2025 Volume 48 Issue 11 Pages 1732-1740
This study investigated the protective effects of metformin against combined high glucose (HG)- and UVA-induced cytotoxicity in fetal rat skin keratinocytes (FRSK cells), a model of diabetic photoaging. HG combined with UVA caused a synergistic loss of cell viability accompanied by marked increases in phosphorylation of AMP-activated protein kinase (p-AMPK), reactive oxygen species (ROS) generation, senescence-associated β-galactosidase (SA-β-Gal) activity, and Sirtuin 1 (SIRT1) expression. HG alone induced moderate cytotoxicity and senescence, whereas UVA alone under normal glucose conditions (NG + UVA) produced negligible ROS and minimal viability loss. Metformin improved cell viability under dual stress conditions in a dose-dependent manner, with maximal protection observed at 8 mM. In UVA-free cultures, metformin increased p-AMPK in both NG and HG, peaking at 8 mM. Under HG + UVA, p-AMPK was higher than in NG + UVA and HG alone, with no additional increase following metformin treatment. ROS accumulation occurred only under HG + UVA and was strongly suppressed by metformin, nearly to baseline at 8 mM. The HG + UVA-induced increases in SA-β-Gal activity and SIRT1 expression were reduced in parallel with ROS suppression. These findings suggest that metformin’s cytoprotective effect in this model is primarily mediated by attenuation of ROS rather than by further AMPK activation, indicating an AMPK-independent antioxidant mechanism.
Diabetes mellitus is a chronic metabolic disorder characterized by sustained hyperglycemia, which impairs multiple physiological systems, including the skin. Emerging evidence suggests that hyperglycemia induces oxidative stress and inflammatory signaling in skin tissues, leading to structural and functional alterations that compromise skin barrier integrity and homeostasis.1) These alterations under hyperglycemic conditions may increase the susceptibility of diabetic skin to environmental insults, including UV radiation.2)
Among UV wavelengths, UVA (320–400 nm) penetrates deeply into the dermis and promotes photoaging through reactive oxygen species (ROS) generation, DNA damage, and mitochondrial dysfunction.3,4) These processes collectively lead to cellular senescence and extracellular matrix degradation. Although the independent effects of hyperglycemia and UVA exposure on skin cells have been studied,5) their combined impact on oxidative stress and cellular aging remains largely uncharacterized.
Metformin, a biguanide derivative, is widely prescribed for the treatment of type 2 diabetes mellitus owing to its glucose-lowering effects through inhibition of hepatic gluconeogenesis and improvement of insulin sensitivity.6) Beyond its metabolic actions, it has demonstrated antioxidant and cytoprotective effects in various cell types.7,8) These protective effects have also been reported in skin cells exposed to high glucose (HG) conditions.5) However, the role of metformin in protecting skin cells under diabetic and photoaging-like stress conditions has not been thoroughly investigated.
In this study, fetal rat skin keratinocytes (FRSK) were used as an in vitro skin cell model to investigate the protective effects of metformin against oxidative stress and cellular senescence induced by combined HG and UVA conditions. To assess metformin’s cytoprotective mechanisms, we examined changes in cell viability, intracellular ROS accumulation, senescence-associated β-galactosidase (SA-β-Gal) activity, and Sirtuin 1 (SIRT1) protein expression—biomarkers indicative of oxidative stress and cellular senescence. ROS accumulation is a central feature of oxidative stress-induced damage, contributing to aging and functional decline in skin cells.9) SA-β-Gal is a widely recognized marker of stress-induced premature senescence, reflecting irreversible cell cycle arrest.10) SIRT1, a NAD+-dependent deacetylase, modulates cellular aging and oxidative stress resistance by regulating transcription factors such as p53 and nuclear factor-kappa B (NF-κB).11) These biomarkers were selected to provide a comprehensive evaluation of metformin’s effects under dual-stressor conditions. Additionally, phosphorylation of AMP-activated protein kinase (p-AMPK), a key mediator of metformin’s pharmacological actions, was assessed to further explore its mechanisms of action under dual-stressor conditions.6) This study aimed to determine whether metformin mitigates dual-stressor-induced skin cell damage, thereby providing insights into its potential therapeutic applications for diabetes-associated photoaging.
