Introduction

Loss of glucose homeostasis can lead to type 2 diabetes (T2D). A hallmark of T2D is persistently elevated blood glucose levels, which can contribute to various complications, such as kidney failure and neuropathy. Given that abnormal gluconeogenic activity accounts for the primary hepatic glucose production (HGP) (1, 2), which is the main source of increased blood glucose concentration in T2D, targeting gluconeogenesis is a reasonable strategy to maintain blood glucose homeostasis.

Fructose 1,6-bisphosphatase (FBP1), a rate-controlling enzyme in gluconeogenesis, catalyzes the conversion of fructose 1,6-bisphosphate to fructose 6-phosphate. Interestingly, it has been reported that a point mutation in FBP1 can significantly reduce the efficacy of metformin, a first-line drug for the treatment of T2D, suggesting that FBP1 is a major contributor to the therapeutic effect of metformin (3). However, it remains unclear how FBP1 responds to metformin treatment.

Ten-Eleven Translocation 2 (TET2) belongs to the TET family, which includes TET1, TET2, and TET3. These enzymes function as DNA dioxygenases, catalyzing the successive oxidation of 5-methylcytosine (5mC) to 5-hydroxymethylcytosine (5hmC) (4), 5-formylcytosine (5fC), and 5-carboxycytosine (5caC) (5, 6). Finally, thymine-DNA glycosylase (TDG) mediates the removal of 5caC, leading to an unmodified cytosine, which participates in DNA demethylation and gene expression regulation (6). Recently, it was reported that the expression of FBP1 can be regulated by promoter methylation (7, 8), leading us to hypothesize that TET2 may play a role in regulating FBP1 expression and gluconeogenesis.

Mutations in TET2 frequently occur in various myeloid cancers. Somatic alterations in TET2 are observed in 50% of patients with chronic myelomonocytic leukemia and are associated with poor outcomes (9). The frequency of TET2 mutations in patients with myelodysplastic syndromes is 19% (10). Moreover, reversing TET2 deficiency facilitates hematopoietic stem and progenitor cell (HSPC) differentiation, blunts HSPC self-renewal, and blocks leukemia progression (11). It has also been reported that TET2 represses mTORC1 and HIF1α signaling, thus suppressing tumor growth in hepatocellular carcinoma (HCC) and clear cell renal cell carcinoma (ccRCC), respectively (12, 13). These studies focused on the function of TET2 in cancer development; however, it is unclear whether TET2 is involved in T2D progression. In this study, we demonstrated that TET2 is recruited by HNF4α to the FBP1 promoter and activates FBP1 expression by demethylation, contributing to gluconeogenesis and T2D pathology, and identified HNF4α-TET2-FBP1 axis as a target of metformin treatment, suggesting that targeting the HNF4α-TET2-FBP1 axis might be a potential strategy for T2D treatment.

Results

TET2 contributes to gluconeogenesis and T2D

To explore the role of TET2 in gluconeogenesis, we developed three mouse models: one subjected to overnight fasting for 16 hours prior to tests, and two subjected to high-fat feeding (HFD) for 11 days and 12 weeks, respectively. The results showed that both fasting and HFD increased the mRNA and protein levels of TET2 in mouse livers compared to the normal chow group (Fig. 1A-E), indicating that TET2 may play an important role in gluconeogenesis. To test this hypothesis, we examined the effect of TET2 on glucose output and found that TET2 overexpression promoted glucose output in HepG2 cells and primary mouse hepatocytes (Fig. 2A and B). Consistent with this, Tet2 knockout (4) impaired gluconeogenesis in HepG2 cells and mouse hepatocytes, even under glucagon treatment (Fig. 2C and D). Together, these data demonstrated that TET2 contributes to gluconeogenesis, prompting us to investigate whether TET2 is involved in T2D progression. Pyruvate tolerance tests (PTT), glucose tolerance tests (GTT), and insulin tolerance tests (ITT) were conducted, and interestingly, the results showed that Tet2 KO markedly increased glucose tolerance and insulin sensitivity compared to the control mice (Fig. 2E-G). In summary, the loss of Tet2 function decreased gluconeogenesis in the liver and may contribute to the treatment of T2D.

Figure 1.

Figure 2.

