Mechanisms of gypenosides in type 2 diabetes mellitus via regulation of m6A methylase METTL3/14 and demethylase FTO

Abstract

Objective

Type 2 diabetes mellitus (T2DM) is characterized by insufficient insulin secretion and insulin resistance. Gypenosides (GPs) are the major bioactive saponins extracted from Gynostemma pentaphyllum. Previous studies suggest that GPs have beneficial effects on T2DM, but the underlying mechanisms remain unclear. This study aims to investigate whether GPs exert therapeutic effects by influencing RNA N⁶-methyladenosine (m6A) methylation modification, thereby regulating the downstream phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) signaling pathway.

Methods

Expression levels and methylation changes of METTL3, METTL14, and FTO in T2DM were analyzed using public databases, and related pathways were explored via gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses. The main active components of GPs were screened from pharmacological databases, followed by compound-target network construction, enrichment analyses, and prediction of potential targets and pathways. A T2DM model was induced in Sprague-Dawley rats using a high-fat/high-sugar diet combined with low-dose streptozotocin. Rats were randomly divided into 4 groups: GPs-treated group (GPG, 400 mg/kg), positive control group (PCG, metformin 100 mg/kg), normal saline control group (CONTROL), and T2DM model group (MODEL). Fasting blood glucose (FBG), oral glucose tolerance test (OGTT)-area under the glucose curve from 0 to 120 min (AUC_0-120 ), fasting insulin (FINS), homeostasis model assessment of insulin resistance (HOMA-IR), serum inflammatory factors, and tissue pathology of pancreas and liver hematoxylin and eosin (HE) staining were assessed. Real-time fluorescence quantitative PCR and Western blotting were used to detect RNA and protein expression levels of METTL3, METTL14, FTO, PI3K, and AKT in pancreatic tissues. Molecular docking was applied to evaluate interactions between GPs’ main components and METTL3, METTL14, and FTO to infer potential binding modes.

Results

Bioinformatic analyses showed downregulation of METTL3/14 and upregulation of FTO in T2DM samples (all P<0.05), with enrichment in pathways related to insulin signaling, PI3K/AKT activation, oxidative stress response, and hormone secretion. Network pharmacology indicated that GPs components may act on targets involved in RNA modification and insulin-related pathways. In diabetic rats, GPs significantly reduced FBG, improved glucose tolerance, decreased HOMA-IR, and decreased the serum tumor necrosis factor-α (TNF-α) and interleukin (IL)-6 levels compared with MODEL (all P<0.05). Pancreatic pathology showed alleviated islet injury and improved cell morphology. GPs treatment up-regulated METTL3/14 mRNA and protein levels, and down-regulated FTO mRNA/protein levels in pancreatic tissue (all P<0.05). PI3K and AKT expression levels increased (both P<0.05), consistent with activation of downstream signals related to glucose uptake and improved insulin sensitivity. Metformin also improved metabolic indices but exhibited a different regulatory pattern on m6A-related enzymes compared with GPs. Molecular docking revealed stable interactions between core GPs saponin structures and methyltransferase-like 3 (METTL3), methyltransferase-like 14 (METTL14), or obesity-associated protein (FTO), suggesting that GPs may directly or indirectly modulate m6A regulatory proteins.

Conclusion

GPs can effectively improve glucose-metabolism disorders, reduce inflammation, and protect pancreatic tissue in T2DM rats. The mechanisms may be associated with METTL3/14 up-regulation and FTO down-regulation, leading to enhanced m6A methylation and subsequent activation of the PI3K/AKT signaling pathway. These findings provide strong evidence for GPs regulation of epigenetic m6A RNA modification and insulin-related downstream pathways, and suggest that natural compounds targeting m6A regulation may be explored in the future for metabolic disease interventions.

Diabetes mellitus (DM) is a chronic metabolic disease characterized by hyperglycemia. Type 2 DM (T2DM) is the most common type of diabetes, accounting for 90% of all DM patients^[1]. Currently, biguanides and insulin are the first-line pharmacological therapy for T2DM. However, biguanides may cause various adverse reactions, including nausea, vomiting, diarrhea, lactic acidosis, and hypoglycemia. Moreover, in patients requiring insulin therapy, prolonged insulin use can also lead to a decrease in insulin receptor sensitivity, resulting in insulin resistance and worsening of the condition.

Gynostemma pentaphyllum, also known as the “Southern Ginseng”, is a traditional Chinese medicinal herb. In 1986, China’s Ministry of Science and Technology listed Gynostemma pentaphyllum as the first “precious Chinese medicine” to be developed in the national “Spark Program” (often translated as “Spark Plan”), reflecting its strategic value for industrial and medical development^[2]. The medicinal value of Gynostemma pentaphyllum is primarily attributed to its purified components, gypenosides (GPs), which have been extensively studied^[3]. GPs have been found to regulate blood pressure, blood lipid, and blood sugar, and aid in weight loss. While GPs show promising potential for the treatment of T2DM and its complications, no obvious side effects have been reported^[4-5]. GPs can reduce the levels of free fatty acids (FFA), tumor necrosis factor-α (TNF-α), resistin, interleukin (IL)-6, and other inflammatory factors in a dose-dependent manner^[6]. In addition, GPs can improve insulin sensitivity in diabetic rats in many ways and protect islet cells^[7-8]. However, the specific mechanism of GPs in DM requires further exploration. N⁶-methyladenosine (m6A) methylation is a widely observed mRNA base modification that primarily regulates RNA stability, splicing, degradation, translation, and other processes^[9]. Reduced expression of m6A methylase [methyltransferase-like 3 (METTL3)and methyltransferase-like 14 (METTL14)], and increased expression of demethylase fat mass and obesity-associated protein (FTO) have been linked to insulin resistance and the progression T2DM^[10-14]. Nevertheless, whether GPs exert their beneficial effects on T2DM by modulating the m6A methylation pathway remains unexplored. This study aims to investigate the therapeutic effects of GPs on T2DM in a rat model and explore their potential mechanisms by examining m6A-related enzymes and their downstream signaling pathways.