FRSK cells were obtained from the JCRB Cell Bank (Osaka, Japan). Cells were cultured in Eagle’s Minimum Essential Medium (Gibco) supplemented with 10% fetal bovine serum (NICHIREI BIOSCIENCES, Tokyo, Japan) and 100 U/mL penicillin–100 µg/mL streptomycin (FUJIFILM Wako Pure Chemical, Osaka, Japan). Cells were maintained at 37°C in a humidified atmosphere with 5% CO2 and were passaged every 3–4 d. A glucose concentration of 25 mM (approximately 450 mg/dL) was employed to simulate hyperglycemic conditions in vitro.1) This concentration exceeds the diagnostic thresholds for diabetes.12)
UVA Irradiation ConditionsCells were exposed to UVA using a dual-wavelength UV + LED light source (Faerie, total output 48 W), which emits 365 nm (UVA) and 405 nm light, a device commonly used for nail curing. The distance between the light source and the cells was set to 5 cm. Based on the device specifications, the estimated UV power density at this distance was approximately 12 mW/cm2, consistent with previously reported intensities of similar UV nail curing devices, which have a median value of 10.6 mW/cm2.13) Irradiation was performed for 15 min at room temperature. This UVA protocol is supported by a previous study demonstrating that nail curing devices emitting 365–395 nm light induce oxidative stress in mammalian cells.13)
Cell Viability AssayCell viability was determined using the Cell Counting Kit-8 (CCK-8; CK04, Dojindo, Kumamoto, Japan) according to the manufacturer’s instructions. Cells were seeded in 96-well microplates at a density of 1 × 104 cells/well and incubated for 84 h to reach confluency before treatment. After 30 min preincubation at 37°C in 5% CO2, cells were exposed to UVA as described above, followed by incubation at 37°C with 5% CO2 for an additional 12–24 h. Throughout the experiment, cells were cultured in normal glucose (NG; 5.6 mM) or HG (25 mM) medium with or without metformin (1, 2, 4, 8, 16, or 32 mM). Following treatment, 10 µL of CCK-8 reagent was added to each well and incubated for 3 h at 37°C. Absorbance was measured at 450 nm using a microplate reader (Infinite® 100 PRO, Tecan Ltd., Switzerland). Cell viability (%) = [(As – Ab)] / (Ac–Ab)] × 100. As = Absorbance of wells with cells, test compound, and CCK-8 reagent. Ac = Absorbance of wells with cells and CCK-8 (no test compound). Ab = Absorbance of blank wells (medium and CCK-8 without cells).
Intracellular ROS MeasurementIntracellular ROS levels were measured using the ROS Assay Kit–Highly Sensitive DCFH-DA (R252, Dojindo) following the manufacturer’s protocol. Cells (1.4 × 105) were seeded in poly-L-lysine-coated 35-mm glass bottom dishes (IWAKI, #4971-040) and incubated for 84 h. After 30 min of preincubation, cells were irradiated with UVA as described above and further incubated for 12 h. Throughout the experiment, cells were cultured in HG (25 mM) medium with or without metformin (2, 4, or 8 mM). Cells were washed twice with Hank’s Balanced Salt Solution (HBSS), incubated with DCFH-DA dye solution for 30 min at 37°C in 5% CO2, and then washed again. Fluorescence images were captured using a fluorescence microscope (BZ-9000, KEYENCE, Osaka, Japan), and ROS intensity was quantified.
SA-β-Gal ActivitySA-β-Gal activity was evaluated using the SPiDER-βGal Cellular Senescence Plate Assay Kit (SG05, Dojindo) according to the manufacturer’s instructions. Cells (1 × 104) were seeded in black, clear-bottomed 96-well plates (Falcon #353219) and incubated for 84 h. After a 30-min preincubation and 15-min UVA irradiation as described above, cells were further incubated for 12 h. Throughout the experiment, cells were maintained in HG (25 mM) medium with or without metformin (2, 4, or 8 mM). Following removal of the supernatant, cells were lysed for 10 min and incubated with SPiDER-β-Gal working solution for 30 min. Fluorescence was measured at 535 nm (excitation) and 580 nm (emission) after the addition of stop solution using a microplate reader (Infinite® 100 PRO, Tecan Ltd., Switzerland).
