DNA2 flox/flox mice (DNA2^fl/fl) were kindly provided by Professor Weiqiang Lin (Zhejiang University). β cell-specific DNA2 knockout (DNA2^INS2−/−) mice were generated by crossing DNA2^fl/fl mice with Ins2-Cre mice (Stock No. 003573). The knockout strategy is shown in Supplementary Figure S1A, and genotyping was performed by PCR on tail and islet DNA. Age-matched littermates were used in all experiments. Mice were housed under specific pathogen-free conditions with controlled temperature, humidity, a 12-h light/dark cycle, and ad libitum access to chow and water. Male DNA2^INS2−/− and DNA2^fl/fl littermates were 6–8 weeks of age at the initiation of the experimental protocol. Age- and weight-matched male DNA2^INS2−/− and DNA2^fl/fl littermates were fed CD (#1025; Beijing Hua Fu Kang) or HFD (60% kcal fat, #D12492; Research Diets) for ≥16 weeks to establish diet-induced obesity. Body weight was measured weekly throughout the experimental period. Food intake was assessed during weeks 13–16 of HFD feeding by measuring weekly food consumption per cage. All procedures were approved by the Institutional Animal Care and Use Committee of Sichuan University.

Genomic DNA was extracted from the mouse tail or islet biopsies for genotyping. The presence of the Dna2 floxed allele was determined by PCR using the following primers: 5′-AGT​TTT​AGC​TAG​CCG​GGA​ACT​C-3′ (Dna2_F2), 5′-CGC​TGC​CTT​AGT​TCT​TTG​GAT​A-3′ (Dna2_R2) (512 bp, WT; 625 bp, flox, Neo-deleted). The presence of the Ins2-Cre transgene was determined by PCR using the following primers: 5′-CAT​ATT​GGC​AGA​ACG​AAA​ACG​C-3′ (Cre_F); 5′-CCT​GTT​TCA​CTA​TCC​AGG​TTA​CGG-3′ (Cre_R) (413 bp, Cre positive). The identification of DNA2^INS2−/− mice was determined by PCR using the following primers: 5′-CTA​ATG​CGG​GAT​CCA​GGA​CTC​AAA​C-3’ (Dna2_F1), 5′-CGC​TGC​CTT​AGT​TCT​TTG​GAT​A-3’ (Dna2_R2) (210 bp).

INS-1 cells (the gift from Prof. Chunbo Teng, Northeast Forestry University) were maintained in RPMI 1640 supplemented with 10% FBS, 1 mM sodium pyruvate, 2 mM glutamine, 50 μM β-mercaptoethanol, 10 mM HEPES, and 1% P/S. Min6 cells were grown in RPMI 1640 containing 10% FBS, 2% B27, 2 mM glutamine, 50 μM β-mercaptoethanol, and 1% P/S. 293T cells (ATCC CRL3216) were cultured in DMEM with 10% FBS and 1% P/S.

For siRNA knockdown, cells were transfected with 50 nM siRNA using HiPerFect (#301705; QIAGEN). Cells were harvested 24 h after transfection for gene expression analysis and 48 h after transfection for functional assays. Two independent siRNAs targeting murine DNA2 (mDNA2) (5′-ATG​GCA​GGT​GAC​AGG​ATT​ATT-3′ and 5′-ACG​CTG​GAG​TCG​CAA​TCT​AAA-3′) and a non-targeting control (5′-UUC​UCC​GAA​CGU​GUC​ACG​UTT-3′) were used.

Murine DNA2 cDNA was cloned into pcDNA3.1 (mDNA2^WT). Point mutants (mDNA2^D278A, mDNA2^K655E, and mDNA2^D278A+K655E) were generated by site-directed mutagenesis (primers in Supplementary Table S1). Plasmids were transfected using Attractene (#1051563; QIAGEN) following the manufacturer’s instructions.

Islets were isolated by collagenase XI (#C7657; Sigma) digestion as previously described (Xu et al., 2020) with slight modifications. Hand-picked islets were cultured in RPMI 1640 supplemented with 10% FBS.

Blood glucose and insulin were measured as described (Xu et al., 2020). Random glucose was assessed in fed mice; fasting glucose after 16 h fasting. Glucose was measured using a portable glucometer (Abbott Laboratories). Serum insulin was quantified with the Ultra-Sensitive Mouse Insulin ELISA Kit (#90080; Crystal Chem). The formula for calculating the homeostasis model assessment of insulin resistance (HOMA-IR) index, is fasting blood glucose (mmol/L) multiplied by fasting blood insulin (μU/mL) and then divided by 22.5.

