversed this trend (Fig. 2D, E).
Fig. 2.
Effect of inhibition of miR-103a-3p on HG-treated ARPE-19 cells. Effects of HG treatment and transfection with miR-103a-3p inhibitor on (A) miR-103a-3p levels, (B) cell viability, (C) apoptosis, (D) MDA levels, and (E) GSH-Px levels in ARPE-19 cells. ***P < 0.001, comparison between HG group versus Control group. ###P < 0.001, comparison between HG + miR-103a-3p inhibitor group versus HG + inhibitor NC group
miR-103a-3p targets MFN2
Binding sites of MFN2 and miR-103a-3p were identified through miRDB database (Fig. 3A). Luciferase assays demonstrated that co-transfection of miR-103a-3p mimic or inhibitor with MFN2-WT significantly decreased or increased luciferase activity, whereas co-transfection with MFN2-MUT showed no significant changes (Fig. 3B). In addition, MFN2 expressions were detected in serum of study subjects, and the results are shown in Fig. 3C. Compared with HC group, serum MFN2 levels were reduced in NDR, NPDR, and PDR. Serum MFN2 levels were also reduced in NPDR and PDR compared to NDR group.
Fig. 3.
miR-103a-3p targets MFN2. (A) Binding sites between miR-103a-3p and MFN2. (B) Dual luciferase assay verified the interaction between miR-103a-3p and MFN2. ***P < 0.001. (C) Expression of MFN2 in the HC, NDR, NPDR, and PDR groups. ***P < 0.001, comparison between diabetic patients (NDR, NPDR and PDR) versus HC group. ###P < 0.001, comparison between diabetic retinopathy patients (NPDR and PDR) versus NDR patients
Effect of inhibition of miR-103a-3p and MFN2 on HG-treated cells
HG treatment reduced MFN2 levels in cells, while transfection of miR-103a-3p inhibitor restored these levels, whereas co-transfection of miR-103a-3p inhibitor and si-MFN2 decreased them (Fig. 4A). Compared with inhibition of miR-103a-3p, inhibition of miR-103a-3p and MFN2 inhibited cell viability and promoted apoptosis (Fig. 4B, C). Under HG conditions, transfection of miR-103a-3p inhibitor and si-MFN2 reversed the tread of inhibiting miR-103a-3p on MDA and GSH-Px levels (Fig. 4D, E). These results indicated that silencing MFN2 effectively reverses the protective effect of miR-103a-3p inhibitor, confirming that MFN2 is a key effector molecule downstream of miR-103a-3p.
Fig. 4.
Effect of inhibition of miR-103a-3p and MFN2 on HG-treated ARPE19 cells. (A) Effects of HG treatment, transfection with miR-103a-3p inhibitor, and transfection with miR-103a-3p inhibitor and si-MFN2 on MFN2 expression. ***P < 0.001, comparison between HG group versus Control group. ###P < 0.001, comparison between HG + miR-103a-3p inhibitor group versus HG + inhibitor NC group. &&P < 0.01, comparison between HG + miR-103a-3p inhibitor + si-MFN2 group versus HG + miR-103a-3p inhibitor + si-NC group. Effects of transfection with miR-103a-3p inhibitor and si-MFN2 on (B) cell viability, (C) apoptosis, (D) MDA levels, and (E) GSH-Px levels. ***P < 0.001, comparison between HG + miR-103a-3p inhibitor group versus HG + inhibitor NC group. ##P < 0.01, ###P < 0.001, comparison between HG + miR-103a-3p inhibitor + si-MFN2 group versus HG + miR-103a-3p inhibitor + si-NC group
Discussion
The pathological process of DR involves complex molecular network regulation, with miRNAs increasingly attracting attention as regulatory factors [20]. This study focused on miR-103a-3p to reveal function and molecular mechanism in DR. Our study showed that serum miR-103a-3p levels in T2DM patients positively correlated with DR severity and were confirmed as an independent risk factor for DR. Mechanistically, we demonstrated that HG promotes miR-103a-3p but suppresses MFN2 expression, thereby contributing to oxidative stress injury, cell viability, and apoptosis. Furthermore, rescue experiments suggested that silencing MFN2 effectively reversed the cytoprotective effects induced by miR-103a-3p inhibitors.