1. Materials and methods

1.1. Ethical statement

This study was approved by the Animal Experimental Ethical Inspection Form of Guizhou Medical University [SYXK (GUI) 2023-0002-2303200] and the ethics guidelines for investigations of conscious animals were followed.

1.2. Analysis of m6A-related genes

This study utilized T2DM-related gene dataset GSE15932 from the Gene Expression Omnibus (GEO) database^[15], which consisted of 8 patients with T2DM and 8 healthy controls. RStudio 4.2.1 was employed to call various expansion packages, such as “Bioconductor” “affyPLM”, to assess the quality of the GSE15932 expression profile chip and generate a box diagram. In cases where a gene corresponded to multiple probes, an average value was considered. The m6A-related genes were selected for analysis, focusing on the differences in expression between the T2DM and normal groups. The internal relationships between these genes were also examined. The analytical workflow is illustrated in Supplementary Figure 1 (https://doi.org/1 0.57760/sciencedb.xbyxb.00171).

Figure 1. Analysis of m6A methylation related gene expression in T2DM patients and healthy controls with GSE15932 dataset.

A: Box plot of GSE15932 dataset prior to homogenization; B: Box plot of GSE15932 dataset following homogenization; C: Expression of m6A related genes in T2DM patients; D: Heat map of m6A methylation-related gene expression in T2DM patients; E: Heat map of m6A methylation-related gene expression in T2DM patients and healthy controls. T2DM: Type 2 diabetes mellitus; m6A: N⁶-methyladenosine.

1.3. Analysis of m6A methylation

The methylated RNA immunoprecipitation sequencing (MeRIP-Seq) dataset GSE120024, which is related to T2DM^[16], was selected from the GEO database, including 8 islet tissue samples from patients with T2DM and seven normal islet tissue samples. The degree of m6A methylation modification of various genes in tissues was obtained. RStudio 4.2.1 was used to normalize GSE120024 dataset and evaluate the quality of data. Differences in standardized microarray expression profiles were analyzed. Differentially expressed genes (DEGs) were screened using the absolute value of log₂[fold change (FC)]>1. Positive or negative log₂FC indicated that the gene was upregulated or downregulated. Through RStudio 4.2.1, “clusterProfiler”“org.Hs.eg.db”“ggplot2”, and “pathview” expansion packages, the obtained DEGs were enriched by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG), so as to screen the biological characteristics and signal pathways related to m6A methylation.

1.4. Network pharmacological analysis

The chemical constituents of the total saponins of G. pentaphyllum were identified using the Traditional Chinese Medicine Systems (TCMSP) database. The chemical constituents and corresponding protein targets were screened by the criteria of oral bioavailability (OB)≥ 30% and drug-likeness (DL)≥0.18 (Supplementary Table 1, https://doi.org/10.57760/sciencedb.34774). The target gene names were standardized to official gene symbols using the UniProt database to ensure consistency in protein identification. We searched for potential disease targets in the time to disease (TTD), Genecards, and online mendelian inheritance in man (OMIM) databases. Combined these database results, we removed the repeated targets, obtained the potential targets related to T2DM, and analyzed the target enrichment by GO and KEGG. Finally, the network diagram of “disease-drug-active ingredient-target” was constructed.

Table 1.

Primer sequences used in RT-qPCR

Gene		Primer sequence (5'-3')
METTL3	Forward:	GATTGAGGTAAAGCGAGGTCT
	Reverse:	TGGTAAAGGTAGCCACTGTAGTC
METTL14	Forward:	TTTGCGAAAGTGGGGTTACAGAA
	Reverse:	AACGGCCTTTGGATCTAGGGTCT
FTO	Forward:	AAGAGCAGAGCAGCCTACAACGT
	Reverse:	CTTTCATCCTCGGAGCCTTCACA
PI3K	Forward:	GCTGGAAGAAATGCTGAACC
	Reverse:	TCCCACCACGACTTGATACG
AKT	Forward:	CTTCTATGGTGCGGAGATTG
	Reverse:	ACAGCCCGAAGTCCGTTA
GAPDH	Forward:	GAAGCTGGTCATCAACGGGA
	Reverse:	GGCGGAGATGATGACCCTTT

RT-qPCR: Real-time fluorescence quantitative PCR.

1.5. Gene verification

The Attie Lab Diabetes database is a publicly available resource containing mouse models of T2DM. It can be used for searching gene expression profiles in 6 key tissues from different age groups (i.e., at 4 and 10 weeks) of BlackandTan, BTBR control mice (lean) and obese mice (ob/ob)^[17]. The related data for METTL3, FTO, PI3K, and AKT genes from this database. were analysis to verify the conclusions of the previous steps.