For GTT, mice were fasted for 16 h, injected intraperitoneally with D-glucose (#G7021; Sigma), and glucose levels measured at 0, 15, 30, 60, 90, and 120 min. For ITT, mice were fasted for 6 h, injected with human insulin (#HI0219; Lilly), and glucose recorded at 0, 15, 30, 60, and 90 min (Xu et al., 2019).

Paraffin-embedded pancreatic sections were deparaffinized, rehydrated, and subjected to antigen retrieval in 10 mM Tris-EDTA (pH 9.0). Sections were blocked with PBS containing 1% BSA and 5% goat serum. For IF, anti-Insulin antibody (#ab181547, Abcam) and Alexa Fluor® 488 Anti-Glucagon antibody (#ab307340, Abcam) were applied overnight at 4 °C, followed by Alexa Fluor 594-conjugated goat anti-rabbit IgG (#A-11012, Invitrogen). Images were captured on a confocal microscope (LSM 880; Zeiss). For IHC, sections were incubated with anti-insulin antibody (#ab181547, Abcam), followed by HRP-conjugated secondary antibody (#31460, Invitrogen) and DAB substrate (#34002, Invitrogen), then counterstained with hematoxylin. ImageJ (v1.54) was used for analysis.

Pancreatic islets were hand-picked after collagenase digestion and immediately preserved in RNAlater (#76104; Qiagen) before RNA isolation with TRIzol (#15596026; Thermo Fisher). cDNA was synthesized using M-MLV Reverse Transcriptase (#28025013; Invitrogen). qRT-PCR was performed in 10 μL reactions with SYBR Green Master Mix (#204145; Qiagen) and specific primers (listed in Supplementary Table S2).

Hand-picked mouse islets were cultured overnight and preincubated for 30 min in Krebs buffer containing 2.8 mM glucose, followed by stimulation with 2.8 mM glucose in the presence or absence of secretagogues (16.7 mM glucose, 20 mM arginine, or 30 mM KCl) for 1 h. Culture supernatants were collected for insulin secretion analysis. Total insulin content was extracted from the same islets using acid–ethanol solution. Insulin concentrations were measured using an Ultra-Sensitive Mouse Insulin ELISA Kit (#90080; Crystal Chem) according to the manufacturer’s instructions, and insulin secretion was normalized to total insulin content.

INS-1 cells (10,000 cells/well) were seeded in XFp plates, transfected for 48 h, and incubated 1 h in assay medium (16.7 mM glucose, 2 mM glutamine, 1 mM pyruvate). Oxygen consumption was measured with the XF Extracellular Flux Analyzer (Seahorse), with sequential injections of oligomycin (4 μM), FCCP (3 μM), antimycin A (0.5 μM), and rotenone (0.5 μM).

Fresh liver, epididymal white adipose tissue (eWAT), brown adipose tissue (BAT), and pancreas were fixed in 10% formalin, paraffin-embedded, sectioned (4 μm), and stained with H&E (Xu et al., 2020). Adipocyte cell size was quantified using the open-source ImageJ software (v1.54), with the AdipoQ plugin (Sie et al., 2022), and lipid-positive areas in Oil Red O–stained liver sections were quantified using ImageJ.

For islet quantification, pancreas was dissected, weighted, fixed, embedded in paraffin, and consecutively cut into 4-μm–thick sections. Sections from the pancreatic tail were stained with H&E, and both islet area and total tissue area were measured using ImageJ software (v1.54).

Fresh liver samples were embedded in optimal cutting temperature compound (#4583; Sakura Finetek), snap-frozen at −80 °C, and sectioned at 8 μm using a cryostat (Leica CM 1950). Sections were fixed in 4% paraformaldehyde, stained with Oil Red O working solution (0.5% in isopropanol, diluted 3:2 with water), washed in 60% isopropanol, counterstained with hematoxylin, and imaged (Olympus VS200).

Cells were stained with 200 nM Mito-Tracker Red CMXRos (C1035; Beyotime) and Hoechst 33342 (C10310-3; RiboBio) per manufacturer’s instructions. Images were acquired with a Zeiss LSM 880 confocal microscope, and signal intensity was quantified in ImageJ (v1.54).

Liver samples (50 mg) were homogenized in ethanol, centrifuged, and supernatants analyzed with commercial kits (Triglyceride, GPO-PAP; Cholesterol, COD-PAP; Nanjing Jiancheng). Absorbance was read at 510 nm (BioTek Synergy H1; Agilent), and values were normalized to protein content.

