β-Cell miRNA-503-5p Induced by Hypomethylation and Inflammation Promotes Insulin Resistance and β-Cell Decompensation Free

Type 2 diabetes is a metabolic disorder characterized by chronic hyperglycemia due to absolute or relative insulin insufficiency. Upon feeding, secreted insulin binds to insulin receptors and activates the phosphoinositide 3-kinase/AKT signaling pathway to maintain postprandial glucose homeostasis (1). The insulin signaling cascade inhibits hepatic glucose production (HGP) and stimulates glucose uptake in the adipose tissue and skeletal muscle. Defects in insulin action and secretion, referring to insulin resistance and β-cell dysfunction, can give rise to insulin signaling disturbances and reciprocally promote diabetes progression (2). However, these defects coexist in the majority of patients with type 2 diabetes, so no concerted insights have yet been reached regarding the cause-and-effect links between them.

Insulin resistance persists for many years before the appearance of frank diabetes, and β-cells adapt to insulin resistance via hypersecretion (3). However, β-cell compensation gradually degenerates to decompensation in response to islet inflammation and the resulting production of proinflammatory factors, principally interleukin 1β (IL-1β) (4). Reports have shown that anti-IL-1β approaches ameliorate tissue inflammation and diabetes in rodents (5). Unfortunately, in the clinical setting, the protective effects only manifest as improved β-cell function in patients with type 2 diabetes, suggesting that the long-term hyperglycemia phenotype may not exclusively involve insulin insufficiency (6). Instead, compensatory remodeling of β-cell mass and function may initiate other epigenetic processes that cannot be disrupted by interrupting IL-1β signaling. Indeed, epigenetic cues, such as DNA methylation, histone modification, and noncoding RNAs (ncRNAs), have comprehensive impacts on the remodeling of metabolically stressed β-cells in the context of chronic inflammation (7,8).

One subset of ncRNAs of particular interest in terms of diabetes, inflammation, and aging-related disorders are the miRNAs. These are endogenous small ncRNAs (∼22 nucleotides) that play important roles in virtually all aspects of biological processes (9). The miRNAs are transcribed as primary miRNAs (pri-miRNAs), which undergo a two-step cleavage by the endoribonucleases Drosha and Dicer to form the mature miRNA. The mature miRNA assembles into the RNA-induced silencing complex and activates the complex, which then targets mRNA for translational inhibition and/or mRNA degradation (9). Numerous studies have demonstrated that miRNAs can be delivered into the bloodstream, acting as hormone-like molecules to facilitate crosstalk among key metabolic organs under obesity and diabetes conditions (10,11). We have discovered that miR-29 exosomes derived from β-cells are regulated by IL-1β and promote the development of diabetes by facilitating monocyte/macrophage activation (12). However, an elevation of miR-29 is observed in almost all metabolic tissues, so whether IL-1β regulates the β-cell–dominant miRNAs linking islet inflammation to the development of type 2 diabetes requires further investigation.

In the current study, we identified miR-503-5p as a master regulator of β-cell decompensation in IL-1β–induced insulitis. Highly expressed miR-503-5p activates c-Jun N-terminal kinase (JNK) mitogen-activated protein kinase (MAPK) and p38 MAPK to decrease glucose-stimulated second phase insulin secretion by targeting JNK-interacting protein 2 (JIP2) within β-cells. Moreover, a probable marker of nanovesicles located within insulin secretory granules indicates the release of miR-503-5p, which has been proven to trigger insulin resistance. The X-linked miR-503 is clustered with miR-424(322) and serves as an intracellular gene regulator to modulate fundamental processes that include cell proliferation, cell differentiation, and tissue remodeling (13). Our investigation identified a previously unappreciated role of miR-503-5p as an intra- and interorgan modulator and defined its contribution to the development of type 2 diabetes using gain-of-function and loss-of-function genetic mouse models, as well as primary human islets.

Blood samples were provided by the Department of Endocrinology, Geriatric Hospital of Nanjing Medical University. A total of 160 individuals were recruited, including 52 healthy adults, 58 elderly individuals, and 50 elderly patients with type 2 diabetes. Fasting blood samples were centrifuged at 3,000 rpm for 20 min to separate serum and blood cells. The serum was used for miR-503-5p content analysis, while the blood cells were used for DNA extraction and methylation-specific PCR analysis. Detailed information of donors is listed in Supplementary Table 1. The primers used for the methylation-specific PCR analysis are listed in Supplementary Table 2. The study was approved by the research ethics committee of Nanjing Medical University (2022006), and all the volunteers gave written informed consent.

Human islets were provided by Tianjin First Central Hospital. Detailed information of donors is listed in Supplementary Table 3. The use of human islets was approved by the research ethics committee of Tianjin First Central Hospital (2018N112KY). Murine islets were isolated as described previously (12).

β-Cell–specific miR-503-322-351 transgenic mice (Cluster TG), miR-503 transgenic mice (βTG), and miR-503-322-351 global deletion mice (knockout [KO]) were generated by GemPharmatech Co., Ltd. (Nanjing, China). Female nonobese diabetic (NOD) mice, male C57BL/6J mice, db/db-C57BLKS (db/db) mice, and their littermates were purchased from GemPharmatech.

All mice were housed in a 12-h light/dark cycle at 23–25°C, and animal studies were approved by the research animal care committee of Nanjing Medical University (IACUC-1707023). KO and βTG mice were fed a high-fat diet (HFD) (60% fat, D12492; Research Diets) at 3–4 weeks old.

Rat insulinoma INS-1 cells (CM-1421; ATCC) were cultured in RPMI media (Invitrogen, Grand Island, NY) containing 10% FBS (12483020; Gibco). Mouse insulinoma MIN6 cells (4) were cultured in DMEM with 15% FBS. Both media were supplemented with 100 mg/mL streptomycin, 100 units/mL penicillin, 10 mmol/L HEPES, and 50 μmol/L β-mercaptoethanol (M6250; Sigma-Aldrich). Trichostatin A (TSA) (S1045; Selleck) and 5-azacytidine (5-Aza) (A1907; APExBIO) were added as indicated.

The mouse 3T3-L1 cells were cultured and differentiated as described previously (14). The primary hepatocytes were isolated from 8-week-old C57BL/6J male mice and cultured as described previously (15). All cells were cultured at 37°C in a humidified atmosphere containing 95% air and 5% CO₂.