1.6. Materials

GPs (active ingredient≥98%) were provided by China Ankang Health Element Pharmtech Co. Ltd. Metformin tablets were purchased from Switzerland Merck Sherano (batch number: 20023370). Insulin (INS) enzyme-linked immunosorbent assay (ELISA) kit (batch number: EW0620R) was pachased from China Guangzhou ELGBIO company. TNF-α ELISA kit (batch number: GM1151) and IL-6 ELISA kit (batch number: GM1154) were pachased from China Wuhan Seville Biotechnology Co. Ltd. Antibody rabbit anti-METTL3 (batch number: 15073-1-AP) was pachased from British Proteintech company; rabbit anti-METTL14 (batch number: ab252562) was pachased from British Abcam company; rabbit anti-FTO (batch number: 45980S) and rabbit anti-β-actin (batch number: 8457S) was pachased from American CST company; mouse anti-phosphatidylinositol 3-kinase (PI3K; batch number: MA1-74183) and mouse anti-protein kinase B (AKT; batch number: #AHO1112) were pachased from American ThermoFisher Scientific company; horseradish peroxidase (HRP)-labeled goat anti-mouse IgG (Batch number: C2225) was pachased from China Beijing Prior Gene Technology Co. Ltd; goat anti-rabbit IgG (batch number: A0208) was pachased from China Shanghai Beyotime company. The fluorescence instrument was purchased from American Applied Biosystems. DYY-5 voltage and current-stabilized electrophoresis instrument and horizontal electrophoresis cell DYY-III 31D were purchased from China Beijing Liuyi Instrument Factory. The Quanity One gel imaging analysis system was purchased from American Bio-Rad Company.

1.7. Establishment of T2DM rat model and drug administration

Thirty-two specific pathogen-free (SPF) male Sprague-Dawley (SD) rats, aged 8 weeks and weighing 180-220 g, were acclimatized under standard laboratory conditions for 7 days with free access to water and a standard diet. After acclimatization, the rats were randomly divided into 4 groups: A GPs-treated group (GPG, 400 mg/kg), a positive control group (PCG, metformin 100 mg/kg), a normal saline control group (CONTROL), and a T2DM model group (MODEL). Except for CONTROL, rats in GPG, PCG, and MODEL were fed a high-fat and high-sugar diet for 8 weeks and then fasted overnight on the last day. After 12 h of fasting, rats in GPG, PC, and MODEL received an intraperitoneal injection of streptozotocin (STZ; 35 mg/kg) dissolved in 1% citric acid buffer. One week after STZ injection, fasting blood glucose (FBG) levels were measured from tail-tip blood samples collected after an overnight fast. A successful T2DM model was defined as FBG>11.1 mmol/L, accompanied by polydipsia, polyphagia, polyuria, and weight loss, as previously described^[18]. Metformin tablets and GPS were dissolved in normal saline to prepare solutions with concentrations of 60 mg/mL and 50 mg/mL, respectively. Previous studies^[19-23] have shown that GPs are dose dependent in T2DM. According to previous experiments, the doses of metformin selected were 400 mg/kg for GPG and 100 mg/kg for PCG. CONTROL and MODEL were intragastrically perfused with normal saline once daily for 6 weeks. After the last administration, all rats were fasted overnight and sacrificed.

1.8. Oral glucose tolerance test and body weight measurement

An oral glucose tolerance test (OGTT) was carried out in rats at the end of the 6th week; after fasting for 12 h, 0 h blood glucose was measured first, and then intragastric glucose solution (2 g/kg weight) was infused. Blood glucose levels were measured at 30, 60, 90, and 120 min, and the area under the glucose curve from 0 to 120 min (AUC _0-120) was calculated according to formula (1) using the linear ladder method^[24]. ${\mathrm{C}}_{\mathrm{0}},{\mathrm{C}}_{\mathrm{30}},{\mathrm{C}}_{\mathrm{60}},{\mathrm{C}}_{\mathrm{90}},\mathrm{a}\mathrm{n}\mathrm{d}{\mathrm{C}}_{\mathrm{120}}$ represent the blood glucose concentrations at 0, 30, 60, 90, and 120 min, respectively. $t_{\mathrm{0}},t_{\mathrm{30}},t_{\mathrm{60}},t_{\mathrm{90}},\mathrm{a}\mathrm{n}\mathrm{d}{t}_{\mathrm{120}}$ represent the timepoint at 0, 30, 60, 90, and 120 min, respectively.

The body weights of rats were measured and recorded weekly, starting from the 1st week of drug administration (i.e., the week after successful model establishment). After fasting for 12 h, FBG was measured once in tail tip blood weekly.

AUC _0-120= $\frac{\left[C_{\mathrm{0}}+C_{\mathrm{30}}\right]·\left[t_{\mathrm{30}}-t_{\mathrm{0}}\right]}{\mathrm{2}}+\frac{\left[C_{\mathrm{30}}+C_{\mathrm{60}}\right]·\left[t_{\mathrm{60}}-t_{\mathrm{30}}\right]}{\mathrm{2}}+$

$$\frac{\left[C_{\mathrm{60}}+C_{\mathrm{90}}\right]·\left[t_{\mathrm{90}}-t_{\mathrm{60}}\right]}{\mathrm{2}}+\frac{\left[C_{\mathrm{90}}+C_{\mathrm{120}}\right]·\left[t_{\mathrm{120}}-t_{\mathrm{90}}\right]}{\mathrm{2}}$$ (1)

1.9. Serological index detection

The levels of fasting insulin (FINS), TNF-α, and IL-6 in rat serum of each group were measured using an ELISA kit. According to the results of FBG and FINS, the insulin resistance index (HOMA-IR) was calculated according to formula (2) ^[25].

$$\mathrm{HOMA-IR}=\frac{FINS\times FBG}{\mathrm{22.5}}$$ (2)

1.10. Histopathological experiment

Liver and pancreas tissues of each group were collected and fixed with 4% neutral formaldehyde. Following fixation, the tissues were dehydrated, embedded in paraffin, and sectioned, and then hematoxylin and eosin (HE) staining was performed. The stained sections were examined under a microscope. Representative images were captured and assessed. Based on the varying degrees of pathological changes in the pancreatic tissue, we assigned corresponding scores to each group; the scoring criteria are provided in the Supplementary Table 2 (https://doi.org/10.57760/sciencedb.34776) and Table 3 (https://doi.org/10.57760/sciencedb.34777).

