Multidimensional Proteomics Reveal Metformin's Impact on Interconnected Regulatory Networks of Protein Turnover, Ubiquitination, DNA Damage, and Cell Cycle

Abstract

Metformin, a first-line therapy for type 2 diabetes, has also been implicated in regulating diverse physiological and pathological processes, including lifespan extension, cancer, and other disease-related conditions. However, its mechanisms of action remain incompletely understood, with many effects still unexplained. In this study, we investigated the impact of metformin on the cellular ubiquitinome and associated protein turnover. Through an integrated analysis combining ubiquitinome profiling with pulsed metabolic labeling, we found that metformin markedly suppresses global protein ubiquitination, including various types of ubiquitin chain linkages, and concurrently inhibits both protein synthesis and degradation. Notably, metformin induces a marked reduction in the ubiquitination of histone H4, a modification closely associated with DNA damage repair. We further establish a mechanistic link whereby metformin regulates DNA damage repair and cell cycle progression through downregulating ubiquitination. Together, our findings demonstrate that metformin modulates ubiquitination and proteostasis, central processes that regulate numerous cellular functions. By identifying histone H4 ubiquitination as a key target, we elucidate a potential mechanism through which metformin influences DNA repair and cell cycle progression. This comprehensive dataset advances understanding of the drug’s multifaceted pharmacological activities and provides a valuable resource for future drug development.

Graphical Abstract

Highlights

• •Pulsed metabolic labeling reveals that metformin suppresses protein turnover.
• •Metformin inhibits global protein ubiquitination and all ubiquitin linkage types.
• •Metformin reduces histone H4-K92 ubiquitination, impairing DNA damage repair.
• •Metformin regulates the cell cycle via inhibition of ubiquitination.

In Brief

Wang et al. employed multidimensional proteomics, including quantitative proteome and ubiquitinome profiling as well as isotope metabolic labeling–based protein turnover analysis, to demonstrate that metformin and phenformin suppress intracellular protein turnover and ubiquitination, fundamental biological processes with broad cellular functions. They further established mechanistic linkages through which these drugs modulate DNA damage repair and cell cycle progression. These findings deepen our understanding of the pharmacological mechanisms of metformin and offer valuable insights for future drug development.

Metformin, a small-molecule compound of the biguanide class, is the most widely prescribed first-line therapy for type 2 diabetes, particularly in overweight or obese patients, with well-established glycemic control efficacy. Beyond its glucose-lowering properties, extensive studies have demonstrated that metformin exerts therapeutic benefits in a variety of conditions, including cancer (1), neurodegenerative diseases (2), and cardiovascular disorders (3). Notably, its use has expanded to include treatment of gestational diabetes (4, 5) and polycystic ovary syndrome (6, 7).

Despite its broad therapeutic potential, the mechanisms of metformin action remain incompletely understood. Metformin is known to lower blood glucose via multiple pathways. At high concentrations, it inhibits mitochondrial complex I of the electron transport chain in hepatocytes, reducing ATP production from oxidative phosphorylation and increasing the intracellular AMP/ATP ratio (8, 9). This shift activates the liver kinase B1 (LKB1)–AMP-activated protein kinase (AMPK) signaling cascade, which suppresses transcription of gluconeogenic genes, such as Pck1 and G6pc by AMPK-mediated phosphorylation and cytoplasmic sequestration of the transcriptional coactivator CRTC2 (10). Since hepatic gluconeogenesis is energetically demanding (requiring six ATP equivalents per glucose molecule), mitochondrial inhibition and subsequent ATP depletion provide a straightforward mechanism for metformin’s suppression of gluconeogenesis (11). Recent findings also reveal that at lower concentrations, metformin acts through the lysosomal PEN2 (presenilin enhancer 2)–AMPK axis to exert its hypoglycemic effects (12).

Given that metformin-mediated inhibition of complex I alters cellular energy status and activates AMPK, a key regulator of the mechanistic target of rapamycin (mTOR) signaling pathway that controls protein synthesis and cell proliferation (13, 14), we hypothesized that metformin might also regulate proteostasis, particularly through the mTOR-governed protein synthesis and ubiquitin (Ub)–proteasome system–mediated protein degradation.

Moreover, other biguanides, such as buformin and phenformin, show similar biological effects. Phenformin, for instance, possesses higher lipophilicity and mitochondrial membrane affinity, resulting in stronger inhibition of oxidative phosphorylation and more potent glycemic control (15). However, because of a greater risk of lactic acidosis, its clinical use has been largely replaced by metformin, which has a more favorable safety profile.

In this study, we systematically investigate the effects of metformin and phenformin on protein synthesis, degradation, and ubiquitination in JHH-7 cells. Our results demonstrate that both compounds broadly suppress proteome-wide protein turnover and ubiquitination. Notably, we identify histone ubiquitination as a key target, linking metformin to regulation of transcription (16), DNA damage repair (17), and cell cycle progression (18, 19)—potential mechanisms underpinning its anticancer properties. As a classical therapeutic with expanding indications, our study provides a valuable resource on how metformin modulates cellular protein ubiquitination and homeostasis, offering new mechanistic insights into its diverse pharmacological actions.