The metabolic cage experiments were performed in the Jiangsu Laboratory Animal Center. For glucose metabolism analysis, mice were fasted for 14–16 h, then fasting blood glucose (FBG) levels or intraperitoneal glucose tolerance test (IPGTT) (1 g/kg body weight i.p.) were performed. Intraperitoneal insulin tolerance test (IPITT) (1 unit insulin/kg body weight i.p.) and in vivo glucose-stimulated insulin secretion (GSIS) (3 g/kg body weight i.p.) were performed as previously described (12). Mice were fasted 4–6 h for IPITT and 12 h for in vivo GSIS. Blood samples were collected from the tail vein.

The 16-week-old βTG mice and HFD-fed KO mice underwent hyperinsulinemic-euglycemic clamps. The intravenous catheterization surgery was performed 1 week before clamp, as described previously. [3-³H]-d-Glucose (3 μCi; Moravek) was given to mice at −90 min of clamp, with an infusion rate of 0.05 μCi/min. A continuous infusion of extraneous insulin (4 mU/kg/min) was started at 0 min to keep the hyperinsulinemic condition. At 75 min of clamp, 10 μCi 2-[¹⁴C]-d-glucose (Moravek) was then administered. Terminally, serum, liver, muscle, and adipose tissues were collected for the determination of radioactivity, the radioactivity was measured and calculated as previously described (16, 17).

Adeno-associated virus serotype 8 (AAV8)-mouse insulin promoter 1 (MIP1)-miR-503-322-351 sponge (βsponge) and control AAV8-MIP1-GFP were constructed by using MIP1-driven expression to implement β-cell specificity. The βsponge sequence was synthesized and inserted to pAAV2/8, and the AAV8 viruses were generated and purified by HanBio biotechnology (Shanghai, China). pAd-JIP2 plasmid was constructed by inserting PCR-amplified mouse Jip2 cDNA into SalI and HindIII sites of pAdTrack-CMV plasmid, and the primary and amplified adenoviruses were purified by HanBio Therapeutics.

Mouse pancreata were fixed with 4% paraformaldehyde in PBS, then embedded in paraffin and cut into slices (5 μm). Hematoxylin-eosin (H-E)–stained sections were used for islet mass analysis by ImageJ software (National Institutes of Health, Bethesda, MD). For islet composition analysis, insulin was costained with glucagon, pancreatic polypeptide, or somatostatin. For islet mass loss analysis, insulin was costained with Ki67, proliferating cell nuclear antigen, and TUNEL kit (A111-02; Vazyme). For JIP2 expression analysis, sections were immunohistochemistry stained with JIP2. We used a confocal laser scanning microscopy system (FV1200; Olympus) or optical microscope (Axiovert A1; Zeiss) to capture and analyze the sections. The antibodies used are listed in Supplementary Table 4.

Islet perfusion assay and insulin detection via radioimmunoassay are previously described (17). Serum hormone levels were measured by ELISA kits according to the manufacturer’s instructions and were as follows: insulin (10-1247-01; Mercodia), C-peptide (80-CPTMS-E01; ALPCO Diagnostics), and glucagon (10-1271-01; Mercodia).

Isolated islets were cultured in a serum exosomal-free culture media (11.1 mmol/L glucose) for 1 week, and the media were replaced and collected each day. Media were first centrifuged at 700g for 5 min to pellet cells and then at 10,000g for 1 h to discard cell debris. The sediment was rinsed and resuspended in PBS and centrifuged at 100,000g for 12 h. The resultant nanovesicles were solved in PBS, and some were labeled with PKH26 or PKH67 (Sigma-Aldrich) for in vivo and in vitro tracing.

For identification of nanovesicles, a transmission electron microscope (TEM) (Tecnai Spirit BioTwin), NanoSight particle tracking analysis (NTA), and Western blot analysis were performed. For colocalization of nanovesicles and insulin granules, immunogold labeling of CD63 (5 nm) and insulin (10 nm) was performed and visualized by TEM. The antibodies used are listed in Supplementary Table 4.

For transient transfection, Lipofectamine 2000 reagent (11668-019; Invitrogen) was mixed with miRNA mimics or overexpression/reporter plasmids as previously described (4). The enhanced GFP (EGFP) or mOrange-tagged neuropeptide Y (NPY) adenovirus was diluted with a serum-free medium at a concentration of 2 × 10⁶ plaque-forming units/mL for dissociated islet cells and MIN6. INS-1 cells (1 × 10³ per well) were transfected with indicated miRNAs, treated with IL-1β, and then added with CCK-8 (Beyotime, Shanghai, China) for cell viability measurement according to the manufacturer’s instructions.

The wild-type (WT) and mutant 3′ untranslated region-luciferase constructs of human JIP2, MEK1, and INSR were generated by annealing and cloning the short sequences into pMIR-REPORT vector (Ambion, Foster City, CA) between the SpeI and HindIII sites. Primer sequences are listed in Supplementary Table 5. Luciferase activities were measured by using the Dual-Glo Luciferase Assay System (Promega, Madison, WI) on a TD-20/20 Luminometer (Turner BioSystems, Sunnyvale, CA) according to the manufacturer’s protocols.

For miRNA microarray, 10⁶ INS-1 cells were treated with 10 ng/mL IL-1β for indicated times, and total RNA was extracted for miRNA microarray (901325, MicroRNA Array 1.0; Affymetrix, Santa Clara, CA). For quantitative PCR (qPCR), total RNA from tissue and cell samples were extracted by using TRIzol (15596026; Invitrogen). cDNA preparation by ReverTra Ace qPCR RT Kit (FSQ-201; Toyobo) and gene expression measurement by THUNDERBIRD Probe qPCR Mix (QPS-101; Toyobo) were described previously (12). Primers for pri-miRNA and miRNA were purchased from Thermo Fisher Scientific.

For absolute PCR, total RNA from mouse sera and nanovesicles were extracted by using TRIzol LS (10296010; Invitrogen). During RNA extraction, 10 μL cel-miR-39 (100 pmol/L) was added in 250-μL liquid samples, and the cel-miR-39 was measured as a systemic control. The reverse transcription and gene expression measurement were the same as above. For absolute calculation, the standard curve of cel-miR-39 was constructed with a 10-fold gradient concentration from 100 fmol/L to 100 nmol/L. Primers were purchased from RiboBio Co., Ltd. (miRB0000010-3-1, MQPS0000071-1-100, MQPS0002912-1-100, MQPS0001690-1-100).