Isolated islets were cultured overnight, fixed in 3% glutaraldehyde at 4 °C for >24 h, post-fixed in 1% osmium tetroxide, dehydrated through graded acetone, and embedded in Epon812. Ultrathin sections (∼60 nm) were examined with a JEM-1400FLASH TEM at 6,000× magnification. Image acquisition and mitochondrial morphology analyses were performed in ImageJ (v1.54).

RNA-seq was performed by Novogene (Beijing, China) using islets from DNA2^INS2−/− and DNA2^fl/fl mice after 5 days of HFD feeding. RNA quality was assessed with an Agilent 5400 Fragment Analyzer. Libraries were prepared with the NEBNext® Ultra™ RNA Library Prep Kit and sequenced on an Illumina NovaSeq 6000 (150 bp paired-end). Reads were quality-checked (FastQC), trimmed (Trimmomatic), and aligned to the mouse reference genome (GRCm39) with HISAT2. Gene counts were obtained with featureCounts, differentially expressed genes (DEGs) were identified with DESeq2 (R). Genes with a p value ≤0.05 and an absolute fold change ≥1.5 were defined as DEGs, while a more stringent cutoff (|fold change| ≥ 2, p ≤ 0.05) was applied for selected downstream analyses. KEGG pathway enrichment was analyzed and visualized with ggplot2.

Data are presented as the mean ± SD from at least three independent experiments. Statistical analysis was performed with GraphPad Prism 10. Group comparisons were carried out using unpaired two-tailed Student’s t-test, two-way ANOVA for multi-factorial analyses, or the Kruskal–Wallis test for non-parametric data. Statistical significance was defined as p ≤ 0.05.

To assess the physiological role of DNA2 in β cell biology, we first examined its expression in islets under insulin-resistant stress. qPCR revealed a significant downregulation of DNA2 mRNA in pancreatic islets isolated from mice subjected to prolonged HFD feeding than that in islets from controls (Figure 1A), indicating a pathogenic role of DNA2 in β cell dysfunction and diabetes.

To test this idea, we generated a β cell-specific DNA2 knockout mouse model (DNA2^INS2−/−) using Ins2-Cre-driven recombinase through homologous recombination-mediated genome editing (Supplementary Figure S1A). As expected, DNA2 expression was specifically downregulated in the islets of DNA2^INS2−/− mice, with no detectable alterations in other tissues, such as the thymus or hypothalamus (Supplementary Figure S1). DNA2^INS2−/− mice were born at Mendelian frequencies and appeared morphologically indistinguishable from their control littermates (DNA2^fl/fl). Under standard CD conditions, DNA2^INS2−/− mice exhibited no significant differences in body weight, organ-to-body weight ratios, or tissue morphology compared with that of DNA2^fl/fl control mice (Figures 1B–F). Interestingly, DNA2^INS2−/− mice had an increased pancreatic islet area, although the difference was not statistically significant (Figures 1G,H). IF microscopy confirmed that β cell-specific DNA2 deletion did not affect the pancreatic islet architecture (Figure 1I). These results indicate that DNA2 deletion in β cells does not affect normal growth or development.

At 14 and 25 weeks of age, blood glucose levels were not significantly different between the two genotypes under either random or fasting conditions (Figures 1J,M). GTT and ITT analyses showed no significant differences in glucose homeostasis between the groups (Figures 1K,L,N,O). Plasma metabolic profiling confirmed that DNA2 deletion in β cells had no significant effect on circulating metabolic parameters, including albumin, urea, creatinine, triglycerides, low-density lipoprotein (LDL), high-density lipoprotein (HDL), and total cholesterol (T-cholesterol) levels (Figures 1P,Q). These data, together with the increased glucose-stimulated insulin secretion observed in vivo, suggest that β cell DNA2 deletion may enhance acute insulin secretion without perturbing systemic metabolic homeostasis under physiological conditions.