Experimental Procedures

Experimental Design and Statistical Rationale

In this study, we employed multiple proteomic approaches to systematically analyze protein abundance, turnover, and ubiquitination in response to drug treatment. Stable isotope labeling by amino acids in cell culture (SILAC)–based quantitative proteomics was used to profile global proteome changes, whereas pulsed SILAC (pSILAC) was applied to quantify newly synthesized proteins. To measure protein degradation, a reversed pulsed SILAC (rp-SILAC) strategy was designed. To minimize labeling-induced bias, swapped-labeling experiments were performed and included in the analysis. For ubiquitinome profiling, tryptic peptides were subjected to K-ε-GG peptide enrichment prior to LC–MS/MS analysis.

All experiments were performed with independent biological replicates, with the specific n values provided in the corresponding figure legends. Statistical analyses were conducted using two-tailed Student’s t tests or one-way ANOVA where appropriate, with p < 0.05 considered statistically significant. Data are presented as mean ± SD unless otherwise specified. Further details on experimental procedures and data analysis pipelines are provided below.

Cell Lines and Cell Culture

JHH-7 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (v/v) fetal bovine serum, 100 U/ml penicillin, and 100 mg/ml streptomycin. Cells were cultured at 37 °C in a humidified environment of 95% air and 5% CO₂.

Intracellular ATP Content Detection

When cells reached 70% confluence, the culture medium was replaced with DMEM containing metformin (1 μM, 10 μM, 100 μM, 1 mM, or 10 mM) or phenformin (100 μM or 1 mM). After 24 h incubation at 37 °C, cellular ATP levels were quantified using the CellTiter-Glo Luminescent Cell Viability Assay (Promega; catalog no.: G9242). Concurrently, total cellular protein content was measured. ATP data for each treatment group were then normalized to total protein content using the ratio [ATP]/[protein], representing relative ATP production per unit protein.

Western Blot Analysis

Cells were lysed in radioimmunoprecipitation assay buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS) supplemented with protease inhibitors. Total protein extracts from cells were separated by SDS-PAGE. For Western blot analysis, the following antibodies were used: AMPK (1:1000 dilution; AF1627; Beyotime), p-AMPK (1:1000 dilution; AF5908; Beyotime), Ub (1:1000 dilution; ENZ-ABS840-0100; ENZO), H2B (1:1000 dilution; ab1790; Abcam), H2B-K121ub (1:1000 dilution; ab1790; Cell Signaling Technology), H2A (1:1000 dilution; ab177308; Abcam), H2A-K119ub (1:1000 dilution; 8240S; Cell Signaling Technology), γ-H2A (p-Ser139) (1:1000 dilution; AF1210; Beyotime), P27 (1:1000 dilution; 25614-1-AP; Proteintech), P21 (1:1000 dilution; A1483; ABclonal), RBX1 (1:1000 dilution; ab221548; Abcam), β-actin (1:10,000 dilution; AC026; ABclonal), horseradish peroxidase–conjugated polyclonal goat anti-rabbit antibody (1:1000 dilution; A0208; Beyotime), horseradish peroxidase–conjugated goat polyclonal anti-mouse antibody (1:1000 dilution; A0216; Beyotime).

siRNA Transfection

The RBX1-siRNA (sequence: CUGUGCCAUCUGCAGGAACCACA) and control-siRNA (sequence: UUCUCCGAACGUGUCACGU) transfection were conducted with Lipofectamine RNAi MAX (Thermo Fisher Scientific, catalog no.: 13778075) reagent according to the manufacturer’s instructions.

Flow Cytometry

To assess the effects of metformin or phenformin on the cell cycle, propidium iodide from the cell cycle analysis kit (Beyotime, catalog no.: C1052) was used. Briefly, upon reaching 70% confluence, cells were treated with metformin or phenformin for 24 h. Subsequently, cells were trypsinized, washed with precooled PBS, and fixed in precooled 70% ethanol at 4 °C for 12 h. Following fixation, cells were incubated with propidium iodide staining solution at 37 °C for 30 min in the dark. Finally, red fluorescence was detected by flow cytometry (Guava easyCyte system) using a 488 nm excitation wavelength.

Protein Sample Preparation for LC–MS/MS

Cells were lysed in radioimmunoprecipitation assay buffer (50 mM Tris [pH 7.4], 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% SDS) supplemented with protease inhibitors using 60 s sonication. Lysates were centrifuged at 16,000g for 15 min at 4 °C. Supernatants were collected, and protein concentration was quantified using a bicinchoninic acid assay kit. Protein samples were then precipitated with ice-cold acetone, centrifuged at 16,000g for 15 min, and pellets dissolved in 8 M urea (100 mM Tris–HCl, pH 8.5).