Table 2.

Molecular docking results

Chemical ingredient	Binding energy/(kcal·mol-1)			OB/%	DL
	FTO	METTL3	METTL14
Gypenoside LXXIX	-6.4	-6.0	-7.2	107.71	0.76
Gypenoside XII	-6.8	-5.3	-7.6	78.72	0.72
Gypenoside XXXII	-6.2	-5.5	-7.1	77.41	0.33

1 kcal=4.186 kJ. FTO: Fat mass and obesity-associated protein; METTL3: Methyltransferase-like 3; METTL14: Methyltransferase-like 14; OB: Obesity; DL: Dyslipidemia.

1.11. Real-time fluorescence quantitative PCR

Total RNA was extracted from rat pancreas tissues using TRIzol reagent. The complementary DNA (cDNA) was then reverse-transcribed using the HiScrip III integrated RT SuperMix kit. It was then amplified using Bio-Rad real-time fluorescence quantitative PCR (RT-qPCR) detection system. The relative expression levels of reduced nicotinamide adenine dinucleotide phosphate (NADPH) were determined using the 2^-ΔΔCt method. The primer sequences used are listed in Table 1.

1.12. Western blotting

After aseptic operation, the pancreatic tissue of rats was collected, the samples were ground with liquid nitrogen, and the protein concentration was measured using a bicinchoninic acid (BCA) kit. Equal amounts of protein (20 µg per lane) were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred onto polyvinylidene difluoride (PVDF) membranes at a constant current of 20 mA for 2 h. The membranes were blocked with 50 g/L skimmed milk for 2 h at room temperature, followed by incubation with primary antibodies (including METTL3, METTL14, FTO, PI3K, and AKT) overnight at 4 ℃. After washing (10 min each time, 3 times) with PBS, he membranes were incubated with HRP-labeled goat anti-mouse IgG for 2 h at room temperature. Protein bands were visualized using an enhanced chemiluminescence (ECL) detection system after exposure. The band intensity was quantified by densitometry using image analysis software.

1.13. Molecular docking

Three main saponins of GPs, Gypenoside LXXIX, Gypenoside XII and Gypenoside XXXII, were selected according to their OB and DL values. The three-dimensional (3D) structures of these components were retrieved from the PubMed Chemical Substances (PubChem) platform, and the 3D structures of METTL3, METTL14, and FTO were obtained from the ProteinData Bank (PDB) database. The obtained saponins and proteins were prepared with PyMOL2.4.0 and then docked with AutoDock vina. The docking results were judged based on the free binding energy. When the free binding energy was less than -5.0 kcal/mol (1 kcal=4.186 kJ), there is good interaction between the compound and the target protein. The lower the binding energy, the more stable the ligand-protein complex and the greater the possibility of interaction between the 2 molecules.

1.14. Statistical analysis

All data were analyzed using statistical software SPSS 27.0.1. Numerical results were expressed as mean±standard deviations. Analysis of m6A methylation related genes data were normalized using the “limma” expansion package and subjected to the τ and Bayesian tests. Numerical data satisfied normal distribution and homogeneity of variance. Statistical significance was evaluated using one-way analysis of variance (ANOVA). P<0.05 is considered statistically significant.

2. Results

2.1. Analysis results of m6A methylation related genes

Quality analysis of GSE15932 dataset revealed that the samples exhibited uniform expression and were all included in the study after homogenization (Figure 1A and 1B). A total of 24 genes related to m6A methylation were analyzed, including ALKBH5, CBLL1, EIF3A, EIF3B, FTO, HNRNPA2B1, HNRNPC, IGF2BP1, IGF2BP2, IGF2BP3, KIAA1429, METTL14, METTL3, METTL4, RBM15, RBM15B, WTAP, YTHDC1, YTHDC2, YTHDF1, YTHDF2, YTHDF3, ZC3H13, and ZNF217. The results indicated that the expression levels of METTL3, METTL14, and FTO were lower in T2DM patients than in healthy controls (Figure 1C and 1D). Correlation analysis among the 24 m6A related genes in T2DM patients revealed that FTO expression exhibited a positive correlation with the expression of EIF3A, HNRNPA2B1, RBM15B, YTHDC2 and YTHDF1, whereas it showed a negative correlation with the expression of WTAP and ZNF217. The METTL14 expression was positively correlated with the expression of EIF3A, HNRNPA2B1, RBM15, YTHDF2, and ZC3H13, but negatively correlated with the expression of ALKBH5 and IGF2BP1. METTL3 expression was positively correlated with the expression of KIAA1429 and YTHDF1 (Supplementary Figure 2, https://doi.org/1 0.57760/sciencedb.34778).

Figure 2. Pathways enrichment analysis of DEGs in T2DM patients with GSE120024 dataset.