Western Blot AnalysisCells were seeded in 35-mm dishes at a density of 5.4 × 105 cells and incubated for 84 h. After a 30-min preincubation, cells were irradiated with UVA for 15 min and further incubated for 12 h. Throughout the experiment, cells were maintained in HG (25 mM) medium with or without metformin (2, 4, 8, or 16 mM). Cells were lysed in lysis buffer, and protein concentrations were determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific (Waltham, MA, USA), #23225). Proteins were separated by sodium dodecyl sulfate- polyacrylamide gel electrophoresis (SDS-PAGE) (5–20%) and transferred to polyvinylidene difluoride (PVDF) membranes (Cytiva #10600086). Membranes were blocked with Blocking One (Nacalai Tesque, Kyoto, Japan) and then incubated overnight at 4°C with primary antibodies against AMPK (1 : 1000, Cell Signaling Technology (Danvers, MA, USA), #2535), SIRT1 (1 : 2000, Abcam (Cambridge, UK), ab110304), or β-actin (1 : 2000, Sigma-Aldrich (St. Louis, MO, USA), A5441). After washing, membranes were incubated with HRP-conjugated secondary antibodies (1 : 2000, Cell Signaling Technology, #7074, #7076) for 1 h. Signals were detected using Chemi-Lumi One Super (Nacalai Tesque) and visualized with a LAS-3000 image analyzer (FUJIFILM, Tokyo, Japan).
Statistical AnalysisData are expressed as mean ± standard deviation (S.D.). Two-group comparisons were performed using the Mann–Whitney U test. For multiple group comparisons, one-way ANOVA followed by Tukey’s post hoc test was conducted. A p-value of <0.05 was considered statistically significant.
We first examined cell viability following metformin treatment at concentrations ranging from 1 to 32 mM under both NG (5.6 mM) and HG (25 mM) conditions, in the absence of UVA irradiation. As shown in Figs. 1a and 1b, metformin had no significant effect on cell viability at concentrations of up to 16 mM under either NG or HG conditions. However, at 32 mM, metformin significantly reduced cell viability in both conditions (p < 0.01 vs. respective controls).
(a) Rat keratinocytes were cultured in NG (5.6 mM) medium with or without metformin (1, 2, 4, 8, 16, or 32 mM) for 12 h. (b) Cells were cultured in high glucose (HG; 25 mM) medium with or without metformin at the same concentrations. Cell viability was assessed using the CCK-8 assay. Data are presented as mean ± S.D. with circles indicating individual data points. n = 10 for NG and metformin at 1–8 mM; n = 12 for metformin at 16 and 32 mM. Statistical comparisons were performed between each metformin-treated group and the NG group in (a) or the HG group in (b). p-Values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.01 vs. NG in (a) and HG in (b). HG: high glucose; NG: normal glucose.
Under NG conditions, cell viability remained at or above 100% at both 12 h and 24 h, indicating no cytotoxic effect (Fig. 2a). By contrast, exposure to HG (25 mM) for 12 h followed by UVA irradiation (15 min) markedly reduced cell viability to approximately 42% compared with NG (12 h) (p = 0.00078). This reduction persisted at 24 h post-irradiation, with no significant difference between the 12 h and 24 h time points, suggesting that the cytotoxic effects of HG + UVA manifested within 12 h and subsequently plateaued.
(a) Time-dependent changes in cell viability under NG (5.6 mM, without UVA) or HG (25 mM, with 15 min UVA). (b) Effects of metformin (1, 2, 4, 8, 16, or 32 mM) on cell viability under HG + UVA conditions. Cell viability was measured at 12 h post-irradiation using the CCK-8 assay. UVA irradiation was performed using a dual-wavelength UV + LED light source, positioned 5 cm above the cells; the UV power density was approximately 12 mW/cm2 (see Materials and Methods). Data are presented as mean ± S.D., with circles indicating individual data points. n = 9 for NG (12 h), HG (12 h) + UVA, and HG (24 h) + UVA groups; n = 4 for NG (24h) group in (a); n = 10 for NG, n = 8 for NG + UVA, HG, HG + UVA, 2 mM, and 4 mM; n = 5 for 8 mM; and n = 4 for 1 mM, 16 mM, and 32 mM; n = 10 for NG (12 h) in (b). Statistical comparisons were performed between the NG and HG + UVA groups at each time point in (a), and between NG and NG + UVA, NG and HG, NG + UVA and HG + UVA, HG and HG + UVA, and HG + UVA vs. each metformin-treated group (1–32 mM) in (b). p-Values were determined using the Mann–Whitney U test in (a) and one-way ANOVA followed by Tukey’s post hoc test in (b). p = 0.00078 vs. NG at 12 h and p = 0.0066 vs. NG at 24 h in (a); p < 0.05 and p < 0.01 for indicated comparisons in (b). HG: high glucose; NG: normal glucose.