Our findings align with recent research on the clinical role of miRNAs in DR. MicroRNAs are increasingly recognized as useful diagnostic and prognostic tools across diseases [21]. For instance, Liu et al. found that abnormal expressions of miR-4328, miR-4422, miR-548z, and miR-628-5p may serve as biomarkers for diagnosing and predicting DR [22]. Additionally, miR-3197 was also recognized as promising biomarker for DR diagnosis [23]. Our findings identified miR-103a-3p as being specifically elevated in the serum of NPDR and PDR patients, positively correlated with disease progression, and served as an independent risk factor. ROC curve revealed that miR-103a-3p had a high diagnostic value for DR. This suggests that miR-103a-3p may as a biomarker for early diagnosis and risk stratification of DR.
We further discovered MFN2 as a principal downstream target of miR-103a-3p. Evidence indicates that miRNAs contribute to disease pathogenesis by modulating downstream genes [24]. For instance, miR-200c-3p contributes to retinal cell injury in DR by promoting pyroptosis through SLC30A7 targeting, and its upregulation serves as a key mechanism in DR cell injury [25]. Additionally, studies have revealed that PDZK1, a key molecule linking mitochondrial dysfunction to apoptosis in the DR, is suppressed by miR-145-5p under HG conditions [26]. This led to impaired mitochondrial function and increased oxidative stress, ultimately exacerbating retinal endothelial cell apoptosis. MFN2 was a key regulator of mitochondrial fusion and exerted role in maintaining mitochondrial function, energy metabolism, and cell survival [27]. The stability of mitochondrial networks is disrupted, leading to excessive reactive oxygen species production and activation of the mitochondrial-dependent apoptosis pathway [28]. Research suggested that mitochondrial dysfunction is one of the early events in retinal neuropathy [29]. We suggested that downregulation of miR-103a-3p leads to elevated MFN2 levels, which explained the phenomena of increased oxidative stress and enhanced apoptosis in our experiments. Our study connects miR-103a-3p with mitochondrial dynamics disruption in DR, which provides a new perspective on the disease’s molecular pathology.
In addition, functional experiments demonstrated that inhibiting miR-103a-3p protected cells from the damaging effects of HG, which was reversed by silencing MFN2. This result is evidenced that MFN2 is a core downstream gene of miR-103a-3p in contributing to pathological process of DR. This finding suggests that miR-103a-3p/MFN2 axis provides a potential theoretical basis for intervening in DR.
Nevertheless, there are several limitations of this study that need further study. First of all, our clinical samples were derived from serum. Future studies should expand the sample size and simultaneously examine miR-103a-3p expressions in vitreous humor and retinal tissue to enable more comprehensive validation. Secondly, although ROC analysis demonstrated that miR-103a-3p exhibited good sensitivity and specificity for distinguishing HC from T2DM and DR from NDR in this study cohort, these results reflect diagnostic performance in predefined groups rather than disease specificity. Future studies should include other disease control groups to more fully evaluate the diagnostic specificity of miR-103a-3p for DR. Notably, the abnormal expression of miR-103a-3p may not be unique to DR, and altered expression may also occur in other microvascular complications such as diabetic nephropathy and diabetic neuropathy. Therefore, future studies should include cohorts with different diabetes complications to assess its DR specificity. Finally, this study primarily focused on ARPE-19 cells, which represent the outer blood-retinal barrier, while DR also involves retinal microvascular endothelial cells. In the future, further validation in multiple cell models and animal experiments will enable a more comprehensive characterization of this pathway’s role in DR.
In summary, this study suggested that miR-103a-3p was significantly upregulated in the serum of DR patients, with its expression level positively correlated with disease severity. The mechanism study confirmed that miR-103a-3p inhibitor reduced oxidative stress, increased cell viability, and suppressed apoptosis of retinal pigment epithelial cells in HG model by targeting and inhibiting MFN2 expression. This study first reveals regulatory role of the miR-103a-3p/MFN2 axis in DR, offering a novel perspective on understanding the molecular pathological mechanism of this disease. This finding not only suggests that miR-103a-3p may serve as a potential diagnostic marker for DR, but also provides a new theoretical basis for therapeutic strategies. However, its specificity as a clinical biomarker and its feasibility as a therapeutic target still require further validation in broader studies.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1 : Fig. S1. Pre-experiments for HG concentration selection. Effect of HG treatment at different concentrations on (A) cell viability and (B) miR-103a-3p. *P < 0.05, ***P < 0.001
Acknowledgements
Not Applicable.