For in-solution digestion, proteins were reduced with 5 mM Tris(2-carboxyethyl)phosphine hydrochloride at room temperature for 20 min, followed by alkylation with 10 mM iodoacetamide in the dark for 15 min. Urea concentration was diluted to 2 M, and 1 mM CaCl₂ was added prior to trypsin digestion at 1:100 (w/w) enzyme-to-protein ratio overnight at 37 °C. Digested peptides were desalted using C₁₈ tips for further LC–MS/MS analysis.

pSILAC and rp-SILAC

pSILAC

Cells were initially cultured in light-DMEM medium. Upon reaching 70% confluence, the medium was replaced with SILAC-DMEM supplemented with medium-labeled l-arginine (¹³C₆) and l-lysine (²H₄) or heavy-labeled l-arginine (¹⁵N₄, ¹³C₆) and l-lysine (¹⁵N₂, ¹³C₆). The experimental group was then switched to the heavy- or medium-labeled medium containing metformin (100 μM or 10 mM) or phenformin (100 μM or 1 mM). Conversely, the control group received the medium-labeled medium (medium or heavy, respectively) containing PBS (vehicle control). After 24 h of treatment, cells were harvested. Equal amounts of protein from the respective groups were combined, subjected to trypsin digestion, and the resulting peptides were desalted. Finally, the desalted peptides were completely dried for subsequent shotgun proteomic analysis by LC–MS/MS.

rp-SILAC

Cells were cultured in SILAC-DMEM (light/heavy) until reaching 70% confluence. After washing three times with PBS, heavy-labeled cells received fresh medium containing biguanides (10 mM metformin or 1 mM phenformin), whereas light-labeled controls received medium containing PBS. Following 24 h treatment, cells were harvested. Equal protein amounts from each group were combined, digested with trypsin, and fractionated by high-pH reversed-phase HPLC prior to LC–MS/MS analysis.

Ubiquitylated Peptide Enrichment

Cells cultured in SILAC medium were treated separately with metformin (100 μM or 10 mM), phenformin (100 μM or 1 mM), and the control group received an equivalent volume of PBS. After 24-h treatment, cells were harvested. Equal protein amounts (1 mg per condition) were combined and subjected to trypsin digestion. Desalted peptides were reconstituted in 400 μl IP buffer (50 mM Mops [pH 8.0], 10 mM Na₂HPO₄, and 50 mM NaCl). The K-GG Ub remnant motif antibody-conjugated beads (PTM Biolabs, catalog no.: PTM-1104) were washed three times with cold PBS and then transferred to a homemade pipette tip packed with C₈ filler as a plug. The peptide solution was loaded into the C₈ tip prepacked with K-GG Ub remnant motif antibody–conjugated beads and centrifuged through the beads by centrifugation at 50g. The peptide solution was loaded onto C₈ tips and passed through the beads by centrifugation at 50g. Beads were washed sequentially: three times with wash buffer (50 mM Mops [pH 8.0], 10 mM Na₂HPO₄, 50 mM NaCl, and 0.3% NP-40); three times with IP buffer, three times with H₂O (all washes performed by centrifugation at 200g). Peptides were eluted three times with 100 μl of 0.15% TFA via centrifugation. Eluates were pooled and completely dried for LC–MS/MS analysis.

High-pH Reverse Phase Liquid Chromatography Fractionation

High-pH reverse phase separation was performed using an Agilent 1260 HPLC system equipped with an XBridge BEH C₁₈ column (1.0 × 150 mm, 3.5 μm; Waters). Mobile phases consisted of (A) 20 mM ammonium formate (pH 10.0) and (B) 90% (v/v) acetonitrile (ACN), 20 mM ammonium formate (pH 10.0). Desalted peptides were reconstituted in mobile phase A and separated at 60 μl/min with a 96-min gradient. Eluate was collected at 1-min intervals into a 96-well plate, followed by noncontiguous pooling into 10 fractions. Pooled fractions were vacuum-dried prior to LC–MS/MS analysis.

LC–MS/MS Analysis of Peptides

Whole-Cell Lysate Peptides

Peptides from whole-cell lysates were analyzed using an online EASY-nLC 1000 system (Thermo Fisher Scientific) coupled to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific). Samples were loaded directly onto a 150 mm × 100 μm i.d. analytical column (packed in-house with 1.9 μm C₁₈ resin; Dr Maisch GmbH). The mobile phases consisted of (A) 0.1% formic acid (FA), 2% ACN; (B) 0.1% FA, 98% ACN. Peptides were separated using a 120-min gradient at a constant flow rate of 300 nl/min: 3% B (0 min), 8% B (5 min), 20% B (92 min), 30% B (113 min), 95% B (115 min), held at 95% B (120 min). Data were acquired in data-dependent acquisition mode (top-20). For MS1, the scan range was set to 350 to 1500 m/z at a resolution of 60,000. The automatic gain control (AGC) target was set as 3e6 with a maximum injection time of 20 ms. For MS2, the resolution was set to 30,000. The AGC target was set to 1e5, and the maximum injection time was set to 45 ms. The isolation window was set to 1.6 m/z. Precursor ions were fragmented using higher-energy collisional dissociation with 27% normalized collision energy.

