HAGE diet exacerbated sarcopenia incidence in mice

Several clinical studies have statistically shown a positive correlation between a HAGE diet or AGE accumulation in the body and the onset of sarcopenia [31,32,33,34]. To investigate the relationship and specific molecular mechanisms between a HAGE diet and the development of sarcopenia, we fed 8-week-old C57BL/6 mice with either a LAGE diet or a HAGE diet, dividing them into two groups. After 24 weeks, we assessed whether sarcopenia had developed in the mice (Fig. 1A). We monitored non-fasting blood glucose and HbA1c levels weekly over the 24-week period. The results showed that non-fasting blood glucose and HbA1c levels were significantly higher in the HAGE group compared to the LAGE group (Fig. 1B and C). Additionally, we measured the levels of free and protein-bound AGEs in the serum and skeletal muscle tissues. The results indicated that levels of free and protein-bound CML and CEL, as well as free MG-H1, were elevated in the HAGE group compared to the LAGE group (Fig. 1D and I and S1). However, the levels of protein-bound MG-H1 in serum and muscle tissues from HAGE mice were similar to those in LAGE mice (Fig. 1I and S1).

Fig. 1

Feeding a HAGE diet increased the incidence of sarcopenia in mice (A) Experimental design. (B) Non-fasting blood glucose levels in mice fed LAGE and HAGE diets (n = 8 per group) at different time points. (C) HbA1c levels in mice fed LAGE and HAGE diets (n = 8 per group) at different time points. (D-I) Concentrations of free and protein-bound (pb) CML, CEL, and MG-H1 in plasma from mice fed LAGE and HAGE diets (n = 8 per group). (D) Free CML in plasma. (E) Free CEL in plasma. (F) Free MG-H1 in plasma. (G) pb-CML in plasma. (H) pb-CEL in plasma. (I) pb-MG-H1 in plasma. (J) Grip strength. (K) Latency to fall as assessed by the rotarod test. (L) Muscle-to-body weight ratio. (M) Representative images of skeletal muscle stained with H&E. Bars = 100 μm. (N) Quantified CSA of muscle fibers (µm2) (n = 8 per group). ns: no significant difference; *P < 0.05; **P < 0.01; *** P < 0.001

To evaluate the incidence of sarcopenia, we measured several relevant parameters. The HAGE group exhibited poorer motor function, as evidenced by reduced grip strength (HAGE: 48.3 ± 5.2 gm × Force, LAGE: 71.3 ± 8.5 gm × Force, P < 0.01) (Fig. 1J) and shorter latency to fall in rotarod tests (HAGE: 10.3 ± 1.7 s, LAGE: 22.5 ± 2.3 s, P < 0.01) (Fig. 1K). Furthermore, the skeletal muscle mass index was significantly lower in the HAGE group compared to the LAGE group (HAGE: 0.48 ± 0.05%, LAGE: 1.16 ± 0.12%, P < 0.001) (Fig. 1L). Histological analysis of skeletal muscle tissues revealed significant muscle damage in the HAGE group, characterized by a reduction in muscle fiber size and an increase in connective tissue (Fig. 1M). Quantification of CSA of muscle fibers revealed a significant reduction in the HAGE group (Fig. 1N). These physiological and pathological indicators suggested that feeding mice a HAGE diet induced the development of sarcopenia.

PRMT1 and key molecules of the intrinsic apoptotic pathway were significantly elevated in HAGE-induced sarcopenic mice