For mass spectrometry, MIN6 cells were washed with ice-cold PBS, collected, and dissolved in lysis buffer (7 mol/L urea, 1% CHAPS). Protein digestion, tandem mass tag labeling, and mass spectrometry analysis then were conducted at the Analysis Center of Nanjing Medical University as previously described (12). Western blotting was performed as previously described (12). The antibodies used are listed in Supplementary Table 6. Stripe intensity was measured by Image.

In vitro experiments were repeated at least three times, and in vivo assays were repeated twice, with the number per condition included in figure legends. Additional data were plotted and analyzed using Prism 8.0 software (GraphPad Software, San Diego, CA). Comparisons were performed using the Student t test between two groups or ANOVA for multiple groups. For IPITT, IPGTT, and serum insulin studies, two-way AVOVA with multiple comparisons was used. Results are presented as mean ± SEM. P < 0.05 is considered statistically significant.

All study data are included in the manuscript and/or supporting information.

The miRNAs that could mimic the effects of IL-1β were revealed by miRNA microarray screening and cell viability comparisons. We identified miR-503-5p as one of the six most upregulated miRNAs in β-cells after treatment with IL-1β (Supplementary Fig. 1A and B). The six altered miRNAs were overexpressed in INS-1 cells for 24 h and then treated with or without a high dose of IL-1β for another 24 h. Cell viabilities were significantly inhibited by expression of miR-25-3p, miR-146a-5p, miR-503-5p, and miR-153-3p, but cells transfected with miR-146a-5p and miR-503-5p showed no further decrease in cell viability after IL-1β administration (Supplementary Fig. 1C). This finding suggests that these two miRNAs are able to function as positive or negative feedback effectors of IL-1β.

Some reports have shown that miR-146-5p acts as a feedback inhibition molecule to suppress IL-1β effects by targeting Irak1 and Traf6 (18). How miR-503-5p functions in the IL-1β signaling was never appreciated. We manipulated miR-503-5p expression with or without IL-1β treatment, and GSIS function and β-cell proliferation were evaluated. Overexpressing miR-503-5p significantly impaired the GSIS function of mouse islets, akin to the effect caused by IL-1β treatment (Supplementary Fig. 1D). On the contrary, reducing miR-503-5p levels under IL-1β circumstances alleviated islet GSIS dysfunction (Supplementary Fig. 1D). The proliferative capacity of β-cells were also impaired by miR-503-5p overexpression and IL-1β treatment, the latter of which could be recovered by reducing miR-503-5p levels (Supplementary Fig. 1E and F). The above observations prompted us to investigate the biological and disease-prone effects of miR-503-5p in β-cells.

Lynn et al. (19) reported that miR-503-5p is enriched in embryonic day 14.5 murine pancreata and necessary for pancreas organogenesis. In the current study, we found that miR-503-5p was enriched in embryo pancreata but significantly decreased postnatally, hence maintaining at a relatively low level in adult islets and β-cells (Supplementary Fig. 1G and H). However, mature miR-503-5p expression was upregulated in INS-1 cells and mouse or human islets treated with IL-1β (Fig. 1A–C). The upregulation was also observed in primary islets from HFD-fed mice, 2-year-old aged mice, diabetic db/db mice, and NOD mice considered as inflammation-associated diabetic rodents (Fig. 1D and E and Supplementary Fig. 1I and J).

DNA methylation and histone modifications are two key epigenetic changes involved in gene silencing and reexpression during development and under disease conditions (20). To confirm that the relevance of IL-1β and miR-503-5p depends on epigenetic modification, MIN6 cells were treated with DNA methylation inhibitor 5-Aza and histone deacetylase inhibitor TSA, resulting in miR-503-5p expression being significantly induced, but inhibition of DNA methylation had a much stronger regulatory effect on miR-503-5p transcription (Fig. 1F). Furthermore, expression of pri-miR-503 was mostly induced in MIN6 cells treated with IL-1β and 5-Aza simultaneously versus separately (Fig. 1F). We analyzed the promoter region of miR503HG and confirmed the extensive CpG sites in region −3,617 to −1,417 base pairs (bp) in both human and mouse genomes (Supplementary Fig. 1K). A search of the human genome methylation database, based on the Illumina HumanMethylation450 BeadChip of >300 subjects, showed hypomethylation of this region of miR503HG in patients with type 2 diabetes (Fig. 1J). Indeed, natural aging triggered a demethylation of miR503HG in the promoter region −2,397 to −2,213 bp in human, but no alterations of promoter methylation were further altered between elderly individuals with or without diabetes (Supplementary Fig. 1L).

The pattern of miR-503-5p in diabetes development was determined by evaluating the elevated serum miR-503-5p levels in diabetic mice (Fig. 1G and H and Supplementary Fig. 1M). The metabolic organ that contributed to the increased serum miR-503-5p levels in the diabetic condition was determined by examining transcripts of miR-503 (pri-miR-503) in HFD-fed mice and in aged mice. The levels of pri-miR-503 increased slightly in white adipose tissue, but increased about threefold in HFD-fed mouse islets and more than fivefold in aged mouse islets (Fig. 1I). These findings suggest that the pancreatic islets, or more specifically the β-cells, might release miR-503-5p, which contributes mainly to the elevation of circulatory miR-503-5p. The upregulation of pri-miR-503 in islets and miR-503-5p content in serum were also confirmed in patients with diabetes compared with healthy human participants (Fig. 1K and L). Serum miR-503-5p levels were positively correlated with FBG levels during aging (Fig. 1M). Taken together, these findings indicate that the IL-1β–induced miR-503 reexpression in islets depends on promotor hypomethylation and may further participate in the deterioration of diabetes in a distal and proximal manner.

miR-503 is one of the clustered miRNAs that is polycistronically expressed with miR-322 (human miR-424) and miR-351. We studied the in vivo role of this cluster in the development of diabetes by constructing a transgenic mouse expressing a β-cell–specific miR-503-322–351 cluster driven by the rat Ins2 promoter (Cluster TG) (Supplementary Fig. 2A and B). Analysis of the phenotype of 12 F1 pups with copy numbers of ∼25 from two different founder mice (Supplementary Fig. 2C) revealed that the Cluster TG mice were weak and defective in growth and had severe hyperglycemia (Supplementary Fig. 2D–F). The F1 pups were unable to survive >34 days (Supplementary Fig. 2G). H-E staining showed a significantly decreased islet mass in the Cluster TG mice due to a defect in β-cell replication (Supplementary Fig. 2H–J). Thus, miR-503-322–351 overexpression in β-cells resulted in growth defects and severe diabetes due to a significant loss of β-cell mass.