Ubiquitinated Peptides

Enriched ubiquitinated peptides were analyzed using an online EASY-nLC 1200 system (Thermo Fisher Scientific) coupled to a Q Exactive HF-X mass spectrometer (Thermo Fisher Scientific). The chromatography analytical column used the same column and a constant flow rate of 300 nl/min. Mobile phases were (A) 0.1% FA; (B) 0.1% FA, 80% ACN. Separation used a 240-min gradient: 3% B (0 min), 11% B (2 min), 28% B (197 min), 40% B (222 min), 50% B (228 min), 100% B (230 min), held at 100% B (240 min). Data were acquired in data-dependent acquisition mode (top-20). For MS1, the scan range was set to 300 to 1800 m/z at a resolution of 60,000, and the AGC target was 1e6. The maximum injection time for the precursor ion was 20 ms. The resolution of MS2 was set to 30,000, and the AGC target was set to 2e4 with a maximum injection time of 120 ms. The isolation window was set to 1.6 m/z. Higher-energy collisional dissociation fragmentation at 27% normalized collision energy was used to fragment the precursor ions.

Mass Spectrometric Data Processing

SILAC quantification was performed using MaxQuant (version 1.6.0.1) (20) with “Match between runs” enabled. Mass spectrometry data were searched against the SwissProt human protein database, including 20,191 protein sequences. The precursor and fragment mass tolerances were set to 20 ppm. Trypsin/P was set as the enzyme, and the maximum missed cleavage was set to two and four for normal and Ub-enriched data, respectively. Metabolic labeling parameters included medium-heavy (Lys4, Arg6) and heavy (Lys8, Arg10) labels. The fixed modification was set as carbamidomethyl (C) (+57.02 Da), whereas variable modifications included methionine oxidation (+15.99 Da) and protein N-terminal acetylation (+42.01 Da). For Ub enrichment samples, lysine diGly modification (+114.04 Da) was additionally included. The false discovery rate was controlled at ≤1% using a decoy database. Protein quantification was based exclusively on the intensity of unique and razor peptides. Peptide elution peak areas were used to quantify peptide abundance. The resulting quantification values were then processed in Perseus (version.1.5.5.3, Max Planck Institute of Biochemistry) (21).

Functional Analysis

Gene Ontology analysis was performed using David bioinformatics resources tool (https://david.ncifcrf.gov/) (21), and gene set enrichment analysis (GSEA) was performed using GSEA (version 4.1.0, Broad Institute) (22).

Statistical Analysis

Statistical analyses were performed using GraphPad Prism 8 (GraphPad Software) or Microsoft Excel. Data are presented as mean ± SD from three independent experiments, unless otherwise specified. Statistical significance was determined by two-tailed Student’s t tests or one-way ANOVA where appropriate, with p < 0.05 considered statistically significant.

Results

Metformin Activates AMPK with Decreased ATP

Existing evidences have indicated that metformin can activate AMPK through various mechanisms (8, 9). The activated AMPK further inhibits mTOR, a master regulator governing protein synthesis. However, how cells maintain proteostasis, when protein synthesis is impaired by metformin treatment, is yet to be answered (Fig. 1A), and thus to be investigated in this study.

Fig. 1.

Metformin downregulates protein turnover.A, working hypothesis illustrating metformin’s role in regulating proteostasis to maintain the balance between protein synthesis and degradation. B, metformin reduces intracellular ATP levels in JHH-7 cells (n = 7, data presented as mean ± SD, one-way ANOVA). C, phenformin reduces intracellular ATP levels in JHH-7 cells (n = 7, data presented as mean ± SD, one-way ANOVA). D, treatment of JHH-7 cells with varying concentrations of metformin and phenformin for 24 h activated the AMPK pathway. E, schematic of the rp-SILAC workflow used to analyze protein degradation. Cells were cultured in L/H-SILAC medium, respectively, the medium was switched to M-SILAC medium when reaching 70% confluence and treated with equal amounts of PBS or biguanide drugs (10 mM metformin/1 mM phenformin) for 24 h. After treatment, the cells were harvested, and samples with equal protein amounts were pooled for sample processing. F, both 10 mM metformin and 1 mM phenformin inhibit global protein degradation in JHH-7 cells. Compared with the control group, the median offset values for the 10 mM metformin and 1 mM phenformin treatment groups were 0.3397 and 0.3497, respectively. Two independent biological replicates. AMPK, AMP-activated protein kinase; rp-SILAC, reversed pulsed stable isotope labeling by amino acids in cell culture.

First, we examined the impact of metformin or phenformin treatments on cellular ATP levels that were measured by using an ATP assay in JHH-7 cells. Following a 24-h treatment with varying drug concentrations, our findings indicated that a low concentration of metformin (<100 μM) had no significant effect on cellular ATP content. However, when cells were exposed to 1 mM and 10 mM metformin, there was a notable decrease in ATP synthesis (Fig. 1B). In contrast, a concentration of 100 μM phenformin was sufficient to significantly reduce cellular ATP synthesis capacity (Fig. 1C), aligning with its reported potent efficacy (23, 24). This suggests that the cellular ATP regulation exhibits greater sensitivity to phenformin compared with metformin. Moreover, we confirmed the activation of the AMPK pathway by both metformin and phenformin, evidenced by the phosphorylation of the catalytic α subunit of AMPK (phosphorylation of AMPKα1 [Thr183] or AMPKα2 [Thr172]) (Fig. 1D).