To identify the key molecules involved in HAGE-induced sarcopenia, we isolated skeletal muscle from mice, extracted the proteins, digested them with trypsin, and labeled them using the iTRAQ method. Mass spectrometry analysis was then performed to identify differentially expressed proteins (Fig. 2A). In skeletal muscle from HAGE-fed mice, we identified 317 DEPs (Table S6), including 198 upregulated and 119 downregulated proteins (Table S6). Figure 2B showed the top 15 most significantly upregulated and downregulated proteins. Among the upregulated proteins, PRMT1 showed the most significant increase, along with a notable upregulation of the receptor for AGEs, RAGE. Additionally, key proteins in the intrinsic apoptotic pathway, such as Bax, Apaf1 (Apoptotic protease activating factor-1), Caspase 3 (Casp3), Caspase 7 (Casp7), and Caspase 9 (Casp9), were significantly elevated (Fig. 2B). Other proteins with marked upregulation included CRTC3, FTO (Fat mass and obesity-associated), MMP2/7/9 (Matrix metalloproteinase 2, 7, and 9), and ADAMTS2/6 (A disintegrin and metalloproteinase with thrombospondin motifs 2 and 6) (Fig. 2B). The five most significantly downregulated proteins were USP2 (Ubiquitin specific peptidase 2), YTHDF3 (YTH N6-methyladenosine RNA binding protein 3), LMOD1 (Leiomodin 1), SLPI (Secretory leukocyte peptidase inhibitor), and LBP (Lipopolysaccharide binding protein) (Fig. 2B). GO analysis of the DEPs revealed that the upregulated proteins were primarily involved in biological processes such as apoptotic signaling pathways, extracellular matrix disassembly, organic acid catabolic processes, cellular response to oxidative stress, glucose transport, and DNA damage response (Fig. 2C). In contrast, the downregulated proteins were mainly associated with muscle tissue development, protein catabolic processes, energy metabolism, regulation of gene expression, and oxidative phosphorylation (Fig. S2).

Fig. 2

PRMT1, CRTC3, and apoptotic molecules were upregulated in sarcopenic muscle tissues. (A) Experimental design for iTRAQ analysis. (B) The top 15 upregulated and top 15 downregulated proteins in skeletal muscle tissues from the LAGE and HAGE groups of mice. (C) Gene Ontology (GO) analysis of biological processes associated with the upregulated proteins. (D) Validation of protein levels for PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, Casp9, USP2, and YTHDF3 in skeletal muscle tissues from the LAGE and HAGE groups of mice (n = 3 per group) using Western blotting

To validate the accuracy of the mass spectrometry results, we examined the expression of several representative proteins, including PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, Casp9, USP2, and YTHDF3, in skeletal muscle tissues from both LAGE and HAGE mice (n = 3 per group). The results confirmed that the expression patterns of all tested proteins were consistent with the iTRAQ quantification data (Fig. 2D).

PRMT1 and key molecules of the intrinsic apoptotic pathway were significantly elevated in AGE-BSA-treated myoblast cells

To determine whether the altered protein expression observed in Fig. 2D was directly regulated by AGEs, we treated C2C12 cells in vitro with different concentrations of AGE-BSA (0, 25, 50, and 100 µM) for 12 h and then assessed the expression levels of PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, Casp9, USP2, and YTHDF3. The results showed that the expression of PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 increased in a dose-dependent manner with higher concentrations of AGE-BSA, whereas the expression of USP2 and YTHDF3 gradually decreased with increasing AGE-BSA doses (Fig. S3). These in vitro findings suggested that PRMT1 and key molecules of the intrinsic apoptotic pathway were directly regulated by AGEs.

Inhibition of PRMT1 in HAGE-fed mice prevented sarcopenia incidence

Given the critical role of PRMTs in the development of various diseases [12,13,14] and the significant upregulation of PRMT1 in the muscle tissues of HAGE-induced sarcopenic mice, we aimed to explore the involvement of PRMT1 in the development of sarcopenia. To investigate whether inhibiting PRMT1 could prevent sarcopenia, we selected TCE5003, a specific PRMT1 inhibitor, and MS023, a type I PRMT inhibitor. In vitro studies confirmed that both inhibitors selectively reduced PRMT1 activity without altering its protein levels (Fig. S4A-S4E).