We conducted a further investigation of the role of miR-503 in β-cells and disease progression by constructing miR-503 transgenic mice with the rat Ins2 promoter (βTG) and using our knowledge that the flanking sequences are critical for miRNA maturation (9). To ensure overexpression of miR-503, 300 bp of both upstream and downstream sequences were included in the construct, and the full sequence of miR-322 was involved. We preserved three lines of βTG mice with different copy numbers as follows: Line50 (5.2), Line58 (12.6), and Line57 (22.8) (Fig. 2A and Supplementary Fig. 2K). The increase of pri-miR-503 in islets was copy number dependent (Fig. 2B), and no elevation of miR-322 was observed in islets (data not shown). The transcription levels of pri-miR-503 were altered in metabolic tissues (Supplementary Fig. 2L–N); however, pri-miR-503 expression was elevated in the hypothalamus of Line57 mice (Supplementary Fig. 2O) owing to endogenous Ins2 expression in the brain (21). Therefore, three genetically stable lines of β-cell–specific miR-503–overexpressing mice were obtained.

The newborn βTG mice were normal, but Line58 and Line57 mice started to lose weight at 4 weeks, in accordance with elevated miR-503-5p levels in mouse sera (Fig. 2C). The decreased body weights resulted from fat expansion deficiencies (Fig. 2D and E). The FBG and random blood glucose (RBG) levels of βTG mice increased in parallel with the increases of serum miR-503-5p levels (Fig. 2F and G) at 6 weeks.

Since Line57 mice had pri-miR-503 elevation in hypothalamus and could not survive for >10 weeks (Supplementary Fig. 2O and P), the metabolic cage experiments were performed for both Line50 and Line58 mice. The Line50 mice revealed no obvious alterations in terms of respiratory exchange ratio (RER), daily food intake, and drinking water at 7 weeks old, whereas the Line58 mice showed a decreased RER during the night (Fig. 2H), suggesting a preferred use of fat as an energy source. Meanwhile, the daily food intake and water consumption were significantly greater in the Line58 mice (Fig. 2I and J). Feeding behavior showed no differences among groups (data not shown). Furthermore, the Line58 mice were dramatically glucose intolerant, with decreased levels of insulin and C-peptide after refeeding, while no alterations of fasting insulin levels were observed (Fig. 2K and L and Supplementary Fig. 2Q). Importantly, the HOMA of insulin resistance analysis confirmed a severe insulin resistance in all three lines of βTG mice (Fig. 2M). Collectively, the results from the Cluster TG and βTG mice strongly support that overexpression of miR-503-5p in β-cells causes a stepwise disease progression, namely insulin resistance, defective insulin secretion, and advanced diabetes.

Aging and overnutrition are two metabolic stresses common to both humans and rodents. The miR-503-5p expression level in islets from Line50 mice resembled those of aged mice and HFD-fed mice. Therefore, Line50 mice were subjected to constant evaluation with age under both normal chow diet (NCD) and HFD feeding conditions. Glucose intolerance in the Line50 mice manifested at 8 weeks (Supplementary Fig. 3A and B), but in vivo GSIS was enhanced in Line50 mice over that in WT mice (Supplementary Fig. 3C), ruling out an involvement of defective insulin secretion. Prolonged elevation of miR-503-5p both in further NCD- and HFD-fed mice did not alter the body weight (data not shown) but strengthened insulin resistance and diminished in vivo GSIS, contributing to the deleterious glucose intolerance (Fig. 3A–D and Supplementary Fig. 3C and D).

According to the timing sequence of the Line50 mouse phenotype, we first ascertained the exact tissues involved in insulin resistance in 16-week-old mice fed an NCD by performing a standard hyperinsulinemic-euglycemic clamp technique using [3-³H]-glucose and 2-deoxy-d-[1-¹⁴C]-glucose as tracers (Fig. 3E). After constant insulin infusion, the blood glucose levels of both groups were clamped at similar levels (Fig. 3F). The glucose infusion rate (GIR) tended to decrease in Line50 mice, and the steady GIR was significantly lower in Line50 mice than in WT mice (Fig. 3F and G), in accordance with the changes of glucose intolerance. The whole-body use of glucose and the clamped glucose disposal rate (GDR) were defective in Line50 mice because of insufficient glucose uptake in the adipose tissue but not in the muscle and liver (Fig. 3H–K and Supplementary Fig. 3E). Basal HGP was not altered in Line50 mice, whereas insulin-inhibited HGP was significantly diminished (Fig. 3L and Supplementary Fig. 3F). No alterations were noted in glycogen synthesis in liver and skeletal muscle (Supplementary Fig. 3G and H). Whole-body lipogenesis was also unchanged (Supplementary Fig. 3I), in line with the similar body weight observed between groups. Together, the clamp data demonstrate that overexpressing miR-503-5p in β-cells causes mouse insulin resistance in the liver and adipose tissue but not in the skeletal muscle.

The islet morphology and GSIS function were ascertained in succession. We found no significant difference in islet mass but reduced islet size in Line50 mice compared with that in WT mice (Fig. 3M and N and Supplementary Fig. 3J). We also observed a slightly decreased number of β-cells and modestly increased number of α-cells and pancreatic polypeptide cells, but no change in number of δ-cells in Line50 mice (Supplementary Fig. 3J–N). These discrepancies in islet size reduction were amplified by HFD feeding because of a suppression of β-cell replication by miR-503-5p (Fig. 3N–P), since no apoptotic cells were observed even in miR-503-5p highly expressed Line57 mouse islets (Supplementary Fig. 3O). Islet perfusion showed a severe impairment of both first and second phases of insulin secretion in Line50 islets, but no alterations of potassium-stimulated insulin secretion and insulin content were observed between groups (Fig. 3Q and R and Supplementary Fig. 3P–R). The static GSIS assay also showed a significant decrease of GSIS in Line50 islets (Supplementary Fig. 3S). As the potassium-stimulated insulin secretion showed an intact effect of membrane depolarization–stimulated insulin secretion (Supplementary Fig. 3Q and R), the decreased intracellular Ca²⁺ level was probably due to the insufficiencies of glucose metabolism and ATP-sensitive potassium channel closure in miR-503-5p–elevated β-cells (Fig. 3S and T and Supplementary Fig. 3T). These results demonstrate that both physiological (aging) and pathological (HFD feeding) metabolic stresses enhance the diabetic phenotype of miR-503-5p overexpression mice, initially leading to insulin resistance in the liver and adipose tissue and ultimately causing β-cell decompensation due to GSIS dysfunction and the inability of compensatory β-cell proliferation.