Metformin Downregulates Protein Turnover

First, quantitative proteomics revealed that treatment with metformin or phenformin only led to minimal changes on the entire cellular proteome. We were able to quantitatively analyze 6135 proteins, and even with high concentrations of metformin or phenformin, only dozens of proteins significantly changed (Supplemental Fig. S1, A and B, Supplemental Table S1). All sample groups exhibit good biological consistency and correlation (Supplemental Fig. S1, C and D). Next, we employed the p-SILAC method (25, 26) to assess cellular protein synthesis in response to metformin and phenformin treatment. Through the labeling of proteins newly synthesized during the designated 24-h period with heavy isotope–labeled amino acids (lysine and arginine) in the culture medium. Cells treated with 100 μM and 10 mM metformin were identified with 1934 and 2024 newly synthesized proteins by single shotgun proteomics analysis, respectively (Supplemental Table S2). We observed that 100 μM metformin had marginal impacts on protein synthesis. In contrast, cells treated with 10 mM metformin exhibited a pronounced inhibition on protein synthesis (Supplemental Fig. S1E). Similarly, cells treated with 100 μM and 1 mM phenformin were identified with 1902 and 1921 newly synthesized proteins, respectively. Phenformin at concentrations of 100 μM and 1 mM also significantly decreased protein synthesis, with the inhibitory effect of 100 μM phenformin being less potent than that of 1 mM phenformin (Supplemental Fig. S1F). The Gene Ontology analysis of these synthesis-downregulated proteins (fold change <0.67 and p < 0.05) indicated their association with protein synthesis (Supplemental Fig. S1, G and H), suggesting that metformin could impair the hemostasis of protein translation machinery, which is consistent with earlier reports (27, 28).

The results of the aforementioned quantitative mass spectrometry data show that metformin can significantly inhibit the protein translation process, but the overall protein level in the cell can remain relatively stable, suggesting that metformin may also affect protein degradation (Fig. 1A). Furthermore, to explore into protein degradation, the opposite pathway to protein synthesis in protein turnover, following drug treatment, we utilized rp-SILAC technology (29) to assess protein degradation (Fig. 1E). In our rp-SILAC data, 7260 and 7270 isotope labeled proteins were identified in cells treated with metformin and phenformin, respectively (Supplemental Table S3). Interestingly, the results indicate that protein degradation is suppressed in cells treated with these two biguanides (Fig. 1F). Overall, our findings support the dual inhibition of protein synthesis and degradation to maintain proteostasis during metformin treatment.

Metformin Decreases the Cellular Protein Ubiquitination and All Types of Ub Linkages

Because protein degradation is tightly regulated by ubiquitination and metformin has been shown to suppress protein degradation, we sought to investigate whether metformin influences the cellular ubiquitination. First, we used antibodies recognizing lysine ubiquitination, the results confirmed that 100 μM metformin did not inhibit cellular ubiquitination greatly, whereas 10 mM metformin and 1 mM phenformin significantly suppressed the global level of protein ubiquitination (Fig. 2A). To detail the impact of metformin treatment on cellular ubiquitination, we conducted a ubiquitinome analysis utilizing a combined approach involving quantitative triple-SILAC and the anti-KGG peptide enrichment method (Fig. 2B). The ubiquitinome analysis identified 5887 ubiquitination sites on 2067 proteins (Supplemental Table S4). The data showed that a low concentration of 100 μM metformin did not induce significant changes in cellular protein ubiquitination, with only 64 ubiquitination sites upregulated in cells treated with 100 μM metformin whereas 12 downregulated (Fig. 2C). In contrast, cells treated with 10 mM metformin exhibited a significant downregulation of 788 ubiquitination sites (Fig. 2, C and D). Similarly, phenformin also demonstrated a remarkable reduction in ubiquitination levels (Fig. 2, E and F). Furthermore, Spearman's correlation analysis revealed a distinct pattern among different treatments, revealing a higher correlation between high concentrations of metformin and phenformin compared with other comparisons (R² = 0.698) (Fig. 2, G and H). Although the aforementioned results suggest that high concentrations of metformin can result in a widespread decrease in the ubiquitination of many proteins, certain ubiquitination sites exhibit consistent downregulation across all treatment conditions. Notably, this includes the histone H4-K92 site, EIF3A-K420 site, and GIPC1-K80 site (fold change <0.67 and p < 0.05) (Supplemental Fig. S2A).

Fig. 2.