For in vivo analysis, mice were divided into four groups: LAGE diet, HAGE diet, HAGE diet with weekly TCE5003 injections, and HAGE diet with weekly MS023 injections, over a 24-week period (Fig. 3A). Throughout this period, non-fasting blood glucose and HbA1c levels in the TCE5003- and MS023-treated groups were comparable to those in the HAGE group and significantly higher than those in the LAGE group (Fig. 3B and C). Additionally, the levels of free and protein-bound CML and CEL, as well as free MG-H1, were elevated in the serum and skeletal muscle tissues of the TCE5003-treated, MS023-treated, and HAGE groups compared to the LAGE group (Fig. 3D and I and S5). However, protein-bound MG-H1 levels showed no significant differences among the groups (Fig. 3I and S5). These results indicated that TCE5003 and MS023 treatments did not impact the accumulation of AGEs in the serum and muscle tissues. Remarkably, treatment with TCE5003 and MS023 significantly mitigated the development of sarcopenia in HAGE-fed mice. Specifically, both inhibitors partially restored grip strength (TCE5003: 58.9 ± 4.8 gm × Force; MS023: 59.7 ± 6.5 gm × Force; HAGE: 47.3 ± 3.9 gm × Force; LAGE: 72.4 ± 6.8 gm × Force) (Fig. 3J), latency to fall (TCE5003: 17.4 ± 1.5 s; MS023: 18.2 ± 2.2 s; HAGE: 10.3 ± 1.0 s; LAGE: 24.6 ± 2.1 s) (Fig. 3K), and skeletal muscle mass index (TCE5003: 0.84 ± 0.07%; MS023: 0.81 ± 0.04%; HAGE: 0.55 ± 0.04%; LAGE: 1.17 ± 0.23%) (Fig. 3L). Histological analysis of skeletal muscle tissues revealed that TCE5003 and MS023 treatment significantly increased muscle fiber size and reduced connective tissue accumulation compared to the HAGE group (Fig. 3M). Quantification of CSA of muscle fibers in TCE5003 and MS023 groups was significantly increased compared to the HAGE group, although these values remained lower than those observed in the LAGE group (Fig. 3N).

Fig. 3

Administration of TCE5003 and MS023 partially prevented the development of sarcopenia in HAGE-fed mice. (A) Experimental design of TCE5003 and MS023 injections. (B) Non-fasting blood glucose levels in LAGE, HAGE, HAGE + TCE5003, and HAGE + MS023 groups of mice (n = 8 per group) at different time points. (C) HbA1c levels in LAGE, HAGE, HAGE + TCE5003, and HAGE + MS023 groups of mice (n = 8 per group) at different time points. (D-I) Concentrations of free and pb CML, CEL, and MG-H1 in plasma from LAGE, HAGE, HAGE + TCE5003, and HAGE + MS023 groups of mice (n = 8 per group). (D) Free CML in plasma. (E) Free CEL in plasma. (F) Free MG-H1 in plasma. (G) pb-CML in plasma. (H) pb-CEL in plasma. (I) pb-MG-H1 in plasma. (J) Grip strength. (K) Latency to fall as assessed by the rotarod test. (L) Muscle-to-body weight ratio. (M) Representative images of skeletal muscle stained with H&E. Bars = 100 μm. (N) Quantified CSA of muscle fibers (µm2) (n = 8 per group). ns: no significant difference; *P < 0.05; **P < 0.01; *** P < 0.001

As PRMT1 functions as a protein arginine methyltransferase, we assessed arginine methylation levels in total cell extracts derived from the muscle tissues of mice in the LAGE, HAGE, TCE5003, and MS023 groups. The results demonstrated significantly elevated levels of ADMA in the HAGE group compared to the other three groups (Fig. S6). While ADMA levels in the TCE5003 and MS023 groups were similar to those observed in the LAGE group (Fig. S6).

Protein levels of CRTC3 and key molecules of the intrinsic apoptotic pathway were dependent on PRMT1

To investigate the target proteins of PRMT1, we transfected C2C12 cells with pCDNA3-empty vector and pCDNA3-PRMT1 plasmids to establish two independent ControlOE and PRMT1OE cell lines (Fig. S7). Using these cell lines, we performed iTRAQ-based quantitative proteomics and mass spectrometry analysis to identify DEPs regulated by PRMT1 (Fig. 4A). The mass spectrometry results revealed that overexpression of PRMT1 led to differential expression of 206 proteins, with 135 proteins being upregulated and 71 proteins downregulated (Table S7). Figure 4B highlighted the top 10 most upregulated and top 10 most downregulated proteins. Interestingly, among the upregulated proteins, we identified CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 (Fig. 4B). We further confirmed the expression levels of CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 in two ControlOE and two PRMT1OE cell lines, observing that CRTC3, Bax, and Apaf1 were upregulated, while Casp3, Casp7, and Casp9 were activated in PRMT1OE cells (Fig. 4C). These findings suggest that overexpression of PRMT1 can lead to the activation of the intrinsic apoptotic pathway. To further validate this, we conducted Annexin V-FITC/PI staining in ControlOE1 and PRMT1OE1 cell lines. The results showed a significant increase in the distribution of cells in the Q3 (late apoptotic cells) and Q4 (early apoptotic cells) quadrants in PRMT1OE1 cells compared to ControlOE1 cells (2.16% vs. 80.2%) (Fig. 4D). These results indicate that overexpression of PRMT1 induces intrinsic apoptosis in C2C12 cells.