The possibility that ablation of miR-503-5p could improve the metabolic disruptions caused by HFD feeding was investigated by global deletion of the miR-503 cluster (KO mice) (Supplementary Fig. 4A). Deletion of this cluster was confirmed by investigation of genomic DNA and gene expression in metabolic organs (Supplementary Fig. 4B and C). The KO mice were healthy and fertile, with no obvious metabolic abnormities. Subjecting the KO mice to HFD feeding for 4 months did not result in any noticeable changes in body weight compared with that of WT mice (Fig. 4A). The FBG levels were also similar between groups; however, blood glucose levels after refeeding were significantly lower in KO than in WT mice (Fig. 4B). Correspondingly, the KO mice showed better glucose tolerance after HFD feeding (Fig. 4C).

Insulin sensitivity and islet function were also monitored to identify the contribution of glucose intolerance. No alterations in fasting serum insulin levels were observed between the groups, but the KO mice had a lower level of postprandial insulin, suggesting an insulin-sensitive state (Fig. 4D). Indeed, IPITTs clearly showed that the KO mice had quick and efficient responses to insulin (Fig. 4E). Islets from HFD-fed mice showed a significantly higher insulin secretion compared with that from NCD-fed mice, but no differences in insulin secretion were noted in islets from KO and WT mice under the HFD condition at that time (Supplementary Fig. 4D and E). Insulin signal transduction determined by phosphorylated AKT(S473) level was significantly improved in KO mice in both adipose tissue and liver (Fig. 4F–H).

The hyperinsulinemic-euglycemic clamp accurately determined increased GIR, whole-body glycolysis, GDR, and adipose tissue glucose uptake and inhibited HGP basally in HFD-fed KO mice compared with WT mice (Fig. 4I–M), opposite of the findings in the βTG mice. We then extended HFD feeding to 5 months, by which time WT mice showed β-cell dysfunction (55). We found that KO mice secreted more insulin than WT mice at 30 min after glucose injection (Fig. 4N). As no alterations of islet and β-cell mass were observed, the enhanced in vivo GSIS was largely attributed to the alleviation of both the first and second phases of insulin secretion at the single-islet level in HFD KO mice (Fig. 4O and P and Supplementary Fig. 4F–I). Together, deletion of the miR-503 cluster alleviates both insulin insensitivity and β-cell dysfunction, thus preventing the canonical pathogenesis of type 2 diabetes.

Analysis of miR-503-5p expression in metabolic tissues of animals, which suffered aging or overnutrition, revealed islets as the main high-expression tissue. However, β-cell dysfunction was not the initial effect in mouse models. To determine whether miR-503-5p reexpression in mature β-cells would lead to β-cell defects, the unbiased proteomic analysis of MIN6 cells was performed. Ingenuity pathway analysis (IPA) revealed that 4 of the top 30 pathways were associated with the MAPK signaling pathway (Fig. 5A). Indeed, miR-503-5p overexpression induced phosphorylation of JNK MAPK, p38 MAPK upregulation, and downregulation of phosphorylated extracellular signal–regulated kinase (ERK) 1/2 MAPK in both MIN6 cells and Line50 mouse islets (Fig. 5B and C and Supplementary Fig. 5A and B). But these changes were reversed in HFD-fed KO mouse islets or by miR-503-5p knockdown in IL-1β–treated MIN6 cells (Fig. 5B and C and Supplementary Fig. 5A and B).

As reported that chronic activations of JNK and p38 repress GSIS function (22), we investigated the involvement of JNK and p38 by treating βTG islets with JNK (SP600125) and p38 (SB23963) inhibitors. The inhibitors significantly improved the second phase of insulin secretion, supporting activations of JNK and p38 causing GSIS dysfunction in βTG islets (Fig. 5D and E and Supplementary Fig. 5C).

To investigate the probable targets of miR-503-5p involved in MAPK pathways, we used miRanda and TargetScan prediction software to narrow down candidates to three: MAPK8IP2 (JIP2), MAP2K1 (MEK1), and GAREM (Supplementary Fig. 5D). It has been reported that reduction of JIP2 promotes JNK phosphorylation, and MEK1 downregulation inhibits ERK1/2 activation (23–25). The regulatory role of miR-503-5p in the expression of Jip2 and Mek1 was verified by luciferase assays (Fig. 5F). We further confirmed the regulatory role of miR-503-5p on JIP2 at the protein level. As shown in Supplementary Fig. 5E–G, the protein level of JIP2 was decreased by miR-503-5p overexpression in MIN6 cells and βTG islets, whereas it was rescued in islets from KO mice fed an HFD and in IL-1β–treated MIN6 cells with miR-503-5p knockdown. Immunohistochemistry staining of JIP2 also confirmed its decrease in mouse islets with high miR-503-5p expression models, such as aging and HFD, and its increase in islets of HFD-fed KO mice (Supplementary Fig. 5H). Notably, we discovered the same regulatory manner of miR-503-5p/JIP2 expression and β-cell dysfunction in human islets (Fig. 5G–I and Supplementary Fig. 5I and J), illustrating the conservation of the miR-503-5p/JIP2 axis among species.

To verify that the decreased JIP2 contributes to the activation of MAPK pathways and GSIS dysfunction in miR-503-5p overexpressed β-cells, we constructed and purified JIP2 adenoviruses. βTG mouse islets infected with Adv.JIP2 significantly recovered the increased levels of phosphorylated JNK and phosphorylated p38 compared with those infected with Adv.Ctrl (Fig. 5J, Supplementary Fig. 5K–M), leading to the improvements of first and second phases of insulin secretion in Adv.JIP2-infected βTG islets (Fig. 5K–M). The above findings support that as a main target of miR-503-5p in β-cells, JIP2 promotes β-cell function via inhibiting JNK and p38 MAPK pathways, and the miR-503-5p/JIP2/MAPK axis is conserved in both mouse and human islet β-cells.