A: Volcano map of m6A methylation modified DEGs; B: KEGG enrichment analysis bubble map of DEGs; C: GO analysis of differential genes enrichment ring map. Blue represents BP, green represents MF, and yellow represents CC respectively. DEGs: Differentially expressed genes; KEGG: Kyoto Encyclopedia of Genes and Genomes; GO: Gene Ontology; BP: Biological process; MF: Molrcular function; CC: Cellular component; FC: Fold change; COVID: Corona virus disease; PI3K: Phosphoinositide 3-kinase; AKT: Protein kinase B; cAMP: Cyclic adenosine monophosphate; cGMP: Cyclic guanosine monophosphate; PKG: Protein kinase G; ECM: Extracellular matrix; IL-17: Interleukin-17.

2.2. Enrichment analysis and construction of “disease-drug-active component-target” network diagram

Using the GSE120024 methylation measurement dataset, we identified 130 DEGs (53 up-regulated and 77 down-regulated), which were associated with m6A methylation modification in T2DM patients (Figure 2A). KEGG pathway enrichment analysis of these DEGs revealed significant enrichment in pathways including neuroactive ligand-receptor interaction, corona virus disease (COVID)-19, and the phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) signaling pathway (Figure 2B). GO analysis found that these DEGs were particularly relevant to biological characteristics, such as glycosaminoglycan binding, hormone secretion, and collagen-containing extracellular matrix (Figure 2C). The analysis of “disease-drug-active component-target” network diagram (Figure 3A) revealed there were 315 genes that were both drug and T2DM targets. Further enrichment analysis implicated key signaling pathways, such as neuroactive ligand-receptor interaction, PI3K/Akt signaling pathway, cyclic adenosine monophosphate (cAMP) signaling pathway, calcium signaling pathways, and proteoglycans in cancer, which were closely associated with important biological characteristics, including protein serine/threonine kinase activity, positive regulation of mitogen-activated protein kinase (MAPK) cascade, and neuronal cell body regulation (Figure 3B, 3C and Supplementary Table 4, https://doi.org/10.57760/sciencedb.xbyxb.00172).

Figure 3. Pathways enrichment analysis via network pharmacology.

A: Venn diagram represents the intersection of T2DM-related genes and targets of GPs components. B: A bubble chart of KEGG enrichment analysis illustrates intersection target genes. C: GO analysis displays enrichment circle diagram of genes corresponding to the intersection targets. Rap1: Ras-proximate-1; AGE: Advanced glycation end-product; RAGE: Receptor for AGE; EGFR: Epidermal growth factor receptor; TRP: Transient receptor potential.

2.3. Gene expression verification in DM database

In the Attie Lab Diabetes database, we examined the expression of METTL3, FTO, PI3K, and AKT in the pancreatic islet tissue of ob/ob mice and lean mice (Figure 4). At 4 weeks of age, the expression levels of METTL3, PI3K, and AKT were significantly reduced in ob/ob mice compared to lean mice, whereas the expression level of FTO showed the opposite trend (all P<0.05). By 10 weeks of age, the expression patterns had changed. While the expression levels of PI3K, AKT, and METTL3 in lean mice remained relatively stable, they all increased to varying degrees in ob/ob mice (all P<0.05), although they remained lower than in lean mice. Concurrently, the expression levels of FTO increased in both groups at the 10th week (P<0.05).

Figure 4. Unraveling the mechanism of GPs in T2DM via Attie Lab Diabetes database.

A-D: The relative expression levels of PI3K (A), METTL3 (B), AKT (C), and FTO (D) in pancreatic islet tissues of ob/ob mice and lean mice at 4 and 10 weeks of age in the Attie Lab Diabetes database. lean: BTBR control mice; ob/ob: Obese mice. *P<0.05 vs lean.

2.4. FBG, OGTT, and body weight measurement results

In FBG test (Figure 5A): At week 8 (pre-modeling), the FBG levels of GPG, PCG, and MODEL were significantly higher than those of CONTROL (all P<0.05), but below the T2DM diagnostic threshold. Following successful model induction at week 9, the FBG levels of GPG, PCG, and MODEL, except CONTROL, rose above the T2DM diagnostic threshold. Subsequently, FBG in the MODEL stabilized at a high level with minor fluctuations. In the GPG, the FBG levels did not significantly change within the first 2 weeks after STZ administration but decreased gradually from week 12. In the PCG, the FBG levels slightly decreased one week after administration and showed a significant decrease by the second week, surpassing the decrease observed in the GPG. By week 15, the FBG levels in the PCG and GPG were comparable, but both remained higher than in CONTROL (all P<0.05). In OGTT (Figure 5B and 5C): Blood glucose peaked at 30 min post-glucose load and gradually declined to the initial blood glucose level in all groups. The MODEL maintained severely elevated glucose levels throughout. PCG showed a sharper glucose decline between 30 min and 60 min compared to GPG, which exhibited a more gradual downward trend. The AUC _0-120 for MODEL was significantly higher than that for CONTROL (P<0.05). Both GPG and PCG exhibited a similar hypoglycemic effect, which was significantly lower than that of MODEL but still higher than that of CONTROL (all P<0.05).

Figure 5. Effects of GPs on glucose metabolism and inflammatory cytokines in T2DM rats.

A: FBG changes of rats in each group; B: Blood glucose change chart of rats in each group in OGTT; C: AUC of blood glucose curve of rats in each group in OGTT; D: Body weight change trend of rats in each group; E: FINS change of rats in each group; F: HOMA-IR results of rats in each group; G to H: Serum contents of TNF-α (G) and IL-6 (H) of rats in each group. *P<0.05, ***P<0.001 vs CONTROL; †P<0.05, †††P<0.001 vs MODEL. CONTROL: Normal saline control group; MODEL: T2DM model group; GPG: Gypenosides-treated group; PCG: Positive control group; OGTT: Oral glucose tolerance test; FBG: Fasting blood glucose; FINS: Fasting insulin; HOMA-IR: Homeostatic model assessment of insulin resistance; IL-6: Interleukin-6; TNF-α: Tumor necrosis factor-alpha; AUC: Area under the curve.