As shown in Fig. 2b, HG alone significantly reduced cell viability compared with NG (p < 0.01), indicating that hyperglycemia contributes to cellular damage. By contrast, UVA irradiation under NG conditions (NG + UVA) caused a non-significant reduction in viability, suggesting minimal cytotoxicity of UVA alone. Notably, the combination of HG and UVA (HG + UVA) produced a greater decline in viability than either HG or NG + UVA alone (p < 0.01 vs. HG; p < 0.01 vs. NG + UVA), reducing viability to approximately 33%. These results indicate that HG exerts a significant cytotoxic effect, UVA alone has minimal impact, and their combination elicits a synergistic cytotoxic response.
To evaluate the protective effects of metformin under these dual-stressor conditions, cells were treated with increasing concentrations of metformin (1–32 mM) under HG + UVA conditions and assessed for viability after 12 h (Fig. 2b). Metformin at concentrations of 2, 4, 8, and 16 mM significantly improved viability compared with the untreated HG + UVA group (p < 0.05, p < 0.05, p < 0.01, p < 0.01, respectively), with viability at 16 mM, which was lower than that at 8 mM. Metformin at 32 mM significantly decreased viability (p < 0.01). These results demonstrate that metformin provides a dose-dependent protective effect of up to 8 mM under combined HG and UVA conditions.
Effects of Metformin on p-AMPK in NG and HG ConditionsUnder NG conditions (Fig. 3b), p-AMPK levels were modestly increased by metformin at 2 and 4 mM compared with the NG group, although these increases were not statistically significant. Significant elevations were observed at 8 mM (p < 0.01) and 16 mM (p < 0.05); however, p-AMPK levels at 16 mM were lower than those at 8 mM.
(a) Representative Western blot images of p-AMPK (62 kDa) and β-actin (42 kDa). Top two rows: p-AMPK and corresponding β-actin in cells treated with 0, 2, or 4 mM metformin (0 mM corresponds to NG). Bottom two rows: p-AMPK and corresponding β-actin in cells treated with 0, 8, or 16 mM metformin (0 mM corresponds to NG). (b) Quantitative analysis of p-AMPK protein levels in cells with or without metformin (2, 4, 8, or 16 mM). Relative p-AMPK levels were measured by Western blotting, normalized to β-actin, and expressed as fold changes relative to the NG (0 mM) metformin group. Data are presented as mean ± S.D., with circles indicating individual data points. Sample sizes were n = 6 for NG; n = 3 for 2–16 mM in (a); n = 6 for NG and HG, n = 3 for 2–16 mM in (b). Statistical comparisons were performed between the NG group and each metformin-treated group (2–16 mM) under NG conditions, and between the HG group and each metformin-treated group (2–16 mM) under HG conditions. p-Values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.05 and p < 0.01 vs. the respective NG group, as indicated in the figure. HG: high glucose; NG: normal glucose.
Under HG conditions (Fig. 3b), p-AMPK levels in the HG group were comparable to those in the NG group. Treatment with metformin significantly increased p-AMPK levels at 4 mM (p < 0.05) and 8 mM (p < 0.01) compared with the HG group. As observed under NG conditions, p-AMPK levels at 16 mM were lower than at 8 mM. The highest p-AMPK levels were consistently observed at 8 mM metformin in both NG and HG conditions. Accordingly, 2, 4, and 8 mM were selected for subsequent experiments.
Effects of Metformin on p-AMPK under Combined HG and UVA ConditionsAs shown in Fig. 4b, p-AMPK levels increased under NG + UVA and HG conditions compared with the NG group, although these increases were not statistically significant. Similarly, p-AMPK levels in HG + UVA exceeded those in the HG control without reaching statistical significance. Notably, the HG + UVA conditions significantly elevated p-AMPK levels compared with the NG + UVA group (p < 0.05), suggesting a synergistic activation of p-AMPK in this model system.
(a) Representative Western blot images of p-AMPK (62 kDa) and β-actin (42 kDa). Top two rows: p-AMPK and corresponding β-actin in cells treated with NG, NG + UVA, HG, and HG + UVA. Bottom two rows: p-AMPK and corresponding β-actin in cells treated with NG and 2, 4, 8 mM metformin + HG + UVA. (b) Quantitative analysis of p-AMPK protein levels in the same groups as in (a). FRSK cells were exposed to HG (25 mM) and UVA irradiation (15 min), with or without metformin (2, 4, or 8 mM). UVA exposure was performed using the same UV + LED light source described in Fig. 2. Relative p-AMPK levels were measured by Western blotting, normalized to β-actin, and expressed as fold changes relative to the NG (0 mM metformin) group. Data are presented as mean ± S.D., with circles indicating individual data points. Sample sizes were n = 6 for NG: n = 3 for NG + UVA, HG, HG + UVA, and 2, 4, or 8 mM metformin + HG + UVA. Statistical comparisons were performed between NG and NG + UVA, NG and HG, NG + UVA and HG + UVA, HG and HG + UVA, and between HG + UVA and each metformin-treated group (2, 4, or 8 mM). p-values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.05 vs. NG + UVA, as indicated in the figure. FRSK: fetal rat skin keratinocytes; HG: high glucose; NG: normal glucose.