The mechanisms underlying insulin resistance induced by β-cell–derived miR-503-5p remain elusive. Careful observation of TEM images from βTG islets revealed a considerable increased number of nanovesicles in mature insulin granules (Fig. 6A and Supplementary Fig. 6A and B). We discovered that elevated serum levels of miR-503-5p accompanied insulin secretion during the fasting and refeeding cycle (Supplementary Fig. 6C–E). We further investigated whether insulin granules might be involved in nanovesicle transportation by dispersing βTG islets into single cells that were infected with EGFP-NPY adenovirus as a tracer for the insulin granules. The addition of PKH26-labled nanovesicles to the culture medium resulted in a clear colocalization with insulin granules (Fig. 6B). We confirmed that the nanovesicles resided within the insulin granules by tracing a nanovesicle membrane marker CD63 with NPY-represented insulin, as well as by observing CD63 within the insulin granules via immunoelectron microscopy (Fig. 6B and Supplementary Fig. 6F). Moreover, we observed a colocalization of immunogold-labeled CD63 and insulin at the single–insulin granule level by TEM in WT islet β-cells, while Line50 islet β-cells had more CD63 protein enriched in a single insulin granule, indicative of an increase in nanovesicle formation in miR-503-5p transgenic mouse (Fig. 6C). The colocalization of NPY and miR-503-cy3 was also observed in MIN6 cells (Fig. 6D), together strongly proving that miR-503-5p–containing nanovesicles exist in insulin granules.

After purification of nanovesicles secreted by mouse islets in ex vivo culture media, their structure and concentrations were analyzed by TEM, NTA, and Western blotting. The TEM images revealed round nanovesicles with a diameter of ∼50 nm (Supplementary Fig. 6G), and NTA confirmed that the nanovesicles were mostly ∼45 nm in size and more concentrated in βTG islets (Supplementary Fig. 6H and I). After adjusting for protein levels, the released nanovesicles were determined to package more miR-503-5p in βTG islets than in WT islets (Supplementary Fig. 6J). The Western blotting assays showed high expression of TSG101 and CD63 in islet-derived nanovesicles, but not GAPDH (Supplementary Fig. 6K). Therefore, miR-503-5p is likely hijacking insulin granules to form nanovesicles, which are transported and released as insulin granule cargos.

The systematic distribution of β-cell–derived miR-503-5p nanovesicles from both WT and βTG islets was determined by analyzing the quantities of miR-503-5p in serum, liver, adipose tissue, and skeletal muscle. The serum miR-503-5p concentrations were higher in 16-week-old Line50 mice and HFD-fed WT mice than in NCD-fed WT mice, leading to accumulation in the liver and adipose tissue (Supplementary Fig. 6L–N). Serum miR-503-5p levels were also found to be positively correlated with fasting insulin levels (Supplementary Fig. 6O). Importantly, we also monitored the distribution and function of miR-503-5p nanovesicles by directly injecting them into C57BL/6J mice via the tail vein. The PKH26-labeled miR-503-5p nanovesicles were also largely accumulated in the liver and adipose tissue but not in the skeletal muscle (Fig. 6E). Unlike the injection of single-strand miR-503-5p, βTG nanovesicles caused both hyperglycemia and hyperinsulinemia without affecting body weights, akin to the effects of insulin resistance inducer S961 (Supplementary Fig. 6P–T).

The target gene of miR-503-5p that induced insulin resistance was identified by a cross analysis of IPA-designed upstream regulators with predicted miR-503-5p target genes using TargetScan and miRanda software (Supplementary Fig. 7A). In addition to affecting MAPK signaling pathway genes, Insr, Igf1r, and Mknk1 had seed sequences of miR-503-5p. Mknk1 was ruled out, as it is a downstream substrate of p38 and ERK1/2 (22), both of which were regulated by miR-503-5p. We confirmed the inhibitory effect of miR-503-5p on Insr via luciferase activities derived from WT and mutant sequences (Supplementary Fig. 7B). The protein levels of INSR were significantly reduced in MIN6 cells, while its level was increased in human islets and βTG islets (Supplementary Fig. 7C and D), probably owing to a high concentration of local insulin that restrains the INSR degradation (26).

Since βTG-NVs tended to accumulate in the liver and adipose tissue (Fig. 6E), we verified a regulatory role of miR-503-5p in primary hepatocytes and mature adipocytes by treating them with culture medium and nanovesicles from Line50 mouse islets, as well as double-stranded miR-503-5p transfection. Hepatic INSR levels were suppressed by miR-503-5p in all cases (Supplementary Fig. 7E and F); however, adipose INSR levels were significantly reduced after administration of insulin (Supplementary Fig. 7G–I). Strikingly, the INSR protein amounts in liver and adipose tissue were also significantly repressed by βTG nanovesicles compared with WT nanovesicles at the in vivo level (Fig. 6F and Supplementary Fig. 7J), in agreement with the accumulation of islet-derived nanovesicles in these two organs. Similarly, protein levels of INSR were also reduced in liver and adipose tissue from βTG mice but not in muscle (Fig. 6G and Supplementary Fig. 7K). Conversely, ablation of miR-503-5p enhanced INSR protein levels in adipose tissue and liver under both NCD and HFD conditions (Fig. 6H and Supplementary Fig. 7L). Taken together, these in vitro and in vivo findings demonstrate that β-cell–derived miR-503-5p nanovesicles are encapsulated in insulin granules, secrete along with insulin, and circulate to accumulate in liver and adipose tissues, where miR-503-5p represses INSR to dampen insulin action.