Metformin downregulates cellular protein ubiquitination.A, metformin and phenformin reduce the overall level of protein ubiquitination in JHH-7 cells. The ubiquitination levels in cell lysate samples were detected. B, workflow of quantitative ubiquitinome analysis using a triple-SILAC strategy. Cells were cultured in L/M/H-SILAC medium, respectively, followed by treatment with equal amounts of PBS, metformin (100 μM or 10 mM), or phenformin (100 μM or 1 mM) for 24 h. After harvesting, lysates containing equal protein amounts were pooled at a 1:1:1 ratio for sample processing, enrichment of ubiquitinated peptides, and subsequent quantitative analysis. The three treatment conditions were rotated across the three differently labeled cell populations. C and D, quantification of ubiquitinome changes upon metformin treatment in JHH-7 cells. The screening criteria for the volcano plot were as follows: |log₂(fold change)| ≥0.585 and p < 0.05, six independent biological replicates, two-tailed Student’s t tests. Compared with the control group, the median offset values for the 100 μM and 10 mM metformin treatment groups were 0.0648 and −0.6464, respectively. E and F, quantification of ubiquitinome changes upon phenformin treatment in JHH-7 cells. The screening criteria and biological replicate number are same as for metformin. Compared with the control group, the median offset values for the 100 μM and 1 mM phenformin treatment groups were −0.2522 and −0.4143, respectively. G, correlation analysis of ubiquitinome data under different metformin and phenformin treatment conditions. H, correlation analysis of ubiquitinome data between 10 mM metformin and 1 mM phenformin treatment conditions. I, effects of metformin and phenformin on different ubiquitin linkage types. The abundance changes of different ubiquitinated lysine sites on ubiquitin proteins were quantified (n = 4–6, data presented as mean ± SD). SILAC, stable isotope labeling by amino acids in cell culture.

Next, we looked into the details on which Ub linkages, playing pivotal roles in different functions of ubiquitination, were affected by metformin. By examining the linkage-specific signature peptides, we showed that the treatment with metformin or phenformin could decrease almost all types of Ub linkage (Fig. 2I). Notably, the significant decrease in K48-linked Ub chains, a key signal in triggering protein degradation, indicates that metformin and phenformin can inhibit protein degradation associated with K48 ubiquitination (Fig. 2I). Moreover, metformin was shown to inhibit all types of ubiquitination linkage, suggesting metformin not only affects protein degradation but also influences many other cellular processes associated with ubiquitination. Since protein ubiquitination is an energy-consuming process, and metformin's impact on decreasing ATP levels likely contributes to its inhibitory effect on this process.

Subsequently, Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis of the downregulated ubiquitinated proteins, which were induced by metformin or phenformin, indicated a predominant association with protein synthesis (Supplemental Fig. S2, B and C), aligning with the conclusions from previous pSILAC experiments. In addition, both the synthesis and degradation of ribosome-related proteins were noticeably suppressed by metformin (Supplemental Fig. S2, D and E). Notably, several translation initiation proteins, including EIF3A, exhibited a significant reduction in ubiquitination levels following metformin or phenformin treatment, despite subtle changes in protein abundance and synthesis (Supplemental Fig. S2F). Collectively, our multiomics data strongly support the notion that metformin can influence the machineries maintaining the proteostasis.

Metformin-Inhibited Protein Ubiquitination is Associated with the DNA Damage Repair

We found that the histone H4-K92 site is one of the three most significantly changed ubiquitination sites under all treatment conditions (Supplemental Fig. S2A). Notably, histone H4-K92 ubiquitination has been implicated in modulating the DNA damage response (30). Given their roles of histone ubiquitination in genome stability and gene expression (31, 32), we further examined changes in other ubiquitination sites of histones. In addition to the histone H4-K92 site, we also identified ubiquitination at H3-K80, H3-K57, and H4-K32 sites on histones H3 and H4. In contrast to the histone H4-K92 site, ubiquitination at other histone protein sites significantly decreased under the treatment of metformin or phenformin, with the exception of the low concentration of 100 μM metformin treatment (Fig. 3A). Moreover, we investigated the ubiquitination of histones H2A and H2B. Our data revealed notable alterations in ubiquitination levels at the K47, K109, K117, and K121 sites of histone H2B in cells treated with biguanide drugs. However, the impact of a low concentration of metformin treatment was mild (Fig. 3B). In addition, the ubiquitination of histone H2A-K119 showed no significant change after treatment with metformin or phenformin (Fig. 3B). Furthermore, we validated that metformin and phenformin inhibit ubiquitination at the H2B-K121 site, with no effect on ubiquitination at the H2A-K119 site using Western blots (Fig. 3C). Moreover, studies have shown that histone ubiquitination is closely related to DNA damage response, such as histone H3 and H4 ubiquitylation, facilitates cellular response to DNA damage (17, 30), and H2B monoubiquitylation facilitates DNA double-strand break repair (33). Therefore, we assessed the marker of DNA damage, γ-H2A, formed by the phosphorylation of histone H2A at serine 139, and observed that the use of metformin indeed upregulated cellular γ-H2A (Fig. 3C), indicating that the high concentration of metformin induces DNA damage in cells.

Fig. 3.

Metformin-regulated protein ubiquitination events are associated with DNA damage repair.A and B, ubiquitination changes on histone H3, H4, H2A, and H2B proteins regulated by metformin. Ubiquitination sites include histone H3-K57/H80, H4-K32/K92, H2A-K119, and H2B-K47/109/117/121. C, metformin and phenformin reduce H2B ubiquitination levels and induce DNA damage. D, GSEA analysis reveals enriched networks among ubiquitinated proteins. GSEA was performed for the differentially ubiquitinated proteins treated with 10 mM metformin. E, metformin-induced ubiquitinome alterations are associated with DNA damage response, repair pathways, and protein ubiquitination. The specific GSEA terms are as follows: “protein monoubiquitination,” “DNA repair,” and “Cellular response to DNA damage stimulus.” GSEA, gene set enrichment analysis.