Author contributions
YD L, Y Z designed this study. JL L and Y Z conducted the experiment and analyzed the data. JL L and Y Z wrote the manuscript. QP Z, M W revised the manuscript. All authors reviewed and approved for publication.
Funding
Not Applicable.
Data availability
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.
Declarations
Ethics approval and consent to participate
The authors state that they have obtained Daping Hospital, Army Medical University review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.
Consent for publication
All patients provided written informed consent.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yidan Liu and Yan Zhao contributed equally to this article.
References
- 1.Kour V, Swain J, Singh J, Singh H, Kour H. A Review on Diabetic Retinopathy. Curr Diabetes Rev. 2024;20(6):e201023222418. [DOI] [PubMed] [Google Scholar]
- 2.Li H, Liu X, Zhong H, Fang J, Li X, Shi R, et al. Research progress on the pathogenesis of diabetic retinopathy. BMC Ophthalmol. 2023;23(1):372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Basir H, Nugrahani ASD, Aman AM, Bakri S, Rasyid H, Umar H, et al. The association between fibroblast growth factor 21 with diabetes retinopathy among type 2 diabetes mellitus patients: a systematic review, meta-analysis, and meta-regression. PeerJ. 2024;12:e18308. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Chandra P, Sachan N, Saraswat N, Vyawahare N. A Detailed Review of Molecular Pathways and Mechanisms Responsible for the Development and Aggravation of Neuropathy and Nephropathy in Diabetes. Curr Mol Pharmacol. 2024;17:e280323215026. [DOI] [PubMed] [Google Scholar]
- 5.Shughoury A, Bhatwadekar A, Jusufbegovic D, Hajrasouliha A, Ciulla TA. The evolving therapeutic landscape of diabetic retinopathy. Expert Opin Biol Ther. 2023;23(10):969–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Xu Y, Huang X, Luo Q, Zhang X. MicroRNAs Involved in Oxidative Stress Processes Regulating Physiological and Pathological Responses. MicroRNA (Shariqah United Arab Emirates). 2021;10(3):164–80. [DOI] [PubMed] [Google Scholar]
- 7.Sadashiv, Sharma P, Dwivedi S, Tiwari S, Singh PK, Pal A, et al. Micro (mi) RNA and Diabetic Retinopathy. Indian J Clin biochemistry: IJCB. 2022;37(3):267–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Yang Y, Liu Y, Li Y, Chen Z, Xiong Y, Zhou T, et al. MicroRNA-15b Targets VEGF and Inhibits Angiogenesis in Proliferative Diabetic Retinopathy. J Clin Endocrinol Metab. 2020;105(11):3404–15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cao J, Zhao C, Gong L, Cheng X, Yang J, Zhu M, et al. MiR-181 Enhances Proliferative and Migratory Potentials of Retinal Endothelial Cells in Diabetic Retinopathy by Targeting KLF6. Curr Eye Res. 2022;47(6):882–8. [DOI] [PubMed] [Google Scholar]
- 10.Kot A, Kaczmarek R. Exosomal miRNA Profiling in Vitreous Humor in Proliferative Diabetic Retinopathy. Cells. 2022;12(1):123. [DOI] [PMC free article] [PubMed]
- 11.Wu L, Zhu X, Pan S, Chen Y, Luo C, Zhao Y, et al. Diabetes Advances Cardiomyocyte Senescence Through Interfering Rnd3 Expression and Function. Aging Cell. 2025;24(6):e70031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Wu Y, Lan H, Zhang D, Hu Z, Zhang J, Li Z, et al. Research progress on ncRNAs regulation of mitochondrial dynamics in diabetes. J Cell Physiol. 2022;237(11):4112–31. [DOI] [PubMed] [Google Scholar]