The expression levels of the miR-503-5p were mostly upregulated in β-cells during aging; therefore, we treated aged mice with MIP1-driven miR-503 cluster sponge AAVs (βsponge), a technique widely used to inhibit miRNA effects in vivo (8). The 60-week-old male mice showed significantly increased body weights, elevated FBG levels, and improved glucose tolerance and insulin resistance compared with the 10-week-old mice (Supplementary Fig. 8A–D). These aged mice were randomly divided into two groups and injected with βsponge viruses or MIP1-GFP controls (Fig. 7A and Supplementary Fig. 8E). Metabolic characteristics were evaluated to show the therapeutic effects starting at 4 weeks after virus injection. No differences in body weights and RBG levels were observed between GFP- and βsponge virus–infected mice (data not shown). However, mice receiving βsponge viruses showed lower fasting glucose and lower refeeding blood glucose levels (Fig. 7B). Aging-induced glucose intolerance was significantly suppressed by blocking miR-503 cluster effects in β-cells due to an increase in GSIS (Fig. 7C and D and Supplementary Fig. 8F and G). An enhanced secretion of insulin in the βsponge-treated mice was frequently detected under random conditions, suggesting a strong recovery of β-cell function (Fig. 7E). This, in turn, might improve insulin sensitivity after 7 weeks posttreatment without much alterations of INSR and IGF-1R protein amounts in insulin-sensitive tissues (Fig. 7F and Supplementary Fig. 8H–J).

The expression levels of miR-503-5p and miR-322-5p were significantly decreased in βsponge mouse islets (Fig. 7G), and their contents in serum nanovesicles were also reduced (Fig. 7H). No differences were detected in islet mass and size distribution (Fig. 7I and J), but JIP2 expression was significantly rescued in βsponge mice (Fig. 7K). However, the use of βsponge viruses showed no improvement in glucose metabolism and β-cell function in HFD-fed mice (data not shown). Taken together, our data suggest that specifically blocking the miR-503-5p effect in β-cells can treat aging-associated diabetes by enhancing insulin secretion.

The results presented here reveal that β-cell–derived miR-503-5p promotes insulin insensitivity and β-cell dysfunction in the context of chronic inflammation, thereby accelerating β-cell decompensation and the onset of diabetes. Early during the inflammatory infiltration of the islets, slight increases in β-cell miR-503-5p results in its partial packaging in insulin secretory granules, which then spill into the bloodstream. These extracellular miR-503-5p nanovesicles circulate to insulin-responsive tissues, particularly the liver and adipose tissues, where it targets insulin receptors and dampens insulin signals. Metabolic stress promotes the accumulation of miR-503-5p–expressing inner β-cells and triggers chronic activation of p38 MAPK and JNK MAPK to drive β-cells from compensatory insulin secretion to decompensatory β-cell defects. Our findings show that pancreatic β-cells could form a metabolic center by secreting insulin if metabolic stress can control insulin resistance and diabetes progression by releasing nanovesicular miR-503-5p.

The elevation of circulating miR-503-5p in elderly people and patients with type 2 diabetes, as well as in HFD-fed and aged mice as shown here, might mainly stem from β-cells that experience a prolonged IL-1β/IL-1 receptor 1 (1R1) activation, since β-cells have the highest numbers of IL-1Rs on their surfaces (27). The promotor region of miR-503-5p is hypomethylated in patients with type 2 diabetes, and activation by IL-1β assures that the β-cells act as generators of miR-503-5p under metaflammation and inflammageing conditions. Numerous reports have shown that persistent stresses from unfolded protein responses and reactive oxygen species–mediated modifications in β-cells creates a proinflammatory islet microenvironment and results in IL-1β–induced insulitis, preceding the onset of frank diabetes (28). Physiological levels of IL-1β stimulate insulin to promote postprandial glucose disposal, whereas supraphysiologic IL-1β can inhibit β-cell function and mass by triggering signaling cascades from membranous IL-1Rs to downstream MAPK and inhibitor of κB kinase effectors (22,29,30). Insulitis also generates IL-1R antagonist (IL-1Ra), which blocks the prolonged IL-1β effects (31). Some reports have shown that plasma levels of IL-1Ra significantly increase in obese humans and rodents and positively correlate with insulin resistance and diabetes onset (32). However, the levels of IL-1Ra are decreased in islets from patients with type 2 diabetes, thus failing to protect β-cell compensatory expansion (33). The seemingly incompatible levels of IL-1Ra in circulation and in islets strongly suggest that other molecules may also contribute to IL-1β–induced β-cell decompensation and systemic insulin resistance. In our opinion, miR-503-5p may serve this function. A gene-regulating effect of miR-503-5p in insulin responsive tissues and β-cells could explain the poor effectiveness of anti-IL-1 approaches in clinical trials, despite a great improvement in β-cell function. Other β-cell molecules, such as miR-26, miR-29, and miR-375, may also contribute to insulin resistance and β-cell dysfunction in response to chronic inflammation (11,12,34).

The miR-503-322 cluster, which is highly expressed in the developing pancreas in embryonic day 16.5 mice, shows gradually decreasing expression postnatally and maintains this low expression level in mature pancreatic islets. The loss of miRNA processing by Dicer deletion uniquely impairs the development of endocrine lineage (35); therefore, embryonic expression of miR-503-322 may be particularly important in β-cells. Indeed, Lynn et al. (19) reported that miR-503-5p is colocalized with the β-cell–determinant Pdx1 in the pancreas. Our findings and those of other groups support the importance of miR-503-5p in β-cell genesis and that its decrease may ensure rapid β-cell expansion and replication at postnatal 2–4 weeks. Moreover, the reexpression of miR-503-5p in mature β-cells from rodents to humans may cause a compensatory decrease in β-cell expansion and function dose dependently by IL-1β–like activations of the MAPK pathways.

The MAPKs, including the JNK, p38, and ERK signaling pathways, contribute to cell morphogenesis and self-renewal during embryo development (36–39). Activation of JNK was confirmed to influence cytoskeletal remodeling, and activation of p38 inhibits PKD1 phosphorylation, both of which cause insulin secretion defects (39–41). The JIPs were first investigated as retrograde regulators of the JNK signaling pathway (23). Compared with the specificity of JIP1 in the regulation of JNK (42), JIP2 was confirmed to function as a scaffold protein for both DLK1 and Tiam1, thereby regulating the JNK and p38 MAPK signaling pathways (43,44). Importantly, the S59N mutant of JIP1 in humans shows reduced Glut2 and Ins expression in β-cells, leading to the onset of the young form of diabetes (45). In this study, we confirmed that downregulation of JIP2 activates JNK and p38 phosphorylation to impair β-cell function and promote the onset of frank type 2 diabetes. The combined results showing downregulation of JIP2 and MEK1 support the conclusion that miR-503 mimics damage the effects of IL-1β in β-cells.