TET2 upregulates FBP1 expression in liver cells

Next, we explored the potential mechanism by which TET2 increases gluconeogenesis. Fructose-1,6-bisphosphatase 1 (FBP1), a rate-limiting enzyme in gluconeogenesis, is crucial for its regulation and was recently identified as a target of metformin (3). This led us to hypothesize that FBP1 might participate in TET2-mediated regulation of gluconeogenesis. The results showed that glucagon significantly increased both TET2 and FBP1 expression levels in HepG2 and primary mouse liver cells (Fig. 3A-C).

Figure 3.

Additionally, fasting and high-fat diet (HFD) also upregulated FBP1 mRNA levels in mouse livers (Fig. 3D and E). Notably, Pearson correlation analysis revealed a positive correlation between TET2 and FBP1 expression in both control and fasting groups (Fig. 3F). To confirm this, Gene Expression Profiling Interactive Analysis (GEPIA) (14) was used to analyze the correlation between TET2 and FBP1 expression in human liver tissue. Consistent with the mouse data, FBP1 expression levels positively correlated with TET2 levels in human livers (Fig. 3G). These findings prompted us to examine whether TET2 regulates FBP1 expression. The results showed that TET2 overexpression promoted FBP1 expression in primary mouse hepatocytes and HepG2 cells (Fig. 3H), while Tet2 KO (4) significantly decreased FBP1 levels in HepG2 and LO-2 cells (Fig. 3I and J). Furthermore, Tet2 KO abolished the glucagon-induced upregulation of FBP1 (Fig. 3K), suggesting that TET2 is required for glucagon-induced FBP1 upregulation. Taken together, these data indicate that TET2 regulates FBP1 in the livers.

To explore the mechanism by which TET2 regulates FBP1, ChIP-qPCR was performed to investigate whether TET2 could bind to the FBP1 promoter and catalyze the conversion of 5mC to 5hmC in HepG2 cells under glucagon treatment. We found that glucagon treatment promoted TET2 binding to the FBP1 promoter in HepG2 cells, which increased 5hmC levels and reduced 5mC levels (Fig. 3L-N). Importantly, Tet2 KO blocked this process and caused the abnormal accumulation of 5mC in the FBP1 promoter (Fig. 3L-N). These data demonstrate that TET2 mediates the transcriptional activation of FBP1 in response to glucagon stimulation.

HNF4α is required for TET2 mediated transcriptional activation of FBP1

Given that TET2 binds to DNA without sequence specificity, we sought to understand how TET2 specifically activates FBP1 expression upon glucagon treatment. Recent genome-wide methylation and transcriptome analyses identified hepatocyte nuclear factor 4 alpha (HNF4α) as a master gluconeogenic transcription factor, playing a critical role in the pathogenesis of diabetic hyperglycemia (15). More importantly, ChIP-seq data suggested that HNF4α binds to the FBP1 promoter and participates in the regulation of gluconeogenesis in adult mouse hepatocytes (16, 17). To determine whether HNF4α is involved in TET2-mediated FBP1 expression, we performed immunofluorescence to examine the colocalization of TET2 and HNF4α with or without glucagon treatment in HepG2 cells. The results showed that the colocalization of TET2 and HNF4α significantly increased under glucagon treatment (Fig. 4A). Consistently, we observed that glucagon increased the interaction between TET2 and HNF4α in HepG2 cells (Fig. 4B). Furthermore, we knocked down HNF4α in HepG2 cells using siRNA (Fig. 4C), which inhibited glucagon-induced TET2 binding to the FBP1 promoter (Fig. 4D), thereby suppressing TET2-mediated FBP1 expression (Fig. 4E) and impairing glucose output under glucagon treatment (Fig. 4F). Notably, the expression of both HNF4α and FBP1 also increased in mouse livers under fasting or HFD treatment compared to the control group (Fig. 4G-I). In conclusion, these results support the notion that TET2-mediated gluconeogenesis is dependent on HNF4α.

Figure 4.