Fig. 4

CRTC3 and apoptotic molecules were dependent on PRMT1. (A) Experimental design for iTRAQ analysis. (B) The top 10 upregulated and top 10 downregulated proteins in ControlOE1 and PRMT1OE1 cells. (C) Validation of protein levels for PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 in two independent ControlOE and PRMT1OE cell lines using Western blotting. (D and E) Flow cytometry analysis. ControlOE1(D) and PRMT1OE1(E) cells were stained with Annexin V-FITC and Propidium Iodide (PI) for 10 min at room temperature in the dark. Stained cells were analyzed using a fluorescence-activated cell sorter. The percentage of cells in different quadrants was indicated

We established ControlKD and PRMT1KD cell lines in C2C12 cells (Fig. S8A and S8B) and treated them with or without 100 µM AGE-BSA for 12 h. Protein levels of PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 were subsequently analyzed. The results showed that PRMT1 depletion failed to increase the protein levels of CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9 (Fig. S8C). These findings suggested that PRMT1 knockdown did not induce intrinsic apoptosis in AGE-BSA-treated C2C12 cells.

PRMT1 methylated CRTC3 in vivo and in vitro

Since PRMT1 is an arginine methyltransferase and both CRTC3 and downstream apoptotic molecules of Bax were dependent on PRMT1, we speculated that PRMT1 may mediate the arginine methylation of these molecules. To test this hypothesis, we conducted IP experiments in skeletal muscle from HAGE-fed mice using IgG and PRMT1 antibodies. We found that PRMT1 bound only to CRTC3 and not to Bax or its downstream apoptotic molecules (Fig. 5A), suggesting that PRMT1 might be involved in the arginine methylation of CRTC3. Next, we co-transfected HA-CRTC3 and various Flag-tagged PRMTs (PRMT1/2/3/4/5) into C2C12 cells and performed IP with anti-HA agarose. Our results showed that CRTC3 bound exclusively to PRMT1 and not to the other PRMTs (Fig. 5B). Additionally, anti-ADMA detection revealed that only PRMT1 could trigger ADMA modification of CRTC3 (Fig. 5B). Furthermore, we co-transfected C2C12 cells with HA-CRTC3 and Flag-tagged PRMT1 constructs containing the catalytic-dead mutants (G98R, E162Q, and E171A). We found that these PRMT1 mutants were unable to bind or methylate CRTC3 (Fig. 5C). In an in vitro methylation assay using recombinant PRMT1 and His-CRTC3, PRMT1 successfully methylated CRTC3 (Fig. 5D). The addition of TCE5003 and MS023 completely inhibited the methylation of CRTC3 in this assay (Fig. 5D).

Fig. 5

PRMT1 interacted with the methylated CRTC3 in vivo and in vitro. (A) PRMT1 interacted with CRTC3 in vivo. Protein extracts from skeletal muscle tissues were immunoprecipitated using IgG and anti-PRMT1-conjugated protein G. Input and immunoprecipitated samples were probed for PRMT1, CRTC3, Bax, Apaf1, Casp3, Casp7, and Casp9. (B) PRMT1 interacted with CRTC3 in vitro. C2C12 cells were co-transfected with Flag-tagged PRMTs (1/2/3/4/5) and HA-tagged CRTC3. After 48 h, cells were lysed and immunoprecipitated using anti-HA agarose. Input and immunoprecipitated samples were analyzed using anti-Flag, anti-HA, anti-ADMA, and GAPDH antibodies. (C) PRMT1 methylated CRTC3 in an enzyme activity-dependent manner. C2C12 cells were co-transfected with HA-CRTC3 and Flag-tagged PRMT1 constructs containing catalytic-dead mutants (G98R, E162Q, and E171A). After 48 h, cells were lysed and immunoprecipitated using anti-HA agarose. Input and immunoprecipitated samples were analyzed using anti-Flag, anti-HA, anti-ADMA, and GAPDH antibodies. (D) TCE5003 and MS023 disrupted the interaction and methylation of CRTC3 by PRMT1 in vitro. An in vitro methylation assay was conducted using recombinant PRMT1 and His-tagged CRTC3, both in the presence and absence of TCE5003 (3.0 µM) and MS023 (90 nM). Immunoblots were performed using anti-ADMA, anti-His, and anti-PRMT1 antibodies. (E) Different regions of the CRTC3 protein: R1 (1-150 aa), R2 (151–525 aa), and R3 (526–619 aa). (F) PRMT1 interacted with the R3 region of CRTC3 in vitro. (G) PRMT1 methylated the R534 site of CRTC3. An in vitro methylation assay was performed using recombinant PRMT1 and His-tagged CRTC3 or its mutants (CRTC3R534K, CRTC3R615K, or CRTC3R618K). Immunoblots were performed using anti-ADMA, anti-His, and anti-PRMT1 antibodies