Exosomes/nanovesicles with diameters of 30–150 nm are enveloped into late endosomes/multivesicular bodies (46). Upon releasing into the extracellular environment, nanovesicles can fuse with live cells and transfer their cargos of proteins, lipids, and RNAs into the acceptor cells (47). Endocrine β-cells release exosomes under normal circumstances while changing their cargos in response to different stimuli (48). Xu et al. (11) reported that islet-derived exosomes promote insulin action and glucose tolerance in local and distal organs via transferring miR-26 cargo, providing direct evidence that deregulation of β-cell exosomes leads to diabetes progression. However, the precise location of exosomes/nanovesicles within the β-cells remains elusive. Our finding that miR-503-5p is encapsulated into nanovesicles and located in insulin granules was serendipitous but exciting. The insulin granule, which has a diameter of 300–350 nm, has been proposed to serve as a signaling hub rather than simply functioning as an insulin container (49). Cosecreted compounds, including amylin and γ-amino butyric acid, also have important metabolic regulatory functions (50). Our TEM, immunogold staining, and live-cell imaging results confirm that nanovesicles, with a diameter of 45 nm, reside within the β-cells in the insulin granule. Primary β-cells contain ∼10,000 insulin granules, corresponding to 10–20% of the total cell volume, thereby providing a large nanovesicle reservoir. Moreover, the insulin granule has a half-life of 3 days, ensuring an opportunity for nanovesicle formation. The process by which the insulin granule encapsulates nanovesicles might be controlled by specific miRNAs based on their sequences, as miR-503-5p overexpression significantly increases the number of nanovesicle-containing insulin granules. Alternatively, miR-503-5p inhibition of ERK signaling may facilitate nanovesicular miR-503-5p formation in the insulin granules, consistent with the literature showing that inhibition of KRAS-MEK signaling boosts Ago2-associated miRNA sorting into exosomes (51). The mechanism by which miRNA and miRNA-associated proteins are sorted into insulin granular nanovesicles requires further study, especially since our finding does not exclude multivesicular bodies as an origin of the nanovesicles in β-cells. Nonetheless, our data indicate that a diabetes-associated nanovesicular miR-503-5p resides within the β-cell insulin granule, and this finding enriches the current knowledge of insulin granule contents.

Nanovesicular miR-503-5p released by mouse islet β-cells directly promotes insulin resistance in liver and adipose tissues, in part by downregulating INSR expression. The discovery of INSR is based on unbiased proteomic data from comparing differential expression of proteins between miR-503-5p–expressing and control miRNA–overexpressing MIN6 cells. The regulatory role between miR-503-5p and Insr was confirmed by 3′ untranslated region seed sequence–based luciferase activity and protein abundance in MIN6 cells and INS-1 cells, two widely used β-cell lines. However, the protein levels of INSR were enhanced, rather than reduced, in primary islets from βTG mice and from miR-503-5p–transfected human islets. The discrepancy in the results between β-cell lines and primary islets may reflect that the insulin secretory abilities enable insulin receptors boosted by autocrine insulin to compensate for the reduction of these receptors in miR-503-5p–overexpressing β-cells (26). Alternatively, the hypersecretion ability of primary β-cells could promote nanovesicular miR-503-5p release, thereby minimizing the cellular miR-503-5p levels and reducing their gene regulatory effect in human and mouse islets.

The fasting serum insulin level of βTG mice is barely altered, regardless of how many β-cells remain, suggesting that insulin receptors in peripheral tissues are unlikely to be affected by circulatory insulin levels. Therefore, the significantly decreased levels of INSR in insulin-responsive tissues solely originate from the elevation of miR-503-5p levels in the βTG mice as well as in HFD-induced mice, at least in the liver and adipose tissues. Mice with tissue-specific deletion of insulin receptors in liver, muscle, adipose tissue, and β-cells display insulin resistance, β-cell decompensation, glucose intolerance, diabetes, and early death, largely akin to the phenotypes of our βTG mice (52,53). The expression of miR-503-5p also determined the extent of preadipocyte maturation together with insulin, as indicated in both the βTG mice and global miR-503-5p cluster KO mice, with the most pronounced abnormalities noted in adipose tissue. This defect likely arises from the loss of IGF-1Rs in the adipose tissue of βTG mice, considering the relatively high expression level of IGF-1R compared with INSR in human preadipocytes (54). Coincidently, Sakaguchi et al. (16) reported that KO mice show a higher body weight due to increased white fat content at a later age, suggesting an essential role of the miR-503-5p cluster in mature adipocytes. Whether the combined loss of insulin and IGF-1Rs is the cause of HFD-induced adipose expansion failure remains to be established.

In summary, our study reveals a metaflammation- and inflammageing-related involvement of miR-503-5p in determining type 2 diabetes that leads to insulin resistance and β-cell decompensation. This miR-503-5p activity also increases the infection risk with age in severe diabetes settings. Thus, therapeutic strategies aimed at blocking miR-503-5p generation in β-cells may prove to be a promising approach for preventing insulin resistance, β-cell decompensation, and diabetes onset, as well as diabetes-associated infection.

Acknowledgments. The authors thank Dr. Zhe-Ming Gu and Long-Meng He from Jiangsu Value Pharmaceutical Services Co. for their technical support in the hyperinsulinemic-euglycemic clamp study using radioactive-labeled glucose.

Funding. This study was supported by National Natural Science Foundation of China grants 82070843, 81870531, and 82270844 (to Y. Zhu) and 81830024 (to X.H.), Natural Science Foundation of Jiangsu Province grant BK20211375 (to W.T.), and Postdoctoral Science Foundation of Jiangsu Province grant 2022ZB718 (to Y.Zho.).

The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. Y.Zho., K.L., Y.Zha., Y.W., Y.S., and Y. Zhu contributed to the investigation. Y.Zho., K.L., Y.S., Y.W., P.S., and Y. Zhu contributed to the methodology. Y.Zho., K.L., and Y. Zhu drafted the manuscript. Y.Zho. and Y. Zhu contributed to the visualization of the experiments. Y.Zho., Y. Zhu, and X.H. contributed to the formal analysis. W.T., Y.L., R.B., R.L., X.C., and S.W. contributed resources. Y. Zhu and X.H. conceptualized the study, provided supervision, acquired funding, and reviewed and edited the manuscript. All authors reviewed and commented on the manuscript. Y. Zhu and X.H. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.