To examine the effects of metformin, HG + UVA-exposed cells were treated with 2, 4, or 8 mM metformin. None of these concentrations significantly altered p-AMPK levels compared with the untreated HG + UVA group, although 8 mM showed a slight upward trend, whereas 2 and 4 mM showed no apparent increase.
Effects of Metformin on Intracellular ROS Levels Under Combined HG and UVA ConditionsIntracellular ROS levels in FRSK cells were assessed by fluorescence microscopy and quantitative analysis using ROS-sensitive green fluorescence (Fig. 5a). Minimal fluorescence was observed in the NG, NG + UVA, and HG groups, indicating negligible ROS production under these conditions. By contrast, cells exposed to HG (25 mM) + UVA exhibited intense green fluorescence (Fig. 5a) and a marked increase in quantitative ROS intensity compared with the NG + UVA and HG groups (Fig. 5b) (p < 0.01 for both), indicating a synergistic increase in ROS levels under dual-stressor conditions.
(a) ROS fluorescence images of NG, NG + UVA, HG, HG + UVA, and HG + UVA + metformin at 2, 4, or 8 mM. (b) Quantitative analysis of ROS fluorescence intensity in the same groups as in (a). FRSK cells were exposed to HG (25 mM) and UVA irradiation (15 min), with or without metformin (2, 4, or 8 mM). UVA exposure was performed using the same UV + LED light source described in Fig. 2. ROS were detected 12 h after UVA exposure using DCFH-DA. Data are presented as mean ± S.D., with circles indicating individual data points. Sample sizes were n = 3 for NG, NG + UVA, HG, and HG + UV + metformin at 8 mM; n = 4 for HG + UV + metformin at 2 and 4 mM; n = 5 for HG + UVA. Statistical comparisons were performed between NG and NG + UVA, NG and HG, NG + UVA and HG + UVA, HG and HG + UVA, and between the HG + UVA and each metformin-treated group (2, 4, or 8 mM). p-Values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.01 vs. NG + UVA, HG, and HG + UVA, as indicated in the figure. FRSK: fetal rat skin keratinocytes; HG: high glucose; NG: normal glucose.
Metformin reduced ROS levels in a concentration-dependent manner. Treatment with 2 mM metformin produced a modest reduction in ROS, 4 mM resulted in a clearer but not statistically significant decrease, and 8 mM markedly decreased ROS levels (p < 0.01), nearly restoring fluorescence signal to baseline levels observed in NG, NG + UVA, and HG groups (Figs. 5a, 5b). These findings indicate that ROS accumulation is strongly induced only under combined HG and UVA stress, and that metformin effectively suppresses this oxidative stress, with substantial effects at 8 mM.
Effects of Metformin on SA-β-Gal Activity Under Combined HG and UVA ConditionsAs shown in Fig. 6, SA-β-Gal activity did not differ significantly between the NG and NG + UVA groups, indicating that UVA exposure alone does not induce cellular senescence under NG conditions. By contrast, SA-β-Gal activity was significantly elevated in the HG group compared with the NG group (p < 0.01), suggesting that HG alone induces a senescence response. Moreover, the combination of HG and UVA further increased SA-β-Gal activity compared with NG + UVA (p < 0.01) and HG alone (p < 0.05), indicating a synergistic effect of HG and UVA on cellular senescence.
SA-β-Gal activity was measured in FRSK cells using the SPiDER-B Gal Cellular Senescence Plate Assay Kit, after exposure to HG (25 mM) and UVA irradiation (15 min), with or without metformin (2, 4, or 8 mM). UVA irradiation was performed using the same UV + LED light source described in Fig. 2. Data are presented as mean ± S.D., with circles indicating individual data points. Sample sizes were n = 5 for NG and metformin at 8 mM; n = 4 for NG + UVA; n = 6 for metformin at 4 mM; n = 8 for HG, HG + UVA, and metformin at 2 mM. Statistical comparisons were performed between NG and NG + UVA, NG and HG, NG + UVA and HG + UVA, HG and HG + UVA, and between HG + UVA and each metformin-treated group (2, 4, or 8 mM). p-Values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.05 vs. HG and HG + UVA; p < 0.01 vs. NG, NG + UVA, and HG + UVA, as indicated in the figure. FRSK; fetal rat skin keratinocytes; HG: high glucose; NG: normal glucose.