To identify the specific arginine methylation site(s) on CRTC3, we divided CRTC3 into three regions [R1 (1-150 aa), R2 (151–525 aa), and R3 (526–619 aa)] (Fig. 5E and S9). The in vitro methylation assay revealed that PRMT1 only methylated the R3 region (Fig. 5F). Within the R3 region, there were three arginine residues: R534, R615, and R618 (Fig. S9). We mutated these three residues to lysine and found that PRMT1 could not methylate the R534K mutant (Fig. 5G), indicating that R534 was the specific site of PRMT1-mediated arginine methylation on CRTC3.

CRTC3 and FOXO3a synergistically regulated the expression of the BAX gene

Since CRTC3 is a transcriptional coactivator [35], we hypothesized that it might be involved in regulating Bax and its downstream apoptotic molecules. To test this hypothesis, we generated ControlKD1, ControlKD2, CRTC3KD1, and CRTC3KD2 cell lines in C2C12 background (Fig. S10A). After treating the cells with 100 µM AGE-BSA, we measured the expression levels of BAX, APAF1, CASP3, CASP7, and CASP9. We found that knockdown of CRTC3 specifically reduced the expression of BAX, while the expression of APAF1, CASP3, CASP7, and CASP9 remained unaffected (Fig. S10A). We also performed ChIP assays to examine whether CRTC3 could bind to the promoter regions of these genes. The results showed that CRTC3 was specifically enriched at the BAX gene promoter (Fig. S10B-S10F).

As a transcriptional co-regulator, CRTC3 required interaction with transcription factors to regulate gene expression. To identify the transcription factors interacting with CRTC3, we conducted IP with anti-CRTC3 in skeletal muscle tissues from HAGE-fed mice, followed by mass spectrometry analysis. This analysis identified 82 proteins that interact with CRTC3 (Table S8), among which only one transcription factor, FOXO3a, was detected (Table S8). Using the same IP products utilized for mass spectrometry analysis, we confirmed that CRTC3 interacted with FOXO3a (Fig. 6A). Additionally, an IP assay performed on the same skeletal muscle tissues using an anti-FOXO3a antibody validated that FOXO3a could reciprocally pull down CRTC3 (Fig. 6B). In C2C12 cells co-transfected with Flag-CRTC3 and MYC-FOXO3a, no interaction between CRTC3 and FOXO3a was observed in the absence of AGE-BSA treatment. However, following AGE-BSA treatment, CRTC3 and FOXO3a formed a direct interaction (Fig. 6C). Interestingly, the CRTC3R534K mutant failed to interact with FOXO3a even under AGE-BSA treatment (Fig. 6D). Furthermore, the PRMT1 inhibitors TCE5003 and MS023 disrupted the interaction between CRTC3 and FOXO3a in AGE-BSA-treated cells (Fig. 6D).