Subsequently, GSEA was performed for the differentially ubiquitinated proteins treated with 10 mM metformin. The result unveiled that the ubiquitination events decreased by metformin were enriched in signaling pathways related to DNA damage, DNA repair, and modification of small proteins (mainly ubiquitination) (Fig. 3D); the detailed Gene Ontology biological process term by GSEA analysis is shown in Supplemental Table S5. Representative GSEA pathway results were also significant, such as “protein monoubiquitination,” “DNA repair,” and “cellular response to DNA damage stimulus” (Fig. 3E). These results were consistent with the conclusion that metformin induced cellular γ-H2A upregulation (Fig. 3C). This implies that metformin might regulate the cellular DNA damage response by modulating the ubiquitination of relevant proteins.

Metformin Affects DNA Replication and Cell Cycle

Previous GSEA results showed that differentially ubiquitinated proteins were significantly associated with the “DNA replication damage” pathway (Fig. 3D); it was subsequently found that differentially ubiquitinated proteins could indeed significantly enrich the “DNA replication” pathway (Fig. 4A). In addition, GSEA analysis of proteomic data treated with 10 mM metformin also found significant enrichment in the “DNA replication” pathway (Fig. 4B). Subsequently, we compared the proteins associated with the “DNA replication” pathway between proteomic and ubiquitinome under 10 mM metformin treatment. The result showed that the alterations in ubiquitination levels were more pronounced (Fig. 4C); it may suggest that metformin may regulate DNA replication or cell cycle by inhibiting the level of protein ubiquitination.

Fig. 4.

Metformin affects DNA replication and cell cycle regulation.A, metformin-induced changes in ubiquitinated proteins are enriched in pathways related to DNA replication. B, metformin-induced changes in the proteome are enriched in pathways related to DNA replication. C, comparison of proteins involved in the “DNA replication” pathway identified by proteomics and ubiquitinome analyses under metformin treatment. D and E, effects of metformin on cell cycle distribution analyzed by flow cytometry (n = 3, data presented as mean ± SD, one-way ANOVA). F, effects of metformin on cell cycle checkpoint proteins P21, P27, and the E3 ligase RBX1. G and H, effects of metformin on cell cycle distribution after RBX1 knockdown by siRNA (n = 3, data presented as mean ± SD, one-way ANOVA). I, metformin-induced expression changes of P21 and P27 are abolished upon RBX1 knockdown.

To validate whether metformin influences the cell cycle, we treated JHH-7 cells with different concentrations of metformin, performed flow cytometry using PI, and found that cells treated with metformin were arrested in the G1 phase (Fig. 4, D and E), consistent with earlier reports (18, 34). Subsequently, we conducted Western blot analysis of cell cycle–related proteins P21 (cyclin-dependent kinase inhibitor 1) and P27 (cyclin-dependent kinase inhibitor 1B) and observed that metformin increased the levels of P21 and P27 in cells (Fig. 4F). Simultaneously, we knocked down the Ub E3 ligase RBX1, closely related to the cell cycle (35). We found that metformin's cell cycle arrest at the G1 phase was abolished (Fig. 4, G and H). In addition, the levels of P21 and P27 were upregulated and were no longer influenced by metformin (Fig. 4I). This indicates that metformin can affect the cell cycle by inhibiting cellular ubiquitination.

Discussion

Proteins are central to virtually all cellular functions, and the maintenance of protein homeostasis is critical for normal cellular physiology. Proteostasis encompasses the intricate network of processes that regulate protein synthesis, folding, trafficking, and degradation, ensuring that proteins achieve and maintain their correct conformations and functional states. Disruption of this delicate balance can lead to the accumulation of misfolded or damaged proteins, which has been implicated in a wide range of pathological conditions, including age-related neurodegenerative disorders, metabolic diseases, and cancer (36, 37, 38). Therefore, identifying small molecules capable of modulating these pathways to preserve protein homeostasis in disease contexts is of great importance.

In this study, we used multiproteomics data to demonstrate how metformin affects and regulates the overall protein homeostasis of cells. First, we confirmed that metformin can inhibit the protein synthesis process through pSILAC, which is consistent with early reports that metformin can inhibit the mTOR signaling pathway and thereby inhibit the protein synthesis process in cancer cells. However, metformin treatment did not cause significant changes in the overall protein level. Therefore, through rp-SILAC experiments, we found that metformin also had a significant inhibitory effect on the overall protein degradation process of cells. The most significant function of metformin itself is to inhibit mitochondrial function and thereby reduce cellular ATP levels. Decreased ATP levels will lead to the inhibition of many energy-consuming pathways in cells, including the ubiquitination process. After linking metformin to reduce ATP levels and inhibit the protein degradation process, we speculate that metformin may inhibit protein degradation by inhibiting the protein ubiquitination process, and finally achieve protein homeostasis. We then conducted in-depth analysis and verification of the ubiquitination proteomics affected by metformin. Whether it was Western blot analysis or ubiquitinome analysis, we found that metformin significantly inhibited all different forms of ubiquitination chains; and these protein functions whose ubiquitination is inhibited are related to protein synthesis and degradation pathways. In the ubiquitinome data, we found that metformin can significantly inhibit the ubiquitination level of histone H4-K92 site, which is related to DNA damage repair, and we verified metformin can inhibit γ-H2A levels, which this may be one of the reasons why metformin exerts its tumor suppressor function. In addition, we found that metformin is also involved in the cell cycle and replication process and can arrest cells in the G1 phase. When we knock down RBX1, a key E3 ligase in the cell cycle, the cell cycle is no longer affected by metformin, which shows that metformin can regulate the cell cycle by inhibiting protein ubiquitination levels.