HNF4α phosphorylation affects its binding to TET2 and FBP1 expression

Our results demonstrated that HNF4α recruits TET2 to the FBP1 promoter and activates FBP1 expression through demethylation, playing a crucial role in the regulation of hepatic glucose output. However, it remains unclear whether the HNF4α-TET2-FBP1 axis responds to metformin treatment. Metformin, a first-line antidiabetic drug widely used to treat hyperglycemia in type 2 diabetes, is also a well-established adenosine 5’-monophosphate-activated protein kinase (AMPK) activator (18). Interestingly, one study revealed that AMPK phosphorylates HNF4α at Ser 313, reducing its transcriptional activity (19). Combining these studies with our findings, we wonder whether metformin can affect the HNF4α-TET2-FBP1 axis. To explore the role of metformin-induced AMPK phosphorylation of HNF4α in FBP1 expression, we treated cells with metformin and assessed the interaction between HNF4α and TET2. The results showed that metformin administration impaired HNF4α’s ability to bind to TET2 (Fig. 5A), leading to a significant reduction in TET2 binding to the FBP1 promoter (Fig. 5B). Consistently, metformin treatment significantly decreased the expression level of FBP1 (Fig. 5C). Notably, metformin also induced high levels of HNF4α phosphorylation at Ser 313 (Fig. 5C). To determine the effect of HNF4α phosphorylation on HNF4α-TET2-FBP1 axis, we transfected HepG2 cells with wild type HNF4α and Ser 313 mutants. The results showed that the phosphomimetic mutation (S313D) of HNF4α impaired its ability to bind to TET2 (Fig. 5D), prevented TET2 from binding to the FBP1 promoter (Fig. 5E), and reduced FBP1 expression at both mRNA and protein levels (Fig. 5F). In contrast, the phosphoresistant mutation (S313A) showed higher activity in interacting with TET2, recruiting TET2 to the FBP1 promoter, and activating its expression (Fig. 5D-F). Taken together, these data demonstrate that metformin-mediated HNF4α phosphorylation suppresses FBP1 expression by preventing TET2 recruitment to the FBP1 promoter by HNF4α.

Figure 5.

Targeting TET2 improves the efficacy of metformin in glucose metabolism in vivo

Increased rates of gluconeogenesis, derived from continuously excessive hepatic glucose production (HGP), result in abnormal glucose homeostasis in type 2 diabetes (T2D). Fasting or high-fat diet (HFD)-induced T2D mice showed increased levels of TET2 and FBP1 in hepatocytes (Fig. 1A-E). However, it remains unclear whether knockdown of TET2, in combination with metformin, would decrease glucose production, thereby improving T2D treatment outcomes. HFD-induced diabetic mice were infused with AAV-scr or AAV-siTet2. To confirm the efficiency of Tet2 knockdown in vivo, Tet2 mRNA levels in mouse livers were examined, revealing that TET2 expression significantly decreased upon AAV-siTet2 treatment (Fig. 6A). Notably, fasting blood glucose and insulin levels were markedly lower in siTet2 group than the control group in response to metformin treatment (Fig. 6B and C). PTT performed on HFD mice indicated that TET2 suppression sharply decreased hepatic gluconeogenesis (Fig. 6D), which was consistent with lower protein levels of FBP1 in mouse livers (Fig. 6E). Moreover, GTT and ITT showed that Tet2 knockdown in HFD mice significantly enhanced glucose tolerance and insulin sensitivity compared to AAV-scr-infused mice (Fig. 6F and G). Of note, the beneficial effects of Tet2 knockdown alone were comparable to those of metformin in lowering glucose levels and improving insulin sensitivity from a therapeutic perspective. Interestingly, the combination of metformin and Tet2 knockdown in HFD mice exhibited better glucose-lowering effects and insulin sensitivity than either Tet2 knockdown or metformin alone. Collectively, these data demonstrated that suppressing TET2 synergizes with metformin to lower glucose production and enhance insulin sensitivity by inhibiting FBP1 expression.

Figure 6.

Discussion

Our study revealed the function of TET2 in the regulation of gluconeogenesis. Notably, gluconeogenesis levels in Tet2 knockout HepG2 cells and primary mouse hepatocytes significantly decreased even under glucagon treatment, demonstrating that TET2 is required for glucagon-induced upregulation of glucose output. Our findings link the novel function of TET2 to the gluconeogenic process. However, the role of TET2 in the pathophysiology of type 2 diabetes (T2D) remains unclear and requires further research. Additionally, the mechanisms by which TET2 is upregulated during gluconeogenesis deserve further exploration.