Fig. 6

CRTC3 interacted with FOXO3a to regulate the expression of BAX. (A) CRTC3 immunoprecipitated with FOXO3a in vivo. Skeletal muscle tissues from HAGE-fed mice were immunoprecipitated using IgG or anti-CRTC3, followed by Western blot with anti-CRTC3 and anti-FOXO3a. (B) FOXO3a immunoprecipitated with CRTC3 in vivo. Skeletal muscle tissues from HAGE-fed mice were immunoprecipitated using IgG or anti-FOXO3a, followed by Western blot with anti-CRTC3 and anti-FOXO3a. (C) CRTC3 interacted with FOXO3a in vitro. C2C12 cells were co-transfected with Flag-tagged and MYC-tagged plasmids, treated with or without AGE-BSA, and analyzed by immunoprecipitation and Western blot. Membranes were simultaneously probed using anti-Flag and anti-MYC. (D) TCE5003 and MS023 inhibited the CRTC3-FOXO3a interaction in vitro. Co-IP assays were performed on cells co-expressing Flag-CRTC3 + MYC-FOXO3 and Flag-CRTC3R534K + MYC-FOXO3. The cells were treated with 100 µM AGE-BSA combined with either 3.0 µM TCE5003 or 90 nM MS023 for 12 h. (E) The occupancy of CRTC3 and FOXO3a on the BAX promoter was dose-dependent on AGE-BSA. C2C12 cells were treated with increasing concentrations (0, 25, 50, and 100 µM) of AGE-BSA for 12 h, followed by ChIP and RT-qPCR to assess promoter occupancy. (F) Knockdown of PRMT1 disrupted the interaction between CRTC3 and FOXO3a. Two independent PRMT1KD cell lines were treated with or without 100 µM AGE-BSA for 12 h. IP assays were performed using anti-CRTC3-conjugated protein A agarose, and the purified proteins were analyzed by Western blot using anti-CRTC3, anti-FOXO3a, and anti-ADMA antibodies. The membranes loading input samples were simultaneously probed with anti-Flag and anti-MYC antibodies. (G and H) TCE5003 and SM023 reduced the occupancy of FOXO3a and CRTC3 on the BAX promoter in a dose-dependent manner. C2C12 cells were treated with 100 µM AGE-BSA and increasing concentrations of TCE5003 (0, 1.5, and 3 µM) (G) or MS023 (0, 30, and 60 nM) (H) for 12 h. ChIP followed by RT-qPCR was performed to quantify the binding of CRTC3 and FOXO3a to the BAX promoter. (I) CRTC3R534K failed to bind to the BAX promoter. Cells, as used in (D), were subjected to ChIP assays with anti-Flag and anti-MYC antibodies, followed by RT-qPCR to assess the binding of CRTC3, CRTC3R534K, and FOXO3a to the BAX promoter. ns: no significant difference. **P < 0.01

We also examined the effects of AGE-BSA on the expression of BAX gene and found AGE-BSA treatment significantly induced the expression of the BAX gene (Fig. S11A and S11B). However, co-treatment with TCE5003 or MS023 alongside AGE-BSA reduced BAX expression in a dose-dependent manner as the concentrations of TCE5003 and MS023 increased (Fig. S11A and S11B). ChIP assays conducted on C2C12 cells treated with varying concentrations of AGE-BSA (0, 25, 50, and 100 µM) revealed enrichment of both FOXO3a and CRTC3 at the BAX gene promoter (Fig. 6E). Using the same PRMT1KD cells in Fig. S8C, IP assays were performed using an anti-CRTC3 antibody and the results demonstrated that CRTC3 could not interact with FOXO3a even under 100 µM AGE-BSA treatment (Fig. 6F). Furthermore, the methylation levels of CRTC3 were completely abolished in PRMT1KD cells treated with AGE-BSA (Fig. 6F). To further evaluate the impact of PRMT1 inhibition on promoter binding, C2C12 cells were co-treated with 100 µM AGE-BSA and increasing concentrations of either TCE5003 (0, 1.5, and 3 µM) or MS023 (0, 30, and 60 nM). ChIP assays revealed a dose-dependent decrease in the occupancy of CRTC3 and FOXO3a on the BAX promoter with both TCE5003 and MS023 treatments (Fig. 6G and H). Using the cells depicted in Fig. 6D, we measured BAX mRNA levels and observed that the CRTC3R534K mutant failed to induce BAX expression following AGE-BSA treatment (Fig. S11C). Furthermore, ChIP assays using anti-Flag and anti-MYC antibodies in these cells demonstrated that CRTC3R534K did not bind to the promoter region of the BAX gene under AGE-BSA treatment conditions (Fig. 6I). Collectively, these results indicated that PRMT1-mediated methylation was critical for the interaction between CRTC3 and FOXO3a. Inhibition of PRMT1 by TCE5003 or MS023 reduced the binding of CRTC3 and FOXO3a to the BAX gene promoter, potentially mitigating BAX gene expression and its downstream apoptotic effects.