Given the pivotal role of Ub -mediated protein turnover in diverse cellular processes, alterations in this pathway may underlie many of the multifaceted effects of metformin reported in various physiological and pathological contexts (3, 39). Beyond the DNA damage response and cell cycle regulation highlighted here, it is conceivable that metformin may also influence other proteostasis-dependent processes, such as stress response, metabolism, and signaling adaptation. Future studies employing additional cellular or animal models will be valuable to delineate the full spectrum of metformin’s regulatory network.

Collectively, our study demonstrates that metformin regulates protein homeostasis and the ubiquitination process—two fundamental cellular mechanisms that orchestrate diverse biological functions and are deeply implicated in human diseases. Our comprehensive proteomic data offer mechanistic insights into how metformin modulates the cell cycle and exerts its tumor-suppressive effects. Importantly, these regulatory pathways are likely to intersect with broader cellular signaling networks, suggesting a wider impact of metformin on human health and disease beyond its known applications. By uncovering these interconnected mechanisms, our work not only deepens the understanding of metformin’s mode of action but also lays a conceptual foundation for the development of novel therapeutic strategies targeting proteostasis and ubiquitination pathways.

Data Availability

The MS proteomics data have been deposited in the ProteomeXchange Consortium via the iProX partner (40, 41) repository with the dataset identifier PXD065302 (ProteomeXchange) and IPX0012314000 (iProX).

Supplemental Data

This article contain supplemental data.

Conflict of Interest

The authors declare no competing interests.

Supplemental Data

Supplemental Table S1

.

Supplemental Figure

Supplemental fig. 1: Quantitative proteomics and protein synthesis analyses of JHH-7 cells upon metformin treatment.A and B, quantitative proteomic profiling of JHH-7 cells treated with metformin (100 μM and 10 mM) or phenformin (100 μM and 1 mM) for 24 h. Cut off: |log₂(fold change)| ≥ 0.585 and p-value < 0.05, three independent biological replicates, two-tailed Student’s t-tests. C and D, correlation analysis of inter-group replicates between the metformin-treated and phenformin-treated groups, with numbers representing pearson correlation coefficient values. E, p-SILAC analysis showing inhibition of protein synthesis in JHH-7 cells upon metformin treatment. Compared to the control group, the median offset values for the 100 μM and 10 mM metformin treatment groups were −0.05872 and −0.3719, respectively. four independent biological replicates, two-tailed Student’s t-tests. F, p-SILAC analysis showing inhibition of protein synthesis in JHH-7 cells upon phenformin treatment. Compared to the control group, the median offset values for the 100 μM and 1 mM phenformin treatment groups were -0.2590 and -0.4506, respectively. Cells were cultured in L-SILAC medium, and the medium was switched to M/L-SILAC medium when reaching 70% confluence. The cells were then treated with metformin (100 μM and 10 mM) or phenformin (100 μM and 1 mM) for 24 h. After treatment, the cells were harvested, and samples with equal protein amounts were pooled for sample processing, data acquisition and analysis. G and H, GO pathway enrichment analysis of proteins with inhibited synthesis after 10 mM metformin and 1 mM phenformin treatment, respectively.

Supplemental fig. 2: Metformin downregulates cellular ubiquitination levels and protein degradation.A, venn diagram comparing downregulated ubiquitination sites under various metformin and phenformin treatment conditions. Downregulated ubiquitination sites: log₂(fold change) ≤ −0.585 and p-value < 0.05. B and C, KEGG pathway analysis of proteins with decreased ubiquitination levels after 10 mM metformin and 1 mM phenformin treatment in JHH-7 cells. D, quantitative changes in ribosomal protein expression levels upon metformin and phenformin treatment. This dataset was derived from the whole cell lysate proteomics, wherein changes in the abundance of these ribosomal proteins are presented relative to the Ctrl group. E, quantitative changes in ribosomal protein degradation following metformin and phenformin treatment. This dataset was derived from cellular degradomics data (rp-SILAC dataset), wherein the changes in degradation of these ribosomal proteins are presented relative to the Ctrl group. F, quantitative changes of EIF3A protein across different omics datasets. The SILAC data represent changes in intracellular EIF3A protein levels; the p-SILAC data reflect alterations in the levels of newly synthesized EIF3A protein; while the ubiquitination data indicate changes in ubiquitinated EIF3A protein levels, which are closely associated with the degradation of EIF3A protein. (n = 3–5, data presented as mean ± SD).