In body weight test (Figure 5D): After an 8-week high-fat/sugar diet, the other 3 groups weighed significantly more than CONTROL. The body weights of MODEL gradually decreased. In contrast, both GPG and PCG showed a gradual weight gain after drug administration, with the weight increase being more pronounced in the PCG. Despite this gain, their final body weights remained lower than those of CONTROL (all P<0.05).

2.5. Results of serological test

The FINS concentration in MODEL was significantly higher than that in CONTROL (P<0.001, Figure 5E). Both GPG and PCG treatments effectively lowered insulin levels compared to MODEL (both P<0.05), with values approaching those of CONTROL (P>0.05). Specifically, The reduction was more pronounced in PCG. The HOMA-IR was calculated based on FINS and FBG (Figure 5F). The HOMA-IR of MODEL was significantly higher than that of CONTROL (P<0.001). Both GPG and PCG treatments exhibited significantly lower HOMA-IR than MODEL (both P<0.001), with values closer to those of CONTROL (P>0.05). The serum levels of the inflammatory cytokines IL-6 and TNF-α were significantly elevated in MODEL compared to CONTROL (both P<0.001, Figure 5H and 5I). Treatment in both GPG and PCG significantly reduced these levels (both P<0.05).

2.6. Results of histological test

Liver tissues from CONTROL demonstrated intact capsules, orderly arranged hepatic cords, and only occasional hepatocyte steatosis without inflammation or fibrosis. MODEL exhibited disrupted lobular architecture, widespread severe hepatocyte steatosis, and punctate necrosis. Both GPG and PCG showed improvements, with only mild steatosis and punctate necrosis present, which was less severe than in MODEL (Figure 6A). Based on the pathological changes of liver sections, we assigned corresponding scores to different degrees of case characteristics: CONTROL, 0; MODEL, 3; GPG, 2; PCG, 2.

Figure 6. Effects of gypenosides on m6A-related gene expression and liver pathological alterations in T2DM rats.

A to B: Pathological staining sections of liver (A) and pancreas(B); C to G: mRNA expressions of PI3K (C), METTL3 (D), METTL14 (E), AKT (F), and FTO (G) in each group. **P<0.01 vs CONTROL; †P<0.05 vs MODEL.

HE staining of pancreas tissues (Figure 6B) revealed that pancreatic tissues from CONTROL exhibited normal architecture with only minimal islet cell vacuolar degeneration and no significant acinar or stromal pathology. In contrast, MODEL displayed marked pathological changes, including islet cell vacuolar degeneration, necrosis, nuclear pyknosis, and cell disintegration. Both GPG and PCG showed similar but notably milder degrees of islet cell injury compared to MODEL. Based on the pathological changes of pancreatic sections, we assigned corresponding scores to different degrees of case characteristics: CONTROL, 0; MODEL, 2; GPG, 1; PCG, 1.

2.7. Results of RT-qPCR

The expression levels of METTL3 and METTL14 mRNA in MODEL were significantly lower than those in CONTROL (both P<0.01, Figure 6C to 6G). The mRNA expression levels of these 2 genes were both rescued in GPG (both P<0.05). However, this recovery was not observed in PCG (P>0.05). Conversely, the expression of FTO was significantly higher in MODEL than in CONTROL (P<0.01). FTO expression were reduced in GPG (P<0.05). Although the PCG showed a downward trend, the difference was not statistically significant (P>0.05). The mRNA expression levels of PI3K and AKT were also significantly lower in MODEL than CONTROL (both P<0.01). GPs treatment significantly increased their expression towards the level of CONTROL (both P<0.05). In contrast, metformin treatment did not significantly alter the expression of PI3K and AKT compared to MODEL (P>0.05, Figure 5C and 5F).

2.8. Results of Western blotting

The Western blotting results (Figure 7) indicated that the protein expression levels of METTL3, METTL14, AKT, and PI3K in MODEL were significantly lower than those in CONTROL; Conversely, FTO level was significantly higher (all P<0.01). In GPG, the levels of METTL3, METTL14, AKT, and PI3K were all significantly higher than those in MODEL, whereas the level of FTO decreased significantly (all P<0.01). This finding aligned with the observed changes in the mRNA content (Figure 5). Interestingly, the levels of METTL3 in PCG were similar to those in GPG than that in MODEL (P<0.001), whereas the AKT level was lower and the FTO level was higher (both P<0.05) than those in MODEL. However, both METTL14 and PI3K level in PCG did not change compared with those in MODEL (both P>0.05).

Figure 7. Effects of gypenosides on m6A-related protein expression and molecular docking analysis in T2DM rats.

A: Western blotting representative images of each group; B to F: Relative protein expression of METTL3 (B), METTL14 (C), AKT (D), FTO (E), and PI3K (F) in each group. **P<0.01, ***P<0.001 vs CONTROL; ††P<0.01, †††P<0.001 vs MODEL. METTL3: Methyltransferase-like 3; METTL14: Methyltransferase-like 14; FTO: Obesity-associated protein.