A recent study showed that metformin treatment can decrease liver glucose production by targeting FBP1 (3), suggesting that FBP1 may be a promising target for ameliorating diabetes. However, the regulation of FBP1 is poorly understood, although several studies have reported that the promoter methylation mediates FBP1 expression silencing in cancer (20, 21). Here, we found that TET2 positively regulates FBP1 upon glucagon treatment, indicating that targeting TET2 may be a promising strategy for treating T2D.

We further explored how TET2 specifically regulates FBP1 expression in response to glucagon stimulation. Like most chromatin-modifying enzymes, which require DNA sequence-specific binding proteins, such as DNA transcription factors, to help them be recruited and regulate specific gene expression (22), TET2 also requires binding partners to regulate particular pathways in a context-dependent manner. This is supported by findings that Wilms tumor protein (WT1) (23) and Smad Nuclear Interacting Protein 1 (SNIP1) (24) are indispensable for TET2 to suppress leukemia cell proliferation and regulate the cellular DNA damage response, respectively. Our data demonstrated that TET2 upregulates FBP1 and further increases gluconeogenesis in an HNF4α-dependent manner. Furthermore, metformin-induced HNF4α phosphorylation at Ser 313 impairs its binding ability to TET2 and reduces TET2 recruitment to the FBP1 promoter, leading to decreased FBP1 expression. More importantly, TET2-FBP1 inhibition has a synergistic effect with metformin in HFD mice.

Intriguingly, a clinical investigation study revealed that TET2 mutation occurred more frequently in the diabetes mellitus (DM) group than the non-DM, which suggested TET2 might be connected with insulin resistance (25). Additionally, inactivating mutations of epigenetic regulator TET2 led to metabolic dysfunction including clonal hematopoiesis, aggravate age- and obesity-related insulin resistance in mice (26). For the HNF4α variants, an analysis study using exome sequencing data showed that human genetic variations in HNF4α destroyed the protein structure and function, impaired insulin secretion and reduced sensitivity to insulin, and increased the risk of T2D in individuals (27, 28). Furthermore, it was reported that a point mutation in FBP1 can reduce the efficacy of metformin treatment (3), providing a genetic evidence for the role of TET2-FBP1 axis in the therapeutic effect of metformin. In summary, our findings uncovered a previously unknown function of TET2 in gluconeogenesis. TET2, together with HNF4α, facilitates FBP1 expression by maintaining the hypomethylation of the FBP1 promoter, which leads to an increase in the gluconeogenic process. Thus, targeting the HNF4α-TET2-FBP1 axis may be a potential strategy to lower blood glucose in T2D.

Methods

Cell Culture and Transfection

HepG2, 293T, and primary mouse hepatocytes were cultured in DMEM medium supplemented with 10% fetal bovine serum (FBS) and incubated at 37°C in a 5% CO2 atmosphere. For transfection with Tet2 plasmids or siRNA, FuGENE HD (Roche) and Lipofectamine 2000 (Invitrogen) were utilized respectively. For lentivirus production, polyethyleneimine (PEI) from EZ Trans, Liji Shengwu, was the preferred agent. All transfection procedures were conducted according to the manufacturers’ instructions. The sequences of HNF4α siRNA were: 5’-AAUGUAGUCAUUGCCUAGGTT-3’ and 5’-UCUUGUCUUUGUCCACCACTT-3’ (21).

Isolation of Primary Mouse Hepatocytes

Primary mouse hepatocytes were isolated by perfusing the liver with pre-warmed Hank’s Balanced Salt Solution (HBSS) at a flow rate of 5-7 mL/minute for 5 minutes after anesthetizing the mouse. Once the liver appeared pale, the perfusion solution was switched to a digestion buffer, maintained at the same flow rate and duration. The liver was then transferred to a 10 cm dish containing 10 mL of digestion medium. Hepatocytes were released by gently cutting and agitating the liver, after which 20 mL of cold PBS was added to halt the trypsin digestion. The cell suspension was filtered through a 70 μm nylon mesh (BD Falcon), centrifuged at 50g for 2 minutes, and the cell pellet was washed twice with cold PBS at 50g for 5 minutes each. The hepatocytes were then ready for subsequent experiments.

qPCR and ChIP-qPCR

Total RNA was extracted using an RNA purification kit (B0004D, EZBioscience). Real-time PCR was conducted using SYBR Green following cDNA synthesis. β-actin was used as an internal control. Chromatin immunoprecipitation (ChIP) assays (Millipore, 17-408) were carried out per the manufacturer’s instructions. Chromatin was sheared by sonication at 4°C. Two micrograms of TET2 antibody (Millipore, MABE 462) were employed to immunoprecipitate chromatin fragments. The ChIP-enriched DNA was subsequently analyzed by qPCR. Primer sequences are provided in the supplementary data.