We assessed the effects of β cell-specific DNA2 deletion on HFD-induced metabolic disturbances. Before HFD feeding, DNA2^INS2−/− mice exhibited comparable body weight, glycemic levels, and glucose tolerance to those of their littermate controls (Supplementary Figure S2). Upon HFD challenge, DNA2^INS2−/− mice gained significantly more body weight than that of DNA2^fl/fl mice, despite similar food intake (Figures 2A–C), indicating an increased susceptibility to diet-induced obesity. Additionally, these mice demonstrated increased fasting blood glucose levels at both 8 and 16 weeks of HFD feeding (Figure 2D; Supplementary Figure S3A). Moreover, DNA2^INS2−/− mice exhibited impaired glucose tolerance and insulin sensitivity, as assessed by GTT and ITT, respectively (Figures 2E–H; Supplementary Figures S3B–E). Notably, the plasma insulin levels were significantly increased in DNA2^INS2−/− mice both under fasting and glucose-inducing conditions (Figure 2I; Supplementary Figure S3F), accompanied by significantly higher homeostasis model assessment of insulin resistance (HOMA-IR) indices (Figure 2J; Supplementary Figure S3G). These findings indicate that β cell-specific deletion of DNA2 exacerbates glucose intolerance, hyperinsulinemia, insulin resistance, and overall metabolic deterioration in HFD-induced obesity.

Because DNA2^INS2−/− mice exhibit exacerbated obesity and insulin resistance, we assessed whether lipid metabolism and distribution were similarly perturbed. Organ analysis demonstrated that HFD-fed DNA2^INS2−/− mice had significantly increased liver, spleen, pancreas, and brown adipose tissue (BAT) weights compared to that of the controls (Figure 2K). H&E and Oil Red O staining revealed significant lipid droplet accumulation in the livers of DNA2^INS2−/− mice, indicating severe hepatic steatosis (Figures 2L–N). Quantitative lipid analysis confirmed increased hepatic triglyceride and cholesterol content in DNA2^INS2−/− mice (Figure 2O). Consistently, plasma lipid profiling revealed increased levels of LDL, HDL, and T-cholesterol in DNA2^INS2−/− mice (Figure 2P), indicating systemic dysregulation of lipid metabolism. Furthermore, BAT from DNA2^INS2−/− mice displayed marked whitening, evidenced by excessive lipid deposition and altered cellular morphology relative to DNA2^fl/fl controls, whereas eWAT remained largely comparable between the two groups (Figures 2Q–T). These findings demonstrate that β cell-specific DNA2 deletion induces ectopic lipid deposition and impairs systemic lipid homeostasis under HFD conditions.

Circulating insulin levels are controlled by a combination of the number of β cells and the insulin secretory capacity of individual β cells. We first determined the effect of DNA2 deletion on β cell mass. We next examined β cell mass and proliferation to investigate whether these functional changes were accompanied by alterations in islet morphology. DNA2^INS2−/− mice consistently exhibited a greater number and larger size of islets compared to that of their littermate controls (Figure 3A). H&E staining and insulin IHC revealed that HFD-induced islet hyperplasia was significantly more pronounced in DNA2^INS2−/− mice (Figures 3B–D). Obesity-induced hyperplasia is commonly associated with increased β cell proliferation (Hu et al., 2022; Kubo et al., 2025; Chen et al., 2025). Therefore, we assessed the expression of hypertrophy- and proliferation-related genes in the isolated islets. Notably, DNA2^INS2−/− islets exhibited significantly increased expression of proliferation markers (Kiel 67 antigen [Ki67], cyclin D2 [Ccnd2], and neurogenin 3 [Ngn3]) following HFD feeding, with only modest increases observed under CD conditions (Figures 3E,F). Consistent with this in vivo phenotype, DNA2 knockdown significantly enhanced the proliferation of Min6 β cells (Figure 3G). Similarly, HFD-fed DNA2^INS2−/− islets significantly upregulated hypertrophy-associated genes (Akt1 and Irs2), whereas under CD conditions, only an increase in Akt1 expression was observed (Figures 3H,I).

These findings indicate that DNA2 acts as a negative regulator of β cell proliferation, and its deficiency facilitates obesity-associated islet expansion, potentially serving as a compensatory response to meet the increased metabolic demand under nutritional stress.

We next evaluated the consequences of DNA2 deficiency on β cell functionality, specifically insulin biosynthesis and secretion. Remarkably, despite the presence of hyperglycemia and elevated circulating insulin in DNA2^INS2−/− mice (Figure 2I; Supplementary Figure S3F), there was no significant alteration in the islet mRNA expression of insulin (Ins-1 and Ins-2) under either CD or HFD conditions (Figures 4A,B). Consistently, ultrastructural analysis demonstrated comparable morphology and abundance of insulin granules between the genotypes (Figure 4C), and the total insulin content per islet mass was unaffected (Figure 4D). Moreover, IF imaging confirmed the preserved islet architecture and α/β cell organization (Figure 4E), indicating that insulin biosynthesis and storage were intact.