- 13.Joaquim M, Altin S, Bulimaga MB, Simões T, Nolte H, Bader V, et al. Mitofusin 2 displays fusion-independent roles in proteostasis surveillance. Nat Commun. 2025;16(1):1501. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Wang Y, Dai X, Li H, Jiang H, Zhou J, Zhang S, et al. The role of mitochondrial dynamics in disease. MedComm. 2023;4(6):e462. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhang R, Garrett Q, Zhou H, Wu X, Mao Y, Cui X, et al. Upregulation of miR-195 accelerates oxidative stress-induced retinal endothelial cell injury by targeting mitofusin 2 in diabetic rats. Mol Cell Endocrinol. 2017;452:33–43. [DOI] [PubMed] [Google Scholar]
- 16.Yeganeh-Hajahmadi M, Moosavi-Saeed Y, Rostamzadeh F. Apelin Receptor Dimerization and Oligomerization. Curr Mol Pharmacol. 2024;17:e180823219999. [DOI] [PubMed] [Google Scholar]
- 17.2. Diagnosis and Classification of Diabetes: Standards of Care in Diabetes-2024. Diabetes Care. 2024;47(Suppl 1):S20–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Platania CBM, Pittalà V, Pascale A, Marchesi N, Anfuso CD, Lupo G, et al. Novel indole derivatives targeting HuR-mRNA complex to counteract high glucose damage in retinal endothelial cells. Biochem Pharmacol. 2020;175:113908. [DOI] [PubMed] [Google Scholar]
- 19.Zhao S, Li T, Li J, Lu Q, Han C, Wang N, et al. miR-23b-3p induces the cellular metabolic memory of high glucose in diabetic retinopathy through a SIRT1-dependent signalling pathway. Diabetologia. 2016;59(3):644–54. [DOI] [PubMed] [Google Scholar]
- 20.Becker K, Klein H, Simon E, Viollet C, Haslinger C, Leparc G, et al. In-depth transcriptomic analysis of human retina reveals molecular mechanisms underlying diabetic retinopathy. Sci Rep. 2021;11(1):10494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Martinez B, Peplow PV. MicroRNAs as biomarkers of diabetic retinopathy and disease progression. Neural regeneration Res. 2019;14(11):1858–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Liu W, Luo Z, Zhang L, Wang Y, Yang J, You D, et al. hsa-mir-(4328, 4422, 548z and – 628-5p) in diabetic retinopathy: diagnosis, prediction and linking a new therapeutic target. Acta Diabetol. 2023;60(7):929–42. [DOI] [PubMed] [Google Scholar]
- 23.Ji H, Yi Q, Chen L, Wong L, Liu Y, Xu G, et al. Circulating miR-3197 and miR-2116-5p as novel biomarkers for diabetic retinopathy. Clin Chim Acta. 2020;501:147–53. [DOI] [PubMed] [Google Scholar]
- 24.Zhao X, Ling F, Zhang GW, Yu N, Yang J, Xin XY. The Correlation Between MicroRNAs and Diabetic Retinopathy. Front Immunol. 2022;13:941982. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Li W, Yang S, Chen G, He S. MiR-200c-3p regulates pyroptosis by targeting SLC30A7 in diabetic retinopathy. Hum Exp Toxicol. 2022;41:9603271221099589. [DOI] [PubMed] [Google Scholar]
- 26.An M, Huang J, Zhao J, Wang L, Liu Y. PDZK1 regulated by miR-145-5p protects against endothelial cell apoptosis and diabetic retinopathy by targeting mitochondrial function. Exp Eye Res. 2025;254:110314. [DOI] [PubMed] [Google Scholar]
- 27.Zaman M, Shutt TE. The Role of Impaired Mitochondrial Dynamics in MFN2-Mediated Pathology. Front cell Dev biology. 2022;10:858286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Wang T, Xing G, Fu T, Ma Y, Wang Q, Zhang S, et al. Role of mitochondria in doxorubicin-mediated cardiotoxicity: from molecular mechanisms to therapeutic strategies. Int J Med Sci. 2024;21(5):809–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Huang CP, Lin YW, Huang YC, Tsai FJ. Mitochondrial Dysfunction as a Novel Target for Neuroprotective Nutraceuticals in Ocular Diseases. Nutrients. 2020;12(7):1950. [DOI] [PMC free article] [PubMed]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1 : Fig. S1. Pre-experiments for HG concentration selection. Effect of HG treatment at different concentrations on (A) cell viability and (B) miR-103a-3p. *P < 0.05, ***P < 0.001
Data Availability Statement
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.