Western Blot and Immunoprecipitation

Cell lysates were prepared using NP40 lysis buffer. Proteins were separated by SDS-PAGE. The following primary antibodies were used: Tubulin (Proteintech, 66031-1-Ig), TET2 (CST, 18950S), HNF4α (SAB, 32591), FBP1 (Sigma, HPA005857), and HA (Shanghai Genomics Technology). Secondary antibodies included anti-rabbit (SAB, L3012) and anti-mouse (SAB, L3032). For immunoprecipitation (10), total protein was incubated with protein A beads (Shanghai Genomics Technology) and TET2 antibody (Millipore, MABE 462) for 3 hours at 4°C, followed by western blot analysis.

Gene Knockout

The CRISPR/Cas9 system was utilized to generate Tet2 knockout cells. Lentiviruses were produced by co-transfecting HEK293T cells with plasmids containing sgRNA (8μg), psPAX2 (6μg), and pMD2.G (2μg) in a 10 cm dish. The collected supernatant was used to infect the target cells. Following selection with puromycin, Western blot analysis was conducted to assess the efficiency of the knockout. The targeting sequence for TET2 sgRNA was 5’-GATTCCGCTTGGTGAAAACG-3’.

Immunofluorescence

For immunofluorescence, antibodies against TET2 and HNF4α were combined and applied to the slides, which were then incubated at 4°C overnight. Subsequently, the slides were stained with secondary antibodies (Alexa Fluor 555 and 488), followed by DAPI staining. Confocal images were captured using a Leica fluorescence microscope.

Glucose Production

Cells were washed three times with PBS before the medium was replaced with glucose-free DMEM (Gibco, A14430-01) supplemented with 20 mM sodium lactate and 2 mM sodium pyruvate. After an 8-hour incubation, glucose levels in the supernatants were measured using a glucose assay kit (ThermoFisher, A22189).

Pyruvate Tolerance Test (PTT), Glucose Tolerance Test (GTT), and Insulin Tolerance Test (ITT)

For the PTT and GTT, mice underwent fasting for 16 hours and 12 hours, respectively, before receiving an intraperitoneal (i.p.) injection of either pyruvate (2g/kg) or glucose (2g/kg). In the ITT, mice fasted for 6 hours in the morning prior to receiving an insulin injection (1U/kg, i.p.). Blood glucose levels were measured at intervals of 0, 15, 30, 45, 60, 90, and 120 minutes after injection using a glucometer during tail vein bleeding.

Animal Study Protocols

All experimental protocols involving animals were evaluated and approved by the Fudan University Animal Care and Use Committee and conducted according to the guidelines. Male C57BL/6J mice, aged 6-8 weeks, were utilized for this study. These mice were housed in specific-pathogen-free (SPF) conditions at 25°C with a 12-hour light/dark cycle and were fed either a normal chow or a high-fat diet (HFD) (Research Diets, 45% calories from fat). To induce insulin resistance, wild-type C57BL/6J mice were placed on a HFD for 2 weeks. For the experiments with AAV-scr or AAV-siTet2, HFD mice were infected with the respective virus for 10 days.

Statistical Analysis

The Student’s t-test was used for comparing two groups, while one-way ANOVA followed by Tukey’s post hoc test was conducted for multiple comparisons when more than two groups were analyzed. Data analysis was performed using SPSS or GraphPad Prism. A p-value of less than 0.05 was considered statistically significant.

Acknowledgements

This work was supported by the National Key R&D Program of China (2020YFA0803400/2020YFA0803402, 2022YFA0807100), the National Natural Science Foundation of China (82172936, 81972620, 82121004, 82372754 and 82073128) and the Fundamental Research Funds for the Central Universities.

Competing interests

The authors declare no competing interests.