Treatment with metformin reduced SA-β-Gal activity in a concentration-dependent manner under HG + UVA conditions. Specifically, 4 mM metformin significantly suppressed SA-β-Gal activity compared with the untreated HG + UVA group (p < 0.05). Notably, 8 mM metformin produced the most pronounced effect (p < 0.01), restoring SA-β-Gal activity to a level comparable with that of the NG + UVA group.
Effects of Metformin on SIRT1 Protein Expression under Combined HG and UVA ConditionsAs shown in Fig. 7b, SIRT1 protein expression under NG + UVA conditions was slightly higher than in the NG group; however, the difference was not statistically significant. HG conditions significantly increased SIRT1 expression compared with the NG group (p < 0.05). HG + UVA conditions further enhanced SIRT1 expression, showing significant increases compared with NG + UVA (p < 0.01) and HG (p < 0.05), suggesting a synergistic upregulation of SIRT1 in this model.
(a) Representative Western blot images of p-SIRT1 (110 kDa) and β-actin (42 kDa). Top two rows: p-SIRT1 and corresponding β-actin in cells treated with NG, NG + UVA, HG, and HG + UVA. Bottom two rows: p-SIRT1 and corresponding β-actin in cells treated with NG, HG + UVA at 2, 4, or 8 mM metformin. (b) Quantitative analysis of SIRT1 protein levels in the same groups as in (a). FRSK cells were exposed to HG (25 mM) and UVA irradiation (15 min), with or without metformin (2, 4, or 8 mM). UVA exposure was performed using the same UV + LED light source described in Fig. 2. Relative SIRT1 levels were measured by Western blotting, normalized to β-actin, and expressed as fold changes relative to the NG (0 mM metformin) group. Data are presented as mean ± S.D., with circles indicating individual data points. Sample sizes were n = 6 for NG; n = 3 for NG + UVA, HG, HG + UVA, and HG + UVA + metformin at 2, 4, or 8 mM. Statistical comparisons were performed between NG and NG + UVA, NG and HG, NG + UVA and HG + UVA, HG and HG + UVA, and between the HG + UVA group and each metformin-treated group (2, 4, or 8 mM). p-values were determined using one-way ANOVA followed by Tukey’s post hoc test. p < 0.05 vs. NG and HG; p < 0.01 vs. NG + UVA and HG + UVA, as indicated in the figure. FRSK: fetal rat skin keratinocytes; HG: high glucose; NG: normal glucose.
In cells exposed to HG + UVA conditions, treatment with metformin at 2, 4, or 8 mM significantly reduced SIRT1 expression compared with the untreated HG + UVA group (p < 0.01 for all). Among these, 8 mM metformin produced the greatest reduction, indicating that it may moderate stress-induced upregulation of SIRT1 and alleviate cellular stress.
This study was designed to investigate diabetes-associated skin vulnerability to UV-induced oxidative stress, reflecting the growing clinical relevance of photoaging in patients with chronic hyperglycemia.9) The rapid onset of cytotoxicity within 12 h after UVA exposure, with no further decline at 24 h (Fig. 2a), suggests that the combined insult triggers acute, irreversible damage rather than progressive injury over time. Combined HG and UVA conditions exerted a synergistic cytotoxic effect on FRSK cells, characterized by marked reductions in cell viability (Fig. 2b), pronounced ROS accumulation (Fig. 5b), elevated SA-β-Gal activity (Fig. 6), and increased SIRT1 protein expression (Fig. 7). HG alone reduced cell viability by 22% compared with NG, whereas UVA alone (NG + UVA) caused a non-significant decline of 14%. By contrast, HG + UVA exposure resulted in a 67% reduction in viability, indicating that combined metabolic and photo-oxidative stress is required to elicit the substantial cellular damage observed in this model. Notably, metformin treatment improved cell viability under HG + UVA conditions in a dose-dependent manner, with maximal protection observed at 8 mM (Fig. 2b). This finding suggests that the cytoprotective effect of metformin extends beyond glycemic control and may counteract dual-stress-induced damage through additional mechanisms. However, the effect was attenuated at 16 mM, and viability markedly decreased at 32 mM, indicating that its cytoprotective potential is confined to an optimal concentration range, beyond which excessive doses may exert off-target or cytotoxic effects that outweigh the benefits.