2.9. Results of molecular docking

The docking results (Table 2) showed that the free binding energies of the three saponins (including Gypenoside LXXIX, Gypenoside XII, and Gypenoside XXXII) and the 3 proteins (including METTL14, METTL3, and FTO) were all less than -5.0 kcal/mol. Additionally, each active component interacts with its corresponding amino acid residues through hydrogen bonds. Notably, the binding energies of the 3 saponins to METTL14 were lower than -7 kcal/mol, which was more prominent than those of FTO and METTL3.

3. Discussion

T2DM is a prevalent and significant global public health issue that poses a serious threat to human health. However, the prevention and treatment of this disease remain challenging. Therefore, there is a critical need to develop drugs with fewer side effects, ease of production, and significant efficacy for the treatment of T2DM^[26]. m6A methylation is the most common chemical modification of mRNA and is highly diverse and conserved in eukaryotic RNA. This modification primarily involves 3 types of protein molecules, methyltransferases (writers), demethylases (erasers), and binding proteins (readers), each with their respective functions. Among these proteins, the methyltransferases METTL3 and METTL14 are responsible for promoting methylation, while FTO is involved in demethylation. These proteins play significant roles and are closely associated with T2DM^[27-28].

Previous studies^[29-31] have demonstrated that METTL3 and METTL14 typically form a catalytic center known as the m6A-METTL complex (MAC) during their functional role. Our integrated approach revealed reduced expression of METTL3, METTL14, and key signaling nodes like AKT in T2DM patients and mouse islets, while FTO was upregulated. This coordinated downregulation of the METTL3 and METTL14 writer complex likely diminishes m6A methylation, a notion supported by their strong positive correlation with other writers and negative correlation with erasers like FTO. Conversely, elevated FTO correlated positively with many reader genes and negatively with writers, suggesting a shift toward a hypomethylated state.

Our bioinformatic analyses further linked T2DM-associated differentially expressed genes to glycosaminoglycan binding and the PI3K/AKT pathway. Following a 6-month treatment with metformin, patients with T2DM exhibited a decrease in urinary sulfate excretion, indicating a regulatory effect of metformin on the modification of the glycosaminoglycan system^[32]. The PI3K/AKT signaling pathway is crucial in insulin signal transduction, as it regulates glucose uptake, cell metabolism, and glycogen synthesis. It plays a significant role in improving insulin resistance and controlling hyperglycemia^[33]. Additionally, this pathway influences serine/threonine (Ser/Thr) kinase, thereby facilitating the transfer of glucose transporter type 4 (GLUT4) to the cell membrane. This, in turn, enhances glucose transportation into the cell and helps reduce blood glucose levels^[34-35]. Multiple drugs have also been shown to improve T2DM by enhancing the expression of this pathway^[36-38].

The enrichment results of network pharmacology revealed the PI3K/AKT, cAMP, and MAPK pathways in the action of GPs against T2DM. Previous experiments^[39] have demonstrated that cyclocarya paliurus triterpenic acids inhibit hepatic gluconeogenesis in the presence of T2DM by activating AMP-activated protein kinase (AMPK)/cAMP signaling pathway, thereby reducing blood glucose levels. β-Arrestin-2 plays a crucial role in regulating irisin-induced glucose metabolism in T2DM through p38-MAPK signaling pathway^[40]. Additionally, it has been discovered that the PI3K/AKT signaling pathway acts as a downstream pathway of m6A methylation. When METTL3 increases, it activates the PI3K/AKT signaling pathway, leading to the promotion of angiogenesis^[41]. In a mouse experiment, it was confirmed that melatonin promotes spermatogonia activity by enhancing m6A methylation and activating the downstream PI3K/AKT signaling pathway^[13]. These findings suggest that the downstream pathways of m6A methylation modification, specifically the PI3K/AKT signaling pathway, are closely associated with T2DM and GPs. Therefore, this should be considered in future studies. Our findings that METTL3, PI3K, and AKT were downregulated in diabetic islets, while FTO was upregulated, are consistent with this regulatory axis.

OGTT, a glucose loading test, has been used to assess the function of islet β-cells and the ability of the body to regulate blood sugar. It serves as a diagnostic test for DM and is widely used in clinical practice and basic research. In animal OGTT experiments, the fasting blood glucose of diabetic rats gradually decreased and approached the levels of normal rats during the administration of GPs. Additionally, the weight loss trend gradually halted and even showed signs of recovery. From a physiological standpoint, it can be inferred that GPs effectively controls the symptoms of T2DM. This confirms that GPs can enhance of islet β-cells function or body regulation ability of blood glucose levels, which is particularly crucial in the treatment of DM. Similar improvements have been observed in previous studies^[42].

Serologically, GPs and metformin comparably attenuated the diabetic increases in inflammatory cytokines (TNF-α, IL-6) and hyperinsulinemia, resulting in significantly lowered HOMA-IR indices. The reduction in elevated insulin levels likely reflects an alleviation of the insulin-resistant state, where compensatory hypersecretion is no longer required. Chronic inflammation, driven by cytokines like IL-6 and TNF-α, contributes to β-cell dysfunction and metabolic dysregulation in T2DM^[43-44] and is linked to its complications^[45-46]. GPs have been found to reduce neuroinflammation and produce antidepressant effects in mouse models. This is achieved by reducing the levels of inflammatory factors such as IL-6, IL-1β, and TNF-α^[47]. Additionally, GPs can decrease the release of IL-6 and TNF-α by macrophages through the MAPK signaling pathway, thereby reducing inflammation^[48]. Our data show that GPs lowered serum TNF-α and IL-6 levels in diabetic rats. GPs treatment also ameliorated pancreatic islet morphology and reduced hepatocyte steatosis and necrosis. Improved islet architecture likely contributes to the observed normalization of insulin secretion and HOMA-IR, indicating restored functional capacity rather than mere cell number preservation. Hyperglycemia and disruption of lipid regulation can worsen liver injury, leading to hepatocyte necrosis and adipocyte formation, and potentially result in conditions such liver disease in the later stages^{[14, 49]}. Previous study^[50] has shown that a polysaccharide complex called Policaptil Gel Retard can effectively improve fatty liver caused by T2DM. In the liver tissue section staining results, it was evident that intragastric administration of GPs improved hepatocyte necrosis and steatosis in diabetic rats. These positive outcomes may be attributed to correction of insulin resistance and endocrine disorders.