Subsequently, we assessed the ex vivo insulin secretory capacity of the isolated islets. Following chronic HFD exposure, both glucose- and KCl-stimulated insulin release were significantly augmented in DNA2-deficient islets compared with that in controls (Figure 4F). Dynamic perfusion analysis revealed that the first and second phases of insulin secretion were amplified in DNA2-depleted islets (Figure 4G). These observations were recapitulated under physiological dietary conditions (Figures 4H,I). Extending these findings to the INS-1 β cell line, siRNA-mediated DNA2 knockdown significantly potentiated insulin secretion, whereas transient overexpression attenuated it (Figures 4J,K). In summary, β cell DNA2 depletion consistently hypersensitizes stimulus-secretion coupling across experimental models without altering insulin biosynthesis and content.

To assess the transcriptional basis underlying the secretory phenotype and capture early molecular responses associated with DNA2 deficiency, we performed transcriptomic profiling to explore the alternative pathways. RNA-seq of islets harvested after 5 days of HFD exposure—a stage that precedes detectable β cell loss—identified 251 DEGs (|fold change| ≥ 1.5, p ≤ 0.05) in DNA2^INS2−/− islets compared to those in DNA2^fl/fl controls. Of these, 99 met a more stringent threshold (|fold change| ≥ 2, p ≤ 0.05) (Figures 5A,D). Subsequent KEGG enrichment analysis of these 99 DEGs revealed that DNA2 deletion was associated with stimulus-secretion-coupling pathways, including insulin secretion, hormonal signaling, and focal adhesion (Figures 5B,C), indicating a crucial role in exocytosis. Notably, five DEGs—lectin, galactoside-binding, soluble 3 (Lgals3), estrogen receptor 1 (Esr1), NADH: ubiquinone oxidoreductase subunit A4-like 2 (Ndufa4l2), vimentin, and matrix metallopeptidase 12 (Mmp12)—were implicated in mitochondrial regulation (Li et al., 2022; Huynh et al., 2024; Kubala et al., 2023; Coppin et al., 2020) (Figure 5E), facilitating further assessment of mitochondrial function in DNA2-deficient β cells.

Because mitochondrial ATP production is crucial for stimulus-secretion coupling, we assessed mitochondrial performance in both islets and INS-1 cells to determine whether DNA2 affects β cell bioenergetics. Glucose-stimulated ATP production was significantly increased in DNA2-deficient islets and INS-1 cells (Figures 6A,B), whereas DNA2 overexpression reduced ATP production (Figure 6Q). Mitochondrial stress tests using Seahorse XF analysis revealed that DNA2-overexpressing INS-1 cells exhibited reduced basal and maximal respiration, along with reduced spare respiratory capacity (Figures 6C–G), indicating compromised oxidative phosphorylation. TEM of pancreatic β cells from DNA2^INS2−/− mice revealed significant mitochondrial hypertrophy, reflected by increased mitochondrial perimeter, aspect ratio, and cross-sectional area (Figures 6H–N), although the mitochondrial density remained unaltered (Figure 6O). Similarly, Mito-Tracker fluorescence intensity was increased in DNA2-deficient INS-1 cells and reduced upon DNA2 overexpression (Figures 6P,R), confirming that DNA2 loss enhanced mitochondrial mass and activity.

To elucidate the enzymatic basis of this regulation, we focused on two catalytic sites of DNA2: the nuclease site D278 and helicase site K655 (Levikova et al., 2013; Pinto et al., 2016; Shen et al., 2022). We generated single-point mutants (mDNA2^D278A and mDNA2^K655A) and a double mutant (mDNA2^D278A+K655A) and expressed them in INS-1 cells to assess the enzymatic activity of each. Both single mutants partially alleviated the inhibitory effects of wild-type DNA2 (mDNA2^WT) overexpression on glucose-stimulated ATP production and mitochondrial activity, whereas the double mutant exerted a more significant rescue effect across all parameters (Figures 6Q,R), highlighting the cooperative requirement for both catalytic domains.

In summary, our findings position DNA2 as a molecular gatekeeper of β cell mitochondrial production. Through its coordinated nuclease and helicase activities, DNA2 limits mitochondrial size and activity, thereby restraining ATP production and insulin secretion in response to glucose. In contrast, DNA2 deficiency results in mitochondrial hypertrophy, hyperactivation, and insulin hypersecretion. These data reveal a DNA2-mitochondria-insulin axis that integrates genome maintenance with metabolic plasticity in β cells under nutritional stress.