Metformin, a well-established pharmacological activator of AMPK,6,14) increased p-AMPK levels in a dose-dependent manner under both NG (Fig. 3b) and HG (Fig. 3b) conditions without UVA exposure, with a maximal response at 8 mM. At 16 mM, p-AMPK levels decreased compared with 8 mM despite preserved cell viability (Figs. 1a, 1b), indicating a nonlinear dose–response relationship. This decline at higher concentrations may reflect feedback regulation of AMPK signaling to prevent excessive activation. Notably, hyperglycemia did not suppress metformin-induced AMPK activation (Fig. 3b), indicating that metformin effectively activates AMPK even under HG conditions without UVA exposure.
Under combined HG and UVA dual stress (Fig. 4b), p-AMPK levels were higher than in either HG or NG + UVA alone, and this difference was statistically significant between NG + UVA and HG + UVA. This pattern suggests that each stressor individually induces a modest AMPK response, whereas their combination elicits a more pronounced activation. Previous studies report that AMPK is activated under energy-deprived conditions, whereas HG suppresses AMPK signaling via MG53-mediated degradation of AMPKα.15) In this study, HG alone elicited a modest, non-significant increase in p-AMPK, suggesting that the impact of hyperglycemia on AMPK varies with experimental conditions. The addition of UVA to HG further increased p-AMPK, indicating a synergistic effect of these two. In HG + UVA-treated cells, metformin at 8 mM produced only a slight upward trend in p-AMPK levels, indicating that its cytoprotective effect under dual-stressor conditions is unlikely to be primarily mediated by AMPK activation.
Oxidative stress is a central mediator of cellular injury, with chronic hyperglycemia enhancing ROS generation via mitochondrial and enzymatic pathways,16) and UVA inducing ROS primarily through photo-oxidative damage to DNA and mitochondria.17) In this dual-stress model (Fig. 5), the marked suppression of ROS by metformin may involve multiple mechanisms, including inhibition of mitochondrial complex I and modulation of endogenous antioxidant systems.18,19) Although the precise mechanism remains to be fully elucidated, the ability of metformin to reduce ROS to near-baseline levels appears to be a key determinant of its cytoprotective action under combined metabolic and photo-oxidative stress.
SA-β-Gal activity is a widely accepted biomarker of cellular senescence, reflecting irreversible cell-cycle arrest in aged or stress-exposed cells.10,20) In this study, HG alone significantly increased SA-β-Gal activity (Fig. 6), consistent with previous studies linking chronic hyperglycemia to senescence-like phenotypes through sustained oxidative and metabolic stress.21,22) This effect was further amplified by UVA exposure, indicating that photo-oxidative stress acts synergistically with hyperglycemia to accelerate senescence pathways. Metformin treatment reduced SA-β-Gal activity in a dose-dependent manner, nearly restoring it to baseline at 8 mM. This finding aligns with previous studies demonstrating that metformin attenuates oxidative stress-induced senescence in various cell types.23,24) Mechanistically, suppression of ROS by metformin may interrupt key stress-induced senescence signaling cascades, including the p16INK4a and p53–p21WAF1/CIP1 pathways,20) suggesting its potential role to delay stress-induced premature senescence.25)
SIRT1, a NAD+-dependent protein deacetylase, plays a central role in cellular stress responses, metabolic regulation, and aging processes.26−31) In our study, SIRT1 expression was markedly elevated under combined HG + UVA dual stress conditions compared with NG + UVA and HG alone (Fig. 7b), suggesting an adaptive upregulation in response to oxidative stress. However, despite this induction, ROS accumulation (Fig. 5) and SA-β-Gal activity (Fig. 6) remained elevated, indicating that increased SIRT1 expression alone was insufficient to prevent cellular injury under excessive stress. This paradox is consistent with previous evidence that oxidative stress can impair SIRT1 enzymatic activity through oxidative modification of cysteine residues and modulate its levels via microRNA-mediated mechanisms.32) Thus, in our model, oxidative stress may have simultaneously triggered compensatory SIRT1 upregulation and compromised its functional activity, contributing to persistent senescence phenotypes. Metformin treatment reduced SIRT1 expression in a concentration-dependent manner, which likely reflects normalization of stress-induced overexpression rather than suppression of its protective function. Consistent with this concept of restoring SIRT1 homeostasis, previous studies in endothelial cells have shown that hyperglycemia decreases SIRT1 protein levels and that metformin can attenuate this reduction while decreasing senescence markers.33) In this study, by lowering ROS levels and attenuating senescence markers, metformin may have reduced the need for compensatory SIRT1 induction, thereby restoring expression toward basal levels. This normalization effect is consistent with previous findings that metformin restores SIRT1 expression and modulates its downstream targets under hyperglycemic conditions.33)