Inhibition of Bax blocked AGE-BSA-induced apoptosis in vitro and prevented sarcopenia incidence in HAGE-fed mice

Given that Bax and its downstream apoptotic signaling pathways were activated in both HAGE-fed mice and AGE-BSA-treated C2C12 cells, we next sought to evaluate the effect of inhibiting Bax on the development of sarcopenia. First, we confirmed that treatment with the Bax inhibitor BAI1 in AGE-BSA-treated cells dose-dependently suppressed the expression of Apaf1 and the activation of Casp3, Casp7, and Casp9 (Fig. S12).

We then assessed the impact of BAI1 on the progression of sarcopenia in HAGE-fed mice. Mice were randomly divided into three groups: one group was fed a LAGE diet, the second group was fed a HAGE diet, and the third group was fed a HAGE diet along with weekly intraperitoneal injections of BAI1 (Fig. 7A). During the 24-week experiment, there were no significant differences in non-fasting blood glucose and HbA1c levels between the BAI1-treated and HAGE groups, although both were markedly higher compared to the LAGE group (Fig. 7B and C). Levels of free and protein-bound CML and CEL, as well as elevated free MG-H1 in serum and skeletal muscle tissues, were significantly increased in both the BAI1-treated and HAGE groups compared to the LAGE group (Fig. 7D and I and S13). However, there was no significant difference in protein-bound MG-H1 between the BAI1-treated and LAGE groups (Fig. 7I and S13). These results suggested that BAI1 treatment did not influence the accumulation of AGEs in the serum and muscle tissues of mice.

Fig. 7

Administration of BAI1 partially prevented the development of sarcopenia in HAGE-fed mice. (A) Experimental design of BAI1 injection. (B) Non-fasting blood glucose levels in LAGE, HAGE, and HAGE + BAI1 groups of mice (n = 8 per group) at different time points. (C) HbA1c levels in LAGE, HAGE, and HAGE + BAI1 groups of mice (n = 8 per group) at different time points. (D-I) Concentrations of free and pb CML, CEL, and MG-H1 in plasma from LAGE, HAGE, and HAGE + BAI1 groups of mice (n = 8 per group). (D) Free CML in plasma. (E) Free CEL in plasma. (F) Free MG-H1 in plasma. (G) pb-CML in plasma. (H) pb-CEL in plasma. (I) pb-MG-H1 in plasma. (J) Grip strength. (K) Latency to fall as assessed by the rotarod test. (L) Muscle-to-body weight ratio. (M) Representative images of skeletal muscle stained with H&E. Bars = 100 μm. (N) Quantified CSA of muscle fibers (µm2) (n = 8 per group). ns: no significant difference; *P < 0.05; **P < 0.01; *** P < 0.001

Remarkably, BAI1 treatment partially inhibited the development of sarcopenia in HAGE-fed mice. Specifically, BAI1 partially improved grip strength (LAGE: 79.1 ± 8.4 gm × Force, HAGE: 45.8 ± 5.2 gm × Force, BAI1: 62.3 ± 7.1 gm × Force) (Fig. 7J), latency to fall (LAGE: 21.5 ± 3.2 s, HAGE: 10.2 ± 1.5 s, BAI1: 15.5 ± 2.4 s) (Fig. 7K), and skeletal muscle mass index (LAGE: 1.15 ± 0.17%, HAGE: 0.53 ± 0.04%, BAI1: 0.87 ± 0.09%) (Fig. 7L). Histological analysis of skeletal muscle tissues demonstrated that BAI1 treatment significantly increased muscle fiber size and decreased connective tissue accumulation compared to the HAGE group (Fig. 7M). Quantitative assessment of CSA of muscle fibers further confirmed a marked increase in the BAI1-treated group relative to the HAGE group, although the CSA remained lower than that observed in the LAGE group (Fig. 7N). These physiological and pathological findings indicated that BAI1 treatment partially prevented the development of sarcopenia in HAGE-fed mice.