After identifying macroscopic pathological changes, we further investigated microscopic alterations at the molecular level. Previous studies^[51-54] have reported that METTL3 expression is downregulated in the islets of patients with DM, and METTL3/METTL14-mediated m6A methylation plays a crucial role in maintaining glucose homeostasis and reducing insulin resistance. In contrast, FTO has been strongly associated with obesity and metabolic disorders, with elevated expression contributing to fat accumulation and insulin resistance^[45].

In this study, the mRNA expression patterns of METTL3, METTL14, and FTO were consistent with predictions derived from bioinformatics analysis. Following GPs administration, METTL3 and METTL14 expression was significantly upregulated, whereas FTO expression was markedly downregulated in diabetic rats. These transcriptional changes were further confirmed at the protein level. Consistent with these molecular alterations, improvements in insulin resistance, FBG levels, pancreatic islet pathology, and hepatic steatosis were observed, supporting the regulatory role of m6A modification in glucose and lipid metabolism.

Furthermore, GPs treatment significantly increased the overall level of m6A methylation and activated the downstream PI3K/AKT signaling pathway, as evidenced by elevated expression of PI3K and AKT. Activation of this pathway enhances insulin signaling and glucose uptake, thereby alleviating diabetes-associated pathological changes^[55-59]. Notably, metformin treatment did not significantly affect FTO expression or activate the PI3K/AKT pathway in pancreatic islets, which is consistent with previous reports^[60-61], indicating that metformin primarily enhances insulin sensitivity in peripheral tissues rather than directly regulating m6A-related enzymes.

Although most studies^[62-63] support elevated FTO expression in diabetes and obesity, a minority of reports have suggested reduced FTO levels in patients with advanced diabetic complications. Such discrepancies may be related to disease stage, complications, glycemic control, or body composition, indicating that FTO expression is dynamically regulated during diabetes progression.

Molecular docking analysis further suggested potential interactions between key GPs components and m6A-related enzymes, providing a structural basis for the observed regulatory effects. Collectively, these findings indicate that GPs alleviate T2DM by modulating m6A methylation through upregulation of METTL3 and METTL14 and inhibition of FTO, thereby activating the PI3K/AKT signaling pathway and improving insulin sensitivity, glucose utilization, and islet function (Figure 8).

Figure 8. Schematic diagram of the action mechanism of GPs in T2DM model rats.

GPs regulate m6A methylase and demethylase, which in turn increases the transcription level of related proteins. Additionally, GPs activate the PI3K/AKT signal pathway downstream of the insulin pathway, leading to increased glucose uptake by the target organs. By reducing insulin resistance and improving the disorder of glucose metabolism, GPs can effectively treat T2DM. GPs: Gypenosides; YTHDF1/2/3: YTH N6-methyladenosine RNA binding protein 1/2/3; GLUT4: Glucose transporter type 4; TBC1D4: TBC1 domain family member 4; YTHDC1/2: YTH domain containing 1/2; IGF2BP1/2/3: Insulin like growth factor 2 mRNA binding protein 1/2/3; METTL16: Methyltransferase like 3/14/16; WTAP: Wilms’ tumor 1 associating protein; ZC3H13: Zinc finger CCCH-Type containing 13; RBM15/15B: RNA binding motif protein 15/15B; IRS1: Insulin receptor substrate 1; EIF3: Eukaryotic translation initiation factor 3; ALKBH5: RNA demethylase alkB homologue 5.

This study demonstrated that GPs can ameliorate the pathological changes associated with insulin resistance and islet β-cell disorder through increasing the degree of m6A methylation in the islets of T2DM rats and activating the downstream PI3K/AKT signaling pathway. These results suggest that targeting m6A modification with natural compounds such as GPs could represent a promising therapeutic strategy for T2DM, warranting further clinical investigation. Future studies could further explore more natural drugs that can effectively treat T2DM with minimal or no side effects, thereby offering medical professionals and patients more and improved treatment options upon diagnosis.

Funding Statement

This work was supported by College Students’ Innovative Entrepreneurial Training Plan Program of Guizhou Province, China (S202210660105, S202310660055).

Conflict of Interest

The authors declare that they have no conflicts of interest to disclose.

AUTHORS’CONTRIBUTIONS

LI Jiayi Research design, experimental operation, paper writing, and paper modification; ZHANG Shaoqian Research design, experimental operation, data collecting and analysis, paper modification; TIAN Renwei Research design, experimental operation, paper writing, data collecting and analysis, paper supervision and revision; TAI Hebei Research design, data collecting and analysis, paper modification; ZHANG Yating Experimental operation, data collecting and analysis, paper modification; HU Mingyi Data collecting and analysis; GONG Guangbin Experimental operation and paper modification; SUN Jianfei and WU Ning Paper supervision and revision; The final version of the manuscript has been approved and read by all authors.

Note

http://xbyxb.csu.edu.cn/xbwk/fileup/PDF/2025101735.pdf