The cytoprotective effects of metformin under combined HG and UVA-induced dual stress are likely mediated by both AMPK-dependent and AMPK-independent pathways, which are well known for regulating oxidative stress, inflammation, and cellular aging.32,34,35) AMPK activation suppresses the mammalian target of rapamycin (mTOR) pathway, thereby promoting autophagy and facilitating the clearance of damaged organelles and macromolecules.32,36,37) Additionally, AMPK enhances SIRT1 activity by increasing NAD+ availability,38) leading to deacetylation and inhibition of transcription factors such as p65/NF-kB, thereby attenuating pro-inflammatory cytokine production39,40) and the senescence-associated secretory phenotype (SASP).41) In parallel, metformin modulates mTORC1 and IKK/NF-kB through AMPK-independent mechanisms.42,43) In keratinocytes, p53 signaling plays a pivotal role in balancing DNA repair and apoptosis following UVA-induced DNA damage.44) Both AMPK-dependent and AMPK-independent actions of metformin have been demonstrated in vivo. For example, in a model of UVB-induced skin damage, metformin and the AMPK agonist AICAR promoted DNA repair in an AMPK-dependent manner while reducing epidermal hyperplasia via AMPK-independent inhibition of ERK signaling.45) Similarly, in a model of UVA-induced skin photoaging, metformin administration (10 mg/kg/d) attenuated skin roughness, epidermal thinning, and collagen degradation, accompanied by suppression of phosphatidylinositol 3-kinase (PI3K)/AKT/mTOR signaling, alleviation of mitochondrial oxidative stress, and decreased expression of mitophagy markers PINK1 (serine/threonine protein kinase) and Parkin (ubiquitin E3 ligase).46) These findings highlight multiple pathways involved in metformin’s protective effects. Taken together, these in vivo data support the physiological relevance of the pathways discussed to cutaneous photoaging; their role in diabetic photoaging is plausible given the coexistence of metabolic and environmental stressors, but warrants confirmation in dedicated models.
Overall, these findings suggest that although AMPK activation occurred under dual-stressor conditions, the magnitude and pattern of metformin’s cytoprotective effects were not directly proportional to p-AMPK levels (Fig. 4). This dissociation, along with the marked suppression of ROS (Fig. 5) and normalization of SIRT1 expression (Fig. 7), supports the view that metformin’s primary protective mechanism in this model may involve AMPK-independent antioxidant pathways. Although the precise mechanisms remain to be elucidated, previous studies have shown that metformin exerts AMPK-independent through direct inhibition of mitochondrial complex I, limiting mitochondrial ROS production and modulating cellular redox status.34) Furthermore, metformin influences metabolic signaling pathway such as mTORC1,42) whose hyperactivation has been associated with impaired autophagy, accumulation of damaged organelles, and promotion of cellular senescence.47,48) These AMPK-independent actions reported across diverse experimental contexts may contribute to the overall antioxidant profile of metformin in addition to its canonical AMPK-dependent effects. This perspective broadens the current understanding of metformin’s versatility in redox regulation, extending its role beyond canonical AMPK activation.
In conclusion, this study demonstrates that in FRSK cells, HG + UVA-induced damage is predominantly mediated by ROS, leading to both acute cell loss and premature senescence. Metformin primarily protects by suppressing ROS accumulation, accompanied by modulation of senescence markers and stress-responsive proteins. The absence of a strong correlation between AMPK activation and cytoprotection under dual-stressor conditions indicates that AMPK-independent antioxidant pathways contribute to metabolic-photo stress. These findings provide mechanistic insight into how metformin mitigates skin cell injury in hyperglycemic states compounded by UV exposure and underscores its potential for preventing photo-accelerated skin aging in diabetic conditions.
Conceptualization, T.W.; Investigation and data analysis, T.T. and R.S.; Manuscript drafting, T.T.; Review and editing, T.W. All authors have read and approved the final version of the manuscript.
The authors declare no conflict of interest.
